ENERGY TRANSPORT, PROTEIN “SYNTHESIS, AND HORMONAL CONTROL OF HEART METABOLISM Fourth USA-USSR Joint Symposium on Myocardial Metabolism Tashkent, USSR September 14-16, 1979 Office of the Director National Heart, Lung, and Blood Institute ication No. August 1980 Pr IS 2) /777 PREFACE SH Ce A / nu KL The Fourth USA-USSR Joint Symposium on Myocardial Metabolism was held September 14-16, 1979, in Tashkent, USSR. These proceedings contain the 30 papers pre- pared for this symposium. Problem Area 3, Myocardial Metabolism, is one of the most active cardiovascu- lar exchange programs conducted under the bilateral Agreement in the field of Medical Science and Public Health between the governments of the United States and the Soviet Union. Collaborative efforts have led to an expansion of sci- entific interests and activities, including the exchange of senior scientists to conduct lectures and seminars, the exchange of young scientists to continue joint laboratory studies and develop joint publications, and the initiation of a series of collaborative research projects. The papers presented at the fourth symposium demonstrate the growth of ideas and significant research findings made in the area of myocardial metabolism under this exchange program. Papers were presented on the regulation of con- traction and energy metabolism of the heart; hormones, prostaglandins, and cyclic nucleotides in the heart; and protein biosynthesis and pathology of the heart. It is hoped that publication of ongoing work in this area will serve to stimulate additional research in both countries. The proceedings are being published simultaneously in Russian in the Soviet Union. Professor Howard E. Morgan of The Pennsylvania State University was the US co- chairman of the symposium and serves as the US coordinator for problem area 3. Academician Eugene I. Chazov and Professor Vladimir N. Smirnov of the All-Union Cardiology Research Center, USSR Academy of Medical Sciences, were the USSR cochairmen of the symposium; Professor Smirnov serves as the USSR coordinator in this problem area. On behalf of the National Heart, Lung, and Blood Institute, I wish to thank the cochairmen of the symposium, Professor Morgan, Academician Chazov, and Professor Smirnov; the US and USSR participants; and all those who partici- pated in the planning of the symposium and editing of the proceedings. Ruth Johnsson Hegyeli, M.D. Assistant Director for International Programs National Heart, Lung, and Blood Institute Bethesda, Maryland US PARTICIPANTS Dr. Howard E. Morgan (US Cochairman) Evan Pugh Professor of Physiology Chairman, Department of Physiology The Milton S. Hershey Medical Center The Pennsylvania State University Hershey, Pennsylvania Dr. John R. Blinks Professor and Chairman Department of Pharmacology Mayo Foundation Rochester, Minnesota Dr. Sidney Fleischer Professor Department of Molecular Biology Vanderbilt University Nashville, Tennessee Dr. Alfred G. Gilman Professor Department of Pharmacology The University of Virginia School of Medicine Charlottesville, Virginia Dr. Ruth Johnsson Hegyeli Assistant Director for International Programs National Heart, Lung, and Blood Institute National Institutes of Health Bethesda, Maryland iii Dr. Robert B. Jennings Professor and Chairman Department of Pathology Duke University Medical Center Durham, North Carolina Dr. Glenn A. Langer Professor of Medicine and Physiology Castera Professor of Cardiology UCLA School of Medicine Los Angeles, California Dr. Philip Needleman Professor and Head Department of Pharmacology Washington University School of Medicine St. Louis, Missouri Dr. Martha Vaughan Chief, Laboratory of Cellular Metabolism National Heart, Lung, and Blood Institute National Institutes of Health Bethesda, Maryland iv USSR PARTICIPANTS Academician E. I. Chazov (USSR Cochairman) Director General All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Professor V. N. Smirnov (USSR Cochairman) Deputy Director General All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Dr. A. E. Bukatina Senior Researcher Institute of Biophysics USSR Academy of Sciences Puschino-on-0Oka Dr. V. I. Burakovsky Director Bakulev Institute for Cardiovascular Surgery USSR Academy of Medical Sciences Moscow Dr. N. A. Fedorova Senior Researcher Laboratory of Protein Chemistry Leningrad State University Leningrad Professor R. Katsenovich Director Uzbekistan Cardiology Institute Tashkent Dr. A. A. Kubatiev Senior Researcher Institute of Morphology USSR Academy of Medical Sciences Moscow Dr. V. V. Lednev Deputy Director Institute of Biophysics USSR Academy of Sciences Puschino-on-Oka Dr. D. O. Levitsky Senior Researcher Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Professor S. N. Lyzlova Department of Biochemistry Leningrad State University Leningrad vi Professor F. Z. Meerson Head, Laboratory of Pathological Physiology Institute of Normal and Pathological Physiology USSR Academy of Medical Sciences Moscow Dr. N. M. Mirsalikhova Senior Researcher Laboratory of Biophysics Institute of Biochemistry Academy of Sciences of the Uzbek SSR Tashkent Professor N. S. Panteléeva Head, Laboratory of Protein Chemistry Leningrad State University Leningrad Dr. 0. I. Pisarenko Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Dr. V. D. Pomoinetsky Senior Researcher Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow vii Dr. L. V. Rozenshtraukh Head, Laboratory of Cardiac Electrophysiology All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Dr. V. A. Saks Senior Researcher Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Dr. D. Saprygin Head, Laboratory of Clinical Chemistry Bakulev Institute for Cardiovascular Surgery USSR Academy of Medical Sciences Moscow Dr. Yu. M. Seleznev Senior Researcher Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Professor E. S. Severin Deputy Director Institute of Molecular Biology USSR Academy of Sciences Moscow viii Dr. V. G. Sharov Senior Researcher Laboratory of Human Pathological Morphology All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Dr. V. Stefanov Department of Biochemistry Leningrad State University Leningrad Dr. V. A. Tkachuk Senior Researcher Department of Biochemistry Moscow State University Moscow Dr. V. P. Torchilin Senior Researcher Laboratory of Myocardial Metabolism All-Union Cardiology Research Center USSR Academy of Medical Sciences Moscow Professor K. Yuldashev Deputy Minister of Health of Uzbekistan Head, Chair of Cardiology and Functional Diagnostics Institute of Advanced Training of Physicians Tashkent ix Dr. R. I. Zhdanov Senior Researcher Laboratory of Physicochemical Methods Institute for Biological Testing of Chemical Compounds Moscow TABLE OF CONTENTS og a = i US Participants tuieeeeeeeseeeeeeeesesosesessssssssssessssssesssassssas iii USSR Participants ....ceeeeeneocecnns Cet ec acacia eec teeta v Papers PART I. Regulation of Contraction and Energy Metabolism of the Heart CELL SURFACE CALCIUM: ITS ROLE IN CONTROL OF MYOCARDIAL CONTRACTION G. A. Langer, K. D. Philipson, and D. M. Bers .....cccveeeeenn 1 MEASUREMENT OF Catt CONCENTRATIONS IN CONTRACTING MUSCLES John R. Blinks, W. Gil Wier, and Kenneth W. Snowdowne ........ 13 QUANTITATIVE ESTIMATE OF THE CALCIUM-TRANSPORTING CAPACITY OF THE SARCOPLASMIC RETICULUM OF THE HEART D. 0. Levitsky, D. S. Benevolensky, and T. S. Levchenko ...... 27 RESOLUTION AND RECONSTITUTION APPROACH TO THE STUDY OF THREE TYPES OF CALCIUM PUMPS FROM MAMMALIAN TISSUES Sidney Fleischer, Paul DeFoor, Brian Chamberlain, Dmitri Levitsky, Klaus Gietzen, and H. Uwe Wolf ..........cvvvuninnnn. 41 REGULATION OF PHOSPHORYLCREATINE SYNTHESIS IN MYOCARDIAL CELLS: THE COUPLING OF TRANSPHOSPHORYLATION TO GLYCOLYSIS AND MITOCHON- DRIAL OXIDATIVE PHOSPHORYLATION AND THE SIGNIFICANCE OF CREATINE KINASE COMPARTMENTATION V. A. Saks, V. V. Kupriyanov, G. V. Elizarova, E. K. Seppet, and W. E. JACObUS +t. iiiiiiieieeeeeeeeeenoeeosonseasssosannnns 65 REGULATION OF MECHANICAL AND ELECTRICAL ACTIVITY OF THE HYPODYNAMIC MYOCARDIUM BY CREATINE PHOSPHATE L. V. Rozenshtraukh, V. A. Saks, V. M. Sharov, I. A. Yuryavichyus, and E. IT Chazov .....ceuiieeneennreneoneananannns 89 xi POSSIBILITY OF PARTICIPATION OF CREATINE KINASE IN THE REGULATION OF CELL METABOLISM : S. N. Lyzlova, V. E. Stefanov, and N. Taame ......... Ceeeaes .e 113 THE FUNCTIONAL ROLE OF STRUCTURAL NONEQUIVALENCE OF F-ACTIN SUBUNITS OF MUSCLE THIN FILAMENTS Ve V. LeANOV ttt ittteeneeeeeeeensoaseseseasossssssssassnnss eee 125 EFFECT OF PHALLOIDIN ON THE UNSTEADY KINETICS OF MUSCLE CONTRACTION A. E. Bukatina and V. N. MOTOZOV +.vueeeeeennnaannns Cetin .. 137 SIMILARITY OF ELEMENTARY STAGES OF ATP HYDROLYSIS IN VARIOUS ATPase SYSTEMS N. S. Panteléeva, E. A. Karandashov, I. E. Krasovskaya, N. V. Kuleva, E. G. Skvortsevich, and L. A. Syrtsova ...... ee. 147 PART II. Hormones, Prostaglandins, and Cyclic Nucleotides in the Heart RESOLUTION, CHARACTERIZATION, AND PARTIAL PURIFICATION OF COMPONENTS OF CATECHOLAMINE-SENSITIVE ADENYLATE CYCLASE Alfred G. Gilman, Paul C. Sternweis, Allyn C. Howlett, John K. Northup, and Murray Smigel .......ccceeveeeeennnn een 157 THE STRUCTURE OF ACTIVE SITES OF ADENYLATE CYCLASE, PROTEIN KINASE, AND PHOSPHODIESTERASE E. S. Severin, T. V. Bulargina, N. N. Gulyaev, M. N. Kochetkova, and V. L. Tunitskaya ......ceeeeeeeeneanenns 173 ADP-RIBOSYLATION AND ACTIVATION OF ADENYLATE CYCLASE Martha Vaughan and Joel MOSS ...veeeeeeenens Cet er ects ee en 187 INTERACTION BETWEEN RB-ADRENERGIC RECEPTORS AND ADENYLATE CYCLASE IN THE HEART V. A. Tkachuk and S. E. Severin ........... Cette eee 201 PROSTAGLANDIN SYNTHESIS AND REGULATION OF VASCULAR TONE AND PLATELET FUNCTION Philip Needleman .......... Cette, ieee eee. 215 PROSTAGLANDINS AND CYCLIC NUCLEOTIDES AS POSSIBLE REGULATORS OF HEART ADAPTATION TO ACUTE AND CHRONIC PRESSURE OVERLOAD V. D. Pomoinetsky, N. G. Geling, A. A. Nekrasova, Ts. R. Orlova, N. M. Cherpachenko, S. A. Kudrjashov, and D. S. BenevolensKy .uo.ueeeeeeeeeeenenennneeennenns ceeeee.. 227 GLUCOCORTICOIDS IN THE HORMONAL REGULATION OF CARDIAC METABOLISM Yu. M. Seleznev, S. M. Danilov, N. G. Volkova, G. V. Kolpakova, and A. V. Martynov ......c.ceeeeeees Chee 243 xii STUDY OF THE INTERACTION OF CARDIOSTEROIDS WITH Na¥t, K+-ADENOSINE TRIPHOSPHATASE BY EPR R. I. Zhdanov, N. M. Mirsalikhova, and Yu. Sh. Moshkovsky .... 261 SOME FEATURES OF THE INHIBITION OF Nat, K+-ATPase IN HEART MUSCLE BY CARDIOTONIC GLYCOSIDES N. M. Mirsalikhova, N. Sh. Palyants, and N. K. Abubakirov .... 269 ACTIVATION OF LIPID PEROXIDATION AS THE DECISIVE LINK IN THE PATHOGENESIS OF STRESS DAMAGE TO THE HEART, AND PREVENTION OF STRESS AND HYPOXIC DAMAGE BY THE ANTIOXIDANT IONOL F. Z. Meerson, L. Yu. Golubeva, V. E. Kagan, M. V. Shimakovich, and A. A. Ugolev ....iiiiiiiiiiienieieneenanannns 277 RELATIONSHIP OF THE GENERATION AND DETOXIFICATION PROCESS OF LIPID PEROXIDES IN THE NORMAL AND HYPERTROPHIC HEART A. A. Kubatiev and S. V. ANdreyev .......eeeeeeeeeeecncnnenens 293 PART III. Protein Biosynthesis and Pathology of the Heart BRANCHED-CHAIN AMINO ACIDS AND THE REGULATION OF PROTEIN TURNOVER IN RAT HEART Balvin Chua, Daniel L. Siehl, Ellen O. Fuller, and HOWard E. MOTZAM tev veenenennenneeeeseeeessesaseaseessessssnsns 305 CHANGE IN ANTIGENIC PROPERTIES OF RAT MYOCARDIAL CHROMATIN UPON ADAPTATION TO HYPOXIA N. A. Fedorova, I. V. Muzurov, and L. B. Nesterchuk .......... 325 STUDY OF NITROGEN METABOLISM IN THE CARDIAC MUSCLE USING THE ISOTOPE 15N 0. I. Pisarenko, A. V. Artemov, and V. N. Smirnov ............ 329 HIGH ENERGY PHOSPHATE AND CELL VOLUME CONTROL IN LETHAL ISCHEMIC INJURY Robert B. Jennings, Hal K. Hawkins, James E. Lowe, Mary L. Hill, and Keith A. Reimer ..........euveeennneennennann 351 COMPARATIVE STUDY OF CARDIOMYOCYTE MEMBRANE PERMEABILITY DEFECTS IN SEVERE MYOCARDIAL ISCHEMIA AND UPON EXPOSURE TO ISOPROTERENOL USING COLLOIDAL LANTHANUM V. G. Sharov, R. B. Jennings, H. K. Hawkins, Yu. M. Seleznev, and A. V. MAaTtyNOV «vues eueneeeeenennneeeennnnseeeeennnneennns 373 BIOCHEMICAL EVALUATION OF DAMAGE TO THE HUMAN MYOCARDIUM WITH COMPLETE ISCHEMIA DURING OPEN-HEART SURGERY V. I. Burakovsky, D. B. Saprygin, and L. S. Kashtelian ....... 391 xiii POSSIBILITY OF USING LIPOSOMES FOR TARGETING OF DRUGS IN THE TREATMENT OF CARDIOVASCULAR DISEASES V. P. Torchilin, V. R. Berdichevsky, Ban-An Khaw, V. M. Zemskov, E. Haber, V. N. Smirnov, and E. I. Chazov ..... 403 MORPHOFUNCTIONAL DESCRIPTION OF THE EFFECT OF CYTOCHROME C ON VARIOUS ZONES OF THE MYOCARDIUM IN EXPERIMENTAL INFARCTION K. A. Zufarov, R. A. Katsenovich, Sh. B. Irgashev, M. F. Khudayberdyeva, and K. N. AzizoV ...cevierennnnnnnnn ee... 415 COMPARISON OF THE EFFECTS OF B-BLOCKERS AND B-STIMULATORS ON MYOCARDIAL FUNCTION B. I. Tkachenko, R. A. Katsenovich, S. Z. Kostko, Kh. A. Khashimov, A. Sh. Kasymkhodzhaev, and Z. Z. Iunusov .......... 427 xiv PART I REGULATION OF CONTRACTION AND ENERGY METABOLISM OF THE HEART CELL SURFACE CALCIUM: ITS ROLE IN CONTROL OF MYOCARDIAL CONTRACTION G. A. Langer, K. D. Philipson, and D. M. Bers SUMMARY The surface of the myocardial cell is composed of a lipid bilayer covered by a glycocalyx composed of mucopolysaccharide, glycoprotein, and glycolipid. The glycocalyx plus phospholipids of the bilayer contain large numbers of fixed negative charges and are sites of calcium (Ca) binding. Studies which compare the amount of Ca bound to the sarcolemmal-glycocalyx complex with force develop- ment support the proposal that this bound Ca controls the amount available to the contractile proteins and thus controls contractility. There are two classes of binding sites, one of high affinity and one of low affinity. The high- affinity sites are relatively few in number and may function in maintenance of the structural integrity of the membrane. The more abundant low-affinity sites are proposed to bind the Ca which modulates force development. Entry of this Ca is via a "pore" or 'channel" and via a "carrier' system. The latter couples the outward movement of sodium (Na) to the inward movement of Ca. The carrier system is activated by increased [Na]j, and interventions which produce this in- crease are found to augment Ca influx and produce a positive inotropic result. INTRODUCTION AND BACKGROUND The Cardiovascular Research Laboratories at UCLA are currently involved in investigation of the location and movement of the Ca which controls force develop- ment in the heart. This work utilizes a variety of preparations including the arterially perfused interventricular septum (1), standard bath perfusion of iso- lated papillary muscles, myocardial tissue culture (2), isolated sarcolemma pre- pared by sucrose gradient technique (3), and sarcolemma prepared by 'high velocity gas dissection" from cells in culture (4). This combination of techniques is designed to relate the isolated membrane and its components to the function of the whole, intact tissue. The present paper summarizes what has been learned recently from these techniques about the process of excitation-contraction (EC) coupling in the heart. From the Cardiovascular Research Laboratories, University of California Medical Center, Los Angeles, California. This work was supported by grant HL-11351-11 and 12 from the National Institutes of Health and by grants from the Castera and Bear Foundations. It has long been recognized that heart muscle is dependent upon extracellular Ca for maintenance of contraction. Removal of Ca from the perfusion medium re- sults in a rapid fall of contractile force in contrast to skeletal muscle in which the decline occurs over the course of many minutes or hours (5). This simple finding clearly indicates a major difference in the EC coupling process in the two tissues and points to the fundamental importance of a rapidly ex- changeable fraction of Ca in the heart. It seems likely that a rapidly ex- changeable fraction might be located at the cellular surface, and structural analysis suggested certain structures as good candidates for Ca binding. Most people think of the cell membrane as the 7.5 nm double track or tri- laminar structure that surrounds the cell. In fact, this structure which we call the "unit membrane' represents only one component of the total sarcolemmal mem- brane complex. Closer examination of even routinely stained electron-micro- scopic sections discloses a fuzzy coat or layer about 50 nm thick that is external to this unit membrane (2,6). This surface layer has two components: an inner, less dense component that is 20 nm thick, and an outer, slightly more dense component that is 30 nm thick. The entire coat has been referred to as the basement membrane, basal lamina, external lamina, surface coat, boundary layer, or Bennett's term, glycocalyx (7). These terms do not indicate that the coat has two structural components and, because each of these components may have a distinct function, they should be referred to separately (8). Thus, we refer to the inner coat as the surface coat, and to the outer layer as the ex- ternal lamina. When referring to the entire complex, Bennett's original term, glycocalyx, is used. The glycocalyx is applied to the external surface of the cell, but in the myocardium the glycocalyx follows the unit membrane as it invaginates to form the transverse tubular (T) system (9,10). As Fawcett and McNutt (11) point out, presence of the glycocalyx is used in identification of T tubules deep within the myocardial cell. Whatever the function of the two coats may be, it is probably safe to assume that this function applies to the T tubule system as well as to the external cellular surface. The glycocalyx does not extend into and fill the T tubules of skeletal muscle cells (6) but, rather, is re- stricted to the surface. Cardiac muscle has, then, a much greater amount of coat material per cell. This material is present on the entire electrically active external surface of the cells, including the T tubules. The glycocalyx on the cellular surface blends with the ground substance of the interstitial space and abuts the basement membrane of closely apposed capil- laries (9). Frank and Langer (9) showed that as much as 36 percent of the cellu- lar surface was in virtual direct contact with capillaries. This makes it likely that significant capillary-cellular exchange occurs through a region in which the entire interstitial space is occupied solely by capillary basement membrane and the glycocalyx of the heart cell. It is therefore likely that there is essen- tially direct exchange between capillary and cell and that interposed basement membrane-glycocalyx layers play a significant role in controlling this exchange. In areas where there is more interposed interstitial space, exchange would be influenced by binding and passage through this area, but substances would be subject to the same influences of the glycocalyx before entrance to or exit from the cell. The glycocalyx and interstitial ground substance have a high concentration of fixed, negatively charged sites due to various mucopolysaccharides, glyco- proteins, and glycolipids. These sites have an affinity for cations, including Ca. The sarcolemma has another region which probably plays a major role in Ca binding. This is the outer surface of the unit membrane from which extend the hydrophilic, polar heads of the phospholipids which also contain negatively charged sites. A strong indication that the Ca bound to these surface sites (glycocalyx and phospholipid) is crucial to EC coupling is the action of the trivalent rare Earth ion lanthanum (La) (2,8). Lanthanum can be localized ultrastructurally to the glycocalyx and sarcolemma and does not penetrate further into the cell. When added to contracting heart muscle in micromolar concentrations it uncouples ex- citation from contraction. Flux studies show that, coincident with uncoupling, displacement of Ca from the cell occurs, indicating that the Ca localized to the surface sites (to which La binding is limited) participates in the EC coupling process. Recent studies in our laboratory have concentrated on further defini- tion of sarcolemmal Ca binding and its role in the EC coupling process. These studies are summarized below. STUDIES IN FUNCTIONAL TISSUE Philipson and Langer (12) used the perfused rabbit interventricular septal preparation (1) to study the kinetics of changes in contractility when the Ca concentration of the perfusate is either raised or lowered. Figure 1 illus- trates the sequence of dT/dt (first derivative of tension) that follows changes in [Calo in the arterially perfused interventricular septum of the rabbits. In (a) [Cal], is abruptly decreased to 50 uM from 5 mM. The ty of decline in dT/dt is 48 seconds. After 7 minutes at 50 uM [Cal], the perfusion is switched back to 5 mM (see b) and the ty of return of dT/dt is 11 seconds or 4.4 times the rate of decline. In (c) another septum, perfused at [Calg = 1.5 mM, La (200 uM) is added to the perfusate at the point indicated. The ty of decline a) [Col:005mM 1 minute —mm c) on § af ' fd 5. conn g EE — FIGURE 1. (a) At the arrow [Ca], was decreased from 5 mM to 50 uM in arterially perfused rabbit interventricular septum. The ty for decline in dT/dt (g/sec) is 48 seconds. (b) At the arrow [Ca], was increased from 50 uM to 5 mM. The ty for return of dT/dt is 11 seconds or 4.4 times the rate of decline in (a). (c) At the arrow 200 pM [La], is added to a septum perfused with [Calg = 1.5 mM. The ty of dT/dt decline is 5.1 seconds. See text for discussion. (Reprinted with permission of J. Mol. Cell. Cardiol.) in dT/dt is 5.1 seconds. It is notable that the rate of loss of tension on re- moval of Ca is significantly slower than the gain upon replacement and that addition of La, a Ca-displacing agent, produces a much more rapid decline than removal of Ca. These results are consistent with the proposal that coupling Ca is bound with a certain affinity at rapidly exchangeable cellular sites from which it would be expected to dissociate less rapidly than it would rebind upon reexposure to Ca. Lanthanum, an ion which competes with and displaces bound Ca, would be expected to be more rapidly effective than removal of Ca from the per- fusate. The results are not consistent with the idea that the component of Ca under study exists in free solution in the interstitial space. If such were the case, the rates of decline and return of tension would be the same and simply dependent upon the interstitial exchange rate. The role of bound Ca in functional tissue can be further examined if the assumption is made that dT/dt is proportional to the amount of Ca bound at sites which release it to activate contraction. Figure 2 represents the relation be- tween dT/dt or Ca bound and Ca concentration in the perfusate. The relation in figure 2 can be plotted as dT/dt/[Ca], vs. dT/dt which, based on the assumption, is equivalent to Ca bound/free Ca vs. Ca bound. This is a Scatchard (13) plot and is illustrated in figure 3. The relationship is nonlinear and can be re- solved into two components indicating two classes of binding sites: one of high affinity (Kp = 3800 M~1l) and one of lower affinity (Kp = 370 M-1). The relative size of each component is given by the intercept on the abscissa and indicates that the number of low-affinity sites is approximately 10 times the number of high-affinity sites. The foregoing analysis provides some support for the existence of rapidly exchangeable cellular-binding sites for Ca and for the proposal that Ca at these 200 e oo + 100 ~N - © 1 1 ] ] 0 3 6 9 12 (Ca*?], (mM) FIGURE 2. The relation between dT/dt and [Ca], in the arterially perfused interventricular septum. Bars represent * 1 SE of the mean. (Reprinted with permission of J. Mol. Cell. Cardiol.) dT/d¥/[Cal, (g/s/M)x 107% H dT/dt (g/s) FIGURE 3. Scatchard plot of data given in figure 2. dT/dt is assumed to be equivalent to Ca bound at releasing sites. Thus, the plot is in the Scatchard mode, with Ca bound/free Ca vs. Ca bound. Note the two components indicative of two classes of binding sites with Kp = 3,800 M-1 and 370 M~1, respectively. The number of sites associated with each component is indicated by the intercept of each component with the abscissa. Note that the number of low-affinity sites is approximately 10 times that of the high-affinity sites. (Reprinted with per- mission of J. Mol. Cell. Cardiol.) sites is important in regulation of force development by the myocardium. Re- cently completed studies in another type of functional tissue, myocardial tissue culture, gives further evidence for surface binding and a new concept of Ca com- partmentalization (14). In these cultured cells, over 90 percent of the exchangeable Ca exchanges as rapidly as the system can be perfused and virtually all of this Ca is dis- placeable by La. This rapidly exchangeable component accounts for 42 percent of the total Ca in the cells. The remainder of the Ca exchanges very slowly-- i.e., no detectable exchange is observed over the course of 1 to 2 hours. These cells increased their exchangeable Ca by 65 percent when the ambient temperature was increased from 24° C to 35° C which also increased the frequency of contrac- tion from 85 bpm to 145 bpm. All of this additional Ca was found in the rapidly exchangeable component and was displaceable by La. It was, therefore, distrib- uted at the cellular surface. Kinetic data indicate that the Ca which is neces- sary for contraction enters the cell but does not exchange with intracellular sites, or exchanges extremely slowly relative to the surface sites. Either of two models is consistent with these results in the cultured cells: (a) Ca is released to the contractile proteins from the sarcolemmal-glycocalyx complex, activates contraction, and is rebound to these surface structures. (b) A smaller amount of Ca is released from the sarcolemmal-glycocalyx complex which triggers release (15) of Ca from subsarcolemmal sarcotubular cisternae. This Ca then activates the myofilaments and is rebound at the surface or within the cisternae. It is presently impossible to determine which of these models is operative for these cells. It should be pointed out, however, that if surface Ca triggers release from an internal store, then the function of surface Ca and internally stored Ca are indistinguishable. This derives from the fact that the time re- quired to return to full contraction amplitude (< 5 seconds) after exposure to 0 [Ca], is independent of the time the cells were exposed to 0 [Ca],. This sug- gests that surface and stored Ca are replenished at the same rate. The kinetics of the two components are also indistinguishable under the present conditions of perfusion. All of the exchangeable cellular Ca exchanges as rapidly as the perfusate reaches the cells. There is no evidence for a more slowly exchangeable component of significant size, and this suggests that surface and exchangeable internal stores (if present) exchange at a similar rate. The results obtained from functional tissue need to be correlated with re- sults obtained from isolated membranes with respect to Ca-binding characteristics. STUDIES OF ISOLATED MEMBRANE Isolated sarcolemma is prepared from rabbit ventricular muscle by the method of Bers (3). This procedure results in a preparation which is tenfold to fifteen- fold enriched (as compared to homogenate) in the sarcolemmal markers Kt-dependent p-nitrophenyl phosphatase, Na,K-ATPase, and sialic acid. Contamination by mitochondria (measured by succinate dehydrogenase activity) and sarcoplasmic reticulum (measured by Ca + MgATPase activity) is minimal. The technique uses low KCl extraction (0.3 M) and sucrose gradient centrifugation. The sarcolemmal membranes are harvested at a density of approximately 1.11 g/ml (27 percent w/w sucrose). They are then centrifuged at 180,000 x g for 90 minutes and the pellets are resuspended in 5 mM Tris (pH 7.6 at 20° C) for the experiments. Calcium ion binding studies, using the technique of Millipore filtration, were done (16). Figure 4 shows the Ca bound to the sarcolemma, or dT/dt versus [Ca]y,, in the presence of 140 mM Na (the variation of external sodium [Na], is discussed below). The dT/dt vs. [Ca], relation is taken from figure 2. Note that the two curves are essentially superimposable indicating an excellent cor- relation between sarcolemmal-bound Ca and contractility of the whole myocardial tissue from rabbit ventricle. As would be expected, a Scatchard plot of Ca bound to sarcolemma/[Ca]o vs. Ca bound is virtually identical to the plot of figure 3, indicating two classes of binding sites. The maximum number of sarcolemmal Ca- binding sites is about 270 nmoles/mg protein. Knowing the number of mg protein/g wet weight (~ 120) and the purification factor of the sarcolemmal fraction (~ 15), an estimate of sarcolemmal binding in whole tissue can be made. This value, 2.2 mmoles of bound Ca/kg wet weight, is 20 times more than the amount required to saturate the myofilament sites and produce maximal contractile force (17). Another series of experiments was done to investigate the relationship be- tween sarcolemmal-bound Ca and contractile force. It has been recognized for many years that the level of [Na], has a major influence on force development 200r s > oO a g ~ 1,3 0 oO 3100+ ~ E 5 E E © & oO oO 9.0 [Ca*] (mm) FIGURE 4. Comparison of Ca bound to sarcolemmal fraction from rabbit ventricle (left ordinate, e) to dT/dt (right ordinate, o) relative to perfusate Ca concen- tration [Cat]. The dT/dt vs. Ca data are taken from figure 2. The incubation medium for sarcolemma and perfusate contained 140 mM Na. Note that the curves are virtually superimposable indicating an excellent correlation between sarco- lemmal-bound Ca and contractility of the functional ventricle in the rabbit. (Reprinted with permission of Am. J. Physiol.) (18). A recent study by Tillisch et al. (19) carefully evaluated the contractile response of rabbit ventricular tissue to perfusion with [Na], between 75 and 200 mM. An initial force transient immediately following the [Na], alteration most likely is based upon the response of sarcolemmal-bound Ca to competition for binding sites by Na. Therefore, the effect of different [Na], on Ca bound to isolated sarcolemma was compared to the effect of different [Na], on the force transient of rabbit ventricle. Figure 5 is a plot of the relationship with 100 percent taken as the amount of Ca bound or the magnitude of force developed at [Na]o = 142 mM. Again, the curves are virtually superimposable. This provides further support for the proposal that sarcolemmal-bound Ca plays a critical role in the control of myocardial contractility in the mammalian heart. A third set of studies (3) directed toward investigation of the role of sarcolemmal Ca has recently been completed. This series correlates the effects of cationic EC uncouplers on functional cardiac muscle with Ca bound to the isolated membranes derived from the same tissue (neonatal rat ventricle). Again, Ca-binding studies using Scatchard analysis demonstrated two classes of sites with characteristics similar to those found for the rabbit ventricle. The ability of a series of divalent and trivalent cations (Mg2t, Cot, cd2+, La3+, Nd3t, and Y3t) to displace Ca from the membranes was measured and compared to the ability of these same cations to EC uncouple in the intact ventricle. The ability to displace and to uncouple was in exactly the same sequence for the divalent and trivalent cations. The most potent displacers and uncouplers were the cations with a nonhydrated ionic radius closest to that of Ca (0.99 A). Therefore, the results show that the selectivity of sarcolemmal Ca-binding sites is the same 125} i & oof a g Q 78} tp 7 142 260 Na* (mM) FIGURE 5. Comparison of Ca bound to sarcolemmal fraction from rabbit ventricle at various Na concentrations (e) to the magnitude of peak transient change in tension which occurs upon changing [Na], in the rabbit papillary muscle (o) [data are taken from Tillisch et al. (19)]. The data are expressed as percent of control values obtained in 142 mM [Na],. The curves are virtually superim- posable indicating that the effect of Na on Ca bound at the sarcolemma is a major factor in the control of force development. (Reprinted with permission of Am. J. Physiol.) as the effective uncoupling sequence. In addition, it was again determined that the amount of Ca bound was many times the Ca required to support force development. These conclusions were again consistent with the proposal that Ca bound to the sarcolemma controls the amount of Ca available to the contractile proteins and thus controls contractility. In summary, there is a striking positive correlation between nonenergy- dependent sarcolemmal Ca binding and the level of contractile force in the mammalian myocardium. It remains to be shown which components of the sarco- lemmal complex are responsible for the binding. Preliminary studies suggest that phospholipids may be of predominant importance. CURRENT MODEL FOR EC COUPLING The preceding studies provide strong evidence that the Ca bound to the sarcolemmal-glycocalyx complex is crucial to the ionic control of myocardial contractility. This work, along with other studies over the past 10-15 years, leads to a conceptual model of Ca movement during the course of the EC coupling process. A schematic diagram of this model is shown in figure 6. The immediate source of the Ca which activates and controls contractile force is Ca bound to surface sites, including the glycocalyx (external lamina 8 INTERSTITIUM INTRACELLULAR _ —_— INTEGRAL PROTEINS PORE Lp No‘ 2Na* | | | canner N : Cot t—ro —> Co**—> —_— EXTERNAL SURFACE LAMINA COAT LIPID UNIT MEMBRANE FIGURE 6. Model for transmembranous Ca movement. Ca is bound to negatively charged sites in the external lamina and surface coat. These sites are in rapid equilibrium with Ca in the interstitial space. They supply the Ca that moves across the sarcolemma via two routes: (a) Through a pore system formed by the integral proteins of the bilayer. This movement would be electrogenic. (b) With a carrier (coupled to outward Na movement) such that movement via this system would be electroneutral. See text for further description. [Reprinted with permission from University Park Press (10).] plus surface coat) and the hydrophilic ends of the bilayer phospholipids (not shown in figure 6). The Ca bound to these structures is in rapid equilibrium with vascular and interstitial Ca. It supplies at least two systems which span the glycocalyx-lipid bilayer region. The system designated "PORE" is modeled on the basis of the fluid-mosaic concept of membrane structure (20) in which integral proteins ''float' in the lipid bilayer and abut to form hydrophilic channels or pores. These pores are visualized to function as selective channels for the electrogenic movement of ions, including Ca. It is the ionic movement through this system which is recorded with voltage clamp technique. The maximum amount of Ca which moves inward through this system with each excitation is 5-10 umoles/kg wet weight (21). This, by itself, is barely sufficient to raise intra- cellular free Ca concentration to mechanical threshold. The other system, designated '"CARRIER,'" is based on data derived from experiments in which Na-Ca interaction was evaluated (19,22-24). Movement of the carrier results in net transport of Na outward and Ca inward without net charge transfer; therefore, it is not electrogenic. It is simplest to think of carrier activation in terms of recruitment of more molecules with the charac- teristics of Na and Ca transport indicated. Since, under physiological condi- tions, [Ca], and [Na], are quite invariant and since [Ca]; is under control of the sarcotubular Ca pump, the level of [Na]j is probably the dominant signal in determination of the activity of the carrier system. An intervention which results in an elevation of [Na]; (sudden increase in excitation rate, digitalis administration) will stimulate the carrier to move Na outward and Ca inward (figure 6). The Ca-inward movement is, then, responsible for positive inotropy which follows upon an intervention which raises [Na]j. The operation of the carrier with its "uphill" transport of Na requires energy. The energy requirements are most likely linked to the large outside- to-inside concentration gradient for Ca that exists across the cell membrane. This gradient is approximately 10% during diastole and 2 x 102 at the peak of contraction. The movement of Ca inward down its concentration gradient could then be tied, in terms of energy, to the movement of Na outward, against its gradient. Whether the Ca derived from surface sites influxes directly to the myofila- ments or whether a subthreshold amount is directed to the sarcoplasmic reticulum from which it triggers release of a larger amount is a point of controversy (25,26) which remains unresolved. There is agreement, however, that the amount of Ca which enters the cell upon excitation grades the contractile response. Thus, whether by directly activating or by grading sarcotubular release, surface- bound Ca is a critical factor in ionic control of myocardial contractility. 10 10. 11. 12. 13. 14. 15. REFERENCES Rau EE, Shine KI, Langer GA: Potassium exchange and mechanical performance in anoxic mammalian myocardium. Am J Physiol 232:H85-H94, 1977 Langer GA, Frank JS: Lanthanum in heart cell culture: Effect on calcium exchange correlated with its localization. J Cell Biol 54:441-445, 1972 Bers DM, Langer GA: Uncoupling cation effects on cardiac contractility and sarcolemmal Cat binding. Am J Physiol, submitted for publication Langer GA, Frank JS, Philipson KD: Preparation of sarcolemmal membrane from myocardial tissue culture monolayer by 'high velocity gas dissection." Science 200:1388-1391, 1978 Rich TL, Langer GA: A comparison of excitation-contraction coupling in heart and skeletal muscle. J Mol Cell Cardiol 7:747-765, 1975 McNutt NS, Fawcett DW: Myocardial ultrastructure. In The Mammalian Myo- cardium, edited by GA Langer and AJ Brady. New York, Wiley, 1974, pp 1-49 Bennett HS: Morphological aspects of extracellular polysaccharides. J Histochem Cytochem 11:14-23, 1963 Frank JS, Langer GA, Nudd LM, Seraydarian K: The myocardial cell surface, its histochemistry and the effect of sialic acid and calcium removal on its structure and cellular ionic exchange. Circ Res 41:702-714, 1977 Frank JS, Langer GA: The myocardial interstitium: Its structure and its role in ionic exchange. J Cell Biol 60:586-601, 1974 Langer GA, Frank JS, Brady AJ: The myocardium. In Cardiovascular Physiology II: International Review of Physiology, vol 9, edited by AC Guyton and AW Cowley. Baltimore, University Park Press, 1976, pp 191- 237 Fawcett DW, McNutt NS: The ultrastructure of the cat myocardium. I. Ventricular papillary muscle. J Cell Biol 60:586-601, 1974 Philipson KD, Langer GA: Sarcolemmal bound calcium and contractility in the mammalian myocardium. J Mol Cell Cardiol, in press Scatchard G: The attraction of proteins for small molecules and ions. Ann NY Acad Sci 51:660-672, 1949 Langer GA, Frank JS, Nudd LM: Correlation of calcium exchange, structure and function in myocardial tissue culture. Am J Physiol, submitted for publication Fabiato A, Fabiato F: Contractions induced by a calcium-triggered release of calcium from the sarcoplasmic reticulum of single skinned cardiac cells. J Physiol (London) 249:469-495, 1975 11 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Philipson KD, Bers DM, Nishimoto AY, Langer GA: The binding of calcium and sodium to sarcolemmal membranes: Relation to control of myocardial contractility. Am J Physiol, submitted for publication Solaro RJ, Wise RM, Shiner JS, Briggs FN: Calcium requirement for cardiac myofibrillar activation. Circ Res 34:525-530, 1974 Daly IdeB, Clark AJ: The action of ions upon the frog's heart. J Physiol (London) 54:367-383, 1921 Tillisch JH, Fung LK, Hom PM, Langer GA: Transient and steady-state effects of sodium and calcium on myocardial contractile response. J Mol Cell Cardiol 11:137-148, 1979 Singer SJ, Nicolson GL: The fluid mosaic model of the structure of cell membranes. Science 175:720-731, 1972 New W, Trautwein W: Inward currents in mammalian myocardium. Pfluegers Arch 334:1-23, 1972 Littgau HC, Niedergerke R: The antagonism between Ca and Na ions on the frog's heart. J Physiol (London) 143:486-505, 1958 Niedergerke R: Movements of Ca in frog ventricles at rest and during con- tractions. J Physiol (London) 167:515-550, 1963 Tillisch JH, Langer GA: Myocardial mechanical responses and ionic exchange in elevated sodium perfusate. Circ Res 34:40-50, 1974 Fabiato A, Fabiato F: Calcium-induced release of calcium from the sarco- plasmic reticulum of skinned cells from adult human, dog, cat, rabbit, rat and frog hearts and from fetal and new-born rat ventricles. Ann NY Acad Sci 307:491-522, 1978 Langer GA: The structure and function of the myocardial cell surface. Am J Physiol 235:H461-H468, 1978 12 MEASUREMENT OF Ca’™ CONCENTRATIONS IN CONTRACTING MUSCLES John R. Blinks, W. Gil Wier, and Kenneth W. Snowdowne SUMMARY The calcium-sensitive bioluminescent protein aequorin was used to measure intracellular free calcium ion concentrations ([Catt]) in isolated frog skeletal muscle fibers and in Purkyné fibers from the right ventricles of dog hearts. In frog skeletal muscle the mean resting [Catt] was estimated to be slightly under 10-7 M; the mean level achieved during a tetanus was approximately 10-5 M. Calcium transients in cardiac muscle are slower than those in skeletal muscle, and do not rise as high. Those in Purkynd fibers have two clearly defined com- ponents which we interpret as reflecting the contributions of calcium entry and calcium release. Both components are subject to modification by inotropic inter- ventions, but the second (release) is far more sensitive than the first (entry). Changes in fiber length have very little influence on the amplitude of the aequorin signal in the Purkyné fiber. INTRODUCTION It is now generally accepted that rapid changes in intracellular [cath] (calcium transients) play a central role in the regulation of contractile activ- ity of muscle. Nevertheless, there is still a great deal of uncertainty about the mechanisms responsible for moment-to-moment control of [Catt] in various types of muscle and about the details of the relation between intracellular [Catt] and mechanical activity under various circumstances. Clearly, a detailed understanding of the role of calcium transients in the regulation of muscle function will depend ultimately on the ability to measure those transients, and to relate them to the mechanical behavior of muscle. Ideally, it should be possible not only to measure changes in calcium concentration from moment to moment within a single contraction, but also from point to point within a single cell. The most promising tool so far available for obtaining information of this sort is the calcium-sensitive bioluminescent protein aequorin. Aequorin, which is extracted from the luminescent hydromedusan Aequorea forsk8lea, was first isolated and characterized by Shimomura, Johnson, and From the Department of Pharmacology, Mayo Foundation, Rochester, Minnesota. This work was supported by grants HL-12186, HL-05600, and HL-07111 from the National Institutes of Health. 13 Saiga in 1962 (1). They found that aequorin was different from all previously known bioluminescent systems in that only a single organic molecular species-- a protein of 21,000 M.W. (2)--was involved in the luminescent reaction and that light emission was not influenced by the availability of free Oj. Moreover, the rate of the luminescent reaction was highly sensitive to the free Catt con- centration, and these workers suggested as early as 1963 (3) that aequorin might prove useful as a biological Catt indicator. Since that time aequorin has been used successfully as an intracellular Catt indicator in at least 25 different kinds of cells (4), including vertebrate striated muscle. This paper provides an overview of results obtained with single frog skeletal muscle fibers and mammalian cardiac muscle. MATERIALS AND METHODS Aequorin was microinjected into single living twitch muscle fibers dissected from the tibialis anterior and semitendinosus muscles of frogs (Rana pipiens or Rana temporaria), and into multicellular strands isolated from the Purkyn¥é system of the right ventricles of dogs. The frog muscle fibers were maintained (at 21° C unless stated otherwise) in amphibian Ringer's solution containing (mM) NaCl 115, KCl 2.5, CaClp 1.8, NajpHPO4 2.15, NaHpPO4 0.85, MgCly 2.0. This solu- tion was used in equilibrium with room air and had a pH of 7.2. The Purkyne fibers were studied at 35° C in a solution containing (mM) NaCl 123, KCl 5.4, CaClp 2.7, NaH2PO4 0.42, NaHCO3 23.8, MgCl; 1.0, glucose 5.5. This solution was equilibrated with 95 percent 02 and 5 percent COp and had a pH of 7.4. Aequorin was extracted, purified, and prepared for injection according to techniques that have been described elsewhere in detail (5). It was dissolved to a concentration of from 1 to 5 mg/ml in a solution containing 150 mM KCl and 5 mM HEPES buffer (n-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid), pH 8. This solution was loaded into fine-tipped micropipettes drawn from glass capil- lary tubing with a filament of the same glass fused to the inside to facilitate filling of the tip. The potential between the electrode and the bathing medium was monitored to determine when a cell had been penetrated. Aequorin was in- jected into the cell by applying gas pressure (up to 10 atmospheres) to the pipette. Light emission was recorded with an EMI 9635B photomultiplier; tension with an AME model AE strain gauge transducer. Methods have been described elsewhere in greater detail (5-9). RESULTS AND DISCUSSION Frog Skeletal Muscle When an isolated frog skeletal muscle fiber has been injected with aequorin, it is possible to measure a very low level of luminescence from the fiber at rest. Photomultipliers of the highest sensitivity are required for this pur- pose, and considerable care must be devoted to construction of the apparatus in order to optimize the optical geometry and to minimize the generation of light from extraneous sources. 14 In order to translate the measured light emission from the muscle fiber in- to [cath], we used calibration curves such as the one illustrated in figure 1 (determined in vitro). In this curve, logarithmic scales are used on both axes. The free [Catt], established both by simple dilutions of CaCly and by calcium- EGTA buffers (EGTA = ethyleneglycol-bis(B-aminoethylether)-N,N'-tetraacetic acid), is plotted on the horizontal axis. The peak light intensities recorded when uni- form amounts of aequorin are mixed with solutions of various [Catt] are plotted on the vertical axis. All values are expressed as fractions (L/Lpax) of the maximum intensity obtained in high (saturating) [Catt]. In this way, all measurements of light intensity are normalized for the amount of aequorin in the system. It may be noted that the calibration curve is very steep in its midportion (maximum slope = 2.5), and that at very low [catt] the curve becomes Or 150mm kei 2mM Mg" -1F pH 7.0 21°C -2r Skeletal Muscle Tetanus | Lr | lo | 9 L max Cardiac Muscle Twitch | -4 hy - = | -5 + | | -6 - | Resting Skeletal Muscle : ! | “7k yy — 1 | il 1 ] ) EGTA -8 -7 -6 -5 -4 -3 -2 log [Ca**] (M) FIGURE 1. Aequorin calibration curve determined in vitro. Double logarithmic plot. Points represent experimental measurements of peak light intensity when standard aliquots of aequorin were injected into solutions of known free [Catt]. All measurements are expressed as fractions of Lpgx, that is, the peak light intensity obtained in solutions of saturating [Catt] (fractional luminescence). ® = Points for which [Catt] was established by simple dilution. O = Points for which calcium-EGTA buffers were used. All solutions contained 150 mM KCl and 2 mM free Mgtt, and were buffered to pH 7.0 with 5 mM PIPES [piperazine- N,N'-bis(ethanesulfonic acid)]. The total concentration of EGTA in the calcium- buffered solutions was 1 mM. The point designated EGTA contained 1 mM EGTA and no added calcium. =--- = Estimates of fractional luminescence determined in aequorin-injected muscle cells (see text for details). Adapted from Allen and Blinks (9). For further details on methodology see Allen et al. (10). 15 horizontal--that is, the light level is independent of [Catt] (10). It should also be pointed out that the position of the curve is rather sensitive to ionic strength and [Mgtt] (5); the curve in figure 1 was determined under conditions assumed to be appropriate to the sarcoplasm of frog skeletal muscle. In order to use calibration curves of this sort to estimate [Catt] in aequorin-injected cells, it is necessary to estimate Lpzx, that is, the peak light intensity that would be obtained if all of the aequorin in the cell were exposed instantly to a saturating concentration of Catt. Because we are not able to measure the amount of aequorin actually expressed from the micropipette during the injection process, we estimate Lpgx by rapidly lysing the cell, and measuring the total amount of light emitted as the aequorin in the cell is con- sumed by exposure to Catt (9). The approach that we used in doing this is illustrated in figure 2. When the intensity of light emission from the aequorin-injected cell has been mea- sured both at rest and during activity, the cell is suddenly exposed to a solu- tion containing a detergent, such as Triton X-100. The emitted light, which must be measured at a very much reduced photomultiplier sensitivity, is integrated over the period (usually less than 30 seconds) required for full discharge of the aequorin. Our best estimate of Lpgx is the product of this integral and the rate constant for the consumption of aequorin (determined in vitro) in the presence of a saturating [Catt] at the temperature of the experiment. In essence, this amounts to taking the peak light intensity of an imaginary light record that subtends the same area as the recorded curve, but has the time course of the light signal that is observed when a sample of aequorin is mixed rapidly with a saturating concentration of Catt. (Implicit in this method are the assumptions that the total amount of light emitted by a given aliquot of aequorin is independent of the speed of mixing with Catt, and that the detergent used to lyse the cell does not alter the light yield. Both of these assumptions have been checked experimentally and were found to be valid.) In 12 experiments, the mean value of log(L/Lpax) for apparently healthy frog skeletal muscle fibers at rest (21° C) was -6.5. This is only slightly above the calcium-independent light emission of aequorin in the presence of 150 mM KCl and 2 mM Mgt+ (figure 1). To verify that the light emission from the resting fiber was indeed above the calcium-independent level, it was important to show that it was possible to reduce the glow through interventions that might be expected to reduce intracellular [Catt]. This was in fact accomplished, either by removing the Catt from the bathing medium with EGTA (keeping free Mgt constant at 2 mM) or by replacing the Ca*tt in the Ringer's solution with srt, On the calibration curve of figure 1 the average resting glow corresponds to a free [Catt] of about 8:10-8 M. When aequorin-injected skeletal muscle fibers are stimulated, the intensity of the light they emit rises abruptly by a factor of 10,000 or more. An example of the light signal recorded during a single twitch is illustrated in figure 3. In this figure the photomultiplier output and tension record are superimposed to facilitate comparison of the timing of the two signals. It can be seen that the light signal rises abruptly shortly after the stimulus, and is followed more gradually by the tension. At 21° C peak light intensity is reached con- siderably before the time that the rate of tension development is maximal, and 16 relative light intensity FIGURE 2. Method used to estimate fractional luminescence (L/Lpgx) of muscle cells injected with aequorin. Tracing A shows photomultiplier output recorded from an intact aequorin-injected muscle fiber at rest (upper level), and the background signal recorded from the same fiber after the aequorin had been discharged with Triton X-100 (lower level). Tracing B shows the light output recorded from a representative fiber during a tetanic contraction (50 Hz). Tracing C shows the light signal recorded when 0.5 percent Triton X-100 was flushed into the bath to discharge all of the aequorin in the cell. The dashed line (---) has been drawn to subtend the same area as the recorded curve, and shows the time course of the flash that would have been obtained if all of the aequorin in the cell had been exposed instantly to a saturating [Catt]. The peak of this curve is our best estimate of Lp, for the amount of aequorin in the cell. The three measurements of light intensity must be made under iden- tical optical conditions and at the same temperature as the calibration curve (figure 1). Note the different scales of light intensity and time in tracings A, B, and C. 17 force light 4 uA 2 mN 100 ms FIGURE 3. Luminescent and mechanical responses in an isometric twitch of an aequorin-injected frog skeletal muscle fiber. This was a single fiber dissected from the tibialis anterior muscle of Rana temporaria. The figure shows a rested- state contraction. The temperature was 21° C, and the sarcomere length was 2.4 ym. The fiber was heavily injected at multiple sites. The time of the stimulus is indicated by the vertical line on the lower tracing. the light signal has subsided well toward the resting level by the time peak tension has been attained. Several points of caution apply to the interpretation of such records. First, the ability of the aequorin reaction to follow rapid changes in catt is limited (11). It may be assumed that the rise of light intensity lags somewhat behind [Catt], although the kinetics of the aequorin reaction have not yet been studied in sufficient detail to allow precise deconvolution of the light signal. Second, the steepness of the midportion of the curve relating light in- tensity to [Catt] means that changes in light intensity within that range will give an exaggerated impression of the changes in [Catt] responsible for them. In the midportion of the curve, light intensity changes approximately in propor- tion to [Catt]2.5) and a 10,000-fold change in light intensity corresponds to only about a fortyfold change in [Catt]. Third, the steepness of the calcium concentration-effect curve leads to another problem in interpretation, namely, that light signals will be dominated by contributions from those regions of the cell with the highest [Catt]. Looked at another way, a given amount of Catt will cause a greater amount of light to 18 be emitted if its concentration is high locally than if it is distributed uni- formly throughout the cell. This is likely to be a significant factor in the genesis of light signals during skeletal muscle twitches because substantial gradients of Catt are almost certain to exist during the twitch. Calcium ions are probably released rather abruptly from the terminal cisternae of the sarco- plasmic reticulum; they then must diffuse past a variety of potential sites of loss (e.g., sarcoplasmic reticulum, parvalbumin, mitochondria) before they reach the troponin C in the myofibrils. It seems probable, therefore, that the aequorin signal in a twitch is dominated by a contribution generated by tran- siently high calcium concentrations immediately adjacent to the sites of re- lease. Thus, changes in the amplitude of the signal are probably a good index of changes in the amount of Catt released, but it does not seem justified to attempt to estimate even a mean intracellular [Catt] in twitches by the method applied above to the resting fiber. The situation is more favorable in tetanic contractions, where a steady level of light emission may be maintained for hundreds of milliseconds (figure 4). Under such steady-state circumstances saturable calcium binding sites will come into equilibrium with the cytoplasmic [Catt], and substantial gradients of [catt] are not likely to exist over distances as short as one sarcomere length. Under such conditions it does seem justified to attempt to estimate the free intracellular [Catt]. We have found that during the plateau of a tetanus at 21° C the mean ratio of L/Lpax is -2.2, a value which corresponds on the curve 5 Hz 10 Hz IU — Msn Duss, TTTTTTTITTTITTTTTT 20 Hr ll [o30a “TN ——————— NN e—— 3s FIGURE 4. Trains of twitches and tetani in an aequorin-injected skeletal muscle fiber. The figure shows isometric contractions of fiber from the semitendinosus muscle of Rana pipiens. The fiber was injected at a single point. The tempera- ture was 15° C, and the sarcomere length was 2.3 um. The trains of pulses at the frequencies indicated were initiated after prolonged periods of rest. 19 of figure 1 to a [Catt] of just under 10-5 M. Thus, in a tetanus we conclude that the free [Catt] rises from slightly below 10-7 M to about 10™2 M. This is approximately the range over which changes in Cat influence tension development in skinned skeletal and cardiac muscle fibers (12). Mammalian Cardiac Muscle It is not possible to introduce enough aequorin into a single heart muscle cell to record signals of the sort shown in figure 3. This is in part because cardiac cells are too small, and in part because the calcium transients are evi- dently of considerably lower amplitude in cardiac than in skeletal muscle. [Allen and Blinks (8) found that for a given amount of aequorin injected, the light signals from frog atrium were only about one-fiftieth as bright as from skeletal muscle of the same species. The level attained is indicated in figure 1, although its interpretation is subject to some uncertainty for the same reason as in the skeletal muscle twitch. We have drawn it on the figure because the very much slower time course of the calcium transient in cardiac muscle suggests that gradients of [Ca'tt] must be considerably less steep than in the skeletal muscle twitch.] For these reasons it is usually desirable to inject more than one cell within a preparation of heart muscle, and even then one must average a series of successive signals in order to get a satisfactory signal-to-noise ratio. This approach has been applied to frog atrial muscle in the past (8), and we are now using it to study mammalian myocardium. Although signals can be recorded from working myocardium such as the papil- lary muscle (13,14), we have found Purkyn& cells particularly favorable because their larger cell size reduces the number of injections required from approxi- mately 50 to no more than 10. Purkyné fibers have another useful feature, in that the light responses from them often have two distinct components, which are much more difficult to resolve in the aequorin signals from working myocardium. These two components are clearly visible in the upper tracing of figure 5. The first phase of the aequorin signal reaches its peak at a time when the tension is just beginning to rise. The peak of the second phase usually occurs very close to the time at which the rate of tension development is at its greatest. The amplitude of the second component of the aequorin signal in the Purkyné fiber is highly subject to change, and when it is particularly prominent, it tends to obscure the first component. This is illustrated in figure 6, which shows the influence of frequency of contraction on the aequorin signal and on tension development. At 100-second intervals, the second component so dominates the light signal that the first component is obscured. As the interval between contractions is increased the contribution of the second component is reduced, making the first component clearly visible and then dominant. We have found closely coupled pairs of beats particularly useful in deter- mining the influence of various factors on the two components of the aequorin signal. If the interval between successive pairs of beats is chosen suitably, the first beat of each pair will usually show both components clearly. If the second beat is closely coupled, its light signal will be dominated by the first component (see top tracing, figure 7). We are currently testing the hypothesis that the first component reflects the entry of calcium into the cell, while the second results from the release of Catt from intracellular storage sites. The 20 Light Force 100ms 0.ImN FIGURE 5. Luminescent and mechanical responses from a dog Purkyné strand. The preparation was injected with aequorin at multiple sites; 512 successive iso- metric contractions were averaged. The temperature was 35° C, and the stimulus frequency was 1.0 sec. The arrow indicates the time of stimulus. idea that the first component reflects calcium entry is reinforced by the fact that its amplitude is altered by changes in stimulus interval in much the same way as is the peak amplitude of the slow inward (calcium) current in voltage- clamped sheep Purkyn¥ fibers (15). The lower tracing of figure 7 shows that both the first and the second components of the light signal are increased by a positive inotropic concentra- tion of a cardiotonic steroid. This finding is consistent with the recent demon- stration by Weingart et al. (16) that strophanthidin enhances the slow inward current of calf Purkyne fibers. One might anticipate that if calcium entry were enhanced, releasable stores would also be more nearly filled, and that increased calcium release would result as well. Catecholamines also dramatically enhance both components of the aequorin signal in Purkyne fibers (not shown). This is consistent with their well-established action to enhance the slow inward (cal- cium) current in cardiac muscle (17). Changes in fiber length are in sharp contrast to the inotropic interven- tions just discussed. Stretch of a Purkynd fiber from slack to optimal fiber length produces only minimal changes in the aequorin signal (figure 8). The 21 100ms WSO 100s n 1s 30s 67s 10s 45s FIGURE 6. Influence of the frequency of stimulation on aequorin signals and isometric contractions of a dog Purkyn& strand. The records were made when the signals had achieved a steady state at each frequency of contraction; the inter- vals between contractions (seconds) are indicated by the numbers at the left of each set of tracings. The signals were averaged. The number of contrac- tions averaged ranged from 8 at the lowest two frequencies to 128 at the highest two. The temperature was 35° C. only consistent change in the aequorin signal that we have noted in experiments of this type is a slight abbreviation of the signal with stretch. A similar observation has been made by Allen and Kurihara (13) in rat papillary muscle. These results from mammalian heart muscle are strikingly different from those reported earlier for frog atrial muscle (8), in which the amplitude of the aequorin signal was markedly reduced as the preparation was stretched. CONCLUSIONS Aequorin, microinjected into cells, provides a satisfactory means of detect- ing changes in intracellular [Catt]. The translation of aequorin signals into absolute [Catt] is indirect, and depends somewhat on assumptions about the com- position of the intracellular fluid. Our best estimate of [Catt] in skeletal muscles at rest is 810-8 M; during a tetanus the concentration rises to approxi- mately 10-5 M. Estimates of absolute [Catt] during twitches are uncertain because of the probable existence of gradients of [Catt] within the cell. How- ever, the amplitude of the aequorin signal in skeletal muscle twitches is prob- ably a good indicator of the amount of calcium released from intracellular stores. 22 Control AcS JpM 200ms FIGURE 7. Influence of acetylstrophanthidin on aequorin signals and isometric contractions of a dog Purkyné strand. The preparation was stimulated to give closely coupled (750 msec) pairs of contractions every 5 seconds; 64 successive pairs were averaged for each set of tracings. The upper set was obtained be- fore addition of 10-7 M acetylstrophanthidin, and the lower set was obtained after the addition. The temperature was 35° C. The calcium transients in cardiac muscle are of lower amplitude than those of skeletal muscle. Aequorin signals recorded from dog Purkyn& fibers have two readily distinguishable components which are probably associated with calcium entry and calcium release. The second component (release) is considerably more sensitive to inotropic interventions than the first; neither is influenced greatly by changes in fiber length. Much of the work presented here is an extension of earlier work performed in this laboratory by David G. Allen (8-10). We wish to acknowledge his contributions and also the expert technical assistance of Gary C. Harrer and Norman K. Lee, without whose participation these experiments might not have been possible. 23 Light Force Slack Length 100ms —_— ee T— No I _ IN Optimal Length FIGURE 8. Influence of fiber length on aequorin signals and isometric contrac- tions of a dog Purkyn& strand. Tracings show responses recorded at four fiber lengths ranging from slack length (top) to the length at which maximum tension was developed (bottom). The stimulus interval was 2.5 sec, the temperature was 35° C, and 128 contractions were averaged. 24 10. 11. 12. 13. REFERENCES Shimomura O, Johnson FH, Saiga Y: Extraction, purification and properties of aequorin, a bioluminescent protein from the luminous hydromedusan, Aequorea. J Cell Comp Physiol 5959:223-239, 1962 Prendergast FG, Mann KG: Chemical and physical properties of aequorin and the green fluroescent protein isolated from Aequorea forskdlea. Biochemistry 17:3448-3453, 1978 Shimomura O, Johnson FH, Saiga Y: Microdetermination of calcium by aequorin luminescence. Science 140:1339-1340, 1963 Blinks JR: Applications of calcium-sensitive photoproteins in experimental biology. Photochem Photobiol 27:423-432, 1978 Blinks JR, Mattingly PH, Jewell BR, van Leeuwen M, Harrer GC, Allen DG: Practical aspects of the use of aequorin as a calcium indicator: Assay, preparation, microinjection and interpretation of signals. Methods Enzymol 57:292-328, 1979 Blinks JR, Prendergast FG, Allen DG: Photoproteins as biological calcium indicators. Pharmacol Rev 28:1-93, 1976 Blinks JR, Ridel R, Taylor SR: Calcium transients in isolated amphibian skeletal muscle fibres: Detection with aequorin. J Physiol 277:291-323, 1978 Allen DG, Blinks JR: Calcium transients in aequorin-injected frog cardiac muscle. Nature 273:509-513, 1978 Allen DG, Blinks JR: The interpretation of light signals from aequorin- injected skeletal and cardiac muscle cells--a new method of calibration. In Detection and Measurement of Free Calcium in Cells, edited by CC Ashley and AK Campbell. Amsterdam, Elsevier North Holland, in press, 1979 Allen DG, Blinks JR, Prendergast FG: Aequorin luminescence: Relation of light emission to calcium concentration--A calcium-independent component. Science 196:996-998, 1977 Hastings JW, Mitchell G, Mattingly PH, Blinks JR, van Leeuwen M: Response of aequorin bioluminescence to rapid changes in calcium concentration. Nature 222:1047-1050, 1969 Fabiato A, Fabiato F: Effects of pH on the myofilaments and the sarco- plasmic reticulum of skinned cells from cardiac and skeletal muscles. J Physiol 276:233-255, 1978 Allen DG, Kurihara S: Calcium transients at different muscle lengths in rat ventricular muscle. J Physiol, in press, 1979 25 14. 15. 16. 17. Wier WG: Intracellular calcium transients accompanying contraction of mammalian cardiac muscle. Fed Proc 38:1389, 1979 Gibbons WR, Fozzard HA: Slow inward current and contraction of sheep cardiac Purkinje fibers. J Gen Physiol 65:367-384, 1975 Weingart R, Kass RS, Tsien RW: Is digitalis inotropy associated with enhanced slow inward calcium current? Nature 273:389-392, 1978 Reuter H: Exchange of calcium ions in the mammalian myocardium. Circ Res 34:599-605, 1974 26 QUANTITATIVE ESTIMATE OF THE CALCIUM-TRANSPORTING CAPACITY OF THE SARCOPLASMIC RETICULUM OF THE HEART D. 0. Levitsky, D. S. Benevolensky, and T. S. Levchenko INTRODUCTION Calculations from a number of laboratories (1-6) indicate that in order to provide normal relaxation of the heart the calcium pump concentrated in 1 g of muscle should eliminate calcium from the myoplasm at 50-150 nmol/200 msec. It is assumed that in the hearts of homoiotherms, as in fast skeletal muscles, the membrane network of the sarcoplasmic reticulum plays a primary role in the accumulation of calcium during relaxation (7-8). There are indications as well that isolated mitochondria, which are highly concentrated in the heart, can under certain conditions also accumulate significant quantities of calcium ions (2,9). Participation of mitochondria in the relaxation of heart muscle is, however, disputed since the calcium transport system of the mitochondria has significantly lower affinity for calcium than the calcium uptake system of the sarcoplasmic reticulum (10). Comparative evaluation of the calcium pump systems of the sarcoplasmic reticulum and the mitochondria is difficult because of the low yield of reticu- lar fraction from the heart and the high lability of isolated preparations of heart microsomes. The calcium-transporting capacity of the sarcoplasmic retic- ulum can be determined by complete extraction of reticular membranes from the heart, while preserving the activity of the calcium pump system. In this paper, we estimate the rate of calcium uptake by the total fraction of sarcoplasmic reticulum separated from the heart and determine the quantity of sarcoplasmic reticulum concentrated in 1 g of heart muscle. MATERIALS AND METHODS Microsomes were extracted from pigeon and guinea pig hearts by two methods. In the first method, the heart was washed in distilled water and in cold extrac- tion medium (1-5° C) containing 30 percent glycerol, 10 mM imidazole (pH 7.2), 1 mM dithiothreitol, and 5 mM sodium azide. Fat and large vessels were removed and, after weighing, the specimens were ground in a Virtis-45 homogenizer. Ho- mogenization of 5-20 g of heart muscle was performed at -5° C in 30 ml or 100 ml From the All-Union Cardiology Research Center, USSR Academy of Medical Sci- ences, Moscow, USSR. 27 glass beakers. Homogenization was carried out at 13,500 rpm for 15 seconds and at 22,500 rpm for 10-15 seconds. The homogenate was centrifuged for 20 minutes at 10,800 g. The precipitate was placed in the cold extraction medium, homoge- nized 10-15 seconds at 22,500 rpm, and then the homogenate was once more sub- jected to low-speed centrifugation. To extract the actomyosin, the obtained supernatant was diluted with an equal volume of 1.2 M KCl. The microsomal membrane fractions were precipitated by centrifugation at 115,000 g for 40-60 minutes. The precipitates produced, containing the microsomal membranes, were suspended in 0.05 M KCl, 20 mM Tris-maleate buffer (pH 6.8). For complete ex- traction of the calcium-transporting membranes, the same tissue specimen was homogenized 10-12 times, producing 10-12 portions of microsomal fragments. The second method of obtaining microsomes drew from the method suggested by Harigaya and Schwartz (2). After washing in 0.9 percent NaCl, 17 g of heart muscle were placed in 68 ml of a medium containing 10 mM sodium bicarbonate and 5 mM sodium azide (pH 7.0), 0° C. The tissue was successively homogenized with the Virtis-45 homogenizer at 13,500 rpm for 10 seconds, and the Polytron PCU-2- 110 (PT-10 generator) for 5 seconds. Cell fragments, myofibrils, and mitochon- dria were precipitated by centrifugation at 8,000 g for 20 minutes. The precip- itate was once more sequentially homogenized using the Polytron in the same mode and centrifuged at low speed. The microsomal fraction was separated from the supernatant fluid by the first method described above. In separating the mitochondrial fraction of the homogenate obtained with the Polytron, centrifugation was performed at 310 g for 10 minutes. An equal volume of solution containing 1.2 M KCl and 20 mM of Tris-maleate buffer (pH 6.8) was added to the supernatant fluid. The mitochondria were precipitated by cen- trifugation at 5,000 g for 20 minutes. The sediment obtained was suspended in 50 mM KCl plus 20 mM Tris-maleate buffer (pH 6.8). The purified fraction of sarcoplasmic reticulum from the guinea pig heart was obtained by a method outlined in an earlier publication (11). Phosphorylation of the Ca2t-dependent ATPase was performed at 0° C in 1 ml of a medium containing 0.1 M KCl, 20 mM Tris-HCl (pH 7.1), 10 mM CaClp, 20 uM [y-32p]-ATP. The reaction was initiated by addition of 0.5-1 mg of microsomal protein and, after 15 seconds, was stopped by the addition of 5 ml of a cold solution containing 6 percent trichloroacetic acid, 0.5 mM ATP, and 1 mM H3PO,. The supernatant fluid obtained after centrifugation at 1,000 g for 15 minutes was removed by aspiration. The sediment was washed twice with 15 ml of 6 percent trichloroacetic acid and once with 15 ml of 10 percent sucrose. Most of the sediment was dissolved at 90° C in 1 ml of 2 M NaOH, after which the concentra- tion of protein and level of radioactivity in the specimens were determined. Some of the sediment was dissolved in 2 percent mercaptoethanol and 2 percent sodium dodecylsulfate, and subjected to electrophoresis in polyacrylamide gels (11). The level of phosphorylated protein also was measured by filtration. In this method, 0.07-0.25 mg of microsomal protein were added to 0.5 ml of the reaction medium. The reaction was stopped by an equal volume of a solution con- taining 20 percent trichloroacetic acid, 1 mM ATP, and 2 mM H4PO,, . After this, 0.5 ml of the suspension of denatured protein was filtered under a vacuum through 28 a Whatman GF/C filter. The membranes adsorbed on the filter were washed 6 times with 3 ml of 6 percent trichloroacetic acid. In the control specimens [y-32p]- ATP was added after the reaction was stopped by a solution containing trichloro- acetic acid. The retention of the radioactive label in the filters was measured in a Mark III scintillation spectrometer. The incubation medium for phosphoryla- tion of Nat, Kt-ATPase contained 100 mM NaCl, 20 mM Tris-HCl (pH 7.1), 5 mM MgCly;, and 2 mM EGTA. Calcium uptake in the presence of potassium oxalate was measured using a pH-metric technique (11). The incubation medium contained 0.1 M KCl, 5 mM ATP, 6 mM MgCl, 5 mM sodium azide, 6.5 mM potassium oxalate, 10 mM Tris-HCl buffer (pH 6.9), and 0.1-0.5 mg microsomal protein. The reaction was initiated by the addition of 25 uM CaCly. The level of Ca2t-dependent ATPase and the rate of ca2t uptake were determined as indicated in our earlier publication (11). The rate of calcium uptake without oxalate, or the ATP-dependent binding of calcium by microsomes, was determined by rapid stopping of the reaction (6). To 0.5 ml of the incubation medium, containing 0.1 M KCl, 6 mM MgClp, 4 mM sodium azide, 20 mM imidazole (pH 6.8), and 22 uM 45CaCly (about 600,000 counts/min/ml), 0.1 mg microsomal protein was added and incubated in a water bath at 22° C or 37° C for 1 minute. After this interval, 20 ul of 25 mM ATP (pH 7.0) were in- jected into the solution while it was constantly agitated on a magnetic stirrer. The concentration of Ca2t in the specimen after the addition of ATP was 24 uM. After a certain time interval, the reaction was stopped by the addition of 0.5 ml of a solution of 50 mM EGTA, 0.1 M KCl, and 40 mM histidine (pH 6.0). The reaction mixture, within 5-10 seconds, was filtered through a 0.45 pm Milli- pore filter. The filter was washed with 5 ml of a solution containing 1 mM EGTA, 0.1 M KC1, 10 mM histidine (pH 6.0), dried and placed in a toluene scin- tillator. The quantity of radioactive calcium bound to the filter was deter- mined in a Mark III scintillation spectrometer. The concentration of protein was determined by a modification of the Lowry method (12). RESULTS Complete Extraction of the Calcium Pump System from Heart Muscle Preliminary experiments established that the activity of isolated micro- somal heart membranes in calcium uptake is decreased significantly with an in- crease in the intensity of homogenization of the muscle. Obviously, in this case, the vesicles extracted from the cells are subjected to further breakdown, which is accompanied by an increase in the permeability of the membranes for calcium ions. We thus preferred gentle homogenization for short time intervals. After centrifugation of the homogenate at 8,000-10,800 g, the microsomal frac- tion, concentrated in the supernatant, was precipitated by high-speed centri- fugation, while the precipitate was once more homogenized. Twelve to 15 such cycles of homogenization and differential centrifugation were performed. The total calcium-transporting activity of the microsomes extracted in this manner was found to be significantly higher than with continuous, intensive homogeniza- tion of the heart muscle. 29 40 ¢ 30+ 201 mg protein /g wet weight | 5 10 15 Number of homogenizations FIGURE 1. Accumulation of protein in the microsomal fraction upon repeated homog- enization of heart muscle. In curve 1, 17 g of guinea pig heart muscle were ho- mogenized as described in "Materials and Methods," using the Polytron homogenizer. In curve 2, 20 g of pigeon heart were homogenized in the Virtis-45 homogenizer. With each isolation cycle, an additional quantity of protein-containing material is precipitated with the microsomal fraction (figure 1). When the Polytron homogenizer is used (curve 1), the microsomal fraction contains approx- imately 1/3 as much protein as when the muscles are disrupted with a Virtis-45 homogenizer (curve 2). Figure 2 shows data on the accumulation of calcium in the microsomal frac- tion of membranes capable of accumulating calcium and forming a phosphorylated product in the presence of [y-32P]-ATP and a high concentration of calcium ions. Upon repeated homogenization of pigeon hearts (figure 2A) and guinea pig hearts (figure 2B), we observe first a decrease and then almost complete cessation of the yield of active membranes in the microsomal fraction. At the same time, significant accumulation of protein occurs even in the last cycles of isolation of the calcium-transporting system (figure 1). The membranes isolated from 1 g of pigeon and guinea pig heart, when no calcium-precipitating anions are present, can bind 66 and 79 nmol of Ca2%t/sec at 37° C, respectively (figures 2A and 2B). When 5 mM oxalate is included in the incubation medium, the rate of calcium accumulation is higher: 151 and 167 nmol Ca2t/sec (figures 2A and 2B). The activity of Ca2+-dependent ATPase in these preparations was 152 and 140 nmol/Pj/sec/g of net weight. The oxalate penetrating through the membrane of the sarcoplasmic reticulum decreases the concentration of calcium within the vesicles and thus supports the linear rate of calcium transport (13). The similarity of the curves shown 30 = A 5 5 1150 [3 2 24 -— Ca uptake [3] 3 oO ~N < 10} oe 100 & Ep rd -— - » © 3 00 © oo 5t 2+ 150 E Ca” binding > N c [1] o ££ & Oo £ — 0 4 8 12 y Number of homogenizations B §° [] 3 24 uw Ca uptake * 6 ° oo * 2 ep c Pd 3? 4 + err Orre0ner 00 Q pad 2+ . . oo Ca binding Q ~~ E > Sc 2f [% o £ Q wn Oo LL a 0 | 5 10 15 Number of homogenizations FIGURE 2. microsomal fraction upon repeated homogenization of heart muscle. heart. B: Guinea pig heart. Each of values determined in the present the presence of oxalate and binding described in "Materials and Methods. bom am b/sy/jowu ‘uondwnsuod 00 + Q 50 WN oO Oo > wn Cc 3 © = {100 > p=] 3 o ~ w ~ 50 “ 1 = ® s @ « = 0 Yield of calcium-transporting and phosphorylating vesicles in the A: Pigeon point on the curves corresponds to the sum and preceding fractions. Calcium uptake in calcium without oxalate were measured as 31 in figures 2A and 2B indicates that in the initial moments of time the calcium accumulated without oxalate has no significant inhibiting effect on the transport system itself. However, as can be seen in figure 3, the curve of calcium con- sumption without oxalate is nonlinear, indicating a progressive increase in the inhibiting effect of the intravesicular calcium or an increase in diffusion of this cation back into the incubation medium. It follows from this that the con- stant rate of calcium accumulation in the presence of oxalate adequately reflects the initial rate of transport measured under more physiological conditions. Nature of Extracted Calcium-Transporting System The total fraction of membranes extracted from 1 g of pigeon heart is capable, at a concentration of free calcium ions of 24 uM, of binding more than 200 nmol of Ca2* (figure 3), which is comparable to the value obtained by Solaro and Briggs for the dog heart (14). This quantity is more than sufficient to support the uptake of calcium localized in the myoplasm and bound to the con- tractile apparatus of the heart in systole (2-6). What structure is responsible for the accumulation of calcium in the prep- arations we extracted? In addition to the membranes of the sarcoplasmic retic- ulum, the myofibrils, mitochondria, and fragments of the sarcolemmal membrane are capable of energy-dependent binding of calcium (2,9,10,15). The possibility of participation of myofibrils in the binding of calcium under our experimental conditions was excluded based on the results of an experiment with the ionophore, A23187. As can be seen from table 1, A23187, which increases calcium flux 250 2007 150 | 100 50 cd’ consumption, nmol / g wet weight a, A AJL 0 0.5 | 5 Time, min FIGURE 3. Accumulation of Ca2t by the total fraction of pigeon heart micro- somes. The total preparation was obtained by combining all microsomal fractions that were isolated as a result of repeated homogenization of heart muscle. 32 through model and natural membranes (16), completely blocks the accumulation of calcium by preparations isolated from guinea pig heart. From this, we conclude that the calcium-accumulating system in this case has a membrane and vesicular structure. Table 1 also contains data on the accumulation of calcium by the mitochon- dria, isolated in the same hypotonic medium as the microsomal fraction. As the data show, the accumulation of calcium by the mitochondria in an azide-containing medium is very minimal. This uptake results most probably from the contamination of mitochondrial fraction by microsomal membranes. In order to estimate participation of fragments of the sarcolemma in calcium uptake by the vesicles which we isolated, we measured the steady- state level of phosphorylated membrane proteins under conditions (17,18) op- timal for the formation of phosphoprotein intermediates of Ca2*-dependent and sodium-potassium ATPases (table 2). In the presence of calcium, any of the preparations obtained in various stages of extraction of calcium-transporting membranes are phosphorylated. A good correlation is observed in this case between the level of phosphoprotein and the rate of calcium accumulation. How- ever, the level of phosphoprotein of Nat, K*-ATPase does not correlate with the calcium-accumulating capability of the vesicles. This indicates that the sarco- lemma fragments do not contribute significantly to the binding of calcium by TABLE 1. Calcium Binding by Preparations of Mitochondria and Microsomes Isolated From Guinea Pig Heart cat Binding Preparation nmol/sec/mg Protein Mitochondria 0.46% Microsomes 7.48% SR4 12.06 SR4 + 10 uM A23187 0.09 SR; , 1.25 SR, + 10 uM A23187 0.02 The mitochondria and microsomes were isolated as described above using a Polytron homogenizer in a medium suggested by Harigaya and Schwartz (2). The composition of the medium for determining ca2t accumulation without oxalate is presented in "Materials and Methods." The medium was incubated at 37° C or 22° C (*). The ionophore A23187 was kindly provided by the Eli Lilly Co. (U.S.). SR3 and SRj2 represent preparations of the sarcoplasmic reticulum obtained after 3 and 12 homogenization cycles of heart muscle. 33 TABLE 2. Rate of Calcium Accumulation by Microsomes From Guinea Pig Heart, and Phosphorylation of Preparations Under Conditions Optimal for the Formation of the Phosphoenzymes of Ca2t-Dependent ATPase and Nat, Kt-ATPase Cat Accumulation Phosphorylation nmol/sec/mg Protein nmol P;/mg Protein + Na' Preparation - Oxalate + Oxalate + cat + EGTA + EGTA SRq 24.9 63.7 1.93 0.31 0.10 SRy 12.2 22.0 0.84 0.19 0.07 SRq 4.8 14.0 0.29 0.10 0.06 SR, 1.5 0.19 0.08 0.02 Ca2t-oxalate preparation 2.80 0.07 0.07 The composition of the incubation media is presented in "Materials and Methods." SR. -SR represent preparations of microsomal membrane obtained after the corresponding number of cycles of homogenization and centrifugation of heart muscle. The purified calcium-pump system of the sarcoplasmic reticulum (CaZt- oxalate preparation) was obtained in accordance with the method described earlier (11). the preparations which we extracted. We can thus conclude that it is the vesicles of the sarcoplasmic reticulum which are the primary, if not the only, structure which accumulates calcium under our experimental conditions. Determination of the Quantity of Sarcoplasmic Reticulum in Pigeon and Guinea Pig Hearts There is as yet no agreement about the concentration of membranes of the sarcoplasmic reticulum in the hearts of homoiotherms. This quantity can be established by comparing the number of Ca2*t-dependent ATPase molecules in prep- arations of sarcoplasmic reticulum which have been purified and completely ex- tracted from the muscle. The number of molecules of Ca2t-dependent ATPase is determined from the steady-state level of phosphorylated protein intermediate formed in one of the stages of the ATPase reaction (17,19). Purified calt- dependent ATPase from the sarcoplasmic reticulum of skeletal muscle contains about 10 nmol P;/mg of protein. In these calculations the enzyme is assumed to have one acylphosphate bond per molecule of enzyme and a molecular weight of 100,000 daltons (20). 34 In order to determine the nature of the component which is phosphorylated in the microsomal preparations which we extracted, we separated them by electro- phoresis in SDS-polyacrylamide gels at pH 2.4. Under these acidic conditions, stabilization of the acylphosphate bond occurs (21). Figure 4 shows the dis- tribution of radioactive phosphate in slabs of gel containing two preparations of pigeon heart microsomes that were separated according to molecular weight. The primary incorporation of the label is observed in the high-molecular compo- nent, which appears in the presence of Coomassie stain and has a molecular weight of 100,000. The small incorporation of radioactive phosphate in the other gel section occurs due to protein with a molecular weight of about 200,000, apparently corresponding to a dimer of the major phosphate acceptor. Thus, the only protein which takes up the terminal phosphate from the ATP molecule under our experimental conditions is the Ca2t-dependent ATPase of the sarcoplasmic reticulum. Accumulation of the phosphorylated product in the microsomal frac- tion extracted in successive cycles of homogenization and differential centrif- ugation (figure 2) reflects the yield of calcium-dependent reticular ATPase in the microsomal fraction. We established that the highly purified membranes of sarcoplasmic reticulum of pigeon heart bind 2.5 nmol of P;/mg of protein (11). When preparations from guinea pig heart are used, this value is 2.8 nmol Pj/mg of protein (table 2). Purified membrane preparations were obtained by a procedure including precipita- tion of active vesicles partially saturated with calcium oxalate and similar in a number of properties to the isolated sarcoplasmic reticulum of skeletal muscle (11). One of the facts confirming the high degree of purity of the calcium- oxalate preparation of heart microsomes is the absence of incorporation of the terminal phosphate of ATP molecule in the preparation under conditions optimal for phosphorylation of Nat, Kt-ATPase (table 2). 40001 SR. counts / min nN o o ° Radioactivity, 7 - Length of gel, cm + FIGURE 4. Electrophoresis in SDS-polyacrylamide gel of phosphorylated guinea pig heart microsomes. The conditions for incubation of the membranes with [Y-32P]-ATP and for electrophoretic separation of protein at pH 2.4 are outlined in "Materials and Methods" and in our earlier publication (11). 35 Comparison of the level of Ca2t-dependent ATPase in purified (2.5 and 2.8 nmol/mg protein) and in fully extracted but not purified (9.7 and 5.9 nmol/g of protein) sarcoplasmic reticulum indicates that 1 g of pigeon and guinea pig heart tissue contains 3.9 and 2.1 mg of reticular protein, respectively. The quantity of sarcoplasmic reticulum can be estimated also by comparing the activity of Ca2t-dependent ATPase in the purified and fully extracted prep- arations of sarcoplasmic reticulum. For microsomes from pigeon heart, these values are 2.6 pmol Pj/min/mg protein (11) and 9.15 umol Pj/min/g heart tissue. According to this estimate, there are 3.5 mg of sarcoplasmic reticulum protein in 1 g of pigeon heart tissue. Keeping in mind the molecular weight of Ca2t-dependent ATPase (100,000), the level of the phosphorylated product in the purified enzyme (10 nmol/mg protein), and the data on phosphorylation of the sarcoplasmic reticulum fully extracted from pigeon and guinea pig hearts (2.5 and 2.8 nmol/g of muscle, respectively), we conclude that 1 g of heart tissue contains from 0.5 to 1 mg of Ca2t-dependent ATPase protein. DISCUSSION The high lability and low yield of sarcoplasmic reticulum from heart muscle homogenate prevent us from studying the role of the reticular membranes in relax- ation of the myocardium. In particular, the contribution of the sarcoplasmic reticulum to the accumulation of calcium upon relaxation of the heart is a sub- ject of extensive discussion (2,4,6-8,10,14). Calculations of the calcium-transporting ability of the sarcoplasmic retic- ulum localized in 1 g of heart muscle are usually based on the following postu- lates and assumptions: 1. For maximum activation of the myofibrils, the contractile apparatus should bind 50-150 nmol Ca2t/g of heart (2-6). 2. The time from achievement of maximum contraction by the heart to onset of full relaxation, i.e., the time during which all of the calcium bound to troponin must be absorbed by the calcium-pump system, is 200 msec (1). 3. The quantity of protein in the sarcoplasmic reticulum of the heart of homoiotherms is 1-5 mg/g of muscle (2,10). 4. The calcium-transporting activity of microsomes isolated from the heart is between 5 and 130 nmol Ca2t/mg of protein/sec (2,6,10,22,23). Based on the rate of calcium uptake by isolated microsomal vesicles as determined by Schwartz (22), 130 nmol/sec/mg protein, and on the upper limit for concentration of sarcoplasmic reticulum protein in the heart (10), it follows that the reticular network of the heart can bind up to 130 nmol calt/g of muscle in 200 msec, i.e., it can achieve full relaxation. We must postulate here that the activity of the entire sarcoplasmic reticulum concentrated in the heart is not less than the activity of the small fraction of isolated microsomal 36 membranes (22). Naturally, the results of such calculations cannot be used in analyzing the role of the sarcoplasmic reticulum in relaxation of the heart until the level of sarcoplasmic reticulum in the muscle is finally established. The first estimate of the quantity of sarcoplasmic reticulum extracted from the heart by standard methods was published in a study by Solaro and Briggs (14). These authors compared the calcium-accumulating capacity of heart homog- enate and isolated fragments of sarcoplasmic reticulum. They showed that 1 g of dog heart contains 6.8 mg of protein of reticular origin. Will et al. (6), using this value, calculated that, at a concentration of free calcium ions of 19 uM, the sarcoplasmic reticulum extracted from 1 g of dog heart can accumu- late 48 nmol Ca2t/200 msec at 37° C. This value is apparently elevated since the preparation of sarcoplasmic reticulum obtained by Solaro and Briggs by in- tensive homogenization of tissue in a Sorvall Omnimixer was not subjected to further purification (4,14). In the present study whose goal was to estimate the calcium-transporting capacity of the sarcoplasmic reticulum, we used a different approach. We selected the phosphorylated intermediate of CcaZ*-dependent ATPase, which is stable at low pH, as the marker of the calcium-pump system. The concentration of phosphoenzyme, in our opinion, reflects better the level of sarcoplasmic reticulum than do parameters such as the rate of calcium accumulation in the presence of oxalate and the calcium-oxalate capacity. The rate and maximum level of calcium accumulation by fragments of sarcoplasmic reticulum depend on the degree of intactness of reticular membrane. It can be assumed that calcium accumulation in the entire heart homogenate, containing sections of undamaged sarcoplasmic reticulum, and in isolated microsomal fragments, occurs at various levels of effectiveness. On the other hand, damage to the membrane by detergent (20) or partial removal of the membrane lipids (24) has practically no influence on the phosphorylation of caZt-dependent ATPase. In microsomal preparations extracted from the heart, the only phosphorylated component is the protein with the molecular weight of 100,000 (figure 4). In- corporation of radioactive phosphate in the preparations was not observed upon binding of calcium by EGTA (table 2). Thus, the phosphate acceptor in the mem- brane preparations in this case is the Ca2t-dependent ATPase. Therefore, the phosphoenzyme formed in the presence of calcium ions can be considered a specific marker of the sarcoplasmic reticulum. Furthermore, the level of phosphoprotein can serve as a reliable criterion of the purity of the preparation. In order to estimate the quantity of sarcoplasmic reticulum in 1 g of heart, we compared the levels of phosphorylated Ca2t-dependent ATPase intermediate in the total fraction of microsomes with the level of phosphoenzyme in a highly purified preparation of the calcium-pump system. It was found that from 2 to 4 mg of sarcoplasmic reticulum protein are concentrated in 1 g of heart tissue. Fluctuation of the phosphoenzyme level in preparations corresponds to changes in the calcium-transporting capacity (figures 1 and 2, and table 2) and in the activity of Ca2t-dependent ATPase (data not presented). This indi- cates that, under gentle homogenization conditions such as those which we used, there is no significant increase in the permeability of the membrane for the calcium accumulated within the vesicles, even in preparations obtained in the 37 later stages of extraction. We therefore concluded that the activity of the calcium-transporting system is not reduced in the long process of extraction. Apparently, the data presented in figure 2 can actually be used to reflect the activity of the calcium-pump system in the intact heart. Based on the results obtained, the rate of calcium uptake by the sarco- plasmic reticulum at 24 uM cat may reach 16 nmol/200 msec/g muscle. We stopped the accumulation of calcium by the microsomes 1 second after its initiation by adding ATP. However, as can be seen in figure 3, the rate of calcium accumula- tion is not linear, thus confirming the data obtained earlier using an analogous method (6) and a spectrophotometric technique using murexide as a calcium ion indicator (2). According to Will et al. (6), the quantity of Ca2t accumulated by the microsomes at 22° C in 200 msec is 50 to 70 percent of the level measured after 1 second. If the ratio of rates of calcium accumulation at 22° C and 37° C remains constant in the time interval from 0.2 to 1 sec, the preparations which we separated from 1 g of muscle can accumulate as much as 30-50 nmol ca2%/200 msec. This corresponds to the rate necessary for all of the calcium to be re- moved from the myoplasm and from the maximally activated myofibrils in the time of relaxation (2-6). 38 10. 11. 12. 13. 14. REFERENCES Langer GA: Ion fluxes in cardiac excitation and contraction and their relation to myocardial contractility. Physiol Rev 48:708-757, 1968 Harigaya S, Schwartz A: Rate of calcium binding and uptake in normal animal and failing human cardiac muscle. Circ Res 25:781-794, 1969 Ebashi S, Endo M, Ohtsuki I: Control of muscle contraction. Quart Rev Biophys 2:351-384, 1969 Katz AM: Contractile proteins of the heart. Physiol Rev 50:63-158, 1970 Solaro RJ, Wise RM, Shiner JS, Briggs FN: Calcium requirements for cardiac myofibrillar activation. Circ Res 34:525-530, 1974 Will H, Blanck J, Smettan G, Wollenberger A: A quench-flow kinetic inves- tigation of calcium ion accumulation by isolated cardiac sarcoplasmic reticulum. Dependence of initial velocity on free calcium ion concentration and influence of preincubation with a protein kinase, MgATP and cyclic AMP. Biochim Biophys Acta 449:295-303, 1976 Katz AM, Repke DI: Quantitative aspects of dog cardiac microsomal calcium binding and calcium uptake. Circ Res 21:153-162, 1967 Ebashi S, Endo M: Calcium ion and muscle contraction. In Progress in Biophysics and Molecular Biology, vol 18, edited by JAV Butler and D Noble. New York, Pergamon Press, 1968, pp 123-183 Patriarca P, Carafoli EA: Study of the intracellular transport of calcium in rat heart. J Cell Physiol 72:29-37, 1968 Scarpa A, Williamson JR: Calcium binding and calcium transport by sub- cellular fractions of heart. In Calcium Binding Proteins, edited by W Drabikowski, H Strzelecka-Golaszewska, and E Carafoli. New York, Elsevier, 1974, pp 547-585 Levitsky DO, Aliev MK, Kuzmin AV, Levchenko TS, Smirnov VN, Chazov EI: Isolation of calcium pump system and purification of Cca2t-dependent ATPase from heart muscle. Biochim Biophys Acta 443:468-484, 1976 Hartree EF: Determination of protein: A modification of the Lowry method that gives a linear photometric response. Anal Biochem 48:422-427, 1972 Makinose M, Hasselbach W: [Influence of oxalate on the calcium transport of isolated vesicles of the sarcoplasmic reticulum] (Ger). Biochem Z 343: 360-382, 1965 Solaro RJ, Briggs FN: Estimating the functional capabilities of sarcoplas- mic reticulum of cardiac muscle. Circ Res 34:531-540, 1974 39 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, Sulakhe PV, Drummond GI, Ng DC: Calcium binding by skeletal muscle sarcolemma. J Biol Chem 248:4150-4157, 1973 Case GD, Vanderkooi JM, Scarpa A: Physical properties of biological membranes determined by the fluorescence of the calcium ionophore A23187. Arch Biochem Biophys 162:174-185, 1974 Yamamoto T, Tonomura Y: Reaction mechanism of the Cat-dependent ATPase of sarcoplasmic reticulum from skeletal muscle. 1. Kinetic studies. J Biochem 62:558-575, 1967 Bastide F, Meissner G, Fleischer S, Post RL: Similarity of the active site of phosphorylation of the adenosine triphosphatase for transport of sodium and potassium ions in kidney to that for transport of calcium ions in the sarcoplasmic reticulum of muscle. J Biol Chem 248:8385-8391, 1973 Hasselbach W, Makinose M: ATP and active transport. Biochem Biophys Res Commun 7:132-136, 1962 Meissner G, Conner GE, Fleischer S: Isolation of sarcoplasmic reticulum by zonal centrifugation and purification of CaZt-pump and ca?t-binding proteins. Biochim Biophys Acta 298:246-269, 1973 Degani C, Boyer PD: A borohydride-reduction method for characterization of the acyl phosphate linkage in proteins and its application to sarco- plasmic reticulum adenosine triphosphatase. J Biol Chem 248:8222-8226, 1973 McCollum WB, Besch HR Jr, Entman ML, Schwartz A: Apparent initial binding rate of calcium by canine cardiac-relaxing system. Am J Physiol 223:608- 614, 1972 Nayler WG, Dunnett J, Burian W: Further observations on species-determined differences in the calcium-accumulating activity of cardiac microsomal fractions. J Mol Cell Cardiol 7:663-675, 1975 Martonosi A: The role of phospholipids in the ATP-ase activity of skeletal muscle microsomes. Biochem Biophys Res Commun 29:753-757, 1967 40 RESOLUTION AND RECONSTITUTION APPROACH TO THE STUDY OF THREE TYPES OF CALCIUM PUMPS FROM MAMMALIAN TISSUES Sidney Fleischer, Paul DeFoor, Brian Chamberlain, Dmitri Levitsky, Klaus Gietzen, and H. Uwe Wolf SUMMARY Intracellular calcium ion concentration serves as a fine-tuning control to regulate cell function. The calcium gradient, i.e., low cytoplasmic (0.1 - 1 uM) vs. high extracellular (1 mM) calcium ion concentration, is maintained by means of membrane-bound, adenosine triphosphate (ATP)-energized pumps which pump calcium ions out of the cytoplasmic compartment. This paper reviews three types of calcium pumping machinery from mammalian tissues using the resolution and reconstitution approach. Skeletal muscle sarcoplasmic reticu- lum (SR) is a highly specialized system whose membrane contains predominantly one protein, the calcium pump protein. Functional membrane vesicles of defined lipid-to-protein ratio have been reconstituted, and the phospholipid composi- tion can be varied in the same range, both lower as well as higher, as that of the normal SR membrane. Such reconstituted membranes are being studied to correlate membrane composition with structure, and structure with function. The calcium pump of heart SR can be modulated by a cyclic AMP-dependent pro- tein kinase which results in the phosphorylation of membrane-bound polypep- tide(s). The calcium pump of the red blood cell plasma membrane is modulated by a calcium-dependent regulatory protein which associates with the membrane when the cytoplasmic calcium concentration is elevated. The latter two calcium pumping systems are being studied to define the nature of regulatory processes at the molecular level. Drs. Fleischer, DeFoor, and Chamberlain are from the Department of Molecular Biology, Vanderbilt University, Nashville, Tennessee, USA; Dr. Levitsky is from the All-Union Cardiology Research Center, USSR Academy of Medical Sciences, Moscow, USSR; and Drs. Gietzen and Wolf are from the Department of Pharmacology and Toxicology, University of Ulm, Oberer Eselberg, Ulm, Federal Republic of Germany. This work was supported by grants from the National Institutes of Health (AM-16432) and from the Muscular Dystrophy Association. Drs. DeFoor and Chamberlain were postdoctoral fellows of the National Insti- tutes of Health (CA-05659) and the Muscular Dystrophy Association, respectively. 41 INTRODUCTION Cytoplasmic free calcium ion concentration has been estimated to be as low as 0.1 um, whereas extracellular calcium ion concentration exceeds 1 mM (1,2). This large differential of calcium ion concentration across the plasma and/or intracellular membranes is modulated by hormonal or excitable events which presumably open selective channels and permit calcium ions to flow across the membrane down the concentration gradient elevating the cytoplasmic calcium ion concentration. This fine tuning of intracellular calcium ion concentration represents an important regulatory mechanism of the cell since a large number of functions involve calcium-modulated reactions (3). The large calcium gradient is generated primarily by ATP-energized calcium ion pumps and is maintained, in part, by the relative impermeability of the membrane to calcium ions. This report details some of our progress in studying the calcium pumping machinery from three different mammalian membranes: oe Skeletal muscle sarcoplasmic reticulum, the most intensively studied calcium ion pump e Cardiac sarcoplasmic reticulum which can be modulated by a cyclic AMP-dependent protein kinase e Human red blood cell plasma membrane which is regulated by intra- cellular calcium ion concentration by way of a cytoplasmic regu- latory protein termed "calcium-dependent regulator," "CDR," or "calmodulin." The basic approach involves isolation and characterization of the pump molecules and reconstitution to form functional membranes of defined composi- tion. Reconstitution makes it possible to simplify and study complex biological systems. Achievement of function in the reconstituted membranes means that the components involved can be identified and characterized and that the pumping machinery can be studied in detail. SKELETAL MUSCLE SARCOPLASMIC RETICULUM Some 10 years ago we initiated a molecular biology approach to the study of SR from rabbit skeletal (4) muscle with the aim to: 1. Isolate and characterize highly purified SR. 2. Disassemble SR into its components and relate the properties of these constituents to SR function. 3. Reconstitute membrane vesicles, which are similar to normal SR and capable of transport function. 4. Correlate composition with structure, and structure with function. Such studies required development of a methodology to reconstitute 42 membranes of defined lipid-to-protein ratio which could be varied in the range of high protein content similar to that of natural membranes. 5. Conduct biophysical studies to obtain detailed membrane structure including motional parameters which are relevant to the nature of lipid-protein interaction. Isolation of highly purified normal SR (N-SR) in sizeable quantity (hun- dreds of mg protein) was achieved by centrifugation using zonal rotors (5). In our preparation of SR, three proteins predominated as judged by polyacryla- mide gel electrophoresis using dissociating conditions (figure 1, gel 2). The calcium pump protein of 119,000 daltons (6) (sometimes referred to as calcium ATPase), the calcium binding protein, and a polypeptide of approximately 55,000 daltons (designated M55), comprise about 75, 10, and 10 percent, respec- tively, of the protein of the purified SR fraction. The morphology of normal SR can be viewed by electron microscopy using three different methods of sample preparation (figure 2). In thin sections, membrane vesicles with trilaminar appearance of the membrane can be visualized (figure 2A). Negative staining (figure 2B) reveals small particles (40 A) at the outer surface of the membrane (7). Freeze-fracture electron microscopy (figure 2C) shows an asymmetric distribution of particles at the hydrophobic fracture faces with most of the intramembranous particles observable in the outer leaflet (concave face) of the membrane (8). TOP OF SEPARATING GEL ~ ca®t pump PROTEIN — ca?t BINDING PROTEIN — Meg roms TRACKING DYE-— GEL 1 2 3 FIGURE 1. Polyacrylamide gel electrophoresis (PAGE) of sarcoplasmic reticulum (SR). The samples in the gels are of: (1) light SR, (2) normal SR, (3) heavy SR. 43 FIGURE 2. Normal sarcoplasmic reticulum as visualized by electron microscopy (A) in thin sections, (B) using negative staining, and (C) by freeze-fracture. Highly purified SR can be further subfractionated into heavy and light SR vesicles based on their isopycnic densities (figure 3). Heavy SR contains electron-dense matter within its compartment (figure 3B), whereas light SR (figure 3A) is practically devoid of such contents (9,10). Polyacrylamide gel patterns of light SR (gel 1) and heavy SR (gel 3) were compared with the normal SR (gel 2) from which they were derived (figure 1). The electron opaque contents of heavy SR are referable mainly to calcium-binding protein. Light SR is essentially all membrane, of which the calcium pump pro- tein comprises 90 percent or more of the protein. Indeed, the SR membrane is highly enriched with regard to the calcium pump protein (9,10). We have purified and characterized the major components of SR (5). The calcium pump protein contains two specific calcium-binding sites, and one ATP- binding site, and is the membrane component which forms the phosphoenzyme inter- mediate (4,11). Normal SR can be dissociated using deoxycholate and then reconstituted to form functional membrane vesicles by removing the detergent (12). Conditions are stringent for reassembly of the membrane to form functionally reconstituted SR (R-SR). Reconstitution must be carried out at 15-20° C in order to achieve good energized accumulation of calcium. The protein of R-SR, like normal SR membranes, consists mainly of calcium pump protein. Using the procedure de- scribed previously, most of the M55 but little of the calcium-binding protein is retained in R-SR. Reconstitution can also be achieved with purified calcium pump protein to yield functional membrane vesicles (13). Thus, the calcium pump protein has all the known characteristics required for calcium pumping and appears to be the calcium pumping molecule. R-SR has a lipid content comparable 44 FIGURE 3. Electron micrographs of isolated (A) light and (B) heavy sarcoplasmic reticulum (SR) and (C) the triad region in situ. A and B: Light and heavy SR were prepared from normal SR by isopycnic centrifugation. C: The triad region is emphasized in the thin section of the rabbit skeletal muscle. The sample was fixed using tannic acid enhancement. The asymmetry of the SR membrane can readily be visualized. TT = T-tubule. TC = Terminal cisternae of SR. LSR = Lateral cisternae of SR (10). 45 to that of the N-SR membrane (not shown). Functionally, the energized calcium pumping rate and capacity of reconstituted SR is somewhat less than one-half that of the original SR (figure 4). The reconstitution procedure which we have described has three key characteristics which differ from that of others (14): (a) R-SR membrane vesicles like N-SR consist mainly of protein (60 percent of the mass) and are not liposomes mildly doped with trace amounts of calcium pump protein, (b) R-SR does not require trapping agents such as oxalate within its compartment during vesicle formation in order to achieve subsequent energized calcium pumping, and (c) R-SR is capable of energized calcium pumping in the absence of a trapping agent. Recently, we developed a procedure to visualize the asymmetry of the SR membrane in thin sections using tannic acid to enhance contrast (10). In normal SR, the outer layer is 70 R wide compared with 20 & each for the middle and inner layers (figure 5A). The asymmetry of the normal SR membrane can thus be visualized in thin sections as well as using negative staining (figure 2B) (7), and by freeze-fracture electron microscopy (figure 2C) (8). The asymmetry of the SR membrane can now be visualized, in situ, in thin sections of muscle using tannic acid enhancement (figure 3C). The broad layer (70 R) is on the outer face of the SR membrane so that calcium ions are pumped from the broad outer face across the membrane to within the compartment. This asymmetry is retained in the isolated SR vesicles (figure 5A). It may also be noted in the structure of the triad (figure 3C) that the compartment of the terminal cisternae of SR contains opaque material, whereas the lateral cisternae of SR is devoid of such contents. These studies suggest that heavy and light SR derive from the terminal and lateral cisternae, respectively (9,10). ORIGINAL SR VESICLES RECONSTITUTED SR VESICLES Ca?* LOADING (pumoles Ca®*/ mq protein) TIME (MIN) FIGURE 4. Time-course of Cat loading of original and reconstituted sarco- plasmic reticulum (SR) vesicles (12). 46 FIGURE 5. Sarcoplasmic reticulum (SR) as visualized in thin sections (A-C) and by negative staining (D-E). A-C: Tannic acid was used to achieve enhanced contrast of the SR membranes (10). A is normal SR; B and C are reconstituted SR of lower and higher protein content, respectively. The 0.1 um bar in C gives the enlargement for A-C. D and E: Negative staining of reconstituted SR. 47 R-SR differs from N-SR in one key aspect, i.e., the asymmetry of the mem- brane has not been retained in the membranes formed in the test tube. This difference can be seen using each of the three methods of sample preparation for electron microscopy. In thin sections, the trilaminar appearance of the R-SR membrane with high protein content is symmetrical, i.e., the trilayer is 70 R, 20 A, and 70 & wide (figure 5C). When the lipid-to-protein ratio of the R-SR membrane is increased, there is less surface material on the membrane surfaces (figure 5B). With negative staining, the surface particles can be visualized on both inner and outer membrane faces of R-SR (figures 5D and 5E). With freeze-fracture, R-SR has a more symmetrical distribution of particles between convex and concave fracture faces (figures 6B through 6F), whereas in normal SR, the particles are observed mainly in the outer (concave) fracture face (figures 2C and 6A). The reconstitution procedure has been modified to prepare R-SR membranes of defined phospholipid content, and the lipid content of the reconstituted SR was varied (15). The lipid content which we found most useful for corre- lative studies ranges from approximately one-half to twice the lipid content of N-SR. This series of preparations has permitted correlation of composition with structure, and is illustrated herein using freeze-fracture electron micros- copy. A series of R-SR membrane vesicles of increasing lipid content is shown in figure 6. The lipid content varies from approximately less than one-half (figure 6B) to the same (figure 6D) to more than twice the lipid content (figure 6F) of normal SR (figure 6A). The concentration of particles observed at the hydrophobic center in the R-SR membrane vesicles of both convex and concave faces was found to decrease proportionately with the protein content of the membrane. Concomitantly, the particle-free area increased with the lipid con- tent. Since the protein content of these membranes consists mainly of one type of protein (90 percent), the concentration of intramembranous particles is di- rectly proportional to the concentration of the calcium pump protein molecules in the membrane. The number of pump molecules calculated in the membrane is greater by at least a factor of 2 than the number of particles observed. Thus, it would appear that the particles, on the average, consist of two or more cal- cium pump protein molecules (15). The ability to prepare functional membrane vesicles of defined composition and with varying phospholipid content, within the range of high protein content of the type described here, makes it possible to study in detail membrane structure including motional parameters. Such studies using X-ray and neutron diffraction and nuclear magnetic resonance are being conducted in collaboration with Leo Herbette, Antonio Scarpa, and Kent Blasie of the University of Penn- sylvania; A. C. McLaughlin of the Brookhaven National Laboratory, Upton, New York; and Joachim Seelig and his colleagues at the University of Basel (16). The studies described above suggest molecular detail for the orientation of the calcium pump protein in the SR membrane (figure 7). The main protein component of the SR membrane is the calcium pump protein. Phospholipid devoid of calcium pump has a symmetrical appearance of the trilayer; the width of each layer is approximately 20 R (figure 7A, frame E). Therefore, the asym- metry of the SR membrane which is observable in thin sections (frame B) and by negative staining (frame A) must be referable to the unidirectional alignment 48 FIGURE 6. Freeze-fracture electron microscopy of normal sarcoplasmic reticulum (SR) (A) and reconstituted SR of varying phospholipid content (in pmoles phos- pholipid per mg protein): (B) 0.38, (C) 0.57, (D) 0.78, (E) 1.03, and (F) 1.29. The value for normal SR is 0.78 (15). 49 RECONSTITUTED-SR PHOSPHOLIPID | ! ! ! 1 I 1 I ! ! \ NORMAL - SR A LB Cc + D E \ \ \ \ | Y TANNIC ACID NEGATIV \ \ \ NEGATIVE E! \ STAINING } FIXATION I STAINING | A NORMAL SR RECONSTITUTED SR P FIGURE 7. Diagrammatic representation of normal and reconstituted sarcoplasmic reticulum (SR) membranes. A: Normal and reconstituted SR as visualized by negative staining and in thin section electron microscopy using tannic acid. The 40 angstroms particles observed with negative staining (frames A and D) and the broadened outer band in thin section using tannic acid extend 50 angstroms from the surface (frames B and C), compared with phospholipid (frame E). The membrane is asymmetric in normal SR (frames A and B) and symmetric in reconsti- tuted membranes (frames C and D). B: Diagrammatic representation of normal SR and reconstituted SR vesicles illustrating the effect of freeze-fracture on these membranes (10). 50 of the calcium pump protein in the membrane, i.e., a portion of the calcium pump protein extends beyond the phospholipid from the outer surface of the membrane. The calcium pump protein must be transmembrane (figure 7B) since calcium is pumped across the membrane from the outside to the inside of the vesicle. This interpretation is supported by X-ray and neutron diffraction studies (16). In R-SR, the pumps are more bidirectionally aligned giving rise to the symmetrical appearance in both thin sections (figure 7A, frame C) as well as by negative staining (figure 7A, frame D). The calcium pump is depicted as transmembrane and as an oligomer for the reasons cited above (figure 7B). The calcium pump from skeletal muscle SR is the most intensively studied system and serves as a prototype for studying pumps from other sources such as the two discussed below. CARDIAC SARCOPLASMIC RETICULUM Cardiac SR has not been isolated to the same apparent degree of purity as SR of skeletal muscle, but recent procedures based on partial loading of vesicles with calcium oxalate have yielded substantially improved preparations. (17). Current evidence indicates that the same general mechanism exists for ATP-energized calcium transport for both cardiac and skeletal muscle SR (18). An added feature of cardiac SR is that it can be modulated by a cyclic adenosine monophosphate (AMP)-dependent protein kinase which in dog heart SR results in phosphorylation of a 22,000-dalton polypeptide designated ''phospho- lamban" (19). The enhancement in calcium pumping is referable to a lowering of the Michaelis constant from 1 pM to 0.3 uM. It has been postulated that modulation occurs via the phosphorylation and dephosphorylation of phospholamban which, in turn, regulates the calcium pump protein (19,20). Such a mechanism might provide the basis for the well-known inotropic effect of catecholamines on the heart. There is, however, no direct evidence linking phospholamban to the modulation of calcium pumping activity. It should be noted that two addi- tional smaller peptides of 14,000 and 11,000 daltons are also labeled in beef (21) and pigeon heart (22) cardiac SR fractions. We are attempting to carry out dissociation and reconstitution of the membrane-bound polypeptides and the calcium pump protein of cardiac SR in order to (a) define which is the modulator polypeptide(s), (b) study the molecular interaction of the modulator with the calcium pump protein, and (c) define the nature of the modulation process. Preliminary studies on the resolution of bovine cardiac SR are shown in table 1. Primary solubilization with low levels of deoxycholate selectively extracts the phosphorylatable polypeptides. The extract is enriched fourfold with respect to these polypeptides. Selective extraction does not occur if the polypeptides in the membrane are phosphorylated, indicating that phosphorylation enhances association of the polypeptides with the membrane (19,21). A second extraction at a higher level of deoxycholate can then be used to solubilize the calcium pump protein. We are trying to work out conditions for reconstitution and are hampered by the greater instability of the calcium pump from heart compared to skeletal muscle SR. This problem is currently under study. 51 TABLE 1. Resolution of Heart Sarcoplasmic Reticulum (SR) Into Regulatory Protein(s) and ca?t ATPase by Successive Deoxycholate Extractions Recovery (% Initial | Ca?t - ATPase | Phosphorylation Sample Protein) (pmol/mgemin) (nmol/mg) 1. Original SR 100 0.40 1.3 2. Primary Solubilization Supernatant 25 0.05 4.9 Pellet 58 0.51 - 3. Secondary Solubilization Supernatant 19 0.90 - ATPase determinations were performed at 25° C. Samples containing 10 to 30 ug of protein were diluted approximately 150-fold into reaction mixtures containing either 1 mM EGTA (basal ATPase) or 50 pM CaClp, plus 25 uM EGTA (total ATPase). The Ca2t-dependent ATPase activity presented is the difference between the total and basal rates. All reaction mixtures contained 0.1 mg/ml Triton X-100. Phos- phorylation reactions were also performed at 25° C. Approximately 250 ug of intact SR or 100 pg of solubilized protein were phosphorylated by a cylic AMP- dependent protein kinase in the presence of v-32p ATP. After precipitation by trichloroacetic acid (TCA), the samples were washed, and the radioactivity incor- porated into protein was determined by scintillation counting. Corrections were made for the autophosphorylation of protein kinase (21). Another aspect of our studies with cardiac SR is more complete. We have compared the calcium pumping systems of skeletal and cardiac SR. Electron micrographs of thin sections of heart SR (figures 8C and 8D) have been compared with that from skeletal muscle (figures 8A and 8B). Heart and skeletal muscle SR have similar morphology. They appear as membrane vesicles in thin section. The insets in the figures are electron micrographs of the same samples observed by negative staining. The 40 R surface structure characteristic of skeletal muscle SR and referable to the calcium pump protein (10) can also be visualized in cardiac SR. The protein profile of rabbit, pigeon, and rat skeletal muscle SR has been compared with rat, rabbit, pigeon, and bovine cardiac SR (figure 9). Skeletal muscle has a simpler protein profile. Nonetheless, the calcium pump protein is the major component of both skeletal muscle and heart SR. 52 FIGURE 8. Sarcoplasmic reticulum (SR) membranes observed by electron microscopy of thin sections. A: Rabbit skeletal muscle (RBSM) SR. B: Pigeon skeletal muscle (PSM) SR. C: Pigeon heart (PH) SR. D: Bovine heart (BH) SR. The insets show the surface structure of the vesicles as observed with negative staining (23). 53 CPP CBP = - Mss = € 1 El u i8 1 kc RB. PRT... RT Re 8B P SM SM SMH i H H FIGURE 9. SDS-polyacrylamide gel electrophoresis of cardiac and skeletal muscle sarcoplasmic reticulum (SR). The concentration of polyacrylamide was 5 percent and 11 percent in the stacking and separating gels, respectively. Gels were run in the presence of 0.1 percent SDS (sodium dodecyl sulfate). The bands were developed using Coomassie brilliant blue R-250. RBSM = Rabbit skeletal muscle SR (30 ug). PSM = Pigeon skeletal muscle SR (25 ug). RTSM = Rat skeletal muscle SR (25 ug). RTH = Rat heart SR (60 pg). RBH = Rabbit heart SR (60 upg). BH = Bovine heart SR (50 pg). PH = Pigeon heart SR (50 ug). The calcium pump protein (CPP), calcium binding protein (CBP), and the protein of molecular weight 55,000 daltons (M55) are indicated for rabbit skeletal muscle SR (23). We have prepared sheep antiserum to purified calcium pump protein from rabbit skeletal muscle SR and have used this antiserum to compare the immuno- logical cross-reactivity of the calcium pump protein of skeletal and cardiac SR from several sources. Immunodiffusion studies using this antiserum gave a single band of identity with purified calcium pump protein from rabbit skeletal muscle SR and with SR from rat and rabbit skeletal muscle (figure 10A). Only partial identity is obtained with pigeon skeletal muscle SR as indicated by the spur or discontinuity in the figure. Cardiac SR prepared from either pigeon, rat, rabbit, or bovine failed to produce a precipitation band by immunodif- fusion, indicating that this cardiac SR was immunogenically different from skeletal muscle, i.e., either it did not cross-react or did so very weakly. 54 _ ANTISERUM DILUTION (1/1400) t Ng PRE-INJECTION SERUM (1/200 I 2 3 4 5 ANTIGEN CONC. (ug PROTEIN/ASSAY VOLUME) 0 PERCENTAGE COMPLEMENT FIXATION H Oo » FIGURE 10. Immunological cross-reactivity of heart and skeletal muscle sarco- plasmic reticulum (SR). A: Immunodiffusion patterns of sheep antiserum to calcium pump protein versus sarcoplasmic reticulum or purified calcium pump protein. Skeletal muscle SR was solubilized with 1 percent Triton X-100, 38 mM Tris, 0.1 M glycine, pH 8.7. The plates were stained with Coomassie blue. The patterns are: calcium pump protein (CPP)-purified from rabbit skel- etal muscle (RBSM) SR (25 ug). PSM = Pigeon skeletal muscle SR (25 pg). RTSM = Rat skeletal muscle SR (25 ug). B: The interaction of SR membrane prepara- tions with sheep antiserum to calcium pump protein as studied by complement fixation. Immune serum was tested against calcium pump protein of rabbit skeletal muscle SR (0——0) and rabbit skeletal muscle SR (6—@) (23). Quantitative complement fixation was used to quantitate the cross-reactivity of each of the skeletal muscle and cardiac SR preparations to the antiserum directed vs. the rabbit skeletal muscle calcium pump protein. A typical exper- iment is shown in figure 10B. A series of titration curves was obtained with a single antigen (in this instance, purified calcium pump protein from rabbit skeletal muscle SR) at several dilutions of the antiserum. The peak values of the percentage of complement fixed were then plotted vs. the log of the antiserum dilution as shown in figure 11. The line on the right in figure 11 contains values derived from these data vs. the original antigen. The index of dissimilarity for the original antigen is designated as 1.00 (24). Rabbit and rat skeletal muscle SR give superimposable lines and are identical with that obtained vs. the calcium pump protein from rabbit skeletal muscle SR, indicating that the calcium pump proteins from the two sources are immunogenically indistinguishable. The lines to the left were obtained by titration with other preparations and express dissimilarity in quantitative terms. Lines displaced more to the left which have greater indices of dissim- ilarity are more dissimilar. Pigeon skeletal muscle SR with an index of dissimilarity of 1.89 is more similar to rat and rabbit skeletal muscle SR than to any of the heart SR preparations, each of which have higher indices of 55 100 H—HEART— FSKELETAL MUSCLE H 17300 1/700 1/1500 1/3000 x g a = 90 \s > £ PCr a c @ << 2 — 6 2.0 2 RB a = 0 20 40 60 Time, min FIGURE 1. Formation of phosphocreatine (PCr) coupled with glycolytic splitting of glucose. The solid line at the top of the figure is the value of the mass ac- tion ratio of this reaction, and it is equal to the equilibrium constant under experimental conditions (Keq = 1.1 x 10-2). In this experiment and other experi- ments, the results of which are shown in figures 2-5, incubation was performed at 30° C in a buffer system containing 40-50 mM Tris-HCl, 20 mM potassium phosphate (pH 7.5), 50 mM KCl, 6 mM Mgt acetate, and 0.33 mM dithiotreitol. In this case, the incubation medium initially also contained 50 mM glucose, 3.9 mM adenosine triphosphate (ATP), 0.5 mM adenosine diphosphate (ADP), 0.2 mM adenosine mono- phosphate (AMP), 0.5 mM NADY, 20 mM creatine (Cr), and 1.1 mg/ml cytosol protein. the creatine kinase reaction. Subsequently, as phosphorylated intermediates of glycolysis accumulate, a linear increase begins in the concentration of lactate and creatine phosphate, and the concentration of ATP begins to return to its initial level. Our attention is drawn to the fact that the concentration of lactate and the change in the sum of concentrations of ATP and phosphocreatine (PCr) in- crease at approximately the same rate, i.e., the quantity A[(PCr) + (ATP)]/A lactate, for the maximum slopes of the curve near the inflection, was 0.6 - 0.7. The slight lag in the increase in the sum of the concentrations of phosphocrea- tine and ATP behind the increase in the concentration of lactate suggests that a steady state is not established in the system, and that the rate of phosphor- ylation of hexose may be more than 1/2 the rate of synthesis of ATP + CP. It is interesting that the mass action ratio for the creatine kinase reaction (PCr) x (ADP)/(creatine) x (ATP), calculated for various points in time, changes little with time and is practically equal to the apparent equilibrium constant of this reaction under the conditions of our experiments (K = 1.1 x 102). The value of the equilibrium constant was experimentally determined under identical 68 conditions with purified CK. This shows that CK is not the limiting step in the entire process, and indicates its capability to effectively trap the ATP formed by glycolysis. Figure 2 shows the results of experiments on the phosphorylation of crea- tine coupled to glycolysis, when the substrate is fructose-1,6-diphosphate. In this case, with comparatively low concentrations of ADP (0.38 mM), the ac- cumulation of phosphocreatine and lactate in the system occurs practically without delay with a constant concentration of ATP (0.35 mM), almost equal to the concentration of ADP added initially. The stoichiometry of the formation of phosphocreatine and lactate is near theoretical [(CP)/(lactate) = 2.0]. The steady-state rate of synthesis of phosphocreatine, determined in the linear section near the inflection, depends hyperbolically on the steady-state concentration of ATP which, in all cases, is practically equal to the concentra- tion of ADP added. This function is linear in a double-reciprocal plot, yielding a value of apparent K for ATP of 0.25 mM and of Vj,y of 0.4 umol/min per mg of protein (figure 3). In kinetic experiments where ATP is regenerated not by glycolysis, but due to an excess of pyruvate kinase and phosphoenolpyruvate, the constants for CK are an apparent Kj of 0.8 mM for ATP and a V,, of 2.5 umol/min per mg of protein. The difference in the values of Vp,. and K for regeneration of ATP by glycolysis (substrate fructose-1,6-diphosphate) and for regeneration of ATP by the excess of pyruvate kinase is understandable. With an excess of pyruvate kinase, the concentration of ADP is practically 0, while the concentration of phosphocreatine during the initial reaction period is also negligible. In 11.0 = g 40 = < £ ° - 5 lo5% oO 420 A "a < \ / a’ oy NN [a & FN < / / 0 FIGURE 2. Synthesis of phosphocreatine (PCr) accompanied with glycolytic split- ting of fructose-1,6-diphosphate. The incubation medium contained 5 mM fructose- 1,6-diphosphate, 0.4 mM ADP, 0.5 mM NADT, 20 mM creatine, and 1.1 mg/ml cytosol protein. The other experimental conditions are given in figure 1. 69 oO E = E = E o 20 . © = 5 3 a - FS = > 10 A |/LATPIst, mM” 20, 60 0 02 0.4 1/[Cr1, mM” FIGURE 3. Effect of a steady-state concentration of ATP and creatine (Cr) on the rate of phosphocreatine synthesis, coupled to conversion of fructose-1,6- diphosphate. Variation of the reciprocal of reaction rate on the reciprocal of Cr concentration with fixed ATP (line 1) and with the reciprocal of ATP concen- tration at a fixed creatine concentration of 20 mM (line 2). The incubation medium contained 5 mM fructose-1,6-diphosphate, 0.5 mM NAD', and 1.1 mg/ml cytosol protein. The remaining experimental conditions are given in figure 1. other words, under these conditions, the characteristics of CK of cytosol are determined by the forward reactions, under conditions where the reverse reaction does not occur. A quite different situation exists in the case of phosphocreatine synthesis coupled to glycolysis. Here, the steady-state rate was determined in the pres- ence of a significant concentration of PCr (from 1.5 to 2.7 mM); ADP was also present in the medium (0.03-0.3 mM). Under these conditions, the reverse reac- tion, or dephosphorylation of PCr, occurs at a significant rate. The presence of significant concentrations of ADP in the medium [(ATP)/(ADP) = 5-10] sug- gests that the glycolytic system ''does not have time to" rephosphorylate the ADP, which reduces the rate of synthesis of phosphocreatine catalyzed by CK, forcing the latter to function at the rate established by glycolysis. For this very reason, the Vp x observed for the synthesis of PCr coupled to glycolysis, is 1/6 the value of Vmax for CK itself. In other words, the reactions which determine the net rate of this process is the glycolytic system. Analysis of the mass action ratio of the creatine kinase reaction, which varies between 0.96+1072 and 1.2-10-2, i.e., always remains close to the apparent equilibrium constant of this reaction (K = 1.1-10-2), leads us to analogous conclusions. 70 Figure 3 also shows the kinetics of formation of phosphocreatine accom- panied with the glycolytic splitting of fructose-1,6-diphosphate as a function of the concentration of creatine in the medium, at low concentrations of added ADP (0.1 mM). This variation also forms a hyperbolic curve and the double- reciprocal plot is shown (figure 3). The value of the apparent K; for creatine = 10 mM. These results show that the concentration of creatine in the medium effectively controls the rate of formation of CP, when coupled to glycolysis, since the value of K; for creatine in the combined system is similar to the value of Ky for the enzyme. This is seen in spite of the fact that the creatine kinase reaction is not the rate-limiting step. The mechanism of this control by creatine occurs by an increase in the concentration of creatine in the medium, accompanied by a decrease in the steady-state concentration of ATP, almost equal to the concentration of the ADP added initially, i.e., always (ATP) g¢ >> (ADP) g¢ - Unfortunately, the low concentrations of ADP in these experiments (= 1-10 uM) prohibits the direct determination of ADP and produces significant experimental errors. Phosphorylation of creatine was conducted with phosphoenolpyruvate as a substrate for the glycolytic system. However, in this case, due to the rela- tively easy reversibility of the reactions of glycolysis between fructose-1,6- diphosphate and phosphoenolpyruvate, the situation is more complex. When creatine is absent and when the concentrations of ADP are catalytic (50 uM), the oxidation of NADH was observed, and to a degree several times larger than the amount of ADP added. This suggests the occurrence of the pyruvate kinase reaction, as well as the presence of some other process expending the ATP that was formed. In the presence of creatine, synthesis of phosphocreatine occurred, but the stoichiometric coefficient [(PCr)/(lactate)] was much lower than the theoretical value of 1. It is possible that, under our conditions, phosphoenolpyruvate may also be used in the formation of 3-phosphoglycerate, which is subsequently phos- phorylated to 1,3-diphosphoglycerate by phosphoglycerate kinase and the ATP syn- thesized in the pyruvate kinase reaction. For this reason, the glycolytic system can oxidize excess NADH in the presence of only catalytic quantities of ADP. Obviously, the stoichiometric coefficient for the formation of PCr was less than 1 as a result of the competition between phosphoglycerate kinase and CK for ATP. The introduction of NaF into the reaction medium at a concentration of 20 mM almost suppressed these reverse reactions by the inhibition of enolase by F~ (18). This concentration of NaF did not, however, influence the activity of pyruvate kinase and the CK reaction. In the presence of NaF, creatine was phosphorylated by phosphoenolpyruvate, with 1 mole of phosphocreatine appear- ing for each mole of lactate formed and NADH oxidized (figure 4A). In other words, the stoichiometry of the reaction corresponds to the theoretically expected. When creatine is absent, the accumulation of lactate or oxidation of NADH occur very slowly, indicating inhibition of enolase by the fluoride ion. The concentration of ATP in the medium, both in the presence of creatine, and when creatine is absent, did not change with time and was approximately equal to the initial concentration of ADP added. It also follows from figure 4A that the exclusion of creatine from the reac- tion mixture leads to a significant decrease in the rate of formation of lactate 71 A = B Lac (+Cr)~ E = = 10 ADHGCO a” 2 c - cretn |= 204 PCr+ ATP 5 eg 0 2 5 - 1 Q Gost 10. Ss a - 0.2 0 | KH ___Oo _ oO 5 Lac (-Cr) | oO -~ ” NADH (-Cr) 0 10 20 0 Time, min Time, min FIGURE 4. Synthesis of phosphocreatine (PCr) coupled with pyruvate kinase and phosphoglycerate kinase reactions in cytosol from myocardial cells. A: Phos- phorylation of creatine (Cr) by phosphoenolpyruvate (PEP). The incubation medium contained 3.0 mM PEP, 50 uM ADP, 2.0 mM NADH, 20 mM NaF, 1 mg/ml cytosol protein, and 20 mM Cr. Lac = Lactate. B: Phosphorylation of Cr by 1,3-diphosphoglycer- ate. The incubation medium contained 5 mM fructose-1,6-diphosphate, 0.12 mM ADP, 2.0 mM NAD*, 20 mM NaF, 20 mM Cr, 2 IU/ml added glyceraldehyde phosphate dehydrogenase, and 0.2 mg/ml cytosol protein. and oxidation of NADH, since under these conditions the constant regeneration of ADP by the CK reaction is eliminated. If, when NaF is present to inhibit enolase, fructose-1,6-diphosphate and an excess of glyceraldehyde phosphate dehydrogenase are added to the reaction mixture instead of phosphoenolpyruvate to assure the rapid synthesis of 1,3- diphosphoglycerate, then when creatine is present, we observe the synthesis of phosphocreatine coupled to the phosphoglycerate kinase reaction (figure 4B). The results presented above illustrate the coupling of the ATP-generating glycolytic reactions and the CK reaction. The combined system leads to the synthesis of phosphocreatine with a stoichiometric coefficient near the theo- retical value, and with a practically constant concentration of ATP in the medium. This coupling is observed in a homogeneous medium and, under these conditions, the enzymes interact in accordance with their kinetic properties, by the establishment of steady-state concentrations of the intermediate products in the medium. Using the kinetic equations for CK and pyruvate kinase and the experimentally determined values of their kinetic constants (10,19), we can mathematically describe the synthesis of phosphocreatine, coupled to the pyru- vate kinase reaction. The results of these calculations and their comparison with experimental data are shown in figure 5. Figure 5A presents the steady-state rates of the process, at various con- centrations of ATP, as a function of the concentration of creatine in the medium. 72 V / Vo A [ATP] To E 04 E @ @ oO £ 3 0.2 > 0 10 B Log C . ia 2. ©.“ 5 mM ATP “ne 5mM ATP En SEE 05" 4x J 205 nd / | mM ATP Re mM ATP 2 re 8 0.5mM ATP ol A a Teor Teel ll . So 8... eee. ‘ SRL 0 5 0 5 10 15 PCr, mM PCr, mM FIGURE 5. Effect of creatine (Cr) and phosphocreatine (PCr) on the rate of PCr synthesis coupled with the pyruvate kinase reaction. A: Variation in the reac- tion rate as a function of Cr concentration with various concentrations of ATP in the medium. The curves on the figure were calculated by computer based on full equations for creatine kinase and pyruvate kinase and numerical values of kinetic constants under our experimental conditions. B and C: Inhibition of CK by PCr at various concentrations of ATP and Cr (B = 10 mM, Cr = 20 mM). V/Vg = Relative reaction rate, where Vo is the reaction rate without PCr. The curves were calculated by computer as for figure 5A. The incubation medium contained 2.0 mM phosphoenolpyruvate (PEP), 20 mM NaF, 0.16 mM NADH, and 0.1- 1.0 mg/ml cytosol protein. The reaction rate was determined spectrophotomet- rically from the decrease in concentration of NADH, recording the decrease in optical density at 340 nm. 73 The experimental data points and theoretical lines practically coincide, demon- strating the stimulatory effect of creatine and ATP on the overall rate of the coupled reactions. This agreement between experimental and theoretical results can also be seen in figure 5A, indicating the suitability of individual enzyme kinetic equations for describing the complex reactions of the homogeneous system. Furthermore, figure 5B illustrates the sensitivity of the soluble system to produce inhibition by phosphocreatine. An increase in the initial concentra- tion of phosphocreatine leads to substantial inhibition of the coupled reactions, when phosphoenolpyruvate is used for the production of ATP. This effect of phosphocreatine agrees with the high affinity of CK for this substrate (10), and also with the quasi-equilibrium position of the CK reaction. Thus, an in- crease in the concentration of phosphocreatine may lead to a proportional de- crease in the steady-state concentration of ADP. This, in turn, leads to a suppression of the pyruvate kinase reaction. With an initial concentration of 15 mM phosphocreatine, the rate of phosphocreatine synthesis by the coupled system is inhibited by approximately 90 percent (figure 5B). It follows from these estimations that, when a high concentration of phosphocreatine is present in the medium, the coupling of glycolysis to phosphocreatine synthesis occurs at a slow rate. The Coupling of Phosphocreatine Synthesis to Mitochondrial Oxidative Phosphorylation Mitochondria isolated from rat heart are characterized by high CK activity. Under conditions of oxidative phosphorylation, the CK reaction can substantially control the rate of utilization of oxygen (acceptor control of respiration). Figure 6 shows polarographic recordings of the consumption of oxygen by rat heart mitochondria. As is shown in this figure, when creatine and small quan- tities of ATP (0.1 mM) are present, the same rate of consumption of oxygen is observed (curve 2) as in the case of direct introduction of ADP (curve 1). When creatine is present, no transition is observed from state 3 back to state 4 (curve 3) after the introduction and phosphorylation of ADP. These data show that, under these conditions, the mitochondrial CK reac- tion supports a rate of synthesis of phosphocreatine and ADP sufficient to approximate the maximum stimulation of oxidative phosphorylation. Analysis of the oxygraph medium showed that, in experiments 2 and 3, phosphocreatine was synthesized at a rate of 0.9 pmol/min/mg of protein. This corresponded to a ratio (phosphocreatine)/A(0y) = 2.8 umol/ug atom. Thus, under these conditions, the majority of the ATP synthesized in the mitochondria is utilized directly for the synthesis of phosphocreatine. Figure 7A presents a comparison of the maximum rate of the forward mito- chondrial CK reaction (synthesis of phosphocreatine and ADP), determined indepen- dently using a spectrophotometric assay system (10) (curve 1), the rate of ATP synthesis in the mitochondrial matrix (curve 2), and the observed rate of phos- phocreatine appearance in the oxygraph medium during oxidative phosphorylation (curve 3) at various temperatures. We can see that, at any temperature, the maximum rate of the CK reaction is greater than the observed rate of the reac- tion of oxidative phosphorylation. The close agreement of curves 2 and 3 74 Cr+ Cr+ | LI NS A Na ADP (105) (324) (324) 20 swoyobu OZ] | min — FIGURE 6. Polarographic recordings of oxygen consumption by cardiac mitochon- dria. Experiments were performed in the medium indicated in '"Materials and Methods" at 37° C and pH 7.4, protein concentration 0.4-0.5 mg/ml. Curve 1: Stimulation of respiration by ADP (0.3 mM) [recording of respiratory control without creatine (Cr)]. Curve 2: Stimulation of respiration by ATP (0.2 mM) with Cr present (25 mM). Curve 3: Recording of respiration on introduction of ADP (0.3 mM) with Cr present (25 mM). Curve 4: Effect of ATP (0.2 mM) with- out creatine. Rates of 02 consumption are given in ng atoms per minute per mg and are shown in parentheses. suggests that, at any temperature in this range, mitochondrial ATP is predomi- nantly utilized for the synthesis of phosphocreatine. It must be noted that, at this concentration of creatine, the maximally possible rate of phosphocreatine synthesis is v . [Cr] max Ky + [Cr] °° where Kg is the Michaelis constant for creatine, while [Cr]/(Kp + [Cr]) repre- sents the saturation factor of the enzyme with creatine. When [Cr] = 25 mM, this factor is 25/[5 + 25] = 0.83. Considering this, we can see from curves 1 and 3 that, at any temperature, more than half of the maximum possible phospho- creatine is produced. Analysis of these curves in semilogarithmic Arrhenius co- ordinates shows (figure 7B) that the mitochondrial CK reaction is characterized by a comparatively high activation energy (96 kJ/mol), whereas for oxidative phosphorylation this quantity is significantly lower (48 kJ/mol). These tem- perature dependencies indicate that, whereas at low temperatures (up to 33° C), the rate of phosphocreatine synthesis is determined by the CK reaction, at higher temperatures (above the inflection on line 3), the limiting factor may 75 A A s B 0.0 on E A E {1.0 05} I. £ E £ 20} o 4 oOo 8 4 {-05 ~ 2 {00 oof . c o | 5 1.0 + | Oo r o 2 5 2 @ -05¢ 3 5 3 x o , . , 25 30 35 32 3.3 34 tC 1/Tx 10° FIGURE 7. A: Temperature dependence of maximum rate of the forward mitochon- drial creatine kinase (CK) reaction (1), rate of oxidative phosphorylation (2), rates of phosphocreatine synthesis coupled to oxidative phosphorylation (3), and K, for ATP in CK reaction (4). Experiments were performed in the medium indi- cated in "Materials and Methods" at pH 7.4. The maximum rate of forward CK re- action (synthesis of phosphocreatine and ADP) and K for ATP were determined by performing full kinetic analysis as described in reference 10 without oxidative phosphorylation. The rate of oxidative phosphorylation (synthesis of ATP in matrix) was determined from the rate of oxygen consumption, multiplied by the ADP/O ratio in each experiment (2.8-3.0). The rate of phosphocreatine synthesis was determined in the presence of 25 mM creatine and 0.3 mM ATP from the appear- ance of phosphocreatine in the medium surrounding the mitochondria. B: Lineari- zation of functions shown in figure 7A in Arrhenius coordinates: log V versus 1/T. The numbering of lines corresponds to the numbering of curves in figure 7A. be the reactions of oxidative phosphorylation (under the given experimental conditions). The main conclusion from these experiments is that, in the mitochondria, the maximum rates of the forward CK reaction may exceed the observed rates of oxidative phosphorylation (synthesis of ATP in the mitochondrial matrix). How- ever, under conditions of oxidative phosphorylation, a large fraction of the maximum rate of CK is observed. Under these conditions, CK is not present in great excess, as is the case in a homogeneous medium such as the cytosol. Curve 4 of figure 7 shows that the Ky for ATP in the CK reaction also de- pends on the temperature. This variation was not observed for the Ky for crea- tine. These changes must be considered in the analysis of temperature dependences 76 of the rate of the CK reaction, determined with substrate concentrations below those required for saturation (20). Another important characteristic of the CK reaction is the pH dependence. Since the pH of the intracellular medium may vary significantly in various meta- bolic situations, there is significant interest in studying the coupling of phos- phocreatine synthesis with oxidative phosphorylation at various pH values. Figure 8A shows the pH dependence for the maximum rate of the forward CK reaction nN o o > ° 23 o » Rate of PCr production, mmoles min mg" Rate of reaction, umoles-minmg? ° Oo o 60 7.0 8.0 60 7.0 8.0 pH pH FIGURE 8. A: pH dependence of the maximum rate of the forward creatine kinase (CK) reaction (1), rate of oxidative phosphorylation in the presence of 0.3 mM ADP (2), and rate of creatine phosphate synthesis in the presence of 25 mM crea- tine and 0.3 mM ATP, determined by stimulation of respiration by creatine and ATP (3), or by measurements of concentration of phosphocreatine formed in the medium (4). The reaction was performed at 33° C as described in the captions of figures 6 and 7. In calculating curve 3, the rates of oxygen consumption, mea- sured after addition of creatine and multiplied by ADP/0O, are characteristic for a given mitochondrial preparation. B: pH dependence of reaction of phosphocre- atine (PCr) synthesis in mitochondrial CK reaction (33° C). 1 = Synthesis of phosphocreatine coupled with oxidative phosphorylation in the presence of 25 mM creatine and 0.1 mM ATP. 2 = Synthesis of phosphocreatine under the same condi- tions after inhibition of oxidative phosphorylation with oligomycin (5 ug/mg) and addition of 3 mM phosphoenolpyruvate and pyruvate kinase (6 IU/ml) as a nonmitochondrial ATP regenerating system. 3 = Same as 2, after addition of rotenone (20 pM) to inhibit electron transfer and deenergize the mitochondrial membrane. The rate of phosphocreatine synthesis was determined from the change in its concentration in the medium surrounding the mitochondria (11). 77 of the mitochondria (curve 1), the rates of oxidative phosphorylation (curve 2), and the rates of accompanying phosphocreatine synthesis (curves 3 and 4). Vmax (curve 1) is characterized by a sharp dependence on pH. At pH 6-6.5, the rate of oxidative phosphorylation (curve 2) is significantly higher than the observed rate of phosphocreatine synthesis (curves 3 and 4). However, at pH 7.0-7.5, the rate of phosphocreatine synthesis is almost equal to the rate of ATP synthesis in the matrix. Comparison of curve 1 with curves 3 and 4 shows that the maximum activity of the enzyme is observed in this pH range. It was demonstrated earlier that the ATP synthesized in the mitochondria supports higher rates of phosphocreatine synthesis than the ATP present in the medium in the same concentration (10,11). As shown in figure 8B, this is correct at any pH value. If an external ATP-regenerating system is present in the medium, but mitochondrial oxidative phosphorylation is inhibited, the rate of phosphocreatine synthesis is significantly lower (curves 2 and 3) than under conditions of oxidative phosphorylation (curve 1). When oxidative phosphoryla- tion is inhibited, an increase in the pH from 7.0 to 8.0 does not result in an increase in the rate of phosphocreatine synthesis to the values observed when oxidative phosphorylation is present. Therefore, the high rates characteristic for the latter situation cannot be totally explained by local changes in pH under conditions of oxidative phosphorylation (9). Furthermore, when oligomycin and substrates are present, the mitochondrial membrane is energized by the transfer of electrons (curve 2); when rotenone and oligomycin are present together, the membrane is de-energized (curve 3). The agreement of curves 2 and 3 in the figure indicate that the energy status (gradient of concentration of protons and membrane potential) has no influencc on the rate of the CK reaction. The increased rate of phosphocreatine synthesis under conditions of oxida- tive phosphorylation is related to a marked increase in the apparent affinity of the system for ATP, as shown in figure 9. Curve 1 of this figure shows the variation in the rate of phosphocreatine synthesis as a function of ATP concen- tration at 25 mM creatine, when oxidative phosphorylation is inhibited and CK interacts with the ATP in the medium. Linearization of this curve in double- reciprocal plots (figure 9B) yields the apparent Ky for ATP, approximately 0.4 mM, which is close to the value of Kp in the CK reaction, determined independently in kinetic experiments (about 0.6 mM) (10). Thus, when mitochondrial CK inter- acts with ATP in the medium, its behavior is determined by its basic kinetic parameters, as derived from the kinetic equation for the individual enzyme (10). However, under conditions of oxidative phosphorylation (figure 9A, curve 2), the variation in the rate of phosphocreatine synthesis with ATP concentra- tion in the medium is significantly altered and characterized by an apparent Ky of about 40 pM (figure 9B). Thus, if the CK reaction of the mitochondria is supported by oxidative phosphorylation, the addition of small quantities of ATP is sufficient for maximum activation of the phosphocreatine synthesis system. Consequently, under these conditions, the utilization of ATP is significantly higher than in the case when oxidative phosphorylation is inhibited. These data, indicating a significant increase in the turnover rate of adenine nucleotides when CK is coupled with oxidative phosphorylation, are 78 To E Tc E 2 ° on : of 1.0 2 i S 9 4.0 3g 7 ° 3 3 a 0.5} > 5 Z 20 oO > ‘S o [] a 0 0.2 0.4 0 20 40 ATP, mM | /ATP, mM” FIGURE 9. A: Variation in the rate of phosphocreatine (PCr) synthesis in mito- chondrial creatine kinase (CK) reaction as a function of ATP concentration in the medium. The rate of PCr synthesis was determined from the change in its con- centration in the medium surrounding the mitochondria at 37° C, pH 7.4, creatine concentration 25 mM. 1 = Oxidative phosphorylation inhibited by oligomycin (5 ug/mg), with concentration of ATP in medium maintained constant by addition of phosphoenolpyruvate (3 mM) and pyruvate kinase (6 IU/ml). 2 = Coupled with oxidative phosphorylation (oligomycin and regenerating system absent). B: Lin- earization of functions shown in figure 2A in double-reciprocal plots. The points on the lines correspond to points on the curves in figure 9A. explained by the functional coupling of mitochondrial CK with the ATP-ADP trans- locase as suggested earlier (see figure 10) (10,11). If this coupling occurs, a small quantity of ATP added to the medium may induce the CK reaction to form phosphocreatine and ADP, the latter being transferred by the translocase into the matrix for resynthesis to ATP. The ATP arriving from the matrix, because of the location of translocase and CK on the mitochondrial membrane, may inter- act with CK without being liberated into the medium. The ATP is thus used for the synthesis of phosphocreatine and the production of ADP. This starts a new transmembrane flux of the adenine nucleotides and their repeated utilization, and supports the effective synthesis of phosphocreatine by the coupled system. According to this suggestion, the specific placement of CK on the mitochondrial membrane allows it to use ATP synthesized in the mitochondrial matrix very effec- tively (9). When oxidative phosphorylation occurs at the maximum rate, CK also functions at a rate close to Vpgx. Of particular significance for energy metabolism in the myocardium is the fact that preferential utilization of mitochondrial ATP for synthesis of phos- phocreatine in the mitochondrial CK reaction is observed at any concentration 79 FDP —---— Glucose EZ ZN — Lac ADP ATP ADP ATP inner membrane / 7 CK cyt CK cyt PCr r P r ATPases = Z ¥ ¥ ¥ § ATP (Supporting = =\Cr ) Cr | contraction £ zk } : and ion Oo PCr ADP transport 7 \ ATP < ATP 7 | NA aoP ADP & myofibrils mitochondria cytoplasm membranes (SR & SL) FIGURE 10. Mechanisms of phosphocreatine (PCr) synthesis in mitochondria and cytoplasm of heart cells. T = ATP-ADP translocase. CKpit = Mitochondrial CK. CKcyt = Cytoplasmic CK. FDP = Fructose-1l,6-diphosphate. PEP = Phos- phoenolpyruvate. Pyr = Pyruvate. Lac = Lactate. 1,3-DPG = Diphosphoglyceride. Cr = Creatine. of phosphocreatine in the medium. Figures 11A and 11B show the results of two types of experiments in which the effect of ATP on the creatine-stimulated consumption of oxygen by the mitochondria was studied with various concentra- tions of phosphocreatine in the medium. Since the ratio (phosphocreatine)/ (05) = 2.5-2.8 umol/pg-atom, the steady rate of oxygen uptake is directly related to the rate of phosphocreatine formation. Figure 11A shows the results of experiments in which cardiac mitochondria with bound CK activity were used. When phosphocreatine is absent, and when creatine is present at a concentration of 25 mM, an increase in the ATP con- centration in the medium from O to 0.2 mM leads to rapid activation of oxygen consumption and phosphocreatine synthesis, in precise correspondence to the data presented in figure 9. An increase in the concentration of phosphocreatine in the medium decreases the effects of ATP on mitochondrial respiration, par- ticularly at low ATP concentrations. However, at an ATP concentration of 3.5 mM, the effect of phosphocreatine is not very great. At a phosphocreatine con- centration of 26 mM in the medium, the rate of oxygen absorption is inhibited by only 30 percent (figure 11A). Figure 11B shows the results of the same experiments in which, instead of cardiac mitochondria, liver mitochondria that do not contain CK were used. In this case, CK was extracted from cardiac mitochondria and added to the system such that the activity of CK was equivalent to that for the bound conditions 80 Vo,, ng atoms. min” B PCr, mM 0.0 _ 300 3a 'c E 6.9 wn 200 5 34 oO o 26.0 2 100 oN o S 0 1.0 20 3.0 4.0 ATP, mM FIGURE 11. Changes in creatine-stimulated rate of oxygen consumption by mitochondria as a function of ATP concentration in the medium at various con- centrations of phosphocreatine (PCr). Temperature 30° C, pH 7.4. Consumption of oxygen is defined as described in "Materials and Methods.'" The concentration of PCr is shown to the right of the curves. The concentration of creatine equals 25 mM. A: Mitochondria of rat hearts; concentration of mitochondrial protein 0.6 mg/ml; activity of mitochondrial CK 1.6 IU/ml (based on reverse reaction, see reference 10). B: Mitochondria separated from rat liver not containing CK; concentration of protein 1.6-2 mg/ml. The soluble CK was extracted from heart mitochondria by sodium phosphate (reference 9), and added to the system to achieve activity of 1.6 IU/ml (determined by reverse reaction). 81 (figure 11A). In this case, the soluble mitochondrial CK interacts with liver mitochondria through ATP and ADP in the medium. At O phosphocreatine concentra- tion, activation of respiration occurs slowly (maximum activation at ATP = 1.0 mM, which is analogous to case 1 in figure 9). In this case, however, phospho- creatine substantially inhibits respiration. When phosphocreatine is present at 26 mM and with 3.5 mM ATP, the consumption of oxygen is inhibited by 70 per- cent. Thus, the soluble mitochondrial CK is significantly more sensitive to inhibition by phosphocreatine than the same enzyme localized on the mitochon- drial membrane. In parallel experiments, we noted that the kinetic properties of CK, defined in reference 10, do not change as it is extracted from the cardiac mitochondria. Thus, these data are not the reflection of enzyme denaturation, but relate to enzyme localization. In figure 12A, the results of figure 11A are presented in double- reciprocal plots. The apparent Ky for ATP at 0 phosphocreatine concentration for the cardiac mitochondria is 37 uM (see figure 9A). An increase in the con- centration of phosphocreatine leads to competitive inhibition of oxygen consump- tion and phosphocreatine synthesis with respect to ATP. At constant Vp,x, phos- phocreatine increases the apparent K for ATP. In a similar manner, the mutual influences of ATP and phosphocreatine were determined for the data presented in figure 11B (liver mitochondria with soluble CK). Figure 12B shows the extrapolated values for the apparent K; for ATP plotted against the concentration of phosphocreatine for the two systems. We see from this figure that the apparent K; for ATP is approximately six to seven times lower for cardiac mitochondria containing bound CK in comparison to the liver mitochondria, in which CK functions as a soluble enzyme. At physiological con- centrations of phosphocreatine (20-25 mM), the rate of its resynthesis in the mitochondria reaches 1/2 Vpax at about 1.2-1.5 mM ATP in the case of cardiac mitochondria, but requires about 7 mM ATP in the case of liver mitochondria with soluble CK. Based on these data, we calculate that a near-maximum rate of phosphocrea- tine synthesis (if phosphocreatine is present at 25 mM) can be achieved by the cardiac mitochondria with approximately 4-5 mM ATP, corresponding to the physiological concentration of ATP in cardiac cells (21). To achieve an equal rate for soluble CK, the concentration of ATP must be increased to 25-30 mM, significantly greater than its physiological concentration. In the experiments which we have described, the behavior of soluble mito- chondrial CK in the presence of hepatic mitochondria was analogous to its behavior when it was bound to the mitochondrial membrane, but oxidative phos- phorylation was inhibited by oligomycin. In these cases CK interacted with ATP in the medium and the K for ATP without phosphocreatine was 200-400 uM (figures 9 and 11B). Comparison of the sensitivity of CK in various systems to inhibition by phosphocreatine is presented in figure 13. The identical and significant degree of inhibition is observed for soluble CK when either a gly- colytic system or hepatic mitochondria are present. In contrast, for membrane- bound CK, the reaction is only slightly inhibited by phosphocreatine when ATP is present at a concentration over 1 mM and when the reactions are coupled to mitochondrial oxidative phosphorylation. Thus, localization of mitochondrial CK on the inner membrane of cardiac mitochondria leads to a great increase in the 82 o on (0) I / Vo, , ug atoms’. min -4 0 4 8 | / ATP], mM KOPP mM | liver mit + 6.0} Cm NN heart mit 20 Q FIGURE 12. A: Linearization of the functions shown in figure 11A in double- reciprocal plots. The concentration of phosphocreatine (PCr) (mM) is indicated in parentheses. B: Variation in apparent K; for ATP in the reaction of PCr synthesis coupled to oxidative phosphorylation as a function of PCr concentra- tion in the medium. KX; is calculated from data analogous to those presented in figure 12A by extrapolation to the abscissa straight lines corresponding to the various concentrations of PCr. CK, = Mitochondrial creatine kinase. 83 3.3mM ATP 05 = | mM ATP ~ > liver mit.+CK ———_ _— I mMATP 1 1 Too* 0 10 20 PCr, mM FIGURE 13. Effect of phosphocreatine (PCr) on the steady-state rate of its syn- thesis in CK reactions coupled to the pyruvate kinase reaction in cytoplasm (@-@, A-A), with oxidative phosphorylation in liver mitochondria in the presence of soluble CK in mitochondria (---), and in the case of coupling of CK reaction with oxidative phosphorylation in cardiac mitochondria (0-0, A-A). The concen- tration of creatine is 25 mM; the concentration of ATP in the medium is shown in the figure. Temperature 30° C, pH 7.4. vp = Rate without PCr in the system. v = Rate of reaction for the given concentration of PCr. the rate of phosphocreatine synthesis from mitochondrial ATP, and to a decrease in the sensitivity to phosphocreatine product inhibition. Both the increase in the rate of phosphocreatine synthesis and the decrease in inhibition of the CK reaction by its product can be explained by the func- tional coupling of CK with ATP-ADP translocase (see figure 10). The latter could transfer a molecule of ATP from the mitochondrial matrix to the active center of CK (10,11), leading to rapid saturation of the latter with ATP and an increase in the rate of the forward CK reaction. Since ATP and phosphocrea- tine cannot be simultaneously bound to CK, due to overlapping of the terminal phosphoryl groups (10,22), rapid saturation of CK with mitochondrial ATP under conditions of oxidative phosphorylation prevents the simultaneous binding of phosphocreatine. This decreases the rate of the reverse reaction and the inhibi- tion of the forward CK reaction by phosphocreatine. This explanation agrees with the observed competitive interaction of ATP and phosphocreatine (figure 12A) and the decrease in apparent Kp for ATP under conditions of oxidative phosphorylation in cardiac mitochondria (figure 12B). CONCLUSIONS Our results indicate that, in cardiac cells, both the glycolytic reactions for the production of energy (ATP) and mitochondrial oxidative phosphorylation 84 may be coupled to the CK reaction. In the presence of creatine, these reactions result in the effective synthesis of phosphocreatine. The processes for produc- tion of phosphocreatine are represented schematically in figure 10. The mecha- nism for the functioning of CK in these two systems (cytosol and mitochondria) is quite different. In the cytoplasm, where CK is present in soluble form, the enzyme is present in excess in relationship to the rates of glycolysis, and functions in a quasi-equilibrium state. In this homogeneous medium, the CK reaction is quite sensitive to inhibition by phosphocreatine. In contrast, because of the localization of CK on the inner membrane of the mitochondria (on its external side) and the close functional coupling with ATP-ADP translocase, the forward CK reaction is accelerated by the preferential utilization of mitochondrial ATP. In this latter case, the CK can function at a near maximum rate, even with low concentrations of ATP in the medium. This functional coupling also results in a lesser degree of inhibition by phospho- creatine, thus permitting the forward reaction to occur under physiological conditions. When the phosphocreatine is present in cardiac cells in normal concentra- tions and the primary energy substrates are fatty acids (23), these properties of the mitochondrial enzyme allow it to synthesize phosphocreatine at a rate equal to the rate of energy utilization for contraction and transport of ions. Thus, it supports a high and constant level of phosphocreatine in the cells (24). However, under these conditions, the equilibrium of the cytoplasmic CK reaction is displaced in the direction of maintenance of a very low cytoplas- mic ADP concentration, and thus glycolysis cannot occur at a rapid rate. Gly- colysis can be activated only by a sharp drop in the level of phosphocreatine in the cells (in anoxia or ischemia) and a rise in ADP. Thus, the cytoplasmic CK reaction may be an important additional factor in the regulation of glycolysis. The authors are grateful to M. V. Yemelin for calculations performed in the modeling of the coupled CK and pyruvate kinase reactions. 85 10. 11. 12. REFERENCES Mommaerts WFHM: Energetics of muscular contraction. Physiol Rev 49: 427-508, 1969 Newsholm EA, Beis I, Leech AR, Zammit VA: The role of creatine kinase and arginine kinase in muscle. Biochem J 172:533-537, 1978 Scholte HR: On the triple localization of creatine kinase in heart and skeletal muscle cells of the rat: Evidence for the existence of myo- fibrillar and mitochondrial isoenzymes. Biochim Biophys Acta 305:413- 427, 1973 Saks VA, Chernousova GB, Voronkov Yul, Smirnov VN, Chazov EI: Study of energy transport mechanism in myocardial cells. Circ Res 34-35 (suppl I1T1):138-149, 1974 Saks VA, Chernousova GB, Vetter R, Smirnov VN, Chazov EI: Kinetic proper- ties and the functional role of particulate MM isoenzyme of creatine phos- phokinase bound to heart muscle myofibrils. FEBS Lett 62:293-296, 1976 Sharov VG, Saks VA, Smirnov VN, Chazov EI: An electron microscopic histo- chemical investigation of the localization of creatine phosphokinase in heart cells. Biochim Biophys Acta 468:495-501, 1977 Levitsky DO, Levchenko TS, Saks VA, Sharov VG, Smirnov VN: The role of creatine phosphokinase in supplying energy for the calcium pump system of heart sarcoplasmic reticulum. Membrane Biochemistry 2:81-96, 1978 Scholte HR, Weijers PJ, Wit-Peeters EM: Localization of mitochondrial creatine kinase and its use for the determination of sideness of submito- chondrial particles. Biochim Biophys Acta 291:764-773, 1973 Jacobus WE, Lehninger AL: Creatine kinase of rat heart mitochondria. Coupling of creatine phosphorylation to electron transport. J Biol Chem 248:4803-4810, 1973 Saks VA, Chernousova GB, Gukovsky DE, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. Mitochondrial isoenzyme of creatine phosphokinase: Kinetic properties and regulatory action of Mg2t ions. Eur J Biochem 57:273-290, 1975 Saks VA, Lipina NV, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. The functional coupling between mitochondrial creatine phos- phokinase and ATP-ADP translocase: Kinetic evidence. Arch Biochem Biophys 173:34-41, 1976 Gudbjarnason S, Mathes P, Ravens KG: Functional compartmentation of ATP and creatine phosphate in heart muscle. J Mol Cell Cardiol 1:325-339, 1970 86 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, Saks VA, Rosenshtraukh LV, Smirnov VN, Chazov EI: Role of creatine phos- phokinase in cellular function and metabolism. Can J Physiol Pharmacol 56:691-706, 1978 Saks VA, Rosenshtraukh LV, Undrovinas AI, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. Intracellular creatine content as a regulatory factor of frog heart energetics and force of contraction. Biochem Med 16:21-36, 1976 Rosenshtraukh LV, Saks VA, Undrovinas AI, Chazov EI, Smirnov VN, Sharov VG: Studies of energy transport in heart cells. The effect of creatine phosphate on the frog ventricular contractile force and action potential duration. Biochem Med 19:148-164, 1978 Vassort G, Ventura-Clapier R: Significance of creatine phosphate on the hypodynamic frog heart. J Physiol (London) 269:86P-87P, 1977 Kupriyanov VV, Seppet EK, Saks VA: [Creatine phosphate synthesis coupled to the glycolytic reactions in heart cells cytosol] (Rus). Biokhimiia 43:1468-1477, 1978 Lehninger AL: [Biochemistry] (Rus). Moscow, Mir, 1974, p 377 In English: Biochemistry: The Molecular Basis of Cell Structure and Function, 2nd ed. New York, Worth, 1975 Kupriyanov VV, Seppet EK, Emelin IV, Saks VA: [The steady state kinetics of pyruvate kinase from muscle] (Rus). Biokhimiia 44:104-115, 1979 Watts DC: Creatine kinase (adenosine 5'-triphosphate-creatine phospho- transferase). In The Enzymes, vol 8, edited by PD Boyer. New York, Academic Press, 1973, pp 383-455 Kubler W, Katz AM: Mechanism of early "pump" failure of the ischemic heart: possible role of adenosine triphosphate depletion and inorganic phosphate accumulation. Am J Cardiol 40:467-471, 1977 Morrison JF, Cleland WW: Isotope exchange studies of the mechanism of the reaction catalyzed by adenosine triphosphate: creatine phosphotransferase. J Biol Chem 241:673-683, 1966 Neely JR, Morgan HE: Relationship between carbohydrate and lipid metabo- lism and the energy balance of heart muscle. Ann Rev Physiol 36:413-459, 1974 Williamson JR, Ford C, Illingworth JA, Safer B: Coordination of citric acid cycle activity with electron transport flux. Circ Res 38 (suppl I):39-51, 1976 87 REGULATION OF MECHANICAL AND ELECTRICAL ACTIVITY OF THE HYPODYNAMIC MYOCARDIUM BY CREATINE PHOSPHATE L. V. Rozenshtraukh, V. A. Saks, V. M. Sharov, I. A. Yuryavichyus, and E. I. Chazov INTRODUCTION The force of contraction of the heart muscle can be regulated by effects on the intracellular creatine phosphokinase systems (1,2). Activation of the mitochondrial creatine phosphokinase reaction by creatine increases the force of contraction of a strip of frog ventricle under conditions such that the in- tracellular content of creatine and the force of the contraction are decreased after long-term perfusion of the muscular strip of the heart with normal Ringer's solution. In parallel with the decrease in the force of contraction and intra- cellular content of creatine, there is a decrease in the intracellular content of creatine phosphate (1,2). These data, recently confirmed by Vassort and Ventura-Clapier (3), indicate that the phenomenon of decreased myocardial con- tractile force, as a result of long-term perfusion, and referred to in the literature as the hypodynamic state (4,5), results from a decrease in the in- tracellular content of creatine phosphate. Thus, these physiological experi- ments (1-3) confirm that energy transport in myocardial cells is performed by creatine phosphate (6,7). In this report, we present data from a further study of the role of crea- tine phosphokinase systems in regulating the contractility of the myocardium. Primary attention is given to studying the effect of creatine phosphate on the force of contraction, the transmembrane action potential, and the slow calcium current of the hypodynamic frog atrial and ventricular myocardium. MATERIALS AND METHODS Recording the force of contraction and transmembrane action potentials. Experiments were performed on small strips of the atrium and ventricles of the frog species, Rana ridibunda and Rana temporaria. The strips were placed in a perfusion chamber and perfused with Ringer's solution of the following composition (mM): Nat 112.8; K+ 2.7; cat 1.8; C1™ 117; HCO3 1.8; pH 7.4. After perfusion in normal Ringer's solution, the strips were perfused in a From the All-Union Cardiology Research Center, USSR Academy of Medical Sci- ences, Moscow, USSR. 89 solution containing 10, 20, 50, and 70 mM creatine phosphate. During the experiment, the strips were stimulated by square-wave pulses (length 3-5 msec; intensity of stimulus 1.5-2.0 times the threshold). The mechanical activity and action potential were recorded by means of a force transducer and glass microelectrodes. Recording of slow inward current. Ionic currents were recorded from iso- lated Rana ridibunda atrial trabeculae. The trabeculae (75-120 ym in diameter and 3-5 mm in length) were placed in a perfusion chamber consisting of 5 sec- tions (8). The central test section (200 pm wide) was isolated by sucrose bridges (400 pm wide) from the 2 outer sections; the end sections were filled with a depolarizing KCl solution. The walls between the sections of the per- fusion chamber were lubricated with vaseline. Low-resistance (less than 5 kQ) extracellular Ag-AgCl electrodes were used to record the membrane potential and current. The test section of the chamber was perfused with a normal Ringer's solu- tion; the end sections of the chamber were filled with Ringer's solution con- taining 140 mM KCl. The sucrose bridge sections were perfused with a sucrose solution (240 mM) which had been purified using ion exchange columns. The membrane potential was recorded using special amplifiers (Dagan 8500, U.S.). The command signals were supplied by an isolated stimulator block (Digipulser 830, Isopulser 850, WPI, U.S.). The transmembrane currents were recorded from the screen of an oscilloscope (Tektronix D13, U.S.) with a polaroid camera, or were recorded on a magnetic tape recorder (Hewlett Packard, 3964A, U.S.) for further analysis. Before the slow inward current was recorded, the atrial trabeculae were treated with tetrodotoxin (10-6 g/ml) to block the rapid inward current. All experiments were conducted at room temperature. Electron-microscopic studies. For the electron-microscopic studies, strips of myocardial tissue were taken and perfused with normal Ringer's solu- tion until the force of contraction decreased to 30 to 40 percent of the initial value (1,2). The control material was obtained from an intact frog heart. The preparation was fixed with a 2.5 percent solution of glutaraldehyde in 0.1 M cocadylate buffer (pH 7.4), and was subsequently postfixed in a 1 percent solu- tion of osmic acid in 0.2 M cocadylate buffer and placed in an epon-araldite mixture. The sections were prepared by ultratome (LKB, Sweden), contrasted with uranyl acetate and lead citrate, and studied under the electron microscope (JEM, 100 V, Japan). Substances used. We used creatine phosphate from the Reanal Co., tetro- dotoxin from the Sankyo Co., and sodium cyanide from the Merck Co. RESULTS Ultrastructure of hypodynamic myocardial myocyte. Long-term perfusion of frog heart strips leads to the hypodynamic state. The force of contraction of the myocardium can be restored to its initial level by adding creatine to 90 the perfusate, thereby restoring the initial level of creatine phosphate in the cells (1,2). For a definitive interpretation of the results, we must determine whether the development of the hypodynamic state was related to changes in the ultrastructure of the cells. Toward this end, control morphological studies were performed, in which we compared the ultrastructure of muscle cells from intact frog myocardium to that of cells from ventricular strips after 8 hours of perfusion with normal Ringer's solution. When these two groups of cells were compared, no differences were found in the structure of the sarcomeres and the mitochondria, or in intercellular contacts. Furthermore, after 8 hours perfu- sion, the cells also retained a significant quantity of glycogen (figure 1). The picture of cell structure agreed with data in the literature (9,10). Thus, the decrease in contractile force (hypodynamic state) of the preparation was not related to disruption of the ultrastructure of the cells. These data sug- gest that the effect of creatine (1,2) is related to its effect on the intra- cellular energy system. If the increase in force of contraction with creatine is explained by an increase in the rate of synthesis of creatine phosphate in the mitochondria, leading to a change in the intracellular content of creatine phosphate and, in the final analysis, to a change in ATP turnover near the myofibrils (11), addi- tion of the energy substrate, creatine phosphate, to the perfusate instead of creatine should lead to a significant increase in the force of contraction. Naturally, an increase in the force of contraction by this mechanism is possi- ble only if the myocardial cells are permeable to creatine phosphate. Permeability of hypodynamic myocardial cells to creatine phosphate. Experiments were performed on an isolated heart preparation, the ventricle of which worked against a hydrostatic pressure of 50 mm of water. After 40 minutes of perfusion with Ringer's solution containing creatine phosphate (10, 20, and 40 mM), its tissue content was determined and the intracellular concentration was calculated (1). The data of these experiments are presented in figure 2 and show that the intracellular concentration of creatine phosphate varies linearly with its concentration in the solution. Effect of creatine phosphate on the force of contraction and transmembrane action potential. The perfusion of the muscle strip with a solution containing 10 mM creatine phosphate caused an increase in the force of contraction during the first 2.5 minutes of perfusion (figure 3A). Subsequently, the force of con- traction remained at a high level throughout the entire 10 minutes of perfusion (figure 3A, curve 2). Subsequent washing with normal Ringer's solution was accompanied by a gradual decrease in the force of contraction which, by the 25th minute of washing, became equal to its initial value (figure 3A, curves 3 and 4). An increase in the concentration of creatine phosphate in the perfusate to 70 mM led to a more complex effect on the force of contraction. During the first minute of perfusion, as in the case of a low concentration of creatine phosphate, the force of contraction increased (figure 3B, compare curves 1 and 2). However, longer perfusion was accompanied by a decrease in the force of contraction which, by the 10th minute, became less than its initial value (fig- ure 3B, curve 3). The suppression of the force of contraction ended when the specimen was washed with a normal Ringer's solution. Furthermore, the force 91 oid FIGURE 1. Ultrastructure of myocyte in a strip of frog ventricle after 8 hours perfusion with normal Ringer's solution. A: Contractile apparatus of cell; magnification x 40,000. B: Accumulation of glycogen (Gl) in myocardial cell; magnification x 40,000. 92 =z E c S121 I] ° = ze $2107 oe £3 £8.01 5 ° oC 6.01 o 3 3 404 : T T T T T — = 10 20 30 40 50 Creatine phosphate in perfusate, mM FIGURE 2. Variation in intracellular concentration of creatine phosphate as a function of its concentration in perfusate. Graph shows mean values and standard deviations obtained in four to eight experiments. of contraction increased during the first 5 minutes of washing and became greater than at the initial stimulation by creatine phosphate (figure 3B, compare curves 4 and 2). The effect of various concentrations of creatine phosphate on the force of contraction of a single preparation can be demonstrated still more clearly by continuous recording of mechanical activity (figure 4). Figure 4A shows a con- tinuous recording of the force of contraction of a strip of muscle perfused with Ringer's solution, containing 10 mM and 20 mM creatine phosphate. At these con- centrations, the creatine phosphate rapidly and significantly increased the force of contraction, which remained at the same level for the entire 10 minutes. During the period of washing with normal Ringer's solution, the force of con- traction was restored by 30 to 40 minutes. As the concentration of creatine phosphate was increased to 50 mM or 70 mM, the contractile response varied in a more complex manner; the increase in sodium concentration was found to have a definite role in the reaction of the strip of myocardium. Figure 4B shows the influence of creatine phosphate in a concentration of 50 mM; the concentration of sodium in the Ringer's solution was maintained at the normal level (113 mM). We can see from figure 4B that, in the first minutes after addition of creatine phosphate, the force of contraction increased; how- ever, it then spontaneously decreased. During the period of washing with nor- mal Ringer's solution, the force of contraction rapidly increased, and then gradually began to decrease to its initial level. Perfusion of the same preparation with Ringer's solution containing 50 mM creatine phosphate and almost double the concentration of sodium (213 mM), led 93 FIGURE 3. Effect of various concentrations of creatine phosphate on the mechani- cal activity of myocardial frog ventricle strip. A: 1. Initial contraction after 8 hours perfusion with normal Ringer's solution. 2. After 2.5 to 10 minutes perfusion with Ringer's solution containing 10 mM creatine phosphate. 3 and 4. After 5 and 25 minutes washing with normal Ringer's solution. B: 1. Initial force of contraction after 5 hours perfusion with normal Ringer's solution. 2 and 3. After 1 and 10 minutes perfusion with Ringer's solution containing 70 mM creatine phosphate. 4, 5, and 6. After 5, 45, and 50 minutes washing with normal Ringer's solution. 94 10 mM CP CP washout 20 mM CP CP washout ey 4 - ee 3 : FO — 1 A mo 50 mM CP CP washout +e ' v me ef meee « pes eee pe Ce | eds me ea FIGURE 4. Effect of creatine phosphate (CP) on the force of contraction of a strip of frog ventricle. A: Continuous recording of the force of contraction after preliminary perfusion for 3.5 hours with normal Ringer's solution. Arrows mark the beginning and end of perfusion with Ringer's solution containing 10 mM and 20 mM CP, respectively. B: Change in the force of contraction of the strip of frog ventricle during perfusion with Ringer's solution containing 50 mM CP. Throughout the entire period of perfusion, sodium concentration was maintained at a constant level of 113 mM. C: Change in the force of contraction of a strip of frog ventricle during perfusion with Ringer's solution containing 50 mM CP. During testing of CP, total sodium concentration in the solution was 213 mM. Recordings A, B, and C were produced during one experiment. The time of perfusion with normal Ringer's solution between recordings A and B was 40 minutes, and between recordings B and C, 45 minutes. to an initial decrease in the force of contraction, followed by a slight in- crease in the contractile response by the end of the 15th minute of the test period (figure 4C). A rapid increase in the force of contraction occurred during the period of washing with normal Ringer's solution not containing creatine phosphate. These data show that high concentrations of sodium do not allow the positive inotropic effect of creatine phosphate to be manifested, possibly due to the influence of sodium on ion transport through the cell mem- brane (12). Therefore, in studying the effect of creatine phosphate as the sodium salt on the force of contraction, one should maintain a normal level of sodium concentration in the Ringer's solution. 95 The data presented in figure 4 show that, during the course of long-term perfusion with a normal Ringer's solution not containing creatine phosphate, the force of contraction of a muscle strip decreases with time. However, addi- tion of various concentrations of creatine phosphate to the perfusate increases the force of contraction in all cases; the maximum force of contraction attained in recordings A and C of figure 4 corresponds approximately for various concen- trations of creatine phosphate. Thus, the effect of creatine phosphate becomes clearer with the period of perfusion with normal Ringer's solution, leading to a decrease in the intracellular concentration of creatine and creatine phos- phate (1,2). Effect of creatine phosphate on the length of the transmembrane action potential. The contraction of heart muscle is preceded by an action potential, during which a calcium current develops that is dependent on the action potential and which plays an important role in the activation of contraction (13-17). Therefore, it was of interest to study the extent to which a change in the force of contraction of the muscle strip under the influence of creatine phosphate is related to a change in the electrical activity of the muscle fibers. For this purpose, special experiments were performed in which the mechanical activity of preparations and transmembrane action potential were recorded simultaneously. Figures 5 and 6 show changes in the force of contraction and action potential under the influence of 20 mM creatine phosphate. As can be seen from these figures, in addition to the increase in the force of contraction by creatine phosphate, a significant increase was observed in the duration of the action potential. Upon subsequent washing with normal Ringer's solu- tion, a simultaneous decrease occurs both in the force of contraction and in the duration of the action potential, which returned to their initial values. Thus, creatine phosphate in a concentration of 20 mM causes a parallel change in both the duration of the action potential and the force of the contraction. These data indicate that creatine phosphate has a significant effect on pro- cesses occurring in the membrane of heart cells related to the phase of its repolarization. An increase in creatine phosphate concentration in the perfusate to 70 mM resulted in different changes in the duration of the action potential and force of contraction of the heart muscle. As is shown in figure 7, during the initial period of perfusion with Ringer's solution containing 70 mM creatine phosphate, an increase was noted both in the duration of the action potential and in the force of contraction of the strip of myocardium. However, after 5 minutes of perfusion with creatine phosphate, the force of the contraction decreased, whereas the duration of the action potential continued to increase. During washing of the preparation with the normal Ringer's solution, the force of con- traction increased significantly, whereas the duration of the action potential remained unchanged or decreased. Thus, in the case of high concentrations of creatine phosphate, the duration of the action potential and the force of the contraction changed in opposite directions. Whereas the duration of the action potential only increased in the presence of creatine phosphate and decreased when it was subsequently washed out, the changes in the force of the contrac- tion were more complex. 96 50mV 150mg 300 msec FIGURE 5. Effect of 20 mM creatine phosphate on the form of transmembrane action potential and individual cycles of contraction of a strip of frog ven- tricle. A: Initial force of contraction (lower curve) and transmembrane action potential (upper curve) after 3.5 hours perfusion with normal Ringer's solution. B and C: After 5 and 10 minutes perfusion with Ringer's solution containing 20 mM creatine phosphate. D and E: After 5 and 10 minutes washing with normal Ringer's solution. In order to compare the effect of high concentrations of creatine phos- phate, the dotted line in figure 7 shows the effect of creatine (92 mM) on the force of contraction of the heart muscle (the data are taken from refer- ence 2). In contrast to creatine phosphate, high concentrations of creatine cause a rapid drop in the force of contraction from the moment they are applied. Since creatine in high concentrations has an inhibitory effect on the force of contraction (2), the decrease in the force of contraction after its initial increase can be explained by the accumulation of creatine within the cell in significant concentrations, due to its rapid liberation from the creatine phosphate. Effect of creatine phosphate on the force of contraction and transmembrane action potential of ventricular myocardium treated with cyanide. As already noted, the force of contraction of the frog ventricle myocardium is most clearly 97 @ 1.6 4 [«}) £ O o 1.4 — g 6 [0 = - 5 12 > [0 2 3 1.0 1 1 I LL Q 10 20 30 40 time of perfusion, min 1 o> Lb a | 1 I 0 with CP without CF FIGURE 6. Effect of 20 mM creatine phosphate (CP) on the duration of the action potential (e) and on the force of contraction (0) of a strip of frog ventricle. The duration of action potential was measured at a level of 90 percent repolari- zation. Relative values of the parameters are normalized to their initial values. The initial duration of the action potential was 430-480 msec. The graph shows the mean values and standard deviations of four to five experiments. regulated with creatine and creatine phosphate, whereas the intracellular con- centration of these substances is decreased by washing as a result of long-term perfusion with normal Ringer's solution (1,2). These experiments lead to the conclusion that the force of contraction and duration of the action potential depend on the intracellular concentration of creatine phosphate, which is the primary mediator of energy transfer from the mitochondria to the myofibrils and the cell membranes. It follows from this conclusion that, if the formation of creatine phosphate in the mitochondria is fully inhibited, the myocardial fibers should not contract and administration of creatine phosphate to the cell should restore their contractile capacity. The permeability of the surface membrane of the myocardial fibers of the frog to creatine phosphate allows its intracel- lular concentration to be increased by addition of creatine phosphate to the perfusion solution. The validity of this argument follows from the results of an experiment illustrated in figure 8A and table 1. After addition of sodium cyanide (1 mM) to the perfusion solution, the force of the contraction decreased rapidly as a result of inhibition of oxidative phosphorylation in the mitochondria. Sub- sequent addition of creatine phosphate in a concentration of 20 mM in the pres- ence of cyanide led to a rapid increase in the force of contraction from 98 N o J on l TTT TTT relative value of parameter 1.0 ~~ 0 20 30 40 \ a ' a 054° / time of perfusion, min Hy 0 wl | vy with CP without CP FIGURE 7. Effect of 70 mM creatine phosphate (CP) on the duration of the action potential (e) and on the force of contraction (0) of a strip of frog ventricle. See figure 6 for details of experiment. ® = Influence of 70 mM creatine on the contractile force (data taken from reference 2). 9 + 3 percent to 50 + 8 percent (M * standard deviation, n = 20) of the initial value. The increase in contractile force can be interrupted by subsequent per- fusion of the muscle strip with Ringer's solution without creatine phosphate, but containing 1 mM sodium cyanide (figure 8A). Figure 8B shows a recording of the action potentials and individual con- traction cycles before the addition of cyanide (curve 1), after 10 minutes of perfusion with Ringer's solution containing cyanide (curve 2), and after 10 min- utes of perfusion with a solution containing creatine phosphate (20 mM) and cyanide. The perfusion of the muscle strip with the solution containing cyanide caused rapid depression not only of the force of contraction, but also of the duration of the action potential, which agrees with the data of Prasad (18). The addition of creatine phosphate in the presence of the cyanide led to a clear restoration of the shape of the action potential (upper curves of figure 8B). These results demonstrate the important role of creatine phosphate in the regu- lation both of the force of contraction and of ion transport, which defines the plateau phase of the action potential. Effect of creatine phosphate on the force of contraction and transmembrane action potential of frog atrial muscle. Since all of the previous experiments were performed on the myocardium of the frog ventricle, it seemed interesting to study the effect of creatine phosphate on atrial tissue, in which one can compare, in the same preparations, changes in the force of contraction and 99 A -1mM NaCN — - 20mM CP 40mVy 600 mg 300 msec FIGURE 8. Effect of creatine phosphate (CP) on the transmembrane action potential and the force of contraction of a strip of ventricle treated with sodium cyanide (NaCN). A: Continuous recording of the force of contraction. The arrows show the time of administration of NaCN (1 mM) and CP (20 mM). B: Recording of the action potential (upper curves) and the cycles of contraction (lower curves) before administration of NaCN (curve 1), after 10 minutes perfusion with Ringer's solution containing NaCN (1 mM) (curve 2), and after 10 minutes perfusion with Ringer's solution containing 1 mM NaCN and 20 mM CP (curve 3). 100 TABLE 1. Effect of 1 mM NaCN on the Force of Contraction and Intracellular Creatine Phosphate and ATP Content. Data of Five Experiments. Normal Ringer's Solution; 20 Min Adaptation; M + SD Ringer's Solution + 1 mM NaCN; 20 Min Action; M + SD Force of contraction (%) 100 9 + 3 Creatine phosphate (mM) 1.1 + 0.1 0.25 + 0.1 ATP (mM) 1.45 + 0.15 1.25 + 0.15 transmembrane action potentials under the influence of creatine phosphate with a change in the slow, inwardly directed calcium flow. The experiments in this series were performed on a strip of atrium weighing 10-15 mg. Strips of atrial tissue were obtained from animals which had been kept for several months in a cooler (4° C). After 5 to 6 months, the intracellular content of creatine phosphate in the atrial fibers spontaneously decreased from 6 * 0.7 mM to 1.1 + 0.1 mM. These hypodynamic hearts were used in the experiments described below. Figure 9A shows the change in the force of contraction of an atrial strip perfused with Ringer's solution containing 20 mM creatine phosphate. As can be seen in the figure, within 5 minutes the creatine phosphate caused a three- fold increase in the force of contraction as compared to the control values after which it spontaneously decreased. The force of contraction increased once again after the creatine phosphate was removed from the perfusate, pos- sibly as a result of washing out of the excess intracellular creatine. During perfusion with normal Ringer's solution, the force of contraction decreased, and after 30 minutes it reached its initial value. Figures 9B through 9D show the action potential (upper curves) and indi- vidual cycles of contraction (lower curves), obtained before addition of crea- tine phosphate (curve 1), at the time of its maximal effect (curve 2), and during the increase in the force of contraction caused by perfusion of the muscle strip with normal Ringer's solution (curve 4). (The time of recording the individual cycles of contraction and action potentials that are shown in figures 9B through 9D is indicated by the asterisks on the continuous recording of contraction given in figure 9A.) Figure 9B shows that, with maximum in- crease in the force of contraction, there was a significant increase in the duration of the action potential and an increase in the level of the action potential plateau. In the case of a spontaneous decrease in contraction (lower curve 3 of figure 9C) in the presence of creatine phosphate, the level of the action potential plateau remained high (upper curves in figure 9C). As the creatine phosphate (and possibly the excess of intracellular creatine) washed out, the force of the contraction increased, but the duration and level of the 101 40mvV} 60mg 100 msec FIGURE 9. Effect of creatine phosphate (CP) (20 mM) on the force of contraction and transmembrane action potential of a strip of frog atrium. A: Continuous recording of the force of contraction of a strip of frog atrium. The arrows show the time of administration of CP. The asterisks show the time of recording of individual cycles of contraction and transmembrane action potentials in B, C, and D. B, C, and D: Transmembrane action potentials (upper curves) and individual contraction cycles (lower curves) before administration of CP (curves 1), 3 minutes after administration of CP (curves 2), 9 minutes after adminis- tration of CP (curves 3), and after 4 minutes washing with normal Ringer's solu- tion (curves 4). 102 FIGURE 9 (continued) 103 action potential plateau decreased (which is seen by comparing curves 3 and 4 of figure 9D). Thus, the experimental data given in figure 9 show that creatine phosphate has an effect on both the action potential and the force of contraction. A similar effect of creatine phosphate was obtained in strips of frog ventricle myocardium (see above). The duration of the action potential and the level of the plateau increased continually in the presence of creatine phosphate and decreased after it was removed. The force of contraction, however, changed in a more complex manner, and the spontaneous decrease in the force of contraction following the addition of creatine phosphate may possibly reflect the accumu- lation of creatine in the region of the myofibrils. Varying changes in the force of contraction and duration of action potential, obtained when creatine phosphate acted on the atrial or ventricular tissue, apparently result from two different sites of action of creatine phosphate on the heart cells: on the myofibrils and on the plasma membrane. Since the inward calcium current through the cell membrane takes part in the formation of the action potential plateau, one can assume that the rise in the level of the action potential plateau in the presence of creatine phosphate results partially from the in- crease in the inward calcium current. In order to confirm this assumption, we took direct measurements of the calcium current in the next series of experiments. Effect of creatine phosphate on slow inward calcium current. These ex- periments were conducted in the following sequence: After the atrial trabecula was treated with tetrodotoxin to block the rapid inward flux of sodium, we determined the membrane potential which initially corresponded to the maximal slow inward current. After this, Ringer's solution containing creatine phos- phate (20 mM) was added to the test section of the chamber, and the calcium current was recorded once more. Curve 1 in figure 10A is a recording of the slow inward current before the action of creatine phosphate. After 1.5 minutes (curve 2) and 2.5 minutes (curve 3) of perfusion with Ringer's solution con- taining 20 mM creatine phosphate, the maximum value of the slow inward calcium current was significantly increased. Similar results were produced in eight analogous experiments, in which the slow current under the influence of cre- atine phosphate increased by more than a factor of 3 in comparison to its ini- tial value. Figure 10B shows that the slow calcium current, increased by creatine phosphate, continues to increase with an increase in extracellular calcium concentration. The sensitivity of the slow inward current to extracellular calcium concentration is a characteristic feature of the calcium current. It must be emphasized that, in the presence of creatine phosphate, a gradual depolarization of the cell membrane is observed (5-7 mV in the first 3 to 5 minutes) (see figure 10). This depolarization may result from a de- crease in the outward Kt current. Therefore, our measurements of calcium current were performed only for the first 3 to 5 minutes of creatine phosphate application. 104 20 msec FIGURE 10. Effect of creatine phosphate on slow inward current of calcium in frog atrial trabeculae. A: 1. Slow inward current before administration of creatine phosphate. 2 and 3. After 1.5 and 2.5 minutes perfusion with Ringer's solution containing 20 mM creatine phosphate. B: 3. Same as 3 in A. 4. After 2.5 minutes perfusion with Ringer's solution containing 20 mM creatine phosphate and 5.4 mM calcium. DISCUSSION Experimental data obtained in this study agree with the conclusion reached earlier (1,2) that creatine phosphate can control the contraction of heart muscle. Furthermore, the regulatory role of creatine phosphate in the transfer of calcium through the surface membrane of heart cells into the myoplasm has been directly demonstrated. 105 Recently, biochemical and physiological data have been obtained that indi- cate to us that creatine phosphate exerts metabolic control over contraction of the heart by acting on a number of processes which play an important role in the regulation of contractile function. The central role of creatine phosphate results from the existence of a creatine phosphate pathway of energy transport (6,7). The hypothesis that creatine phosphate transports energy from the mito- chondria to the point of utilization in the muscle cell was formulated by Bessman (19), Gudbjarnason et al. (20), and other authors (21,22). This hy- pothesis was experimentally confirmed in studies of creatine phosphate synthe- sis in heart mitochondria (23,24), in studies of the localization and functional role of isoenzymes of creatine phosphokinase in myocardial cells (11,25,26), and also in in vivo experiments on the effect of creatine on the energy status and contractile force of the myocardium (1,2,11). The isoenzymes of creatine phosphokinase (MM) were detected in all cell structures which actively utilize energy--the myofibrils, the membrane of the sarcoplasmic reticulum, and the plasma membrane (2,13,25-27). These isoenzymes of creatine phosphokinase support effective utilization of creatine phosphate for the resynthesis of ATP from ADP, formed as a result of ATPase reactions. Since creatine phosphate controls the rate of ATP regeneration due to the creatine phosphokinase reactions, these reactions play a central role support- ing the effect of creatine phosphate within the cell. Myofibrillar creatine phosphokinase plays the most significant role in supporting contractions through the creatine phosphate pathway. Thus, in spite of the fact that the molecular mechanism of contraction includes hydrolysis of ATP by the myosin ATPase (28), muscle contraction is always accompanied by a decrease in the level of creatine phosphate with a constant level of ATP (29). The use of ATP in muscle contraction was observed only after inhibition of creatine phosphokinase (30). In 1954, Perry (31) showed that, in the pres- ence of creatine phosphokinase and creatine phosphate, trace quantities of ADP (3 uM) cause contraction of glycerinized myofibrils; this degree of contraction was not obtained by addition of more than 60 uM ATP to the myofibrils. Strong binding between myosin and creatine phosphokinase was found in vitro (32-35). It has been shown that, in skeletal muscle, creatine phosphokinase is an inte- gral part of the structure of the M line of the sarcomere (34). The important role of myofibrillar creatine phosphokinase in supplying the energy of contrac- tion is illustrated also by studies in which it was shown that, after inhibition of oxidative phosphorylation by cyanide, the intracellular content of creatine phosphate decreases and the force of contraction drops (see table 1). However, the administration of creatine phosphate can support contraction of heart mus- cle at a high level (figure 8A). The results obtained in the present study and those of earlier published data (1,2), have shown that, in addition to its important role in supplying energy to the contractile apparatus, creatine phosphate takes part in the regu- lation of the entry of calcium ions into the myoplasm. The effect of creatine phosphate on the slow inward current was demonstrated in direct experiments (figure 10). We know that calcium current is an important component of electro- mechanical coupling in the heart (13,14,36), and that it depends on the metabolic 106 state of the cell (36-39). Sperelakis and Schneider (39) stated the view that the protein of the slow calcium channel of the membrane surface is phosphory- lated by cAMP-dependent protein kinase using ATP as an energy donor. Phosphory- lation of protein is necessary for maintenance of the channel in the active state, i.e., in the state in which the potential across the membrane is capable of controlling the permeability of calcium ions (38,40). This hypothesis was developed based on studies of the effect of catechol- amines (35,41,42). In this case, one can assume that the increase in the in- flux of calcium through the cell membrane results from activation of adenylate cyclase, an increase in the content of cAMP, and activation of cAMP-dependent protein kinase (35,41,42). Thus, new slow channels develop in the presence of catecholamines. This hypothesis has been further confirmed by the studies of Reuter and Scholtz (38,40). Detailed analysis of the slow inward current under the influence of catecholamines has shown that, when they are present, the equilibrium potential of the current does not change, since the ion selectivity of the slow channels does not change. Based on data from the present study, one can assume that the protein kinase is related to creatine phosphokinase which is localized on the plasma membrane of heart cells (11,27). The protein kinase system can be linked by means of this creatine phosphokinase through creatine phosphate with the intra- cellular metabolic reactions. Figure 11 shows schematically one possible rela- tionship between protein kinase and creatine phosphokinase on the cell membrane. In accordance with this system, the creatine phosphokinase regenerates ATP for the protein kinase reaction and, thereby, the influence of creatine phosphate on the slow inward current is explained by its capability to determine the rate of ATP regeneration. As a result, the intracellular concentration of creatine phosphate can influence the number of slow cation channels in the active state. The increase in the slow inward calcium flow contributes to an increase in the duration of the action potential in the presence of creatine phosphate. However, the duration of the action potential is determined not only by the calcium cur- rent, but also by other ion currents (16,43) which may be changed by creatine phosphate. Future studies will address the question of the possible effect of creatine phosphate on processes of electroneutral entry of calcium into the cells (12) and on the energy-dependent exit of calcium from the cells (44). One important conclusion which can be drawn on the basis of the present study is that the decrease in creatine phosphate content during anoxia (or ischemia) leads not only to a decrease in the energy supply of the contractile apparatus of the cell, but also to a decrease in the entry of calcium through the surface membrane into the cell. The latter helps to explain why the myo- fibrils of heart muscle are in the relaxed state and do not form rigor complexes when contraction is interrupted due to a shortage of energy in anoxia (45). 107 plasma — membrane ch.p. chp’ ATP ~ ADP FIGURE 11. Diagram of possible interaction between phosphorylation of proteins in slow cation channel and creatine phosphokinase reactions in heart cells. P.K. = Protein kinase. CPK = Creatine phosphokinase of membrane. Ph-ase = Phosphatase. ch.p. = Nonphosphorylated protein of channel. ch.p.* = Phosphory- lated protein of channel. ATP = Adenosine triphosphate. ADP = Adenosine di- phosphate. Cr = Creatine. CP = Creatine phosphate. cAMP = Cyclic adenosine monophosphate. - = Movement of caZt through slow channel. 108 10. 11. REFERENCES Rozenshtraukh LV, Saks VA, Undrovinas AI, Smirnov VN, Chazov EI: Mechanism of energy transport in cardiac cells. II. Regulation of myocardial con- tractility via influence on intracellular energy transport. In Proceedings of the Second US-USSR Joint Symposium on Myocardial Metabolism, Sochi, USSR, May 28-30, 1975. Washington, D.C., U.S. Department of Health, Edu- cation, and Welfare, Public Health Service, National Institutes of Health, DHEW Publication No. (NIH) 77-924, pp 163-174 Saks VA, Rozenshtraukh LV, Undrovinas AI, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. Intracellular creatine content as a regulatory factor of frog heart energetics and force of contraction. Biochem Med 16:21-36, 1976 Vassort G, Ventura-Clapier R: Significance of creatine phosphate on the hypodynamic frog heart. J Physiol (London) 269:86P-87P, 1977 Chapman RA, Niedergerke R: Effects of calcium on the contraction of the hypodynamic frog heart. J Physiol (London) 211:389-421, 1970 Clark AJ: The action of ions and lipids upon the frog's heart. J Physiol (London) 47:66-107, 1913 Saks VA, Rozenshtraukh LV: Current problem of cardiac cells energetics. Ter Arkh 49(1):120-132, 1977 Saks VA, Rozenshtraukh LV, Smirnov VN, Chazov EI: Role of creatine phos- phokinase in cellular function and metabolism. Can J Physiol Pharmacol 56:691-706, 1978 Rozenshtraukh LV, Jurevichus IA, Undrovinas AI, Chikharev VN, Yushamova AV, Lyskovtsev VV: Effects of ethmozin on the contractile force, trans- membrane action potential, and sodium current in frog auricular muscle. In Proceedings of the USA-USSR First Joint Symposium on Sudden Death, Yalta, USSR, October 3-5, 1977. Washington, D.C., U.S. Department of Health, Education, and Welfare, Public Health Service, National Institutes of Health, DHEW Publication No. (NIH) 78-1470, pp 291-308 Kisch B: Electronmicroscopy of the frog's heart. Exp Med Surg 19(2-3): 104-142, 1961 Staley NA, Benson ES: The ultrastructure of frog ventricular cardiac muscle and its relationship to mechanisms of excitation-contraction coupling. J Cell Biol 38(1):99-114, 1968 Saks VA, Rozenshtraukh LV, Sharov VG, Emelin IV, Chazov EI: Role of the creatine-phosphokinase reactions in the energy metabolism of cardiac cells. In Proceedings of the Third US-USSR Joint Symposium on Myocardial Metabolism, Williamsburg, Virginia, May 9-11, 1977. Washington, D.C., U.S. Department of Health, Education, and Welfare, Public Health Service, National Insti- tutes of Health, DHEW Publication No. (NIH) 78-1457, pp 289-321 109 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. Langer GA, Frank JS, Brady AJ: The myocardium. In International Review of Physiology, vol 9, edited by AC Guyton and AW Cowley. Baltimore, University Park Press, 1976, pp 191-237 Beeler GW Jr, Reuter H: Membrane calcium current in ventricular myocardial fibres. J Physiol (London) 207:191-209, 1970 Beeler GW Jr, Reuter H: The relation between membrane potential, membrane currents and activation of contraction in ventricular myocardial fibres. J Physiol (London) 207:211-229, 1970 Fozzard HA, Gibbons WB: Action potential and contraction of heart muscle. Am J Cardiol 31:182-192, 1973 Coraboeub E: Membrane ionic permeabilities and contractile phenomena in myocardium. Cardiovasc Res (suppl I):55-63, 1971 Morad M, Goldman Y: Excitation-contraction coupling in heart muscle: membrane control of development of tension. Prog Biophys Mol Biol 27:257- 313, 1973 Prasad K: Cardiac metabolism and electromechanics of human heart. In The Metabolism of Contraction, edited by PE Roy and G Rona. (Recent Ad- vances in Studies on Cardiac Structure and Metabolism, vol 10.) Baltimore, University Park Press, 1975, pp 119-137 Bessman SP: A molecular basis for the mechanism of insulin action. Am J Med 40:740-749, 1966 Gudbjarnason S, Mathes P, Ravens KG: Functional compartmentation of ATP and creatine phosphate in heart muscle. J Mol Cell Cardiol 1:325-339, 1970 Gercken G, Schlette U: Metabolic states of heart in acute insufficiency due to l-fluoro-2,4-dinitrobenzene. Experimentia 24:17-19, 1968 Ndgle S: [The significance of creatine phosphate and adenosine triphos- phate in regard to energy supply, transport, and utilization in the normal and insufficient myocardium] (Ger). Klin Wochenschr 48:332-341, 1970 Jacobus WE, Lehninger AL: Creatine kinase of rat heart mitochondria. Coupling of creatine phosphorylation to electron transport. J Biol Chem 248:4803-4810, 1973 Saks VA, Chernousova GB, Gukovsky DE, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. Mitochondrial isoenzyme of creatine phos- phokinase: kinetic properties and regulatory action of Mg2t ions. Eur J Biochem 57:273-290, 1975 Levitsky DO, Levchenko TS, Saks VA, Sharov VG, Smirnov VN: The role of creatine phosphokinase in supplying energy for the calcium pump system of heart sarcoplasmic reticulum. Membrane Biochemistry 2:81-96, 1978 110 26. 27. 28. 29, 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. Saks VA, Lipina NV, Sharov VG, Smirnov VN, Chazov E, Grosse R: The localization of the MM isozyme of creatine phosphokinase on the surface membrane of myocardial cells and its functional coupling to ouabain- inhibited (Nat, K¥)-ATPase. Biochim Biophys Acta 465:550-558, 1977 Sharov VG, Saks VA, Smirnov VN, Chazov EI: An electron microscopic histo- chemical investigation of the localization of creatine phosphokinase in heart cells. Biochim Biophys Acta 468:495-501, 1977 Taylor EW: Chemistry of muscle contraction. Ann Rev Biochem 41:577-616, 1972 Mommaerts WFHM: Energetics of muscular contraction. Physiol Rev 49: 427-508, 1969 Cain DF, Davies RE: Breakdown of adenosine triphosphate during a single contraction of working muscle. Biochem Biophys Res Comm 8:361-366, 1962 Perry SV: Creatine phosphokinase and the enzymic and contractile proper- ties of the isolated myofibrils. Biochem J 57:427-431, 1954 Botts J, Stone MJ: Kinetics of coupled enzymes: creatine kinase and myosin A. Biochemistry 7:2688-2696, 1968 Mani RS, Kay CM: Physicochemical studies on the creatine kinase M-line protein and its interaction with myosin and myosine fragments. Biochim Biophys Acta 453:391-399, 1976 Walliman T, Turner DC, Eppenberger HM: Localization of creatine kinase isoenzymes in myofibrils. I. Chicken skeletal muscle. J Cell Biol 75(1):297-317, 1977 Williamson JR, Schaffer S: Epinephrine, cyclic AMP, calcium, and myocardial contractility. In The Sarcolemma, edited by PG Roy and NS Dhalla. (Recent Advances in Studies on Cardiac Structure and Metabolism, vol 9.) Baltimore, University Park Press, 1976, pp 205-223 Carmeliet E: Cardiac transmembrane potentials and metabolism. Circ Res 42:577-587, 1978 Luttgau HC: New trends in membrane physiology of nerve and muscle fibers. J Comp Physiol 120:51-70, 1977 Reuter H, Scholz H: The regulation of the calcium conductance of cardiac muscle by adrenaline. J Physiol (London) 264:49-62, 1977 Sperelakis N, Schneider JA: A metabolic control mechanism for calcium ion influx that may protect the ventricular myocardial cell. Am J Cardiol 37: 1079-1085, 1976 111 40. 41. 42. 43. 4a, 45. Reuter H, Scholz H: A study of the ion selectivity and the kinetic prop- erties of the calcium dependent slow inward current in mammalian cardiac muscle. J Physiol (London) 264:17-47, 1977 Tsien RW, Giles W, Greengard P: Cyclic AMP mediates the effects of adrenaline on cardiac Purkinje fibres. Nature [New Biol] 240:181-183, 1972 Tsien RW: Mode of action of chronotropic agents in cardiac Purkinje fibers. Does epinephrine act by directly modifying the external surface charge? J Gen Physiol 64:320-342, 1974 Noble D: The Initiation of the Heart Beat. Oxford, Clarendon Press, 1975 Jundt H, Reuter H: Is sodium-activated calcium efflux from mammalian cardiac muscle dependent on metabolic energy? J Physiol (London) 266: 78-79P, 1977 Kubler W, Katz AM: Mechanism of early "pump" failure of the ischemic heart: possible role of adenosine triphosphate depletion and inorganic phosphate accumulation. Am J Cardiol 40:467-471, 1977 112 POSSIBILITY OF PARTICIPATION OF CREATINE KINASE IN THE REGULATION OF CELL METABOLISM S. N. Lyzlova, V. E. Stefanov, and N. Taame INTRODUCTION Data are being increasingly accumulated which indicate that creatine kinase (ATP:creatine-phosphotransferase EC 2.7.3.2), the enzyme which was thought to be exclusively responsible for maintenance of a constant level of ATP in muscle and nerve tissue, actually may play a more universal role. We now know that creatine kinase and its substrate are widespread in tissue and localized in various subcellular structures. Reports also indicate that substrates of the creatine kinase reaction participate in regulation of the activity of series of enzymes and that, at the same time, various metabolites influence the activity of creatine kinase. Evidence has been presented which indicates that creatine kinase has an allosteric nature. In the present paper data are analyzed which indicate broad involvement of the creatine kinase reaction in processes of cell metabolism and the significant role which creatine kinase may play in regulation of these processes. DISTRIBUTION OF CREATINE KINASE IN TISSUE AND SUBCELLULAR STRUCTURES Detailed information on the distribution of creatine kinase is contained in a number of works (1,2, and others). Attention may be drawn to the fact that creatine kinase has been detected in leukocytes (3). Most probably, the primary physiological function of creatine kinase in these cells is to facilitate cer- tain stages of phagocytosis. Questions about the physiological functions of the creatine kinase system are also raised by the detection of all components of this system in adipose tissue (4). The level of creatine kinase and its substrates in brown adipose tissue is particularly high and comparable to the level of these substances in heart and nerve tissue where large quantities of mitochondria are present. Creatine kinase is also present in the membranes of synaptosomes and synaptic vesicles (5). Recently, creatine kinase was found in the nuclei of heart and skeletal muscles at sites containing chromatin (6). The function of the nuclear enzyme has not yet been studied but, obviously, it is determined by its location and From the Department of Biochemistry, Leningrad State University, Leningrad, USSR. 113 may be that of providing chemical energy for intranuclear processes. It can also be assumed that the creatine kinase system influences certain stages in protein biosynthesis. This has been suggested indirectly by information (7) on the effective utilization of ATP that is regenerated in the creatine kinase reaction for protein synthesis in vitro, and, to a great extent, by data on the indirect stimulation of GTP synthesis when the rate of creatine production is increased (8). PARTICIPATION OF GUANIDINE SUBSTRATES OF THE CREATINE KINASE REACTION IN THE REGULATION OF METABOLIC TRANSFORMATIONS Creatine phosphate and creatine are substrates of the creatine kinase reaction, the content of which changes most sharply during muscular activity. One important distinguishing feature of the regulatory mechanism in question is the fact that, in the muscle cell, creatine and creatine phosphate are not substrates of any other reaction except the creatine kinase reaction. Thus, in principle, they can function as an undistorted signal, coordinating the processes of formation of high-energy compounds, and converting them to reserve form and subsequent mobilization at the points of utilization of the chemical energy. Adenosine diphosphate (ADP), which controls respiration in liver and kidney cells, does not have these advantages, since the level of this substance in the muscle is also determined by the presence of highly active adenylate kinase. The regulatory role of the guanidine substrate of creatine kinase has been experimentally confirmed (8-11). It has also been shown that creatine phosphate inhibits the enzymes of glycolysis: phosphofructokinase (12), pyruvate kinase (13), and 3-phosphoglycerinaldehyde dehydrogenase (14). Fructose-1,6-diphos- phatase is activated (15). Creatine, and the inorganic phosphate formed upon decomposition of creatine phosphate, are necessary for activation of glycolysis, mitochondrial respiration, and oxidative phosphorylation (8,16). Experiments with a purified preparation of creatine kinase have shown that creatine is an activator of the creatine kinase reaction (10). A highly effective regulatory mechanism, which is based on the interaction of two creatine kinase systems, plays a key role in regulation of the energy supply of working muscle and related processes. In the muscle cell, there are at least two morphologically separate enzymes: the cytoplasmic and mitochon- drial creatine kinases. Due to the spatial separation of the two enzymes, which catalyze reactions moving in opposite directions, they can interact with each other. Actually, there is a direct relationship between the rate of creatine production in the reaction catalyzed by the cytoplasmic enzyme, and the intens- ity of its phosphorylation in the reaction occurring in the mitochondria (9). The physiological significance of this regulation is indicated by the fact that, in ischemia, a decreased level of creatine phosphate is observed with a practically unchanged level of ATP (17). Thus, the combination of data from various researchers working at the cellular and organ level, and the purified preparations of the enzyme, indicate that creatine is the effector which con- trols the production of ATP and creatine phosphate in the mitochondria and cytoplasm of heart and skeletal muscles. 114 In turn, the high effectiveness of the creatine kinase system of the mitochondria results both from the significant content of the enzyme, and enhancement of its function by the adenine nucleotide translocase (18). Con- sequently, there is every reason to believe that the traditional function of the enzyme--supplying the ATP to working muscle--is related to global partici- pation of the substrates of the creatine kinase reaction in feedback systems controlling glycolysis and oxidative phosphorylation. EFFECT OF METABOLITES OF GLYCOLYSIS AND PYROPHOSPHATE ON CREATINE KINASE ACTIVITY The rate of the creatine kinase reaction is regulated by substrates, and a definite role may be played by local changes in pH occurring, for example, upon accumulation of lactate in a working muscle. However, there is particular interest in the possibility of modulating the activity of the enzyme by metabo- lites differing from the substrates of the catalyzed reaction. This possibility usually results from presence in the enzyme molecule of an allosteric center for binding the regulatory ligand which, with rare exceptions (19), correlates with the subunit structure of the protein. In an enzyme with a quaternary structure there is the capacity for homotrophic and heterotrophic interaction of the ligand binding centers, and these interactions significantly increase the regulatory potential of the enzyme. Creatine kinase, purified from a number of sources, has a subunit (dimer) structure. However, the interactions between subunits are masked in the kinetics of the reaction occurring at the optimal pH and saturating concentrations of a given substrate. Due to the absence of any deviations in the kinetics of the reaction from the Michaelis kinetics, the question of the possible allosteric nature of creatine kinase did not arise for a long time (20). Reconsideration of the traditional concepts was facilitated by experiments in which interaction was found between enzyme subunits. For example, for crea- tine kinase from skeletal muscles of the rabbit, negative cooperativity was found in the binding of ADP in the presence of creatine and nitrate, resulting in a transient state conformation (21). Several other sources indicate that the subunits interact, based on analysis of the blocking of thiol groups and the accompanying inactivation of the enzyme (22,23). Similarly, we have suc- ceeded in revealing interaction of the subunits of creatine kinase from white muscles of the catfish. For the enzyme from this same source, we also demon- strated a significant deviation in the substrate dependences of the rate of the creatine kinase reaction from the hyperbolic at pH differing from optimal (figure 1). We have also shown that S-shaped curves describe the variation in the rate of the reaction catalyzed by creatine kinase from muscles of the pike, as a function of the concentration of the one substrate at low concentrations of the other substrate of the enzyme. This function normalizes with increasing concentration of the other substrate and becomes hyperbolic when its concentra- tion reaches saturation (figure 2). During measurements of the rate of the reaction over a broad range of concentrations, anomalies were noted in the kinetics of the reaction catalyzed by creatine kinase from human skeletal 115 0.15 ® 0.10 [ £ E 9 S = 005F > o ~~ 1 1 1 1 ) 5 1 2 3 4 [MgADP]™, mm’ FIGURE 1. Effect of pH on the shape of Lineweaver-Burk graphs for creatine kinase reaction. The pH is as follows: O = 7, A = 6.75, OD = 6.5, @ = 6.25, V=26.0, ® =7.5 A= 28.0. Concentration of creatine phosphate at saturation. Cr = Creatine. muscles (24). The authors of this study consider their results to be indicative of the interaction of binding centers of the substrates in the enzyme molecule. The presence in the creatine kinase molecule of an intramolecular regula- tory mechanism that is related to the interaction of subunits raises the question of which metabolites may function as signals that modify the creatine kinase activity. Information on the effectors of the creatine kinase reaction is very limited and partially contradictory. There has been a description of an in- hibiting effect on creatine kinase of skeletal muscles of the rabbit by glyco- lytic intermediates: glucose-6-phosphate, fructose-1l,6-diphosphate, and phosphoenolpyruvate (25), as well as pyruvate and lactate (26). However, the possibility of nonspecific influence of phosphate esters, resulting from their anionic nature, makes it difficult to physiologically interpret these interest- ing data in a definitive way. The glycolytic potential of the white muscles of fish is significantly higher than that of the skeletal muscles of homoiotherms; 116 1 — o T mo 10 1 1 | J 05 1 15 2 25 2 4 6 8 10 [CP], mm [MgADP]", mm FIGURE 2. Effect of concentrations of a fixed substrate on the shape of Lineweaver-Burk graphs. A: Fixed concentrations of creatine phosphate (CP) are as follows: @ =4 mM, A = 2.4 mM, OD =1.6 mM, V = 1.2 mM, O = 0.8 mM. B: Fixed concentrations of MgADP are as follows: @ = 1 mM, A = 0.8 mM, OQO= 0.6 mM, V = 0.4 mM, O = 0.2 mM. under normal conditions, anaerobic glycolysis is the most important source of chemical energy for them. Certain changes in external conditions, e.g., a decrease in temperature and the resultant increase in accessibility of oxygen, may enhance the breakdown of fatty acids. Keeping this in mind, we investigated the effect of the glycolytic inter- mediates: glucose-6-phosphate, fructose-1,6-diphosphate, and phosphoenolpyru- vate, as well as pyrophosphate, on the catalytic properties of creatine kinase from the white muscles of catfish. The nature of cellular metabolism in the white muscles of fish enables us to reproduce certain aspects of metabolism in other cells, and to model certain pathological situations. For this reason, information on the regulatory properties of creatine kinase should not only be of special, but also of general theoretical interest. We showed that glucose-6-phosphate, fructose-1,6-diphosphate, and phospho- enolpyruvate noncompetitively inhibit creatine kinase in the white muscles of catfish. Saturation concentrations of the first two metabolites cause only partial, and not over 50 percent, inhibition (figure 3). Variation in the reaction rate as a function of substrate concentrations showed no deviation from the Michaelis kinetics when the above inhibitors were absent, but underwent significant transformation to an S-shaped curve when the inhibitors were pres- ent (figures 4 through 6). We must note that lack of data on the content of glycolytic intermediates in the white muscles of fish (we have information only on the content of fructose-1,6-diphosphate and glucose-6-phosphate in the white muscles of Gadus marhua, 250 and 59.2 mmol/100 g of tissue, respectively) (27) prevents us from estimating the true physiological significance of the effects that we observed. This difficulty is increased by the need to keep in mind 117 100 A activity (%) 3S % 10 30 50 [G-6-P] mM 100 B 100 C 2 $ 8 £ 50 £ 50 > > 0 0 0 20 60 100 01 2 3 4 5 [FDP], mM [PEP], mM FIGURE 3. Inhibitory effect of glycolytic intermediates glucose-6-phosphate (G-6-P) (A), fructose-l,6-diphosphate (FDP) (B), and phosphoenolpyruvate (PEP) (C) on creatine kinase activity of fresh (0) and aged (A) preparations of the enzyme. the effective local concentrations which may differ significantly from average concentrations. Still, one can assume that the potential capacity of the creatine kinase system to respond specifically to glycolytic signals, rather than to some inte- gral indicator of energy status, such as adenylate energy charge, is of great significance for many species of fish. This is particularly true for those which can be considered facultative anaerobes. Some of them, e.g., the carp, have a system of metabolic regulation (at the level of the LDH reaction) which allows the accumulation of lactate to be avoided under anoxic conditions for 2 to 3 months (28). Obviously, when a shift is made from aerobiosis to anaerobiosis and back, the functioning of the glycolytic system of such animals changes 118 05 10 15 [CP], mM [MgADP], mM FIGURE 4. Effect of glucose-6-phosphate (G-6-P) on the substrate dependence of creatine kinase at saturating concentrations of creatine phosphate (CP) (A) and MgADP (B). The concentrations of G-6-P are as follows: @ = 0 mM, O= 12.5 mM, A = 25 mM, O = 37.5 mM. 1.0 A 10r 8 < 05 < ost 0 0 0 1 2 3 4 5 0 05 10 15 [CP], mM [MgADP], mM FIGURE 5. Effect of fructose-1,6-diphosphate (FDP) on substrate dependence of creatine kinase at saturating concentrations of creatine phosphate (CP) (A) and MgADP (B). The concentrations of FDP are as follows: @ = 0 mM, O= 25 mM, A= 50 mM, O = 75 mM. radically, and the intensity of the flow along the glycolytic path also under- goes sharp changes. Important changes in the concentrations of glycolytic intermediates, and the ability of these to modulate the catalytic properties of creatine kinase, can support coordinated operation of creatine kinase and the glycolytic system. The effect of pyrophosphate on the activity of creatine kinase has a bi- phasic nature (figure 7): activation at low concentrations and inhibition at 119 10 A 10r > < 0s) < os} 0 SL 0 0 1 2 3 4 5 0 05 10 15 [CP], mM [MgADP], mM FIGURE 6. Effect of phosphoenolpyruvate (PEP) on substrate dependence of creatine kinase at saturating concentrations of creatine phosphate (CP) (A) and MgADP (B). The concentrations of PEP are as follows: © = 0 mM, O= 1 mM, A=2.5mM, O = 5 mM. 120 c & i 60 05 10 1. 0 [CP] mM [MgADP], mM (PP), mM FIGURE 7. Variations in rate of the creatine kinase reaction as a function of the creatine phosphate (CP) concentration (A) and MgADP concentration (B) for various concentrations of pyrophosphate (PP): @ = 0 mM, O= 2.5 mM, A=5mM, O=10 mM. C: Activity of fresh (0) and aged (A) preparations of creatine kinase when PP is present. 120 higher concentrations. The inhibition, caused by pyrophosphate, was complete and one can assume that it results from binding in the active center and is competitive in nature. In contrast, the activating effect of this metabolite, observed at lower concentrations, is probably related to interaction in the allosteric center. This assumption is confirmed, in particular, by the nature of desensitization of the enzyme to the effect of pyrophosphate as a result of storage of the enzyme preparation. With a slight change in activity, the enzyme fully loses its ability to be activated by pyrophosphate; the in- hibiting effect of the pyrophosphate remained practically unchanged. The activity of a fresh preparation of creatine kinase, as a function of pyrophosphate concentration (figure 7C), is characterized by a sharp peak at low concentrations and indicates that the pyrophosphate regulation of creatine kinase may have physiological significance. This may be the mechanism which coordinates the breakdown of fatty acids with increased availability of oxygen and with the more rapid storage of energy as creatine phosphate. There is a known relationship between glycolysis and oxidation of fatty acids. Glycolysis is inhibited by the oxidation of fatty acids (29), while creatine phosphate, as we have noted, inhibits the activity of a number of the enzymes of glycolysis, but activates fructose-1,6-diphosphatase. Creatine phosphate can thus be considered the factor which switches from glycolysis to gluconeogenesis when the muscle receives the energy necessary for contraction by the breakdown of fatty acids. Analysis of the relationship of the creatine kinase reaction to B-oxidation of fatty acids may be useful in studying the creatine kinase system of myocardial cells, for which fatty acids are an important source of energy (30). In turn, information about the proper- ties of regulation related to the creatine kinase reaction in nonmuscle cells can lead to deeper understanding of regulation in muscle cells. 121 10. 11. 12. 13. REFERENCES Lyzlova SN: [Phosphagen Kinases] (Rus). Leningrad, State University Press, 1974 Saks VA, Rozenshtraukh LV, Sharov VG, Emelin IV, Chazov EI: Role of the creatine-phosphokinase reactions in the energy metabolism of cardiac cells. In Proceedings of the Third Joint Symposium on Myocardial Metab- olism, Williamsburg, Virginia, May 9-11, 1977. Washington, D.C., U.S. Department of Health, Education, and Welfare, Public Health Service, Na- tional Institutes of Health, DHEW Publication No. (NIH) 78-1457, pp 289- 322. Nersesova LS, Lyzlova SN, Ashmarin IP, Potapenko LM: [The supply of energy for leukocyte mobility] (Rus). In [Nonmuscular Forms of Mobility], edited by GM Frank. Pushchino, USSR Academy of Sciences, 1976, pp 159- 166 Berlet HH, Bonsman NJ, Birringer H: Occurrence of free creatine, phospho- creatine and creatine phosphokinase in adipose tissue. Biochim Biophys Acta 437:166-174, 1976 Friedhoff AJ, Lerner MH: Creatine kinase isoenzyme associated with synaptosomal membrane and synaptic vesicles. Life Sci 20:867-873, 1977 Yerashova NS, Saks VA, Sharov VG, Lyzlova SN: [Creatine kinase from muscle cell nuclei] (Rus). Biokhimiia 44:372-378, 1979 Hickey ED, Weber LA, Baglioni C: Nuclease activity in preparations of creatine phosphokinase: effect on mRNA stability. Biochem Biophys Res Commun 80(2):377-383, 1978 Seraydarian MW, Abbott BC: The role of the creatine-phosphoryl-creatine system in muscle. J Mol Cell Cardiol 8(10):741-746, 1976 Seraydarian MW, Artaza L: Regulation of energy metabolism by creatine in cardiac and skeletal muscle cells in culture. J Mol Cell Cardiol 8: 669-678, 1976 Chetverikova EP, Rozanova NA: [Allosteric properties of muscle creatine kinase] (Rus). Biokhimiia 42(3):481-489, 1977 Seraydarian MW, Artaza L, Abbott BC: Creatine and the control of energy metabolism in cardiac and skeletal muscle cells in culture. J Mol Cell Cardiol 6:406-413, 1974 Krzanowski J, Matschinsky FM: Regulation of phosphofructokinase by phosphocreatine and phosphorylated glycolytic intermediates. Biochem Biophys Res Commun 34:816-823, 1969 Kemp RG: Inhibition of muscle pyruvate kinase by creatine phosphate. J Biol Chem 248:3963-3967, 1973 122 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Oguchi M, Gerth E, Fitzgerald B, Park JH: Regulation of glyceraldehyde-3- phosphate dehydrogenase by phosphocreatine and adenosine triphosphate. IV. Factors affecting in vivo control of enzymatic activity. J Biol Chem 248: 5571-5576, 1973 Fu JY, Kemp RG: Activation of muscle fructose 1,6-diphosphatase by creatine phosphate and citrate. J Biol Chem 248:1124-1125, 1976 Belitser VA: [Chemical Conversions in Muscle] (Rus). Moscow-Leningrad, Medgiz, 1940 Dhalla NS, Yates JC, Walz DA, McDonald VA, Olson RE: Correlation between changes in the endogenous energy stores and myocardial function due to hypoxia in the isolated perfused rat heart. Can J Physiol Pharmacol 50: 333-345, 1972 Saks VA, Chernousova GB, Gukovsky DE, Smirnov VN, Chazov EI: Studies of energy transport in heart cells. Mitochondrial isoenzyme of creatine phos- phokinase: Kinetic properties and regulatory action of Mg2t ions. Eur J Biochem 57:273-290, 1975 Singh D, Tamao J, Blakley RL: Allosterism regulation and cooperativity: the case of ribonucleotide reductase of lactobacillus leichmanmi. Adv Enzyme Regul 15:81-100, 1976 Watts DC: Creatine kinase (adenosine-5'-triphosphate-creatine phospho- transferase). In The Enzymes, edited by PD Boyer, vol 8, 3rd ed. New York, Academic Press, 1973, pp 383-455 McLaughlin AC: The interaction of 8-anilino-l-naphthalene-sulfonate with creatine kinase. J Biol Chem 249:1445-1452, 1974 Kumudayalli I, Moreland BH, Watts DC: Properties and reaction with iod- acetamide of adenosine-5'-triphosphate-creatine phosphotransferase from human skeletal muscle. Further evidence about the role of the essential thiol group in relation to the mechanism of action. Biochem J 117:513- 523, 1970 Price NC: Effects of temperature and reagent size on the reaction of the thiol groups of rabbit muscle creatine kinase. Biochem Soc Trans 5:764-— 765, 1977 Lee CS, Nicholson GA, O'Sullivan WJ: Some properties of human skeletal muscle creatine kinase. Aust J Biol Sci 30:507-517, 1977 Rozanova NA, Chetverikova EP: [Effect of phosphates of sugars, phospho- enolpyruvate, and adenylic acid on creatine kinase of the skeletal muscles, brain, and heart] (Rus). Biokhimiia 40:1299-1304, 1975 Moldoveanu N: Influence of glycolysis final terms, iodoacetamide and 1-fluoro-2,4~dinitrobenzene on the miokinase and creatinkinase activity. Rev Roum Biochem 7:123-129, 1970 123 27. 28. 29. 30. Malkolm LR: [The Chemical Biology of Fish] (Rus). Moscow, Pishevaya Promyshlennost Press, 1976 Hochachka PW, Somero GN: Strategies of Biochemical Adaptation. Phila- delphia, Saunders, 1973 Newsholme EA, Start C: Regulation in Metabolism. London, New York, Wiley, 1973 Neely JR, Morgan HE: Relationship between carbohydrate and lipid metabo- lism and the energy balance of heart muscle. Ann Rev Physiol 36:413-459, 1974 124 THE FUNCTIONAL ROLE OF STRUCTURAL NONEQUIVALENCE OF F-ACTIN SUBUNITS OF MUSCLE THIN FILAMENTS V. V. Lednev SUMMARY According to the model of actin-containing filaments suggested earlier by G. M. Frank and the author, thin filaments of skeletal muscle can be considered as a linear polymer chain composed of structural units each of which can adopt either of two different structural states—-I (inhibited) and A (activated). The structural unit or monomer can be represented by a protein block consisting of two complementary functional units of the thin filaments. Data available indicate that the length of the unit in the A state is slightly longer than its length in the I state. In this case, the relative number of functional units in the A and I states in thin filaments must depend not only on the concentra- tion of calcium within the sarcomere, but also on the external force pulling the thin filaments in contracting muscle. This paper shows that this property of thin filaments may provide the molecular basis of the well-known Hill equa- tion, as well as the mechanism of self-adaptation of the contractile apparatus of the muscle to external load. It is suggested that the presence of myosin regulation in muscles, and also in nonmuscular motile systems, is necessary to assure relaxation of the contractile apparatus. INTRODUCTION It has been shown earlier that the appearance of forbidden meridional re- flections in the X-ray diffraction patterns of oriented gels of actin-containing filaments (1,2) indicates that F-actin subunits are capable of existing in two different structural states (2). A model of thin filaments taking into account this structural feature of F-actin agrees with data available on the structure and physical-chemical properties of actin-containing filaments and enables us to explain some experimental findings which have not been previously understood. In this paper we analyze the relationship between certain structural properties of thin filaments and the Hill equation, and the mechanism of coordinating the work of the contractile apparatus of muscles with the externally applied load. [see also the paper of Lednev and Frank (3)]. From the Institute of Biophysics, USSR Academy of Sciences, Puschino, USSR. 125 STRUCTURAL PROPERTIES OF THIN FILAMENTS The elementary structural unit (or monomer) of a thin filament is a linear section of filament with the length of about 385 A consisting of two complemen- tary functional units (FU) which we would call a "block" for brevity. A func- tional unit is a protein block consisting of 7 F-actin subunits covered with a tropomyosin-troponin pair. In one possible structural state of a block, the two FU are "off," and in the other structural state, only one FU is "on." The struc- ture of muscle thin filaments at rest can be considered to be a linear chain consisting of blocks which are "off." When Ca?t is added to the medium, the structure of FU transforms into the "on" state, a process which continues until the ratio of "on" or activated FU (the A state) and the "off" or inhibited FU (the I state) reach equilibrium. If the projection of a block on the longitudinal axis of thin filament in one of the two structural states is longer than its projection in the other state, the equilibrium between the two types of blocks, for a given concentra- tion of Cat, may be shifted with lengthening of the thin filament by an external load or, in general, when an external force is applied. It is not clear a priori, however, in which direction the equilibrium be- tween the A and I states of the FU will be shifted. Comparison of the param- eters of the F-actin helix in thin filaments of muscles at rest and during contraction (4), and also in Mg2t paracrystals obtained at concentrations of 1072 M < Cat < 10-9 M (5) indicate a possible slight increase in the peach of the F-actin helix in the contracting muscle. It may therefore be considered that the "on" pi state of the FU corresponds to the more extended structural state of the subunits. Clearly, the presence of an external tensile force should "switch on'" an additional number of FU for a given concentration of CaZt, The thermodynamic properties of certain simple models of this type of structure have been studied by a number of authors in calculations of elasticity and other characteristics of linear polymers (6-7). In analysis of the vari- ation of activation of thin filaments with concentration of Cat and external load, one must consider the presence of positive and negative cooperativity in the interaction between actin subunits (2,3), the presence of several sites of binding of cat to troponin, possible interaction between them, and other factors which have not yet been described quantitatively. An analytical equation for the number of "on" FU in thin filaments as a function of the external load on the muscle fiber can, however, be obtained by an indirect method based on the assumption that the regulation of work of the mechanical-chemical converter (MCC) of the muscle fiber, which provides the hyperbolic relation between the velocity of muscle contraction and external load, is based exclusively on the properties of thin filaments described above, i.e., on the degree of activation of thin filaments both with Ca?% concentration and with the external force. The relationship between the velocity of isotonic shortening v and load P (P - v relation) of skeletal muscle can be described by various equations, the best known of which is the Hill equation (8): (P + a)(v + b) = Po, + a = (Vpax + b)a, where P, is the isometric tension and a and b are constants. An analo- gous equation also describes the rate of release of additional (in comparison to isometric contraction) energy in the form of work and heart during isotonic contraction of the muscle. 126 The formal similarity of the thermal and mechanical equations of Hill, and the agreement of the values of the thermal and mechanical constants (according to Hill's data obtained in 1933) (8), led to the suggestion that Hill's equation "represents something more than a simple mathematical description of the experi- mental results" (9). In 1964, however, Hill showed, using more precise measure- ments, that constant "a'" which was obtained by thermal measurements and was equal to the heat released as the muscle shortened by 1 cm (heat of shortening) differs slightly from the corresponding mechanical constant (10). Nevertheless, it is still too early to stop a search for the conceptual sense of Hill's equation. In the mid-1960's, Caplan, using the thermodynamics of irreversible processes, demonstrated that Hill's equation describes the rela- tionship of output characteristics of the simplest class of energy transducers which adjust themselves to external loads (11-13). Regulation of the output in these energy transducers can be achieved by two different mechanisms, acting separately or jointly. In the terms of the concept of "sliding filaments" and of the cross-bridge theory of A. Huxley, these possible types of regulation can be described as fol- lows. With the first type of regulation (14,15), the external load changes the number of activated bridges, each of which can be considered as an elementary linear energy MCC. In this case an efficiency of transformation of chemical energy to mechanical energy both by an individual bridge and by the entire mus- cle does not change. With regulation of the second type (11-13), the change in external load leads to a change in the effectiveness of transformation of free energy from ATP hydrolysis to mechanical energy. Wilkie and Woledge (16), as well as Born- horst and Minardi (14,15), showed that Caplan's model (17) agrees poorly with the experimental data. Comparison of the results of these theoretical works with the concepts outlined above allows us to assume that a muscle fiber can be looked upon as a MCC, whose nonlinearity of output is provided by the regulator of the contractile filaments interaction, the function of which is performed by thin filaments. These filaments provide the dependence of the number of switched on FU on variations in external load on the muscle. It is this structural fea- ture of thin filaments which determines the hyperbolic P - v relation expressed by Hill's equation. By comparing Onzager's phenomenological equation for muscular MCC with Hill's equation, we can obtain an analytical expression of the variation in number of working bridges with external load (see equation 19 in reference 14): . a/P, + P/P, . . = al? 1 quation |? Io where n and n, are the number of attached bridges corresponding to the loads P and P,, respectively. We presume that this variation results from the structural-functional properties of thin filaments. It must be noted, however, that in developing equation 1, the physical nature of the variation in number of working bridges with external load, i.e., the nature of the MCC regulator, was not specified. Expression 1, obtained on the assumption that Onzager's 127 equation was applicable to the description of the muscular mechanical converter, can also be obtained by assuming that the force of internal friction of the muscles is Newtonian in nature (18,19), or by introducing the concept of the existence of "breaking' bridges, whose number is determined by the kinetic con- stant which is proportional to the velocity of muscle shortening (20,21). As can be seen from equation 1, for a given load P, the number of FU which are switched on in the thin filaments and, consequently, the number of working bridges, is unambiguously determined by the value of parameter "a" (we recall that equation 1 describes muscle contraction upon tetanic stimulation, i.e., with a concentration of free Cat within the sarcomere of 10™2-10" M). Expres- sion 1 can be written as 3 -x+ (1 - xE- Equation 2 o oO 2 where the quantity x = (a/Py)/ (1 + a/P,) corresponds to the share of FU (or bridges) switched on following tetanic stimulation of the unloaded muscle (P = 0). The second term in the right part of equation 2 expresses the fraction of actomyosin bridges formed in an isotonically contracting muscle by activation of the thin filaments by external tensile force P. As is known, the value of constant a/P, for frog skeletal muscle is approximately 0.25, so that equation 1 means that n/n, = 0.2 + 0.8 P/P,. Thus, upon tetanic stimulation of skeletal muscle, isotonically contracting under zero load, only 0.2 of the total number of FU recruited during an isometric tetanic contraction is switched on. Data available in the literature confirm equation 2. Thus, the variation in stiffness* of an isotonically contracting single frog skeletal muscle fiber with external load (figure 6 in reference 23) can be well approximated by equa- tion 2 with a value of parameter x = 0.2. An analogous formula can be used to describe data on the change in stiffness of the entire frog sartorius muscle in the developmental phase of isometric tension (figure 3a in reference 24). Thus, the two structural states of FU in actin-containing filaments pro- vide feedback between the contractile apparatus and the external load on the muscle. This concept allows us to pursue a new approach in understanding coordination of the work of the muscle's executive and energy-supplying system. COORDINATION OF MUSCLE EXECUTIVE AND ENERGY-SUPPLYING SYSTEM The term "concerted regulation" of muscular contraction is increasingly seen in the literature, and is usually used in the strict sense of the well- established fact of synchronous startup of the interaction of contractile *The stiffness of a muscle is the ratio AT/AL, where AT is the change in muscle tension with a change in its length by AL, and is a measure of the number of bridges connected to the thin filaments at any given moment (22). 128 filaments and glycogenolysis with an increase in the intrasarcomere concentra- tion of Cat (25). Obviously, for efficiency in the muscle's work, the execu- tive and energy-supplying systems must not only be synchronized at startup, but the force developed by the muscle must also be proportional to the external load, and the quantity of fuel both utilized and synthesized must be proportional to the sum of work and heat. This conclusion can be reached based on the following rather general consideration: The evolutionary selection of systems for ATP synthesis with a high yield of ATP per unit of substrate utilized makes sense only if one assumes economical utilization of ATP by cellular energy converters and, in particular, the muscle's contractile apparatus. Correlation between the energy release by a muscle, on the one hand, and the utilization and synthesis of high-energy compounds, on the other hand, has been established in a number of experiments. When oxidative phosphorylation and glycolysis are blocked, the quantity of creatine phosphate split (and ATP since creatine phosphate is used to replenish ATP stores) is proportional to the total energy (work plus heat) released by the muscle in various working modes (26-29). On the other hand, study of the relationship between lactate production and the energy released in the intact muscle working under anaerobic conditions (28,30) indicates that the muscle is capable of providing direct pro- portionality not only between energy release and ATP utilization, but also pro- portionality between energy release and the rate of synthesis of the high-energy compounds. Recently, Mommaerts et al. (31,32) demonstrated direct proportion- ality between the quantity of activated phosphorylase and the total energy re- leased by the semitendinosus muscle of the frog, working in various modes under anaerobic conditions. These authors, however, could not explain the results which they found. It follows from these arbitrarily selected data that there is actually a connection in the muscle which assures proportionality between the external load and the degree of activation of the contractile apparatus, allowing the muscle to use its store of ATP economically. There is also a relationship in the muscle between the contractile apparatus and the energy-supplying system which assures proportionality between synthesis of high-energy compounds and their utilization as the contractile apparatus works. No explanation of these mechanisms is offered in the literature. As we have shown previously, the structural properties of actin-containing filaments can in principle assure a linear dependence between the external load on the muscle and the number of switched on elementary force generators (see equation 1), which, in turn, leads to a linear variation in the rate of energy release of the muscle with the external load in corresponding models of muscle contraction (see equation 27 in reference 14). The molecular mechanism of regulation of the energy supply system of the muscle is less clear. As we know, ATP production in skeletal muscle under aerobic conditions occurs sequentially in the process of glycolysis, then in the Krebs cycle and in the electron transport chain, and the initial substrate for the last two links in the energy system is the product of glycolysis-- pyruvate. We show below that the mechanism for coordinating the work of these systems may again be based on variation in the structural state of the thin filaments with external load. 129 Over the past 10 years, it has been established that most enzymes in the glycolytic cycle can, at least under certain conditions, specifically bind with pure F-actin and with reconstituted thin filaments (33). There is good reason to assume that some of these enzymes, frequently called soluble, are actually bound in vivo with the thin filaments (33). In any case, the relationship be- tween the total concentration of glycolytic enzymes and actin shows that all of the glycolytic enzymes can be adsorbed by the thin filaments (33). Phospho- fructokinase, considered on the basis of in vitro data to be a key enzyme in regulation of the rate of glycolysis and glycogenolysis, has the greatest af- finity for actin of all glycolytic enzymes (33,34). A number of authors have stated the opinion that the binding of glycolytic enzymes with thin filaments may somehow influence the kinetic parameters of these enzymes and the rate of glycolysis. It is usually assumed that regulation of the kinetic parameters of the glycolytic enzymes can be achieved both by adsorption and by desorption of these enzymes on the thin filaments, and by changing the interaction of the glycolytic enzymes with the tropomyosin-troponin component of the thin filaments (33,35,36). A new model of thin filament structure can explain the results obtained in a study of the interaction of glycolytic enzymes with thin filaments (on the other hand these results can be seen as confirming the new model of thin filaments) and, in addition, allows us to suggest a specific mechanism for the correlation of energy expenditure of muscle with the rate of glycolysis (and glycogenolysis). According to the new model, F-actin subunits in the filaments of pure F-actin and in the reconstructed thin filaments (when [caZt] > 1072 M) take on two different conformations, whereas if [cat] < 10-8 M, the actin subunits in the thin filaments all have the same conformation. Based on this, we can expect that the binding of glycolytic enzymes with pure F-actin and with the F-actin of thin filaments (where [Ca2t] 2 10-3 M) might be characterized by two different binding constants. The model predicts that, with complete saturation of actin by any enzyme, half of the entire quantity of the bound enzyme will be bound to it with one binding constant, and the other half will be bound to it with another binding constant. This situation is observed experimentally in studying the binding of aldolase with pure F-actin (33,35,36-38), although the authors explain their results by assuming the existence of two different centers of binding with actin on the surface of aldolase (33,36). Based on the new model of thin filaments, we can expect that the binding of aldolase with thin filaments in a medium not containing Ca * will be characterized by a single binding con- stant; this can be easily tested experimentally. The change of the structural state of F-actin subunits making up the thin filaments that is dependent upon a change in Ca2t concentration provides the cast sensitivity of the kinetic parameters of the glycolytic enzymes; there is no need to introduce the concept of dissociation-association of glycolytic en- zymes and actin as a function of the physiological state of the muscle. We suggest that this is how the data of Walsh et al. (39) should be interpreted. These data show that the apparent binding constant of the substrate with aldo- lase, attached to the thin filaments, increases by a factor of 5 with an in- crease in the concentration of Ca? between 10-8 M and 1072 M, as a result of which the catalytic activity of aldolase increases greatly. 130 Obviously, the fraction of activated glycolytic enzymes may depend on the external load on the muscle, just as the number of F-actin subunits (or FU of the thin filaments) switched on depends on the external load. This dependence may be the basis for maintenance of a quantitative correlation between the energy utilization of the muscle and the intensity of ATP synthesis. These considerations show that glycolysis regulation at the level of indi- vidual glycolytic enzymes can be performed by changing the concentration of caZt simultaneously with the force of the muscular contraction, i.e., by a parallel mechanism. It was assumed earlier that glycolysis regulation can be performed only by a consecutive series of events, i.e., nerve pulse - contraction - change in concentration of adenine nucleotides = change in intensity of glycolysis (40). Regulation of the synthesis of high-energy compounds described above is not the only possible mechanism. Variation in the ratio of the number of FU which are on and off, and consequently, the Ca2t binding properties of the thin filaments, both with a given concentration of Ca2t and with an external load, means that the complete set of thin filaments of a given sarcomere has the prop- erties of a Cat buffer, whose effective capacity in a contracting muscle is determined by the external load: By changing the external load, one can change the level of free Cat within the sarcomere and, correspondingly, regulate the intensity of ca2t-dependent processes. It is thus possible to regulate the quan- tity of activated phosphorylase kinase and, correspondingly, the intensity of glycogenolysis, as well as the activity of various CaZt-dependent kinases and the regulatory proteins that are phosphorylated by them, which determine, in turn, the processes of protein synthesis in the muscle cell. In this way, the muscles can adapt over a long period of training to increased loads or, under pathological conditions such as cardiac muscle hypertrophy, the type of muscle fiber can change with denervation and cross-innervation. Other means of changing the rate of energy production of the muscle are also possible in principle, particularly by allosteric interaction of metabolic products (ADP, NH3, etc.) with the enzymes of the glycolytic cycle (the exis- tence of this type of regulation in vivo is frequently postulated without suf- ficient experimental data obtained in vitro). FUNCTIONAL SIGNIFICANCE OF THE VARIOUS TYPES OF REGULATORY SYSTEMS IN MUSCLES Comparative study of the distribution of various regulatory systems in the muscles of animals has shown that actin regulation is characteristic of the skeletal muscles of vertebrates, myosin regulation is characteristic of the muscles of mollusks and the smooth muscles of vertebrates, and dual (i.e., both actin and myosin) regulation is characteristic of the muscles of most invertebrates (41). Data have been accumulated which indicate the possible presence in verte- brate skeletal muscle of myosin regulation in addition to actin regulation (42). If these data are confirmed in future studies, then the presence of myosin type of regulation can be considered an essential characteristic of the contractile apparatus of all types of muscles, while in many cases muscle may function with- out regulation of the actin type. Finding the reason for evolutionary selection 131 of different regulatory systems in the muscles should simultaneously explain these two fundamental facts that were obtained from comparative study of the regulatory systems of muscles. It seems possible that the presence of ca2t switches on the myosin bridges and is necessary for relaxation of the muscles. Actually, as we showed earlier in this paper, during isometric contraction of the muscle, most of the FU are turned on by the extension of thin filaments with external load, and cat may act only as a trigger in the activation of the thin filaments. Let us assume that the movement of the myosin head from the backbone of the thick filament toward the thin filament is induced by switching on of the FU located opposite this head, leading to a change in the balance of long-range forces of attraction and repulsion between them. In this case, even after pumping of cat from the isometrically contracting sarcomere, there is always a sufficient number of FU switched on to support interaction of the bridges with the thin filaments and correspondingly development of force by the muscles until the ATP reserves are exhausted. It is well known, however, that in a single isometric twitch, relaxation begins soon after maximum tension is developed. The ability of the muscle fiber to relax can be easily explained if we assume that the interaction of the bridge with a thin filament, or at least its movement away from the backbone of the thick filament towards the thin filament, is possible only when the cat is bound by the light chains of the myosin head (bridge); correspondingly, loss of cat from the light chains is accompanied by a return of the myosin head to the back- bone of the thick filament. This mechanism can explain why the skeletal muscle relaxes even when FU of thin filaments are switched on due to stretching by an external load, and also why muscles with only myosin regulation are capable of relaxing (we recall that, in thin filaments without troponin, i.e., filaments which represent a complex of F-actin with tropomyosin, the FU are always switched on, if we consider FU as a complex of 7 monomers of actin covered with a molecule of tropomyosin). We can also assume that phosphorylation and dephosphorylation of the light chains of myosin in certain types of muscles and in nonmuscular motile systems are also needed not only to start up the interaction of the contractile myofila- ments but also and mainly to provide relaxation of the contractile apparatus. Let us analyze the reasons for the loss of actin regulation by certain muscles. It is possible that the presence of myosin regulation alone in verte- brate smooth muscle, muscles of mollusks, and certain actomyosin nonmuscular motile systems, results from the fact that, in these muscles, the molar actin/ myosin ratio is significantly higher than in skeletal muscles. A significant portion of the actin filaments in these systems does not interact with the myosin at all, but rather performs more of a structural role, providing, for example, the plastic properties of the smooth muscles. In these situations, localization of the Ca2t receptors on the myosin requires a smaller quantity of cat for activation of the contractile apparatus, and additionally, allows less powerful pumps to be used to regulate the level of Ca2t within the muscle cells. Thus, in evolutionary selection of a type of regulation for vertebrate smooth muscle 132 and muscles of mollusks, such factors as the limited store of mobile cat, on the one hand, and the economy of energy expended in the work of cat pumps, on the other hand, may have been decisive. We note in conclusion that variation in the degree of activation of thin filaments of skeletal muscles with external load should be considered in studying a number of other phenomena such as the binding of caZt by various troponin- containing systems, mechanical activation and deactivation of muscle contraction, and the reaction of muscles to changes in their length. More detailed analysis of these and related problems have been recently presented in a review (43). 133 10. 11. 12. 13. 14. 15. REFERENCES Hanson J, Lednev V, O'Brien EJ, Bennett P: Structure of the actin- containing filaments in vertebrate striated muscle. Cold Spring Harbor Symp Quant Biol 37:311-318, 1972 Lednev VV: [The nature of forbidden meridional reflections in x-rays of actin-containing filaments] (Rus). Dokl Akad Nauk SSSR 225:1430-1433, 1975 Lednev VV, Frank GM: [Structural nonequivalence of F-actin subunits and its possible significance in the regulation of ATPase activity and the development of tension in skeletal muscles] (Rus). Biofizika 22:376, 1977 Huxley HE, Brown W: The low-angle x-ray diagram of vertebrate striated muscle and its behavior during contraction and rigor. J Mol Biol 30:383- 384, 1967 O'Brien EJ, Gillis JM, Couch J: Symmetry and molecular arrangement in paracrystals of reconstituted muscle thin filaments. J Mol Biol 99:461- 475, 1975 Hill TL: Thermodynamics for Chemists and Biologists. Reading, Massa- chusetts, Addison-Wesley, 1968 Burte H, Halsey G: A new theory of non-linear viscous elasticity. Textile Res J 17:465-476, 1947 Hill AV: The heat of shortening and dynamic constants of muscle. Proc R Soc Lond [Biol] B126:136-195, 1938 Pringle J: [Muscle Models] (Rus). In [Models and Analogies in Biology]. Moscow, Foreign Literature Press, 1963, pp 85-125. Printed in English by Cambridge University Press, 1960 Hill AV: The effect of load on the heat of shortening of muscle. Proc R Soc Lond [Biol] B159:297-324, 1964 Caplan SR: A characteristic of self-regulated linear energy converters. The Hill force-velocity relation for muscle. J Theor Biol 11:63-86, 1966 Caplan SR: Autonomic energy conversion. I. The input relation: pheno- menological and mechanistic considerations. Biophys J 8:1146-1166, 1968 Caplan SR: Autonomic energy conversion. II. An approach to energetics of muscular contraction. Biophys J 8:1167-1193, 1968 Bornhorst WJ, Minardi JE: A phenomenological theory of muscular contrac- tion. I. Rate equations at a given length based on irreversible thermo- dynamics. Biophys J 10:137-154, 1970 Bornhorst WJ, Minardi JE: A phenomenological theory of muscular contrac- tion. II. Generalized length variations. Biophys J 10:155-171, 1970 134 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. Wilkie DR, Woledge RC: The application of irreversible thermodynamics to muscular contraction. Comments on a recent theory by SR Caplan. Proc R Soc Lond [Biol] B169:17-29, 1967 Caplan SR: Nonequilibrium thermodynamics and its application to bio- energetics. In Current Topics in Bioenergetics, vol 4. New York, Academic Press, 1971, pp 1-79 Volkenstein MV: Muscular contraction. Biochim Biophys Acta 180:562-572, 1969 Akasawa K, Yamamoto M, Fujii K, Mashima H: A mechanochemical model for the steady and transient contractions of the skeletal muscle. Jpn J Physiol 26:9-28, 1976 Deshcherevsky VI: [Two models of muscle contraction] (Rus). Biofizika 13:928-935, 1968 Deshcherevsky VI: A kinetic theory of striated muscle contraction. Biorheology 7:147-170, 1971 Huxley AF: Review lecture. Muscular contraction. J Physiol 243:1-43, 1974 Julian FJ, Sollins HR: Variation of muscle stiffness with force at in- creasing speeds of shortening. J Gen Physiol 66:287-302, 1975 Bressler BH, Clinch NF: The compliance of contracting skeletal muscle. J Physiol 237:477-493, 1974 Fischer EH, Becker JU, Blum HE, Kerrick WGL, Lehky P, Malencik DA, Pocinwong S: Concerted regulation of glycogen metabolism and muscle con- traction. In Molecular Basis of Motility, 26 Colloquium der Gesellschaft fiir Biologische Chemie, 10-12 April, 1975. Mosbach/Baden, edited by LMG Heilmeyer Jr, JC Riiegg, and Th Wieland. Berlin, New York, Springer-Verlag, 1976, p 137 Wilkie DR: Heat work and phosphorylcreatine break-down in muscle. J Physiol 195:157-183, 1968 Wilkie DR: Energy transformation in muscle. In Molecular Basis of Motility. 26 Colloquium der Gesellschaft fiir Biologische Chemie, 10-12 April, 1975, Mosbach/Baden, edited by LMG Heilmeyer Jr, JC Riiegg, and Th Wieland. Berlin, New York, Springer-Verlag, 1976, pp 69-78 Woledge RC: Heat production and chemical change in muscle. Prog Biophys Mol Biol 22:39-74, 1971 Woledge RC: In vitro calorimetric studies relating to the interpretation of muscle heat experiments. Cold Spring Harbor Symp Quant Biol 37:629- 634, 1972 135 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. Wilkie DR: Energetic aspects of muscular contraction (frog). In Sym-— posium on Muscle, Budapest, Hungary, September 12-16, 1966. Symp Biol Hung 8:207-224, 1968 Mommaerts W: [Contemporary problems of myocardial metabolism] (Rus). In [Myocardial Metabolism]. Moscow, Meditsina, 1975, pp 8-25 Mommaerts WFH, Vegh K, Homsher E: Activation of phosphorylase in frog muscle as determined by contractile activity. J Gen Physiol 66:657-669, 1975 Clarke FM, Masters CJ: Interactions between muscle proteins and glycolytic enzymes. Int J Biochem 7:359-365, 1976 Arnold H, Henning R, Pette D: Quantitative comparison of the binding of various glycolytic enzymes to F-actin and the interaction of aldolase with G-actin. Eur J Biochem 22:121-126, 1971 Arnold H, Pette D: Binding of glycolytic enzymes to structure proteins of the muscle. Eur J Biochem 6:163-171, 1968 Arnold H, Pette D: Binding of aldolase and triosephosphate dehydrogenase to F-actin and modification of catalytic properties of aldolase. Eur J Biochem 15:360-366, 1970 Clarke FM, Morton DJ: Aldolase binding to actin-containing filaments. Biochem J 159:797-798, 1976 Morton DJ, Clarke FM, Masters CJ: An electron microscope study of the interaction between fructose diphosphate, aldolase and actin-containing filaments. J Cell Biol 74:1016-1022, 1977 Walsh TP, Clarke FM, Masters CJ: Modification of the kinetic parameters of aldolase on binding to the actin-containing filaments of skeletal muscle. Biochem J 165:165-167, 1977 Newsholm EA, Start C: Regulation in Metabolism. London, New York, Wiley, 1973, pp 146-147 Lehman W, Szent-Gyorgyi AG: Regulation of muscular contraction. Distribu- tion of actin control and myosin control in the animal kingdom. J Gen Physiol 66:1-30, 1975 Lehman W: Thick filament linked calcium regulation in vertebrate striated muscle. Nature 274:81, 1978 Lednev VV: [Some aspects of the regulation of muscular contraction] (Rus). In [Structural Principles and Regulation of Biological Mobility]. Moscow, Nauka, 1979, pp 221-270 136 EFFECT OF PHALLOIDIN ON THE UNSTEADY KINETICS OF MUSCLE CONTRACTION A. E. Bukatina and V. N. Morozov SUMMARY It is demonstrated that the addition of phalloidin (2-4+10"2 M) leads to multiphasic changes in isometric tension of glycerinated rabbit muscle fibers. These changes are accelerated as temperature increases. When phalloidin is present, the fibers retain their capacity for contraction and the Ca’ sensi- tivity of the force which was developed. Fibers treated with phalloidin lose their capacity to work in the self-oscillating mode at 20° C, but can work in this mode at 30° C. Phalloidin was not found to have any influence on the mechanical properties of fibers in rigor. The data obtained indicate that binding of phalloidin with fibers causes changes in the parameters of the cycle of interaction of actin and myosin during ATP hydrolysis. It is possible that these changes result from stabilization of the F-actin structure. INTRODUCTION The role of actin in the mechanochemical processes of muscle has not been fully defined. Based on structural and biochemical data, it is generally ac- cepted that actin is a collector of forces and an allosteric myosin modifier. According to the results of X-ray diffraction studies, the parameters of the actin helix do not change significantly upon contraction of the muscles (1). Therefore, conformational changes of actin are usually not analyzed in the modeling of muscle contraction. However, the relative stability of the struc- tural parameters of the thin filaments does not exclude the possibility of small conformational changes in the actin. There are experimental indications of conformational mobility of actin, and this property has been considered by a number of authors to be a necessary condition for the transformation of energy in muscle (2-4). In order to study the role of actin's structural mobility in the elementary mechanochemical process of muscle (the cycle of attachment and detachment of the myosin bridge), we conducted an experimental study of changes in the dynamics of muscle contraction under the influence of agents which change the structural From the Institute of Biophysics, USSR Academy of Sciences, Puschino, USSR. 137 stability of F-actin. Of particular interest is the cyclical peptide phalloidin (Ph) studied by Wieland et al. (5) because this peptide bonds specifically with actin, accelerates polymerization of G-actin (6), and increases the resistance of F-actin to many influences: pH, ultrasound (7), cytochalasin B, KI (8), and high temperature (9). On the other hand, Ph does not cause great changes in the functioning of the actomyosin system in solution. According to data derived from steady-state kinetics, Ph does not influence the activation of myosin by actin and does not change the Catt regulation of actomyosin containing regula- tory proteins (6). Therefore, we undertook a study of the effect of Ph on the contractile properties of glycerated muscles. MATERIALS AND METHODS The study involved the psoas muscle of the rabbit. The method of gly- cerinization has been described earlier (10). The fibers were stored until studied in a solution of 50 percent glycerin, 1 mM EDTA, 6.7 mM KH,PO4, pH 7.0, at -20° C for 10 months. The following solutions were used to study the mechanical properties of the fibers: 1. A rigor solution: 80 mM KCl, 15 imidazole, 5 mM MgClj, and 4 mM EDTA. 2. A relaxing solution: 80 mM KCl, 15 mM imidazole, 5 mM MgCl,, 5 mM ATP, and 2 mM EDTA. 3. A contracting solution: The relaxing solution plus 1.8 mM CaClp. The pH of all solutions was 6.8. The Catt concentration was calculated according to Portzehl et al. (11). In solution 2, pCa 2 8.5. Before the experiments, the fibers were washed in a solution of 20 percent glycerin and 1 mM EDTA for 1 hour at 0° C, and then in the rigor solution for at least 1 hour. The experiments used thin bundles of fibers less than 0.1 mm in diameter. To measure the mechanical properties, the ends of the bundle of fibers, 7-9 mm in length, were held in clamps, one of which was rigidly con- nected to a type GMK-1 electromechanical transducer, while the other was con- nected to an optical force transducer with a natural frequency of 50 Hz and a rigidity of 1.35 um/mg. The length of the sarcomeres of the fibers in the rigor solution was about 2.5 pm. The mounted fibers were placed in the relaxing solution at 0° C and shortened to a length of 2.2 um as the solution was heated. After the desired temperature of the solution was reached, the contracting solu- tion was added until the desired Ca’ concentration was reached. Measurements were performed with continuous agitation of the solution. In the experiments, the isometric tension and length-force characteristics were measured with an amplitude of length change of 0.1-0.5 percent of the length of the muscle at frequencies of 0.3-5 Hz. In order to study the influence of Ph on the mechanical properties of the fibers, an aqueous solution of 1 mM Ph was added to the cuvette, or the fiber was placed in a cuvette to which Ph had previously been added. In some cases, the fibers in rigor solution were first treated with 4+10~3 M Ph for 15 minutes at room temperature. 138 RESULTS After addition of the contracting solution, the isometric tension of the fibers increased and the work performed by them during the course of a cycle became positive, which reflected the ability of the muscles to work in a self- oscillating mode. Usually, it was sufficient to increase the concentration of Cat to 110-7 M to achieve this effect (12). In some cases, it was also nec- essary to stretch the fibers slightly. Positive values of work per cycle were observed in the range of 0.5-1.5 Hz at 20° C. Increasing the temperature by 10° C shifted this interval into the area of higher frequencies (usually 1-3 Hz) and increased the isometric tension by about 2.5 times. The influence of Ph on the contraction properties of the fibers depended on temperature. The results of a typical experiment of the effect of Ph on the contraction of fibers at 20° C are presented in figure 1. We can see that Tension (mg) min FIGURE 1. Changes in isometric tension and length-force characteristics during an experiment at 20° C. Initially, the fiber was placed in a relaxing solution. Numbers with asterisks indicate the amount of increase in [catt]-108 M as a result of the addition of the contracting solution. Arrows without numbers (=>, >>, — =>) indicate addition of the contracting solution, leading to an increase in [Catt] by 0.44108 M, 1.1-10-8 M, and 2.2:10-8 M, respec- tively. '"Phalloidin" indicates addition of phalloidin to a concentration of 2.54103 M. "Relax" indicates transfer to the relaxing solution. The numbers in circles indicate the moment of recording of the length-force characteristics, presented in the inset beneath the same number. On the inset the frequency of changes in length is 0.8 Hz, and the direction of movement of the representative point is indicated by the arrows. Counter-clockwise movement corresponds to performance of positive work during the period. 139 addition of Ph to the fibers, which developed active tension and performed positive work in the cycle at [Cat] = 10-7 M, led to multiphased changes in isometric tension and qualitative changes in the dynamic length-force charac- teristics. These characteristics were observed throughout the experiment. In all 10 experiments performed, the work cycle became positive after addi- tion of Ph, and the change of sign of work coincided in time with passage of the isometric tension through its minimum. We can see that, although the ten- sion returns approximately to its initial value, this does not lead to a restor- ation of relaxation properties. We can also see from the results that the fibers retain their sensitivity to Catt: They relax to the initial level when placed in the relaxing solution and develop tension with subsequent addition of Catt. The fibers also retain their capacity for contraction. At the end of the experiment, after they were released, they contracted to one-half of the length before release. The sensitivity of Ca’t and the capacity for contraction were retained also in solutions containing Ph. However, we did not succeed in produc- ing positive work per cycle at 20° C throughout the entire range of frequencies after addition of Ph, either by addition of Catt to 2+10-7 M, or by washing out the Ph in the relaxing or rigor solutions, or by stretching the fibers. Neither could the self-oscillating mode be produced in fibers treated with Ph without stretching in the rigor solution (see Materials and Methods). However, after any of the treatments used, the fibers performed positive work in the cycle at temperatures above 30° C in the frequency interval characteristic for these temperatures. The overall course of the change in isometric tension after addition of Ph at elevated temperature remained as before, although various stages were not accelerated in proportion. This can be seen in figure 2, which shows the changes in tension after treatment with Ph that were obtained in two bundles of fibers at various temperatures. After addition of Ph at 33° C, we sometimes observed narrowing of the ellipse, but the work per cycle remained nonnegative and it could always be made positive by means of fiber stretching. In order to study the influence of Ph on the mechanical properties of the rigor fibers (figure 3), a bundle of fibers was stretched in rigor solution (without preliminary incubation in solutions containing ATP) to a force level comparable with that developed in the contracting solution. After stretching, relaxation of the force began. Throughout the entire experiment, the muscle was subjected to sinusoidal changes in length with a frequency of 0.5 Hz and an amplitude of 0.1 percent of the fiber length. The width of the line on the force recording serves as a measure of the dynamic modulus of the fiber. We can see that the addition of Ph has practically no effect on either the tension or the speed of relaxation, nor on the dynamic modulus. DISCUSSION The main result of the studies performed is the demonstration of a change in the unsteady kinetics of muscle contraction under the influence of Ph. The high specificity of binding of Ph with actin allows us to assume that the effects 140 33°C 22°C phalloidin © E 40f C 0 phalloidin on C PP 20 + 5 min pry Time FIGURE 2. Variation of changes in isometric tension evoked by phalloidin (4+105 M) as a function of temperature with [cath] = 0.9-10-/ M. Excursions on curves at the moment of change of the composition of the solution are an artifact caused by transfer of the fiber to a new cuvette. observed result from the interaction of Ph with the thin filaments, although data concerning the interaction of this substance with the other components of muscle fiber are not available. The irreversibility of the effect of Ph--the properties of the fiber are not restored upon washing away Ph--also favors this assumption. We know that Ph binds very tightly with F-actin: One molecule of Ph per 20 molecules of G-actin is not exchanged (6). Furthermore, we know that Ph competes with destabilizing influences in its action on actin. This agrees with the observation that an increase in temperature restores the capacity for self-oscillation. We did not perform a quantitative study of the relaxation properties of the fibers over a long time interval. However, we know from the literature that the relaxation spectrum of the activated contracting system of the rabbit muscle is discrete, and that the time constants differ significantly (13,14). Since, in fibers treated with Ph, no positive work per cycle was observed throughout the entire range of frequencies studied at 20° C, while at 30° C all that was observed was a decrease in the quantity of work per cycle without a significant shift in the frequency interval of the zone of positive work, we can assume that Ph causes a change (decrease or even change in sign) of the 141 water addition | [ phalloidin addition © E Cc 0 2 o 20 + —— stretch 5 min TIME FIGURE 3. Effect of phalloidin on the mechanical properties of fibers in rigor solution. Phalloidin was added to a concentration of 4°10~° M, and water and the phalloidin solution were added in equal volumes. amplitude of the process responsible for development of the holding tension, i.e., Ph influences the magnitude of tension activation of the muscle. The relaxation parameters of the muscle, within the framework of models based on the sliding-filament hypothesis, are determined by the values of the kinetic constants of the interaction of actin and myosin, and their variation with the coordinate and force function of the individual cross-bridge (14-16). Absence of the effect of Ph on the mechanical properties of a muscle in rigor indicates (assuming that they reflect the mechanical properties of the attached bridge of an active muscle) that Ph influences the kinetic parameters of inter- action of actin and myosin during the ATPase cycle. The great magnitude and nonmonotonic nature of changes in isometric ten- sion after addition of Ph, and the varying temperature sensitivity of the vari- ous stages of these changes, apparently indicate that Ph influences several parameters. Since Ph stabilizes the structure of F-actin, we can assume that the change of these parameters depends on the conformational mobility of actin. However, it is also possible that the changes observed are the result of mod- ification of actin. In order to answer this alternative hypothesis, we must 142 study the changes in the relaxation spectra of the active muscle under the in- fluence of a broad range of substances which change the structural stability of actin. The authors are grateful to M. E. Sakson, who initiated our interest in study- ing the effect of Ph on muscles and who provided the preparation for performing this work, and to S. E. Shnol for useful discussions. 143 10. 11. 12. 13. REFERENCES Huxley HE, Brown W: The low-angle X-ray diagram of vertebrate striated muscle and its behavior during contraction and rigor. J Mol Biol 30: 383-434, 1967 Laki K: Actin as an energy transducer. J Theor Biol 44:117-130, 1974 Oosawa F, Fujime S, Ishiwata S, Mihashi K: Dynamic property of F-actin and thin filament. Cold Spring Harbor Symp Quant Biol 37:227-285, 1972 Lednev VV, Frank GM: [Structural non-equivalency of F-actin subunits and its possible significance in regulation of ATPase activity and the devel- opment of tension in skeletal muscles] (Rus). Biofizika 22:376-388, 1977 Wieland Th, Govindan VM: Phallotoxins bind to actins. FEBS Lett 46: 351-353, 1974 Dancker P, Low I, Hasselbach W, Wieland Th: Interaction of actin with phalloidin: polymerization and stabilization of F-actin. Biochim Biophys Acta 400:407-414, 1975 Low I, Dancker P, Wieland Th: Stabilization of actin polymer structure by phalloidin: ATP-ase activity of actin induced by phalloidin at low pH. FEBS Lett 65:358-360, 1976 Low I, Dancker P, Wieland Th: Stabilization of F-actin by phalloidin reversal of the destabilizing effect of cytochalasin B. FEBS Lett 54: 263-265, 1975 De Vries JX, Schafer AJ, Faulstich H, Wieland Th: Protection of actin from heat denaturation by various phallotoxins. Hoppe-Seylers Z Physiol Chem 357:1139-1143, 1976 Bukatina AE: [Variation in tension and ATPase activity of glycerinized muscle fibers with relative fiber length] (Rus). Biofizika 16:52-59, 1971 Portzehl H, Caldwell PC, Riiegg JC: The dependence of contraction and relaxation of muscle fibres from the crab Maia squinado on the internal concentration of free calcium ions. Biochim Biophys Acta 79:581-591, 1964 Riegg JC, Steiger GJ, Schadler M: Mechanical activation of the contractile system in skeletal muscle. Pfluegers Arch 319:139-145, 1970 Kawai M: Head rotation or dissociation? A study of exponential rate processes in chemically skinned rabbit muscle fibers when Mg ATP concen- tration is changed. Biophys J 22:97-103, 1978 144 14. 15. 16. Abbott RH, Steiger GJ: Temperature and amplitude dependence of tension transients in glycerinated skeletal and insect fibrillar muscle. J Physiol 266:13-42, 1977 Deshcherevsky VI: [Mathematical Models of Muscle Contraction] (Rus). Moscow, Nauka Press, 1977 Julian FJ, Sollins KR, Sollins MR: A model for the transient and steady- state mechanical behavior of contracting muscle. Biophys J 14:546-562, 1974 145 SIMILARITY OF ELEMENTARY STAGES OF ATP HYDROLYSIS IN VARIOUS ATPase SYSTEMS N. S. Panteléeva, E. A. Karandashov, I. E. Krasovskaya, N. V. Kuleva, E. G. Skvortsevich, and L. A. Syrtsova In studying the molecular principles of energy transfer in living systems, two aspects attract attention. The first is the similarity of elementary mechanisms of energy conversion in different functional systems, involving the participation of adenosine triphosphate (ATP). The second aspect is the sur- prising economy of energy processes, i.e., the ability of nature to make do with a limited number of elementary stages and reactions, the combinations of which support an amazing variety of manifested vital activity. We are confronted here with the phenomenon of unification in the economy of energy in living sys- tems, side by side with the great variety of forms of energy transformation. In 1979, we celebrate the 40th anniversary of the discovery of adenosine triphosphatase (ATPase) properties of myosin by Engelgardt and Lyubimova (1,2). During the intervening years, hundreds of experimental works and fundamental reviews have been published on the intermediate stages of ATP hydrolysis by myofibrillar, transport, and other ATPases. However, only in recent years have we begun to approach the concept of the 'coupling stage,'" that stage in which the energy bound in ATP is utilized in the form of mechanical and other types of work. The process is achieved by energizing proteins (ATPase), and by conservation of ATP energy in isomers, or conformationally altered struc- tures, with its subsequent liberation in the performance of work (2-7). In this process, "Of all types of energy transformation, only its degradation in the form of heat occurs directly, without the participation of any specialized mechanisms. In all other cases, definite physical tools of widely varying degrees of complexity are required" (V. A. Engelgardt). Adenosine triphosphatases are enzymes involved in the utilization of ATP energy, and are distinguished by a unique property: Their subunit structure and spatial organization are quite varied (figure 1). However, at the same time, the final chemical manifestation of the resultant enzymatic reaction is the same for all ATPases, i.e., splitting of the single oxygen-phosphate bond at the gamma-phosphoryl group of the ATP. Doubtless, the structural differences result from the specific type of biological work performed. Although they split From the Department of Biochemistry, Leningrad State University, Leningrad, and from the Institute of Chemical Physics, USSR Academy of Sciences, Chernogolovka, USSR. 147 00 ATP + H,0 0D ONS ADP + P; or () l pd FIGURE 1. Structural complexes of different forms of adenosine triphosphatase (ATPase). A = Mysoin ATPase (HMM subfragment 1). B = Nat, Kt-ATPase. C = Mitochondrial HY-ATPase (F; factor). D = catt, Mgtt-ATPase. E = Nitrogenase (E-1 and E-2, ATPase centers). ADP = Adenosine diphosphate. the very same bond in the ATP molecule, the ATPases of various structures cannot replace each other. Myosin ATPase, for example, although it splits ATP, cannot transport Catt against a concentration gradient. ATPases are strictly specific, not only in terms of the chemical bond they split, but also with respect to the process for which they supply energy. Here we have an illustration of the phe- nomenon of "unity in diversity." In this paper, we demonstrate this phenomenon on a number of ATPases which we studied using the method of 180 exchange-reactions. The method of 180 exchange has been used increasingly in recent studies of the mechanism of action of ATPases. It allows us to obtain information on the possible covalent and noncovalent intermediates, on the sequence of reactions of attachment and splitting of substrates and reaction products in the catalytic center, and on stages limiting the rate of hydrolysis as a whole (3,5,6,8-10). 180-exchange reactions have been found in systems related both to the storage and utilization of ATP energy. 18¢ exchange is accompanied by trans- formation of ATP and orthophosphate in systems of myofibrillar ATPases, in sys- tems of oxidative and photosynthetic phosphorylation, and in the functioning of transport ATPases (3-13). 148 At the present time, two types of reactions of oxygen isotope exchange, catalyzed by ATP phosphohydrolases, are arbitrarily distinguished. The first is intermediary 18¢ exchange, which occurs either in the intermediate stage of firmly bound ATP with the enzyme (E-ATP 2 Hy180). The second type of reaction is exchange with the environment, or direct 18¢ exchange, occurring with free Pi, added to the medium or formed as a result of hydrolysis (H3PO4 2 H2180 or H3P180, 2 H0). Both types of 18p-exchange reactions have been found in systems utilizing ATP for widely varying purposes. We have studied myofibrillar and transport ATPases, as well as the ATPase center of the nitrogenase enzyme, which fixes atmospheric nitrogen and reduces it to ammonia. As can be seen from table 1, myosin from various species and various types of muscles catalyzes intermediary 180 exchange (in the process of ATP hydrolysis and in the presence of Mgtt), whereas direct 189 exchange (with no hydrolysis) is quite low. This reaction is stimulated only by ADP and Mott, and also during complex formation with actin (8,14). The available data indicate that there is a quantitative difference in the values of intermediary 180 exchange for myosin from various types of muscles: For skeletal myosin from the rabbit and carp (15), the metabolic values are significantly higher than for cardiac (16) and smooth muscle myosin (17). The quantitative difference doubtless reflects differences in the rates of equivalent intermediate stages of ATP hydrolysis (table 2), catalyzed by skeletal, cardiac, and smooth muscle myosin. TABLE 1. Intermediary and Direct 18, Exchange in the System of Myofibrillar ATPases (in the presence of Mgtt) Atoms of 180 per Molecule of Pj Direct 180 Exchange Intermediary 189 Exchange Protein Preparation Mgt Mott + ADP Skeletal myosin (HMM or HMM SI), rabbit 2.1; 2.8 0; 0.2 2.9; 3.1 Actomyosin (or actoHMM SI), rabbit 0.10; 0.20 0.8; 2.8 —— Skeletal myosin, carp 1.48; 2.56 0.28; 0.32 0.55; 1.22 Smooth muscle myosin from calf small intestine 0.16; 0.82 0.10 0.26; 0.64 149 TABLE 2. Rate Constants of Certain Intermediate Stages in the Hydrolysis of ATP, Catalyzed by Myosin From Various Types of Muscles Rabbit Chick _1 Skeletal Ox Cardiac Smooth Muscle Stage (6) and ky (s 7) Myosin (6) Myosin (18) Myosin (19) k 2 400 100 100 -M! 0 a Q 200 = I o g co °3 Si =F GTP+ t = Isoproterenol > < E 100- x —0 r2 E a - © O GTP 5 x A th Lr o# T T T | 0 0.025 0.05 0.10 0.15 [B82 Extract] (mg/ml) FIGURE 1. Reconstitution of catecholamine-stimulated adenylate cyclase activity with cye™ S49 cell plasma membranes and Lubrol extracts of B82 cell plasma mem- branes. Adenylate cyclase activity in reconstituted mixtures was assayed in the presence of 100 uM of guanosine triphosphate (GTP) (A), 1 uM of GTP plus iso- proterenol (J), or 10 mM of NaF (0). Stimulation by isoproterenol above that by GTP is also shown (e). In this experiment increasing amounts of B82 extract were mixed with a fixed amount of cyc~ membranes. [Reprinted by permission of Proc Natl Acad Sci USA from Ross and Gilman (8)] Resolution of Multiple Components of Adenylate Cyclase If this simple explanation were true, the adenylate cyclase activity of the reconstituted mixture should mirror any destructive manipulation of activity in the donor extract. The data in figure 2 show that this is not the case. If the donor extract (from wild type S49 membranes in this experiment) is incubated at 30° C prior to reconstitution, its adenylate cyclase activity declines with a ty of about 4 minutes (dashed lines, figure 2). However, such incubation causes only a slight decrease in the activity that is observed after subsequent reconstitution of incubated extracts with cyc~ membranes (figure 2). Notably, the addition of a completely inactive extract (i.e., heated 30 minutes) to cyc~ membranes reconstitutes basal activity as well as fluoride-, Gpp(NH)p-, and hormone-stimulated enzymatic activities. (An analogous experiment performed with donor extract that has been inactivated by N-ethylmaleimide yields essen- tially identical results.) Reconstitution of basal and fluoride- or Gpp (NH)p- stimulated activity can also be observed if a heated (or N-ethylmaleimide inactivated) extract of either wild type or B82 membranes is mixed with a 161 100 Donor at 30° 7 OE 02 S 30 Gpp(NH)p a £ Oo ~~ 7 Oo a © 20 o £ S L m—C] a oN INE+GTP \ 2 10 8 “ oo AN “ > SN O AN o SF so 3 > c Q “© << 2.5 OS N.Y ——Gpr(NH)p o AN NaF “ “ \ “ SON 1 1 1 Ss 1 1 1 0 5 10 15 20 25 30 Time (min) FIGURE 2. Reconstitution of hormone-sensitive adenylate cyclase by mixture of cyc™ membranes with heat-inactivated extracts of wild type membrane. Detergent (Lubrol 12A9) extracts of wild type membranes were heated at 30° C for the times shown on the abscissa, chilled, and mixed with cyc™ membranes. --- = NaF- and Gpp(NH)p-stimulated adenylate cyclase activities of the incubated extracts (prior to reconstitution). Aliquots of the reconstituted mixtures, prepared with these incubated extracts, were assayed in the presence of guanosine tri- phosphate (GTP) (o), isoproterenol (INE) plus GTP (0), NaF (A), or Gpp(NH)p (V). [Reprinted by permission of the Journal of Biological Chemistry from Ross et al. (7)] detergent extract of cyc™ membranes, since these activities of the system can be measured in detergent solution (18). The above data, and data from several other experiments (7,18), suggest that at least two factors are necessary for expression of MgATP-dependent adenylate cyclase activity. The cyc™ membranes or a Lubrol extract of cyc™ membranes supply a heat-labile, N-ethylmaleimide-sensitive factor that is inactivated in the wild type extract by incubation at 30° C or by treatment with the sulfhydryl reagent. This hypothesis is supported by the fact that the ability of a cyc~ extract to reconstitute enzyme activity is lost at 30° C with the same kinetics as is total wild type adenylate cyclase activity. The reconstituting factor from cyc™ also shows identical sensitivity to 162 N-ethylmaleimide as does wild type adenylate cyclase. Both of the factors that are necessary for reconstitution of any adenylate cyclase activity (the thermolabile factor in cyc™ membranes and the thermostable factor in heated wild type extract) appear to be proteins, since they are inactivated by a variety of proteases and by N-ethylmaleimide (7,18). Identification of the Thermolabile Factor from cyc~™ S49 Cells Since two proteins are required for adenylate cyclase activity, one must be the catalyst and the other must serve to activate and to regulate. Since we obviously wanted to assign roles to the resolved proteins, it was of interest to discover that cyc~ plasma membranes and their Lubrol extracts are capable of catalyzing the synthesis of cyclic AMP if MnATP, rather than MgATP, is used as the substrate (7). Evidence strongly suggests that the thermolabile protein from cyc™ that is necessary for reconstitution of MgATP-dependent adenylate cyclase activity is the same as the protein that can synthesize cyclic AMP with MnATP as substrate. We thus feel that the protein from cyc~ that is required for reconstitution is the catalytic subunit of adenylate cyclase and, for con- venience, will refer to it as C. The data that support this hypothesis have been reported (7). The most important of these experiments demonstrate that the two activities cofractionate by gel filtration (figure 3) or by sucrose density gradient centrifugation. These studies have also allowed calculation of hydrodynamic parameters for C (table 2). Properties and Partial Purification of the Thermostable Component of Adenylate Cyclase The minimal role of this component must be to confer upon C the ability to use MgATP as substrate. In addition, since the MnATP-dependent activity of cyc™ membranes is not stimulated by NaF, Gpp(NH)p, or hormone, the more thermostable protein is necessary for regulation of catalytic activity by these ligands. Experiments described below indicate that, in fact, this component of adenylate cyclase is a molecular site of action of guanine nucleotides and sodium fluoride; for this reason we have, for now, designated this component as G/F. The data presented above indicate that G/F may be resolved from C by incu- bation (at 30-37° C) of plasma membrane extracts from either wild type S49 cells or B82 cells, and that C is resolved from G/F in cyc~ cells by genetic alteration. The generality of these observations would be supported by the resolution of G/F and C by other techniques and from other tissues, and by the isolation of pheno- typically adenylate cyclase-deficient variant clones that lack C but retain G/F. G/F may be resolved from C by heating of plasma membranes from several tissues (rat brain, rat or rabbit liver, and turkey or pigeon erythrocytes) (7). As mentioned above, it can also be resolved from C by chemical means, since C is almost 100-fold more sensitive to the sulfhydryl reagent N-ethylmaleimide than is G/F. This chemically resolved G/F can interact with cyc~ membranes to reconstitute each of the relevant Mg2t-dependent adenylate cyclase activities with efficiency comparable to that of G/F prepared by heat treatment. Work is in progress to attempt to separate G/F and C in active states from wild type S49 163 - yA a(nm) y oO ~ < a > 5 © © 4 140 3 @ 5 2 3 - i NS 23 3 o 3 oO 3 73° o. © © EN <> r 3 a w= > 0 cE 2 120 ©3 Ts | =2 23 fo 0 E 1 410 | © oa | a +! 0 = % ~ 2 Ss | 1 1 ] © 10 20 30 40 50 Fraction Number FIGURE 3. Gel filtration of C activity (the catalytic subunit of adenylate cyclase) and MnATP-dependent adenylate cyclase from a cyc~ S49 plasma membrane extract. Aliquots were assayed for adenylate cyclase activity either in the presence of 10 mM MnCly (0) or, after reconstitution with a heat-inactivated wild type extract, in the presence of 10 mM MgCly plus 0.1 mM Gpp(NH)p ((J). The upper panel is a calibration curve for determination of the Stokes radius. The arrows show the elution volume of blue dextran (left) and 2,4-dinitro- phenylglycine (right). [Reprinted by permission of the Journal of Biological Chemistry from Ross et al. (7)] cells or from other sources. This is complicated, in particular, by the lability of C. The clonal HC-1 hepatoma cell line, which is also nearly completely devoid of adenylate cyclase activity, serves as an example of genetically resolved G/F. A detergent extract of HC-1 plasma membranes, which has no adenylate cyclase activity when assayed with either MnATP or MgATP, can combine with cyc” mem- branes to yield fluoride-, Gpp(NH)p-, or hormone-stimulated enzymatic activity (7). 164 Physical characterization of G/F by gel filtration and sucrose density gradient centrifugation yields values of hydrodynamic parameters that are shown in table 2. Identical values are obtained with G/F resolved from C by heat treatment, exposure to N-ethylmaleimide, or genetically (HC-1). It was hoped that comparison of the physical properties of C, G/F, and Gpp(NH)p-activated (stabilized) adenylate cyclase would yield hints about the mechanisms of regulation of catalytic activity; at the moment the answer is not clear. The molecular weight of the Gpp(NH)p-activated enzyme is not simply re- lated to the sum of the molecular weights of C and G/F. While it must be recog- nized that these studies of crude extracts are subject to errors and that the calculations involve assumptions, the discrepancy in molecular weights may suggest a complex molecular explanation for the persistent activation and stabilization of adenylate cyclase by Gpp(NH)p or it may be indicative of self-association of G/F or C in detergent extracts. Exposure of G/F to either Gpp(NH)p or fluoride results in "activation'" of the protein and changes in its physical properties, indicating that G/F is a site of action of these ligands. Thus, when G/F is incubated with Gpp(NH)p and Mg2+t and the mixture is diluted prior to the addition of C (such that the concen- tration of Gpp(NH)p is now below that required for activation), enzymatic ac- tivity equivalent to that of the Gpp(NH)p-activated state is observed initially. There is no effect of similar exposure of C to the guanine nucleotide. Analogous experiments with fluoride yield essentially the same result. However, activation of G/F by fluoride also requires the presence of Mg2t and a nucleotide; ATP is preferred. Activation of G/F by Gpp(NH)p or by fluoride plus ATP is reversible. The activated (and presumably liganded) state decays within a few minutes when exposure to fluoride is terminated; reversal of the effect of Gpp(NH)p is slower. G/F is also altered in other properties following exposure to fluoride plus ATP or to Gpp(NH)p. In this state it is more readily incorporated into cyc~ membranes (19), and its sedimentation coefficient is reduced (figure 4). Assess- ment of hydrodynamic parameters of G/F in the presence of Gpp(NH)p or ATP plus fluoride suggests that it has lost mass (as much as 40,000 daltons). However, the change is completely reversible, even after attempts to remove a hypothetical component that may have dissociated during treatment with the activating ligands. At the moment, therefore, our understanding of the mechanism of this change is rudimentary. Because of its greater stability, we have concentrated our initial attempts to purify components of adenylate cyclase on G/F. To date, the protein (from rabbit liver) has been purified several thousandfold (relative to the activity of purified membranes) with good overall recovery (~ 25 percent). The specific activity of the preparation of this stage (assayed following reconstitution with cyc™ membranes) is 1-2 umol/min/mg protein. Techniques that have proven useful include chromatography on DEAE cellulose, Ultrogel AcA34, heptane-substituted agarose, hydroxylapatite, and GTP-substituted Sepharose (20). 165 10 Adenylate Cyclase (pmol /min/fraction) 5 10 15 20 Fraction Number FIGURE 4. Sedimentation of the regulatory protein G/F through sucrose density gradients in the absence (eo) or presence (A) of MgATP plus NaF. Lubrol extracts of wild type S49 cell membranes, which had been incubated at 37° C for 15 min- utes to inactivate C, were then incubated at 30° C for 15 minutes in the ab- sence or presence of 10 mM MgCly, 2 mM ATP, and 10 mM NaF. Extracts were layered on top of linear gradients of sucrose (5 to 20 percent) and centri- fuged for 15 hours at 48,000 rpm in a Type SW50.1 rotor (Beckman). Aliquots were assayed for G/F activity, following reconstitution with cyc™ membranes, in the presence of NaF. In the absence of ligands, the sedimentation coefficient is 4.8S; in their presence, the coefficient is 3.8S. Reconstitution of cyc™ and UNC Membranes With Cholate Extracts Solubilization of wild type S49 cell membranes with 25 mM sodium cholate results in an extract that has largely lost detectable adenylate cyclase ac- tivity and in which the remainder is rapidly destroyed by incubation at 25° C. However, these extracts retain G/F activity and are very useful for reconstitu- tion of both cyc™ and UNC membranes (figure 5) by a specific protocol described elsewhere in detail (9). When the total concentration of added detergent is held constant in this procedure, titration of increasing amounts of active ex- tract results in a nearly linear increase of reconstituted (hormone-stimulated) activity that becomes maximal as optimal levels of extract are added. Membranes that have been reconstituted by this technique have properties that are essentially indistinguishable from those derived from wild type cells. Thus, reconstituted adenylate cyclase activity in these membranes can be stimu- lated by B-adrenergic agonists, PGE], guanine nucleotide analogs, and fluoride. 166 1 I 1 I 400 T T ] yaa A 300 4 — - o 0 2° “. a c NN ° \ o oO , AN 8 200 /’ N - £ ’ “ , Q ¢ NS E ’ o, J \ 9 ’ \ 2 100 | J (N N\ u 2 ¢ “ 2 J “ oe 0, ~ %0 -S —e—" SS. ‘“~ £ 0 be + — £ > oO__n0 Oo. — a © 300F Oo vg 0 B | E Le o wv o 3 i 3 200 | 2+ a 2 3 9 . > o c z [7] I 2 100 A x S 1.4 Extract Added (mg/mg UNC protein) FIGURE 5. Reconstitution of uncoupled (UNC) membranes with a cholate extract of wild type membranes. A: UNC membranes were mixed with increasing amounts of extract and were reconstituted as described by Sternweis and Gilman (9). B: The total amount of extract added to the reconstitutions was kept constant by supple- menting increasing amounts of active extract with decreasing amounts of extract that had been inactivated at 60° C for 20 minutes. The inset is a recalculation of the data in part B as indicated. Adenylate cyclase assays contained 50 uM GTP (e,0,A), 2 uM isoproterenol (o,A), 2 uM propranolol (A), or 10 mM NaF (0). [Reprinted by permission of the Journal of Biological Chemistry from Sternweis and Gilman (9)] 167 Specific activities of the enzyme and the dependence of enzymatic activity on the concentrations of activating ligands are all normal. The Kp for the binding of isoproterenol to the recoupled B-adrenergic receptors, the ratio of the Kp for isoproterenol to its value of K,.. (see 13, 21), and the ability of guanine nucleotides to cause agonist-specific alterations in receptor affinity all re- semble the properties of membranes from wild type cells. Using this technique as an assay for the factor(s) necessary for reconsti- tution of the UNC membrane, we are attempting to determine the nature of this interesting lesion. At the present time we can say that the defect is correct- able by the appropriate addition of one or more factors to UNC membranes, and that a crucial factor is temperature-sensitive and susceptible to inactivation by N-ethylmaleimide and by trypsin (9). The factor is indistinguishable from G/F in its thermal lability and sensitivity to N-ethylmaleimide, or by sucrose density gradient centrifugation. Furthermore, the most purified fractions of G/F that are available are fully active in the reconstitution of UNC membranes. Finally, it should be noted that cyc~ (i.e., G/F”) and UNC variants are not complementary in vitro, indicating that their lesions overlap. The evidence to date thus hints strongly at a close relationship between G/F and the factor necessary for reconstitution of UNC membranes. We feel that there are two likely hypotheses to explain the UNC defect. One hypothesis is that there may be two separate components--G/F and a factor that is missing in UNC. If this is the case, it seems likely that the novel factor necessary to reconstitute UNC may be tightly bound to G/F, relatively small, and dependent on G/F for incorporation into the membrane. Equally possible is the hypothesis that G/F is subject to post-translational modification; a defect in a component necessary for such covalent modification of G/F or its reversal could readily explain the uncoupled phenotype. DISCUSSION We have largely discussed the significance of our own data above and will conclude with speculation on the relationship between the components that we have described and those that have been characterized by others. Pfeuffer (20) has performed affinity chromatography of detergent-solubilized pigeon erythrocyte membranes on GTP-substituted Sepharose. Adenylate cyclase that passes through the column has decreased Gpp(NH)p- and fluoride-stimulated activity. A protein that is eluted from the Sepharose by GTP or Gpp(NH)p reverses this loss of ac- tivity; this protein appears to have a molecular weight of 42,000 by sodium dodecyl sulfate polyacrylamide gel electrophoresis. It is reasonable to suggest a relationship between this protein and G/F, although it is not yet possible to compare the molecular weight with our data that indicate a somewhat larger size for native G/F. Cassel and Selinger (22,23) have described a GTPase activity of avian eryth- rocyte membranes that is stimulated by B-adrenergic agonists, and they have developed a plausible model that envisions the binding of GTP (or the release of GDP) as a crucial step in enzymatic activation and the hydrolysis of GTP as essential for the return of activity to the basal state. Hormone-stimulated GTPase activity is inhibited by cholera toxin, and toxin-treated enzyme is 168 activated by GTP in a manner similar to that normally seen with Gpp(NH)p (24). Experiments performed by Moss and Vaughan (25) and by Gill and Meren (26) in- dicate that cholera toxin catalyzes the ADP-ribosylation of a membrane-bound protein that is important for the regulation of adenylate cyclase; observations of Cassel and Pfeuffer (27) and Gill and Meren (26) indicate that the 42,000 dalton GTP-binding protein is a substrate. Our own experiments (19) and those of Johnson et al. (28) demonstrate that the activity of G/F is modified following exposure to cholera toxin, and it thus may be a substrate. The probability of interrelationship is thus strong. It may be hypothesized that G/F has a subunit of 42,000 daltons, that this is a binding site for GIP and a site of cholera toxin-catalyzed ADP-ribosylation, and that it constitutes at least a portion of a GTPase activity that is crucial for the regulation of adenylate cyclase. Proof will require completion of the purification of G/F. 169 10. 11. 12. 13. REFERENCES Daniel V, Litwack G, Tomkins GM: Induction of cytolysis of cultured lymphoma cells by adenosine 3':5'-cyclic monophosphate and the isolation of resistant variants. Proc Natl Acad Sci USA 70:76-79, 1973 Coffino P, Bourne HR, Friedrich U, Hochman J, Insel PA, Lemaire I, Melmon KL, Tomkins GM: Molecular mechanisms of cyclic AMP action: a genetic approach. Recent Prog Horm Res 32:669, 1976 Bourne HR, Coffino P, Tomkins GM: Selection of a variant lymphoma cell deficient in adenylate cyclase. Science 187:750-752, 1975 Insel PA, Maguire ME, Gilman AG, Bourne HR, Coffino P, Melmon KL: Beta adrenergic receptors and adenylate cyclase: products of separate genes? Mol Pharmacol 12:1062-1069, 1976 Haga T, Ross EM, Anderson HJ, Gilman AG: Adenylate cyclase permanently uncoupled from hormone receptors in a novel S49 lymphoma cell variant. Proc Natl Acad Sci USA 74:2016-2020, 1977 Ross EM, Maguire ME, Sturgill TW, Biltonen RL, Gilman AG: The relationship between the B-adrenergic receptor and adenylate cyclase. Studies of ligand binding and enzyme activity in purified membranes of S49 lymphoma cells. J Biol Chem 252:5761-5775, 1977 Ross EM, Howlett AC, Ferguson KM, Gilman AG: Reconstitution of hormone- sensitive adenylate cyclase activity with resolved components of the enzyme. J Biol Chem 253:6401-6412, 1978 Ross EM, Gilman AG: Reconstitution of catecholamine-sensitive adenylate cyclase activity: interaction of solubilized components with receptor- replete membranes. Proc Natl Acad Sci USA 74:3715-3719, 1977 Sternweis PC, Gilman AG: Reconstitution of catecholamine-sensitive adeny- late cyclase: reconstitution of the uncoupled (UNC) variant of the S49 lymphoma cell. J Biol Chem 254:3333-3340, 1979 Salomon Y, Londos C, Rodbell M: A highly sensitive adenylate cyclase assay. Anal Biochem 58:541-548, 1974 Aurbach GD, Fedak SA, Woodard CJ, Palmer JS, Hauser D, Troxler F: Beta- adrenergic receptors: stereospecific interaction of iodinated beta block- ing agents with high affinity sites. Science 186:1223-1224, 1974 Maguire ME, Wiklund RA, Anderson HJ, Gilman AG: Binding of [1251)i0do- hydroxybenzylpindolol to putative B-adrenergic receptors of rat glioma cells and other cell clones. J Biol Chem 251:1221-1231, 1976 Maguire ME, Ross EM, Gilman AG: B-adrenergic receptor: ligand binding properties and the interaction with adenylyl cyclase. Adv Cyclic Nucleotide Res 8:1-83, 1977 170 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Limbird LE, Lefkowitz RJ: Resolution of B-adrenergic receptor binding and adenylate cyclase activity by gel exclusion chromatography. J Biol Chem 252:799-802, 1977 Haga T, Haga K, Gilman AG: Hydrodynamic properties of the B-adrenergic receptor and adenylate cyclase from wild type and variant S49 lymphoma cells. J Biol Chem 252:5776-5782, 1977 Granner D, Chase LR, Aurbach GD, Tomkins GM: Tyrosine aminotransferase: enzyme induction independent of adenosine 3',5'-monophosphate. Science 162:1018-1020, 1968 Brunton LL, Maguire ME, Anderson HJ, Gilman AG: Expression of genes for metabolism of cyclic adenosine 3':5'-monophosphate in somatic cells. B- adrenergic and PGE; receptors in parental and hybrid cells. J Biol Chem 252:1293-1302, 1977 Ross EM, Gilman AG: Resolution of some components of adenylate cyclase necessary for catalytic activity. J Biol Chem 252:6966-6969, 1977 Howlett AC, Sternweis PC, Macik BA, Van Arsdale PM, Gilman AG: Reconsti- tution of catecholamine-sensitive adenylate cyclase: association of a regulatory component of the enzyme with membranes containing the catalytic protein and B-adrenergic receptors. J Biol Chem 254:2287-2295, 1979 Pfeuffer T: GTP-binding proteins in membranes and the control of adenylate cyclase activity. J Biol Chem 252:7224-7234, 1977 Howlett AC, Van Arsdale PM, Gilman AG: Efficiency of coupling between the beta adrenergic receptor and adenylate cyclase. Mol Pharmacol 14:531-539, 1978 Cassel D, Selinger Z: Catecholamine-stimulated GTPase activity in turkey erythrocyte membranes. Biochim Biophys Acta 452:538-555, 1976 Cassel D, Selinger Z: Mechanism of adenylate cyclase activation by cholera toxin: inhibition of GTP hydrolysis at the regulatory site. Proc Natl Acad Sci USA 74:3307-3311, 1977 Levinson SL, Blume AJ: Altered guanine nucleotide hydrolysis as basis for increased adenylate cyclase activity after cholera toxin treatment. J Biol Chem 252:3766-3774, 1977 Moss J, Vaughan M: Mechanism of action of choleragen. Evidence for ADP- ribosyltransferase activity with arginine as an acceptor. J Biol Chem 252:2455-2457, 1977 Gill DM, Meren R: ADP-ribosylation of membrane proteins catalyzed by cholera toxin: basis of the activation of adenylate cyclase. Proc Natl Acad Sci USA 75:3050-3054, 1978 171 27. 28. Cassel D, Pfeuffer T: Mechanism of cholera toxin action: covalent modifi- cation of the guanyl nucleotide-binding protein of the adenylate cyclase system. Proc Natl Acad Sci USA 75:2669-2673, 1978 Johnson GL, Kaslow HR, Bourne HR: Reconstitution of cholera toxin- activated adenylate cyclase. Proc Natl Acad Sci USA 75:3113-3117, 1978 172 THE STRUCTURE OF ACTIVE SITES OF ADENYLATE CYCLASE, PROTEIN KINASE, AND PHOSPHODIESTERASE E. S. Severin, T. V. Bulargina, N. N. Gulyaev, M. N. Kochetkova, and V. L. Tunitskaya INTRODUCTION Recently, increasing attention has been given to the role of cyclic nucleo- tides in the regulation of cellular metabolism, especially the mechanism of action of enzymes participating in the metabolism of cyclic adenosine 3',5'- monophosphate (cAMP) and the manifestation of its biological activity. Metabolic transformation of cAMP involves two enzymes--adenylate cyclase, the cAMP biosynthetic enzyme, and phosphodiesterase, the enzyme which splits cAMP to 5'-AMP. As a result of the hormone's effect on the corresponding recep- tor protein, activation of the catalytic component of adenylate cyclase occurs, catalyzing the formation of cAMP from ATP. The biological effect of cAMP is realized through the activation of a spe- cial class of enzymes--cAMP-dependent protein kinases. When cAMP binds with the regulatory subunit of protein kinase, the active catalytic subunit of the enzyme is released. This subunit catalyzes the phosphorylation of a number of functionally important proteins, and may be important to the functional activity of the cell. Because of the unusually broad participation of cAMP in the regulation of widely varied cellular processes (1), study of the mechanism of action and structure of the active sites of adenylate cyclase, protein kinase, and phos- phodiesterase is of interest. Especially valuable to these studies is the use of ATP and cAMP analogs containing reactive groups in their molecules. Because of their structural similarity to ATP or cAMP, such compounds have increased af- finity for active or regulatory sites and, in the presence of reactive groups, can form covalent bonds with these portions of the enzymes. This interaction is highly selective and allows us to study the structure of the active sites of enzymes which are difficult to purify to a homogeneous state, e.g., phospho- diesterase, or which have never been purified in an active and homogeneous form, e.g., adenylate cyclase. From the Department of Biochemistry, Moscow State University, Moscow, USSR. 173 In this paper, we report on a study of the structure of the active sites of adenylate cyclase and protein kinase using ATP analogs, and the active site of phosphodiesterase using cAMP analogs. Most of the compounds that were used contained active groups in their molecule that were capable of covalently block- ing the functional groups of the enzyme. Specifics of the structure of ATP- binding portions of adenylate cyclase and protein kinase are discussed, as well as the similarity and differences in the structure of cAMP-binding sites of phosphodiesterase and the regulatory subunit of protein kinase. INTERACTION OF ADENYLATE CYCLASE WITH ATP ANALOGS We know that in eukaryotic cells the catalytic component of adenylate cyclase is firmly bound to the plasma membrane of the cell and is easily in- activated during purification of the enzyme. Because of the lack of active and homogeneous preparations of adenylate cyclase, it is impossible to study the adenylate cyclase complex at the molecular level. Therefore, studies directed toward the structural-functional nature of individual components of this com- plex recently have become more significant. In the present work, we studied the ATP-binding portion of solubilized adenylate cyclase from rabbit heart, using ATP analogs which we synthesized and which had functionally active groups in various positions of the molecule (2-4). The ATP analogs which were selected (see table 1) can be divided into three groups: compounds with substitution only at the triphosphate portion of the ATP molecule (I-IX, XI, and XII); compounds with substitution at the C(8) position of the purine ring and with modification of the triphosphate portion (XIII-XVI); and, a compound with a modified ribose ring and triphosphate frag- ment (X). All of the compounds, in spite of the presence of alkylating groups in their structure, were found to be reversible and competitive (with ATP) adenylate cyclase inhibitors (table 1). For compounds I-VI, there was a definite increase in affinity for the ATP- binding portion of the enzyme with increasing numbers of phosphorus atoms in the molecule, indicating the importance of the triphosphate fragment of ATP for binding with adenylate cyclase. Lack of any irreversible inhibition in the presence of IX and XII analogs, with an alkyl halide substituent in the a,B- and B,y-methylene links, respectively, may indicate that either the conformation of the phosphate fragment of IX (or the triphosphate portion of ATP) in the active site is such that the substituent (or the a,B-bridge oxygen atom of ATP) is directed away from the catalytic group, or that the enzyme lacks a func- tional group in the area of binding of the y-phosphate of ATP. Furthermore, the presence of any functional group of the enzyme in the area of binding of the 2'-0H group of the ribose fragment is improbable, since compound X, containing a 2',3'-0-isopropylidene group, has high affinity for the active site. All 8-substituted analogs of ATP (XIII-XVI), which have preferentially the syn- conformation, showed low affinity for the ATP-binding site of adenylate cyc- lase, from which we can conclude that the latter apparently is more specific for the anti-conformation of the substrate. Based on these results, we also can suggest a structure for the Mg2+-ATP complex in the active adenylate cyclase site, in which the hydroxyl at the 174 TABLE 1. Inhibitory Effect of ATP Analogs on the Activity of Solubilized Adenylate Cyclase From Rabbit Heart Rs Ry Compound (mM) Compound (mM) I. (BrCH,CH,)pA 2.12 X. p(CH)pA-2'3'-0-isopropylidene 0.192 II. (BrCH,CH,)ppA 1.8% NHCOCH,, Br III. (BrCH,CH,) pppA 0.5% XI. p(CH,)ppA 0.29% Iv. (CICH,)pA 2.2% | XII. p(CH)ppA 0.202 V. (CLCH,)ppA 1.72 NHCOCH,, Br VI. (CICH,)pppA 0.19" | XIII. p (CH.,) ppA-8-NH (CH,) [NHCOCH,C1 0.91% VII. (CLCH,CH,O0)pA 1.0% | xiv. PPPA-8-NH (CH, ) [NHCOCH, C1 1.6" VIII. pp(CH,)pA 0.09% | xv. pA-8-NH(CH,) ,NHCOCH, C1 1.9° IX. p(CH)pA 0.42% | XVI. pppA-8-Br 1.8" NHCOCH, Br 8Enzyme preparation with K; = 0.075 mM bEnzyme preparation with Kj = 0.100 mM a-phosphorus atom is close to the 3'-OH group of the ribose group. The Mg2+ apparently forms a complex with the hydroxyl groups at the B- and y-phosphorus atoms and are located on the same side of the triphosphate fragment of ATP as the hydroxyl group of the a-phosphorus atom which, in turn, is directed oppo- site from the o,B- and B,y-bridge oxygen atoms (see figure 1). INTERACTION OF THE CATALYTIC SUBUNIT OF PROTEIN KINASE WITH ATP ANALOGS The structure of the active site of cAMP-dependent protein kinases from skeletal muscle has not been studied extensively at the present time. The ATP analogs used by previous authors for this purpose contained no active groups in the structure of the molecules (5,6), which made it impossible to establish the nature of the groups in the active site. We studied the interaction of the catalytic subunit of protein kinase from pigeon breast muscle with the ATP analogs mentioned earlier (table 1), or specially prepared by us (7). The ATP analogs were modified at the triphosphate ATP fragment, and most of them con- tained reactive groups in this portion of the molecule (tables 1, 2, and 3). 175 anti-conformation N of adenosine ring e Hydrogen @ Carbon O Oxygen Phosphorus FIGURE 1. Tentative structure of Mg2+-ATP complex in the active site of adenylate cyclase. In compounds I-IV, VI-VIII, XI, XVII, and XVIII, inhibition was found to be reversible and competitive with ATP, indicating the specificity of the bind- ing of these analogs in the active site of the catalytic subunit (table 2). As can be seen from table 2, in the sequence of analogs I-III and IV-VI, compound II had the greatest affinity for the active site of the enzyme, since this compound has two phosphate groups in its molecule. However, increasing the number of phosphate groups to three, and simultaneously introducing a sub- stituent at the y-phosphorus atom (compounds III and VI), led to a drop in the affinity by a factor of approximately 10 to 12. Thus, substitution at the terminal phosphate of the ATP has a negative effect on the affinity of compounds for the active site of the catalytic subunit. Of great significance for inter- action with the enzyme are modifications by substitution of a,B- and B-y-bridge oxygen atoms of ATP on the methylene links (compounds VIII and XI). The great affinity of these analogs for the ATP-binding site is apparently explained by the different structure of their complexes with Mg2+t. In contrast to the above substances, inhibition of the catalytic subunit of protein kinase by compounds V, IX, XII, and XIX-XXII over time achieved practically complete inactivation of the enzyme (table 3). ATP, in concentrations of 0.2-1 mM, protected the enzyme from inactivation by analogs, indicating a modifying reaction at the active site. The protective 176 TABLE 2. Inhibitory Effect of ATP Analogs on the Activity of the Catalytic Subunit of cAMP-dependent Protein Kinase Compound | Ky (uM) pp(CH,)pA (VIII) 2.3 p(CH,)ppA (XI) 500 (BrCH,CH,) pA (1) 200 (BrCH,CH,)ppA (II) 70 (BrCH,CH,) pppA (111) 1000 (C1CH,)pA (IV) 100 (C1CH,)pppA (VI) 790 (C1CH,CH,0)pA (VII) 70 (BrCH, CONHCH,, CH,0) ppA (XVII) 530 (Mes) *pA (XVIII) 340 ATP Kg = 5 uM CHj *Mes = CHj c(0)o- CH3 effect of ATP is shown in figure 2 where compounds V and XX are taken as ex- amples. In studying the pH dependence of the rate of modification of catalytic subunits under the influence of compound V, a sigmoid curve was produced with an inflection in the area of pH 6.6-6.8 (figure 3). This suggests covalent blocking of the imidazole ring of the histidine group in the active site. Con- sidering that, upon binding of compound V with the enzyme, its alkylating ClCH, group should be located in the immediate vicinity of the functional group in- volved in the splitting of the y-P-O bond of ATP, we can assume that a histidine group may participate in the catalytic action of the enzyme. Similar results were obtained earlier in a study of the ATP-binding portion of cAMP-dependent protein kinase from pig brain (2,8). By analyzing the interaction of ATP analogs with adenylate cyclase and protein kinase we can draw certain conclusions about the similarities and dif- ferences in the structure of the active sites of these enzymes. The triphosphate 177 TABLE 3. ATP Analogs that Irreversibly Inhibit the Activity of the Catalytic Subunit of cAMP-dependent Protein Kinase _1 Concentration of Compound Kapp (min 7) Inhibitor (mM) (C1CH,)ppA Vn) 0.017 5.0 (p=F0,SC,H,0)pA (XX) 0.043 0.2 (p-F0,5C.H,CO)A (XXI) 0.019 0.5 (m-F0,SC H,CO)A (XXII) 0.034 0.5 (Mes) ppA (XIX) 0.046 5.0 p(CH)ppA (XII) 0.027 5.0 NHCOCH,, Br p(CH)pA (IX) 0.013 8.0 Nicos, pr portion of ATP is important in binding both with adenylate cyclase and with protein kinase. We find, however, that, for adenylate cyclase, substitution at the y-phosphorus atom has no significant effect on the affinity for the ac- tive site, while an unsubstituted y-phosphate must be present in the case of protein kinase in order to interact with the ATP-binding portion. Furthermore, the conformation of the triphosphate fragment of ATP in the active site of protein kinase is apparently extended, whereas the closeness of the o-phosphate hydroxyl and 3'-OH group of ATP to the active site of adenyl- ate cyclase presumes a different and more compact structure of this portion of ATP. Finally, the active site of adenylate cyclase in the area of binding of the y-phosphorus atom does not have the functional group necessary for manifes- tation of the activity of the enzyme, whereas in the active site of protein kinase in the region of binding of the y-phosphate there is a catalytic group-- apparently the imidazole ring of the histidine group. INTERACTION OF PHOSPHODIESTERASE WITH cAMP ANALOGS Extensive data have been accumulated about the effects of various compounds on cyclic and nucleotide phosphodiesterase (1,9-11). Nevertheless, the results of these studies are not sufficient to determine the structure of the active site of this enzyme. Therefore, we studied the interaction of partially purified phosphodiesterase from rat liver with cAMP analogs which we synthesized and 178 20 40 TIME (min) 40 80 TIME(min) FIGURE 2. Influence of (C1CHy)ppA (V) and p-FO02SCgH40)pA (XX) on the activity of a catalytic subunit and the protective effect of ATP. The ordinate shows residual activity of the catalytic subunit expressed in percent. A: Catalytic subunit incubated in 50 mM potassium phosphate buffer, pH 6.5, in the presence of 0.2 mM of compound XX and 2 mM MgClo (0-@), and in the presence of 0.2 mM of compound XX, 4 mM MgCly, and 2 mM ATP (0-0). B: Catalytic subunit incubated in 50 mM Tris-HCl buffer, pH 7.4, in the presence of 5 mM of compound V and 10 mM of MgCly (0-0), and in the presence of 5 mM compound V, 20 mM MgClo, and 1 mM ATP (0-0). which contained reactive or charged groups in various positions of the molecule (12-14). Some of these compounds were used by us earlier to study the cAMP- binding portion of protein kinase from pig brain (8,12,13). The structure of the cAMP analogs is presented in tables 4 and 5. In studying the inhibitory effect of cAMP analogs with phosphodiesterase, it was found that some of these analogs are reversible and competitive enzyme inhibitors with respect to cAMP (table 4). In order to evaluate the 'correct" stereospecificity of the binding of cAMP analogs in the active phosphodiesterase center, we also studied the substrate properties of some of these analogs with respect to the enzyme (tables 4 and 5). As can be seen from table 4, all 8-substituted cAMP derivatives with syn- conformation are rather strong phosphodiesterase inhibitors, but are practically incapable of hydrolysis by the enzyme. At the same time, 1-N-substituted cAMP analogs, which primarily have the anti-conformation, manifested a slightly weaker inhibitory effect, but served as phosphodiesterase substrates. These results indicate, on the one hand, the importance of the purine base for the hydrophobic interaction of the enzyme and, on the other hand, the preference of the anti-conformation of cAMP for "correct" seating in the active site. In protein kinase from pig brain, the cAMP-binding portion is apparently specific for the syn-conformation of the cyclic nucleotide (13). 179 (min -1) 20 - 10 + ¥ v v v v v 6.0 7.0 80 PH FIGURE 3. Variation in the rate of modification of catalytic subunits by (C1CH2)ppA (V) as a function of pH. The ordinate shows values of Kapp. The catalytic subunit was incubated in 50 mM potassium phosphate buffer in the presence of 5 mM of compound V and 10 mM of MgCljp. 2'-0-acyl derivatives of cAMP showed great affinity (greater than theo- phylline) for the active site of phosphodiesterase (table 4). It appears that, in the active site of phosphodiesterase, in contrast to the regulatory center of protein kinase (12), no functional group is responsible for the binding of the 2'-OH group of the ribose group. In the active site of phosphodiesterase, in contrast to the cAMP-binding portion of protein kinase (12,13), no cationic group participates in the binding of the phosphate hydroxyl of the cAMP, since the neutral bromoethyl (XXI) and charged carboxymethyl (XXII) esters of cAMP bind rather well with the enzyme (table 4). However, in this portion of the active site a catalytic group probably takes part in the splitting of the cyclophosphate cAMP ring, since compound XXII is not subject to hydrolysis by phosphodiesterase (table 4). This is related to steric screening of the 3'-P-0 bond by the carboxyl group included in compound XXII. Among the cAMP analogs studied, we found compounds which irreversibly in- hibit the activity of phosphodiesterase (table 5). Cyclic AMP, in a concentra- tion of 1 uM, protected the enzyme from inactivation by these analogs, indicating their interaction with the active site. In a number of inhibitors which had irreversible effects, two compounds (VII and XIV) had the same active group at various positions on the adenine base. We studied the pH dependence of the 180 TABLE 4. Inhibitory and Substrate Properties of cAMP Analogs with Respect to Phosphodiesterase Compound Ki (mM) | K,;/K theory a) 8-Br-cAMP (I) 0.049 0.22 19 8-NH(CH,) ,OH~-cAMP (11) 0.235 1.07 0 8-NHCH,, COOH~-cAMP (111) 0.550 2.5 - 8-NH(CH,) ,NH, cAMP (IV) 0.780 3.54 0 8-5 (CH,) ,NH, -cAMP (V) 0.075 0.34 0 8-NH (CH, ) ,NHCOCH ,—cAMP (VI) 1.000 4.54 0 8-N-OCH,-cAMP (XI) 0.450 2.04 98 1-N-0(CH,) ,NHCOCH ;-cAMP (XIII) 2.080 9.45 100 1-N-0(CH,) ,NHCOC H, SO, F-cAMP (XVI) 0.288 1.31 10 8-Br, 1-N-0(CH,) ,NHCOC H, 50,G- -cAMP (XVII) 0.075 0.34 0 2'-0-COCH, C1-cAMP (XVIII) 0.043 0.20 — 2'-0-COCH=CH,~cAMP (XIX) 0.180 0.82 —— N°—COC H, 2" o- -COC,H_- -cAMP (XX) 0.038 0.17 — cAMP- -OCH, CH, Br (XXI) 0.910 4.15 —— cAMP-OCH, COOH (XXII) 0.400 1.82 0 Theophylline 0.220 1.00 — effect of these inhibitors and discovered that the maximum inhibition in both cases occurs at pH 7.3-7.4 (figure 4), indicating modification of the same amino acid residue. Based on these results and data from the literature, we suggest one possible model for the active phosphodiesterase site (figure 5). Placement of cAMP in the agiiye site occurs due to interaction of the adenine ring and the exo-NH) group (Nb-acyl derivatives of cAMP are not hydrolyzed by phosphodiesterase) (9) with the aromatic amino acid group and proton donor group (X), respectively, located in the hydrophobic slot of the enzyme. The "correct" binding is 181 TABLE 5. Irreversible Phosphodiesterase Inhibitors From Liver om pH of app Hydrolysis/ Incubation Compound (min-1) cAMP (%) Medium 8-NH(CH,) ,NHCOCH,, C1-cAMP (Vii) 0.008 11 7.5 8-5 (CH,) ,NHCOCH, C1~CcAMP (VIII) 0.0065 -_ 7.5 8-NH(CH,) ,NHCOC H, SO, F~cAMP (IX) 0.010 — 6.8 8-5 (CH, ) ,NHCOC H, SO, F~cAMP (xX) 0.011 - 6.8 1-N-0(CH,) ,Br-cAMP (X11) 0.007 100 6.8 1-N-0(CH,) ,NHCOCH,C1-CcAMP (X1V) 0.0023 -_ 7.5 8-Br, 1-N-0(CH,) ,NHCOCH,, C1-cAMP (XV) 0.0093 0 7.5 *Analogs were tested at 0.1 mM concentration. } (min 1) 80 - 60 1 @ 40 - 20 1 69 71 73 75 77 pH FIGURE 4. pH dependence of inhibition of phosphodiesterase by 8-NH(CH2) pNHCOCH2C1-cAMP (VII) (0) and 1-N-0(CH9) gNHCOCHC1-cAMP (XIV) (®). The enzyme was incubated in the presence of 0.1 mM analog, 1 mM MgClp, and 0.75 mM mercaptoethanol in 50 mM potassium phosphate buffer (pH 4.8-7.6) or in 50 mM Tris-HCl buffer (pH 7.3-8.0). 182 FIGURE 5. Hypothetical diagram of the structure of the active site of phospho- diesterase. determined by the anti-conformation of the cyclic nucleotide. Finally, in the area of binding of the cyclophosphate ring there is a catalytic group (Z) and one more nucleophilic group (Y) interacting with the 5'-phosphoester oxygen atom (5'-deoxy-5'-methylene analogs of cAMP do not have affinity for phospho- diesterase) (11). CONCLUSIONS We studied the effect of various ATP analogs on solubilized adenylate cyclase from rabbit heart and the catalytic subunit of protein kinase from pigeon breast muscle, as well as the effect of cAMP analogs on phosphodiesterase from rat liver. Certain patterns were found in the binding of these analogs with these enzymes, and we have made some suggestions about the structure of the ATP- binding portion of adenylate cyclase and protein kinase, and the cAMP-binding portion of the active phosphodiesterase site. The characteristics of the active sites of adenylate cyclase, protein kinase, and phosphodiesterase which we have presented are far from complete. Further studies, including those based on specific, covalent blocking of certain portions of the active site of these enzymes, may significantly add to our under- standing of the functional topography of these active sites. 183 10. 11. 12. REFERENCES Vasiliev VYu, Gulyaev NN, Severin ES: [Cyclic adenosine monophosphate, its biological role and mechanism of action] (Rus). Zh Vses Khim O-va 20(3):306-322, 1975 Gulyaev NN, Tunitskaya VL, Baranova LA, Nesterova MV, Murtuzaev IM, Severin ES: [Investigation of the structure of the active site of the catalytic subunit of histone kinase] (Rus). Biokhimiia 41:1241-1249, 1976 Gulyaev NN, Sharkova EV, Dedyukina MM, Severin ES, Khomutov RM: [Selective inhibition of acetyl-CoA-synthetase by structural analogs of acetyladenylate] (Rus). Biokhimiia 36:1267-1273, 1971 Tunitskaya VL, Khropov YuV, Baranova LA, Mazurova LA, Gulyaev NN: Synthesis of analogs of adenosine cyclic 3',5'-phosphate and adenosine triphosphate containing reactive groups. In Abstracts, lst All-Union Conference of Nucleoside and Nucleotide Chemistry. Riga, 1978, pp 66-67 Hoppe J, Marutzky R, Freist W, Wagner KG: Mechanism of activation of the protein kinase I from rabbit skeletal muscle. Eur J Biochem 80:369-372, 1977 Bechel PJ, Beavo JA, Krebs EG: Purification and characterization of catalytic subunit of skeletal muscle adenosine 3':5'-monophosphate- dependent protein kinase. J Biol Chem 252:2691-2697, 1977 Bulargina TV, Grivennikov IA, Severin SE: Study of the reaction of adenosine triphosphate analogs with protein kinase from the breast muscle of the dove. In Abstracts, lst All-Union Conference on Nucleoside and Nucleotide Chemistry, Riga, 1978, pp 83-84 Severin ES, Nesterova MV, Gulyaev NN, Shlyapnikov SV: Brain histone kinase: the structure, the substrate specificity and the mechanism of action. Adv Enzyme Regul 14:407-444, 1976 Simon LN, Shuman DA, Robins RK: The chemistry and biological properties of nucleotides related to nucleoside 3',5'-cyclic phosphates. In Advances in Cyclic Nucleotide Research, vol 3, edited by P Greengard and GA Robison. New York, Raven Press, 1973, pp 225-353 Meyer RB Jr, Miller JP: Analogs of cyclic AMP and cyclic GMP: general methods of synthesis and the relationship of structure to enzymic activity. Life Sci 14:1019-1040, 1974 Amer MS, Kreighbaum WE: Cyclic nucleotide phosphodiesterases: properties, activators, inhibitors, structure-activity relationships, and possible role in drug development. J Pharm Sci 64:2-37, 1975 Severin ES, Nesterova MV, Sashchenko LP, Rasumova VV, Tunitskaya VL, Kochetkov SN, Gulyaev NN: Investigation of the adenosine 3',5'-cyclic phosphate binding site of pig brain histone kinase with the aid of some 184 13. 14. analogs of adenosine 3',5'-cyclic phosphate. Biochim Biophys Acta 384: 413-422, 1975 Gulyaev NN, Tunitskaya VL, Nesterova MV, Mazurova LA, Murtuzaev IM, Severin ES: [Interaction of 8-substituted derivatives and adenosine 3',5'-cyclic phosphate esters with pig brain protein kinase] (Rus). Biokhimiia 42:2071-2078, 1977 Gulyaev NN, Tunitskaya VL, Masurova LA, Severin ES: Synthesis, physico- chemical properties and kinetics of the rearrangement of 1-N-substituted derivatives of adenosine 3',5'-cyclophosphate. Bioorgan Khim 5, 1979, in press 185 ADP-RIBOSYLATION AND ACTIVATION OF ADENYLATE CYCLASE Martha Vaughan and Joel Moss SUMMARY Choleragen, the enterotoxin of Vibrio cholerae, activates adenylate cyclase from essentially all vertebrate tissues. Thus, knowledge of the mechanism of action of choleragen should aid in understanding the regulatory properties of the adenylate cyclases through which hormones, neurotransmitters, and pharma- cological agents influence the function of so many tissues including the myo- cardium. We have shown that choleragen possesses NAD glycohydrolase activity and catalyzes the NAD-dependent ADP-ribosylation of arginine, other guanidino compounds, and several purified proteins. Another structurally distinct bac- terial toxin, the heat-labile E. coli enterotoxin, exhibits the same enzymatic activities as choleragen, although it differs somewhat in kinetic properties. Other workers have shown that the E. coli enterotoxin activates adenylate cyclase in an NAD-dependent manner. Recently, we purified from turkey erythrocytes an ADP-ribosyltransferase that catalyzes the hydrolysis of NAD as well as the NAD- dependent ADP-ribosylation of arginine, related guanidino compounds, and purified proteins (histone, lysozyme, polyarginine). This enzyme differs in several char- acteristics from choleragen and the heat-labile E. coli enterotoxin but, like the two toxins, activates adenylate cyclase in the presence of NAD. Whether or not the ADP-ribosyltransferase plays a role in regulation of adenylate cyclase activity in the intact erythrocyte remains to be determined. Nevertheless, the evidence that these three proteins from different sources all catalyze the NAD- dependent activation of adenylate cyclase and the same model ADP-ribosyltrans- ferase reactions supports the view that adenylate cyclase activation can result from ADP-ribosylation of a protein component of the system. INTRODUCTION It is now generally believed that the effects of choleragen on vertebrate cells result from accumulation of intracellular cyclic AMP (adenosine 3',5'- monophosphate) secondary to activation of adenylate cyclase by the toxin. Choleragen is a complex molecule (table 1) consisting of five B subunits (11,600 daltons) in noncovalent association with one A subunit. The A subunit contains two peptides, Aj and Ap, of 23,500 and 5,500 daltons, respectively, From the Laboratory of Cellular Metabolism, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland. 187 TABLE 1. Choleragen: Structure and Function* A -s=s-A, (B) 1 Peptide MH Function Ay 23,500 activation of adenylate cyclase A, 5,500 unknown B 11,600 binding to ganglioside C1 on cell surface *Based on sources cited in references 1 to 6. **Molecular weight. joined by a single disulfide bond. Work in several laboratories (reviewed in references 1 to 6) has established that the B subunits are responsible for choleragen binding to the cell surface through interaction with the oligosac- charide moiety of ganglioside Gy; in the plasma membrane. In broken cells, however, activation of adenylate cyclase by choleragen requires neither Gym nor the B subunits. It does require NAD (nicotinamide adenine dinucleotide) as first shown by Gill (7) and is brought about by the Aj; peptide (8) which is effective only after liberation from the A subunit by reduction. In studying the mechanism through which choleragen interacts with NAD to activate adenylate cyclase, we found in 1976 that the toxin exhibits NAD glyco- hydrolase and ADP-ribosyltransferase activities (table 2). We suggested that choleragen activates adenylate cyclase by catalyzing ADP-ribosylation of an arginine or related amino acid residue in a protein which is the cyclase itself or a regulatory component of the cyclase system (14-16). Further support for the view that ADP-ribosylation can increase adenylate cyclase activity was pro- vided by subsequent demonstration that the heat-labile E. coli enterotoxin, which activates adenylate cyclase in an NAD-dependent fashion (17), also possesses ADP-ribosyltransferase activity (18) and by the isolation of an ADP-ribosyl- transferase from turkey erythrocytes that activates adenylate cyclase (9). Some characteristics of this avian enzyme and two in bacterial toxins that catalyze NAD-dependent activation of adenylate cyclase are discussed in this paper. EXPERIMENTAL PROCEDURE Materials L-Arginine methyl ester di-HCl, lysozyme (18,000 U/mg), polyarginine, histone (Type IIA, calf thymus), snake venom phosphodiesterase, and NAD were purchased from Sigma; [carbonyl-14C]NAD (50 mCi/mmol) and [adenine-U-14C]NAD (280 mCi/mmol) from Amersham/Searle. AG 1-X2 (Bio-Rad), 100-200 mesh (in the 188 TABLE 2. Reactions Catalyzed by Choleragen, E. coli Enterotoxin (Heat-Labile), and the ADP-Ribosyltransferase From Turkey Erythrocytes I. NAD glycohydrolase NAD + HOH —— ADP-ribose + nicotinamide + Nd II. ADP-ribosyltransferase + NAD + arginine* ADP-ribosylarginine + nicotinamide + H Eu NAD + protein —— ADP-ribosylprotein + nicotinamide + H *Several other compounds containing a guanidino moiety can also serve as ADP-ribose acceptors. ADP-ribosylation of purified proteins, e.g., lysozyme, polyarginine, histone, by all three enzymes (9,10) and ADP-ribosylation of cellular proteins by choleragen (11-13) have been demonstrated. chloride form), was washed with 0.5 M NaOH, water until neutral, 0.5 M HCl, and water until neutral. Columns (0.5 x 4 cm) were prepared and washed with 2 ml of 20 mM Tris-HCl, pH 7.5, before they were used to separate [carbonyl-l4C]- nicotinamide from unhydrolyzed [carbonyl-l4C]NAD. Choleragen was purchased from Schwarz/Mann. The polymyxin-released heat- labile E. coli enterotoxin is described in references 10 and 18. The ADP- ribosyltransferase was purified from the supernatant fraction (100,000 x g for 30 minutes) of washed turkey erythrocytes suspended in 50 mM potassium phosphate buffer, pH 7.0. Following acid precipitation, ethanol extraction, and chroma- tography successively on cellulose phosphate and phenyl-Sepharose, the enzyme exhibited one major band on polyacrylamide gel electrophoresis (9). NAD Glycohydrolase and ADP-Ribosyltransferase Assays Details of these procedures have been published (14,15). For both assays [carbonyl-14C]NAD was used as substrate; for assay of ADP-ribosyltransferase activity an ADP-ribose acceptor, usually arginine or arginine methyl ester, was also present. Buffers and concentrations of substrates were varied depending on the enzyme source (i.e., choleragen, E. coli toxin, or the erythrocyte ADP- ribosyltransferase) and the nature of the experiment. Standard assays with choleragen contained 400 mM potassium phosphate buffer (pH 7.0), 20 mM dithiothreitol, and 2 mM [carbonyl-l4CINAD (~ 40,000 cpm) with or without 75 mM arginine methyl ester for ADP-ribosyltransferase and NAD glyco- hydrolase activities, respectively. Assays were initiated by addition of choleragen in 50 mM Tris-HCl (pH 7.5), 200 mM NaCl, 1 mM EDTA, and 3 mM NaNj. After incubation at 30° C for the specified time, duplicate samples from each 189 tube were transferred to columns (0.5 x 4 cm) of Dowex 1 (prepared as described under Materials"). The [carbonyl-l4C]nicotinamide was eluted with five 1-ml portions of 20 mM Tris-HCl (pH 7.5) and collected for radioassay in a liquid scintillation spectrometer. The effectiveness of this procedure for quantita- tive isolation of [carbonyl-14C]nicotinamide and its separation from [carbonyl- 14CINAD was confirmed by thin-layer chromatography (14). Standard assays with the E. coli heat-labile enterotoxin were like those with choleragen except that the concentration of potassium phosphate buffer (pH 7.0) was 50 mM and incubation was at 30° C (18). Standard assays for the ADP-ribosyltransferase purified from turkey erythrocytes contained 50 mM potassium phosphate buffer (pH 7.0), 30 uM [carbonyl-14C]NAD (~ 40,000 cpm), 75 mM arginine methyl ester, and 0.3 mg of ovalbumin in a total volume of 0.3 ml (9). Arginine methyl ester was omitted for assay of NAD glycohydrolase activity. Assays were initiated with the addition of the enzyme and incubated at 30° C for 5 hours before isolation of [carbonyl-14C]nicotinamide as described above. ADP-Ribosylation of Proteins Incubations with choleragen, E. coli enterotoxin, or the turkey ADP- ribosyltransferase were carried out as described for assay of NAD glycohydrolase activity with each of these enzymes except that [adenine-l4C]NAD replaced [carbonyl-14C]NAD and the potential ADP-ribosyl acceptor protein was present at a concentration of 5 mg/ml. At the end of the incubation, one-third volume of cold 20 percent trichloroacetic acid was added and, after 30 minutes at 0° C, the samples were transferred to 0.45 pM Millipore filters. The assay tubes and filters were washed three times with 2 ml of cold 5 percent trichloroacetic acid. Dried filters were placed in vials and dissolved in 2 ml of Piersolve (Pierce). Radioactivity was determined in a liquid scintillation spectrometer after addition of counting solution (9). RESULTS AND DISCUSSION Choleragen Our initial attempts to demonstrate that choleragen possessed enzymatic activity were based on the knowledge that diphtheria toxin, which causes NAD- dependent inhibition of protein synthesis, does so by catalyzing ADP- ribosylation of elongation factor II (19). No analogous protein acceptor was known for the postulated reaction catalyzed by choleragen. It had been shown, however, that in the absence of elongation factor II diphtheria toxin displays NAD glycohydrolase activity (20), and we found that choleragen simi- larly catalyzes the hydrolysis of NAD to yield ADP-ribose and nicotinamide (reaction I, table 2). A number of amino acids commonly present in proteins were then tested for their ability to serve as ADP-ribose acceptors when incubated with NAD and choleragen. Only arginine was effective; nicotinamide and a compound with the properties of ADP-ribosylarginine were produced in equimolar amounts (15). 190 The stereospecificity of the ADP-ribosyltransferase reaction was established by Oppenheimer (21). The product of the choleragen-catalyzed reaction of B-NAD with L-arginine is a-ADP-ribosyl-L-arginine, shown in figure 1, which anomerizes relatively rapidly to a mixture of a- and B-forms (21). In addition to arginine, several other guanidino compounds, including guanidine itself, can serve as ADP- ribose acceptors (10). Arginine methyl ester and agmatine are, in fact, more effective than arginine (15,22). With the carboxyl moiety blocked or absent, respectively, these compounds may resemble more closely than does free arginine the natural acceptor, which we presume to be an arginine residue in a protein, and thus probably has both carboxyl and amino groups involved in peptide linkages. As shown in table 3, demonstration of the enzymatic activities of choleragen requires the presence of relatively high concentrations of thiol, probably to reduce the disulfide bond in the A subunit and liberate the active A] peptide (14). In other experiments (J Moss, SJ Stanley, and MC Lin, unpublished), it has been shown that the NAD glycohydrolase and ADP-ribosyltransferase activities of choleragen are intrinsic to the A] peptide, which is also responsible for activation of adenylate cyclase (8). Dithiothreitol is without effect on the enzymatic activities of the reduced and alkylated Aj peptide (J Moss, SJ Stanley, and MC Lin, unpublished). Several purified proteins can serve as ADP-ribose acceptors in the choleragen- catalyzed transferase reaction (9). As shown in table 4, the addition of 75 mM arginine methyl ester decreased protein ADP-ribosylation by about 50 percent in this experiment. (The K, for arginine methyl ester as an ADP-ribose acceptor 0 XX 0 H, 7 IN N H, NZ ADP- ADP-0 0 0 i . FIGURE 1. Stereospecificity of the guanidine-dependent ADP-ribosyltransferase reaction (see reaction II, table 2). 191 TABLE 3. Effect of Thiol-Reducing Agents on NAD Glycohydrolase and ADP-Ribosyltransferase Activities of Choleragen Thiol Added* Concentration Enzyme Activity Hydrolase Transferase nmol/min/mg Choleragen Dithiothreitol, 0.3 3 30 Glutathione, 0.3 mM 3 mM 30 mM mM mM mM 0.67 0.44 2.9 8.2 17 52 1.1 0.67 0.67 3.6 7.6 24 *Essentially no activity was detectable in the absence of added thiol. TABLE 4. ADP-Ribosylation of Proteins Catalyzed by Choleragen Inhibited by Arginine Methyl Ester (ArgOMe) ADP-Ribose Incorporated Protein No ArgOMe 75 mM ArgOMe 5 mg/ml nmol/min/mg Histone 2.7 1.4 Lysozyme 1.3 0.6 Polyarginine 6.5 3.2 192 with choleragen is ~ 50 mM). When lysozyme that had been ADP-ribosylated in the presence of choleragen and [adenine- 4CINAD was incubated with snake venom phos- phodiesterase, essentially all of the radioactivity was released and migrated on thin-layer chromatography with a mobility indistinguishable from that of 5'"-AMP. This is the result expected with an ADP-ribosylated, but not with a poly (ADP-ribosylated), protein. The amino acids that are ADP-ribosylated in lysozyme and histone have not been identified. It is notable, however, that arginine methyl ester (table 4) or arginine can inhibit this process, whereas other amino acids which are not ADP-ribose acceptors (15,22), and presumably do not interact with the catalytic site of choleragen, do not. All of these observations are consistent with the view that the acceptor amino acid in pro- teins ADP-ribosylated by choleragen is arginine. Despite considerable evidence for the hypothesis that the activation of adenylate cyclase by choleragen results from ADP-ribosylation of a component of the cyclase system, it has been difficult to identify this specific protein. To do so requires demonstration of altered function of the ADP-ribosylated protein, because choleragen catalyzes the ADP-ribosylation of so many proteins, at least in the model system thus far studied. The work of Cassel and Pfeuffer (11) provides strong support for the view that it is the GIP-binding protein associated with the adenylate cyclase that is ADP-ribosylated. Gill and Meren (12) have demonstrated ADP-ribosylation of what is probably the GTP-binding pro- tein. Cassel and Selinger (23) earlier proposed that choleragen acts by in- hibiting a specific GTPase which hydrolyzes bound GTP and thereby limits adenylate cyclase activity. The findings of Johnson et al. (13) are also con- sistent with a role for an ADP-ribosylated protein in the activation of adenylate cyclase by choleragen. Heat-Labile Enterotoxin From E. coli The heat-labile E. coli enterotoxin, like choleragen, causes NAD-dependent activation of adenylate cyclase (17). It also possesses NAD glycohydrolase activity and catalyzes the ADP-ribosylation of arginine, a number of other guanidino compounds, and several proteins (10,18,24). These reactions are shown in table 2. The stereospecificity of the reaction with arginine as acceptor and the nature of the ADP-ribosylated product of the reaction with lysozyme are like those of the analogous choleragen-catalyzed reactions (table 5). Although both toxins require dithiothreitol (or other thiol) for maximal activity and have similar affinities for NAD, there are significant differences in their enzymatic properties (table 6). The E. coli toxin has a much lower affinity for arginine methyl ester than does choleragen, and the relative effectiveness of certain guanidino compounds or proteins as ADP-ribose acceptors is somewhat different for the two toxins (table 6). The fact that two structurally distinct proteins [which do exhibit some similarities in structure (25) and immunological reac- tivity (26-28)] both catalyze the NAD-dependent activation of adenylate cyclase and possess NAD glycohydrolase and ADP-ribosyltransferase activity provides strong support for the view that activation results from the ADP-ribosylation of a protein component of the adenylate cyclase system. 193 TABLE 5. Characteristics of Reactions Catalyzed by Choleragen, Heat-Labile E. coli Enterotoxin, and the Turkey Erythrocyte ADP-Ribosyltransferase 1. PB-NAD is the substrate. 2. o-ADP-ribosylarginine is the product of reaction with arginine. 3. When lysozyme is acceptor, the product is ADP- ribosyl-lysozyme with no evidence of formation of poly (ADP-ribosyl)-protein. TABLE 6. Comparison of Physical and Enzymatic Properties of Choleragen, Heat-Labile E. coli Enterotoxin, and the ADP-Ribosyltransferase From Turkey Erythrocytes Heat-Labile ADP-Ribosyl- Property Choleragen E. coli Toxin transferase Molecular weight 87,000 = 20,000 ~ 25,000 Subunits A1,A9,B ? ? DTT requirement yes yes no Kp for NAD 4 mM 8 mM 30 uM K, for arginine methyl ester 50 mM 240 mM 2 and 50 mM Effectiveness of arginine arginine arginine guanidine com- pounds (75 mM) as substrate Effectiveness of proteins (5 mg/ml) as ADP- ribose acceptor methyl ester > L-arginine ~ D-arginine > guanidine polyarginine > histone > lysozyme methyl ester > L-arginine ~ D-arginine ~ guanidine polyarginine >> histone ~ lysozyme methyl ester > L-arginine ~ D-arginine > guanidine histone > lysozyme ~ poly- arginine 194 ADP-Ribosyltransferase From Turkey Erythrocytes To investigate whether animal cells might employ an ADP-ribosylation mecha- nism to activate adenylate cyclase, we looked for an enzyme that, like choleragen and the heat-labile E. coli enterotoxin, could catalyze the synthesis of ADP- ribosylarginine. Such an ADP-ribosyltransferase was purified from the super- natant fraction of turkey erythrocytes as described under "Experimental Procedure" (9). The stereospecificity of the reaction with arginine is identical to that of the two toxin-catalyzed reactions shown in figure 1 (J Moss, SJ Stanley, and NJ Oppenheimer, unpublished). The enzyme also exhibits NAD glycohydrolase activity and catalyzes the ADP-ribosylation of polyarginine, lysozyme, and histone (table 2). The product of the reaction with lysozyme is an ADP-ribosyl-, not a poly (ADP-ribosyl)-protein (9). As shown in table 7, histone was a much better ADP-ribose acceptor than was lysozyme or polyarginine (9). This is in contrast to the situation with the two bacterial toxins where polyarginine was clearly more effective than histone (table 6). The ADP-ribosyltransferase also differs in other ways from choleragen and the heat-labile E. coli enterotoxin (table 7). The turkey enzyme has a molecular weight of about 25,000, which is somewhat greater than that of the catalytically active units of the toxin, and its activity is not increased by dithiothreitol. Its affinity for NAD is much greater than that of the toxins, and it exhibits anomalous kinetics with arginine methyl ester as substrate (table 7). Despite these differences in properties, the ADP-ribosyltransferase from turkey erythro- cytes appears to be functionally equivalent to choleragen and the heat-labile E. coli enterotoxin. It catalyzes the same model reactions and activates adeny- late cyclase in an NAD-dependent manner (9). Whether it plays a role in the activation of adenylate cyclase in the intact erythrocyte remains to be deter- mined. TABLE 7. ADP-Ribosylation of Proteins Catalyzed by Purified Transferase From Turkey Erythrocytes Inhibited by Arginine Methyl Ester (ArgOMe) ADP-Ribose Incorporated Protein No ArgOMe 75 mM ArgOMe 5 mg/ml nmol/min/mg Histone 130 2.6 Lysozyme 55 0.85 Polyarginine 55 5.5 195 CONCLUSIONS Three distinct proteins from different sources, choleragen, the heat-labile E. coli enterotoxin, and an ADP-ribosyltransferase purified from turkey erythro- cytes, can bring about NAD-dependent activation of adenylate cyclase. These three proteins differ in structure and enzymatic characteristics but each cata- lyzes the hydrolysis of NAD and the NAD-dependent ADP-ribosylation of arginine, several other guanidino compounds, and proteins. All of our observations are consistent with the hypothesis that the two bacterial toxins and the avian enzyme activate adenylate cyclase by catalyzing the ADP-ribosylation of an arginine, or similar amino acid residue, in a protein component of the cyclase system. 196 10. 11. 12. 13. 14. 15. REFERENCES Finkelstein RA: Cholera. CRC Crit Rev Microbiol 2:553-623, 1973 Van Heyningen WE: Gangliosides as membrane receptors for tetanus, toxin, cholera toxin and serotonin. Nature 249:415-417, 1974 Bennett V, Cuatrecasas P: Mechanism of action of Vibrio cholerae enterotoxin. J Membr Biol 22:1-28, 1975 Gill DM: Mechanism of action of cholera toxin, Adv Cyclic Nucleotide Res 8:85-118, 1977 Van Heyningen S: Cholera toxin. Biol Rev 52:509-549, 1977 Moss J, Vaughan M: Activation of adenylate cyclase by choleragen. Annu Rev Biochem, in press Gill DM: Involvement of nicotinamide adenine dinucleotide in the action of cholera toxin in vitro. Proc Natl Acad Sci USA 72:2064-2068, 1975 Gill DM, King CA: The mechanism of action of cholera toxin in pigeon erythrocyte lysates. J Biol Chem 250:6424-6432, 1975 Moss J, Vaughan M: Isolation of an avian erythrocyte protein possessing ADP-ribosyltransferase activity and capable of activating adenylate cyclase. Proc Natl Acad Sci USA 75:3621-3624, 1978 Moss J, Garrison S, Oppenheimer NJ, Richardson SH: NAD-dependent ADP- ribosylation of arginine and proteins by E. coli heat-labile enterotoxin. J Biol Chem, in press Cassel D, Pfeuffer T: Mechanism of cholera toxin action: Covalent modifi- cation of the guanyl nucleotide-binding protein of the adenylate cyclase system. Proc Natl Acad Sci USA 75:2669-2673, 1978 Gill DM, Meren R: ADP-ribosylation of membrane proteins catalysed by cholera toxin: Basis of the activation of adenylate cyclase. Proc Natl Acad Sci USA 75:3050-3054, 1978 Johnson GL, Kaslow HR, Bourne HR: Genetic evidence that cholera toxin substrates are regulatory components of adenylate cyclase. J Biol Chem 253:7120-7123, 1978 Moss J, Manganiello VC, Vaughan M: Hydrolysis of nicotinamide adenine dinucleotide by choleragen and its A protomer: Possible role in the activation of adenylate cyclase. Proc Natl Acad Sci USA 73:4424-4427, 1976 Moss J, Vaughan M: Mechanism of action of choleragen. Evidence for ADP- ribosyltransferase activity with arginine as an acceptor. J Biol Chem 252:2455-2457, 1977 197 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. 26. Moss J, Osborne JC Jr, Fishman PH, Brewer HB Jr, Vaughan M, Brady RO: Effect of gangliosides and substrate analogues on the hydrolysis of nico- tinamide adenine dinucleotide by choleragen. Proc Natl Acad Sci USA 74: 74-78, 1977 Gill DM, Evans DJ Jr, Evans DG: Mechanism of activation of adenylate cyclase in vitro by polymyxin-released, heat-labile enterotoxin of Escherichia coli. J Infect Dis 133 (suppl):S103-S107, 1976 Moss J, Richardson SH: Activation of adenylate cyclase by heat-labile Escherichia coli enterotoxin. Evidence for ADP-ribosyltransferase activity similar to that of choleragen. J Clin Invest 62:281-285, 1978 Honjo T, Nishizuka Y, Kato I, Hayaishi O: Adenosine diphosphate ribosyla- tion of aminoacyl transferase II and inhibition of protein synthesis by diphtheria toxin. J Biol Chem 246:4251-4260, 1971 Kandel J, Collier RJ, Chung DW: Interaction of fragment A from diphtheria toxin with nicotinamide adenine dinucleotide. J Biol Chem 249:2088-2097, 1974 Oppenheimer NJ: Structural determination and stereospecificity of the choleragen-catalyzed reaction of NADT with guanidines. J Biol Chem 253: 4907-4910, 1978 Moss J, Vaughan M: Role for ADP-ribosylation in the activation of adenylate cyclase by bacterial toxins and avian enzymes. In Proceedings of the Josiah Macy, Jr., Foundation Conference on Receptors and Human Diseases, New Orleans, Louisiana, December 4-6, 1978. Edited by JZ Bowers and EF Purcell. In press Cassel D, Selinger Z: Mechanism of adenylate cyclase activation by cholera toxin: Inhibition of GTP hydrolysis at the regulatory site. Proc Natl Acad Sci USA 74:3307-3311, 1977 Moss J, Garrison S, Oppenheimer NJ, Richardson SH: ADP-ribosylation catalyzed by E. coli heat-labile enterotoxin: Possible role in the activa- tion of adenylate cyclase. In Proceedings of the US-Japan Joint Cholera Conference, Karatsu, Japan, September 27-29, 1978. Edited by K Takeya and Y Zinnaka. Tokyo, Fuji Printing Co., Ltd., 1979, pp 274-281 Robertson DC, Kunkel SL, Gilligan PH: Purification and characterization of heat-labile enterotoxin produced by ENTY E. coli: Application of hydro- phobic chromatography and use of defined media (abstr). The 14th Joint Conference US-Japan Cooperative Medical Science Program Cholera Panel, Karatsu, September 27-29, 1978, 75(abs) Gyles CL, Barnum DA: A heat-labile enterotoxin from strains of Escherichia coli enteropathogenic for pigs. J Infect Dis 120:419-426, 1969 198 27. 28. Smith NW, Sack RB: Immunologic cross-reactions of enterotoxins from Escherichia coli and Vibrio cholerae. J Infect Dis 127:164-170, 1973 Holmgren J, S&derlind O, Wadstrom T: Cross-reactivity between heat labile anterotoxins of Vibrio cholerae and Escherichia coli in neutralization tests in rabbit ileum and skin. Acta Pathol Microbiol Scand [B] 81:757-782, 1973 199 INTERACTION BETWEEN B-ADRENERGIC RECEPTORS AND ADENYLATE CYCLASE IN THE HEART V. A. Tkachuk and S. E. Severin The effects of catecholamines on the myocardium are manifested through the a-adrenergic and B-adrenergic receptors. The mode of action and physiological effects of these two types of receptors are significantly different, and there is a strict selectivity of a- and B-receptors with respect to antagonists. Phentolamine and phenoxybenzamine block the action of a-adrenergic receptors, whereas dichloroisoproterenol, propranolol, alprenolol, practolol, and butox- amine block the action of B-adrenergic receptors. After careful study of the effects of catecholamines and their antagonists on adenylate cyclase in the myocardium, Sutherland and his colleagues (1) con- cluded in 1962 that synthesis of cAMP in the heart is controlled by the B- adrenergic receptors. To explain the interaction of B-receptors with adenylate cyclase, a model was suggested (1) in which the adenylate cyclase system consists of regulatory and catalytic elements that are closely interrelated and surrounded by a lipid matrix. The regulatory element is the hormone receptor, located on the outside of the membrane, and the catalytic element--adenylate cyclase itself-- is located on the inside of the membrane. Until the classic work of Sutherland, hormone receptors were thought to be some component of the cell which recognized and bound the hormone, and then participated in manifesting its effect on the cell. At the present time, the function of the receptor is limited only to recognition and binding of the hor- mone molecules. Thus, the only direct method of studying the properties of this protein is to measure the binding of the hormone with the receptor. For peptide hormones and cholinergic agents, the method for measuring binding with the re- ceptor was discovered over 10 years ago. Correct measurement of binding of catecholamines with their receptors became possible only after 1974, when labeled catecholamine antagonists began to be used (2-4). Classic pharmacological theory holds that the agonists and antagonists compete for the same binding sites; therefore, the displacement of a labeled antagonist can be used to indicate the binding of a hormone with a receptor. As can be seen from figure 1, binding of 3H-alprenolol, an antagonist of B- adrenergic receptors, to the membrane of a rabbit heart preparation occurs rapidly and reversibly. After addition of isoproterenol, a B-adrenergic From the Department of Biochemistry, Moscow State University, Moscow, USSR. 201 10M isoproterenol V | Oo 1500 |- ' \ = \ £ \ Q \ 2 \ © \ c S 1000 e Q \ Qo \ S \ o \ 5 \ £ } ° % T 500 Oo \ ® ™ . ol Seas ® ® ——— 0 1 1 1 1 J 0 2 4 6 8 10 min FIGURE 1. Specific binding of 3H-alprenolol with rabbit heart plasma membranes. Five minutes after beginning incubation of membranes with 3H-alprenolol, the medium was divided into two portions, and isoproterenol was added to one of them (the addition of isoproterenol solutions changed the volume of the medium by 7 percent). receptor agonist, to the incubation medium, rapid release of 3H-alprenolol from the membrane occurs. Nonradioactive antagonists also release 3H-alprenolol from binding with the receptor; this process depends on the steric configuration of the antagonist molecule. Figure 2 shows that the (+)-alprenolol reduces the binding of 2H-alprenolol at concentrations 100 times greater than the (-)- stereoisomer of alprenolol. The a-adrenergic blocker phentolamine has an effect only at concentrations several orders of magnitude greater than those necessary for manifestation of its physiological action. The specific binding of 3H-alprenolol with heart membranes shows saturation kinetics, and expression of the kinetics of binding in double-reciprocal or Scatchard plots yields a single straight line. The dissociation constant for the 3H-alprenolol-receptor complex, obtained in experiments on rabbit heart membranes, was 5.7 nM. This value is similar to the dissociation constants ob- tained for other membrane preparations (4). The kinetics of binding, its re- versibility and stereospecificity, led us to conclude that the binding that was measured reflected the interaction of the ligand with B-adrenergic receptors of the heart. 202 ~ ol T N a T % of 3H-alprenolol bound 8 T 0 yy) 1 1 1 1 | o 10° 10% 107 10% 105 10% 103m catecholamines or their antagonists FIGURE 2. Effect of nonradioactive compounds on the specific binding of 35- alprenolol with rabbit heart plasma membranes. Y-V = D,L-isoproterenol. 0-80 = L(-)-alprenolol. 0-0 = D(+)-alprenolol. @8 = Phentolamine. In the literature on pharmacology and endocrinology, some investigators believe that the binding of a hormone with its receptor leads to a proportional increase in the biological response, and that this response becomes maximal if all of the molecules of the receptor are occupied by hormone molecules. With direct methods of testing hormonal receptors it is now possible to check this hypothesis. Figure 3 presents the results of measuring the dependence of adenylate cyclase activity in the heart on isoproterenol concentration with and without guanylyl imidodiphosphate [Gpp(NH)p], a nonmetabolized GTP analog, as well as the dependence of binding of the hormone on its concentration, also measured with and without Gpp (NH)p. Obviously, this nucleotide increases the affinity for isoproterenol, based on the activation of adenylate cyclase, and decreases the affinity for isopro- terenol, as measured by the release of 3H-alprenolol. If these results are expressed as the dependence of adenylate cyclase activity on the degree of occu- pation of the receptors, we find that, when the nucleotide is absent, there is a proportionality between the binding of the hormone and activation of adenylate cyclase, and when Gpp(NH)p is present, occupation of only 5 to 10 percent of all the receptors of the membrane leads to an increase in adenylate cyclase activity by 50 percent (figure 4). In the case of the peptide hormones, it has been observed that maximum activation of adenylate cyclase can occur upon occupation of only 1 to 5 percent of the hormonal receptors (5). Binding with the receptors is a necessary, but not sufficient, condition for a physiological effect. This is well illustrated by the antagonists, whose binding with the receptor causes no other effects except for competition with the hormone molecules. There are substances (the antibiotic philipin and SH- reagents) which do not affect the binding of catecholamines with receptors but which do prevent their effect on adenylate cyclase (5). Consequently, the bio- logical effect is not necessarily directly dependent on the quantity of hormone- receptor complex or on the rate of its formation. Apparently, there is an 203 100 % of AC activation 3 % of 3H-alprenolol bound 0 0 0 107 10% 105% 10M 0 107 10% 10° 10*Mm D,L-isoproterenol D,L-isoproterenol FIGURE 3. Variation of degree of activation of adenylate cyclase (AC) (left side) and binding of 3H-alprenolol (right side) with concentration of isopro- terenol. One hundred percent activation of adenylate cyclase was taken as the effect of 10~%4 M isoproterenol. One hundred percent binding of 3H-alprenolol was taken as the quantity of antagonist which specifically binds without isopro- terenol. 0-0 = Without guanylyl imidodiphosphate [Gpp(NH)p]. @-0 = In the presence of 10% M Gpp (NH)p. 100 ® © o ° 0 +Gpp(NH) 2 pp Pp 2 -Gpp(NH)p § olf" gS = | Q | © | Q | 4 | p ° | o | | 0 1 J 0 50 100% R-occupation by iso FIGURE 4. Variation between specific binding of isoproterenol (ISO) (calcu- lated by extraction of 3H-alprenolol) and activation of adenylate cyclase (AC) of rabbit heart. Gpp(NH)p = Guanylyl imidodiphosphate. 204 excess of receptors or "reserve" receptors in the membrane, and the variation of the final effect as a function of the occupation of receptors is subject to a number of influences. As we have already noted (in figures 3 and 4), the guanyl nucleotides in- fluence the binding constant of antagonists with receptors. The binding of the antagonist (alprenolol) with B-receptors of the rabbit heart was insensitive to GTP and Gpp(NH)p. Similar results were obtained in a study of B-adrenergic receptors in erythrocytes (6). This observation forces us to use yet another postulate of pharmacology and endocrinology, namely that the agonist and antag- onist are bound more cautiously at the same site on the molecule of a receptor. It would be more correct to state that the agonist and antagonist compete for a binding site on the receptor. Actually, there is no basis at present for ex- cluding the hypothesis that the agonist and antagonist bind with different sites between which there is negative interaction. Apparently, Gpp(NH)p does not interact directly with B-adrenergic receptors. This is indicated by the following facts. Specific binding of 3H-alprenolol with heart membrane is suppressed by trypsin. Interestingly, the sensitivity of B-receptors, located on the outer surface of the membrane vesicles, to trypsin is significantly less than is the case with adenylate cyclase which is located on the inside of the membrane (figure 5). This property of the proteins can be used to uncouple the B-receptors and adenylate cyclase, i.e., to produce mem- branes free of enzyme activity but retaining their capacity to bind the hormone. p-adrenergic receptors 150] 100 adenylate cyclase activity pmol cyclic AMP/mg protein/min (0—0) (e—e) © w (62) uigjoid Bw /punoq lojouaidie-H jowd 0 1:500 1:250 150 trypsin/membrane protein FIGURE 5. Variation of activity of adenylate cyclase and specific binding of 3H-alprenolol with rabbit heart membrane as a function of trypsin concentration. The membranes were incubated with trypsin for 10 minutes at room temperature. Then the trypsin protein inhibitor was added, and the activity of adenylate cyclase or binding of 3H-alprenolol was measured. O-0 = Basal activity. 0-0 = Activity in the presence of 10-4 M isoproterenol. 205 After preincubation of heart membrane with trypsin, when full inactivation of adenylate cyclase has occurred, and the quantity of B-receptors has decreased by 26 percent, we found no influence of Gpp(NH)p, either on the binding of alprenolol or on the binding of isoproterenol. Thus, the effect of guanyl nucleotides on the binding of a hormone with its receptor involves either adenylate cyclase or other proteins more sensitive to hydrolysis by trypsin than the receptor protein. Interest in studying the effect of guanyl nucleotides on adenylate cyclase arose after a series of remarkable experiments in Rodbell's laboratory (7-8), conducted on membrane preparations of liver. These studies established that purine nucleotides (GTP most effectively) increased the activation of adenylate cyclase by the hormone. Rodbell's data were confirmed by other workers, par- ticularly on membrane preparations of heart adenylate cyclase (9). Presently, the overwhelming majority of investigators consider that guanyl nucleotides participate in transmission of the signal from the receptor to the adenylate cyclase. In the opinion of Rodbell (10), the activation of adenylate cyclase involves hydrolysis of GTP. In order to transform adenylate cyclase to a state with high catalytic activity, it is necessary that the medium simultaneously contains the hormone and guanyl nucleotides. To confirm this theory, a GTP- binding protein has been separated from erythrocyte membranes, and it has been shown that this protein is hormone-activated GTPase (11-12). Without this protein, adenylate cyclase loses its sensitivity both to hormones and guanyl nucleotides. We succeeded in extracting this protein from rabbit heart membranes. Removal of the protein from the membranes decreases the activation of adenylate cyclase by isoproterenol and guanyl nucleotides (table 1). TABLE 1. Effect of Factors Extracted From Rabbit Heart Membrane on Adenylate Cyclase Activity in These Membranes Without Factor After Addition of Factor Additives Specific Activity* % Specific Activity* % Basal activity 57.2 + 7.0 100 67.0 + 4.8 100 10-4 M isoproterenol 65.0 + 5.0 114 96.0 + 4.2 143 10~% M Gpp (NH)p 66.0 + 9.5 115 104 + 3 155 107% M isoproterenol + 10-4 M Gpp(NH)p 115 + 8 201 170 + 5 254 *pmol cAMP/mg proteins min Gpp(NH)p = Guanylyl imidodiphosphate 206 The data presented above were obtained on membrane preparations of hearts from random-bred, gray rabbits. Activation of adenylate cyclase by catechol- amines in these membranes occurred without guanyl nucleotides, and reached 200 to 300 percent. In studies of the membrane preparation from New Zealand white rabbit hearts, we found that adenylate cyclase is practically insensitive to the effects of catecholamines. Addition of GTP to the incubation medium led to activation of adenylate cyclase; however, in the presence of GTP, isoprotere- nol activated adenylate cyclase only by 20 to 30 percent (figure 6). Activation of adenylate cyclase by guanyl nucleotides was significantly suppressed by al- prenolol. These results gave us a basis for assuming that, in this line of rabbits, the adenylate cyclase system of the heart is desensitized with respect to catecholamines. Usually, desensitization of tissue develops as a result of increased hormone content in the blood. It might be thought that reserpine, which blocks the release of catecholamines into the blood, would return the adenylate cyclase system to a state which is sensitive to catecholamines. In fact, after injec- tion of rabbits with reserpine (24 and 2 hours before decapitation), we succeeded in separating a membrane preparation in which adenylate cyclase was highly sen- sitive to the effects of isoproterenol. As can be seen in figure 6, adenylate cyclase of rabbit hearts treated with reserpine is slightly sensitive to the activating effect of GTP, while isoproterenol plus GTP more than doubles the activation of adenylate cyclase. A change does not occur in the affinity of adenylate cyclase, either for isoproterenol (figure 7) or for guanyl nucleotides (figure 8). Under the combined effect of isoproterenol and guanyl nucleotides, practically identical adenylate cyclase activity is produced in the membranes of both control rabbits and those treated with reserpine (see figures 6 through 8). N 3 \N g SRLLHMIIHTHitirY pmol cyclic AMP/mg protein/min 8 T 100 - 7 sof | 2 . 2 211s] |e] |5 8 sll] |3 0 control reserpine pretreatment FIGURE 6. Activity of adenylate cyclase in heart membranes of control rabbits and rabbits treated with reserpine. (JJ= Without guanosine triphosphate (GTP). B= + 104 GTP. iso = isoproterenol. alp = alprenolol. 207 3 £ +iso £ NS A v -iso Oo 5 200 o £ ~N a. = < -iso oO > o © £ o oL—r—1! | l L 0 107 10% 10% 10% Gpp(NH)p (M) FIGURE 7. Variation of activity of adenylate cyclase in heart membranes of control rabbits (0-0 and 0-0) and rabbits treated with reserpine (0-0 and @-8 ) as a function of isoproterenol (iso) concentration. +Gpp(NH)p Ww 3 ; 1 = > og “orpiHp Sy 0 —/ / | 1 1 1 0 107 10% 10% 10% D,L-isoproterenol (M) FIGURE 8. Variation of activity of adenylate cyclase in heart membranes of control rabbits (0-0 and @-@) and rabbits treated with reserpine ((O-0O and ®-@ ) as a function of guanylyl imidodiphosphate [Gpp(NH)p] concentration. pmol cyclic AMP/mg protein/min 208 Under the effect of reserpine, the quantity of B-adrenergic receptors in heart membranes increases, as determined from the binding of 3H-alprenolol (figure 9). If the receptors of the control membranes contain bound molecules of the hormone, and the receptors of the membranes treated with reserpine are in the free state, then titration of the membranes with 3H-alprenolol would reveal an increase in the apparent affinity of this antagonist for the receptor, and the maximum level of binding would be identical (due to the displacement of molecules of the hormone by alprenolol). However, reserpine resulted only in a change in the maximum number of receptors, while the affinity for the antag- onist remained unchanged. This is indicated by the parallel curves on the Scatchard plot for binding of alprenolol with control and reserpine-treated membranes (figure 9). Consequently, reserpine did not result in removal of the molecules of the hormone from the receptor, but rather revealed additional molecules of B-adrenergic receptors. Based on the effects outlined above--increase in the activating effect of isoproterenol on adenylate cyclase and increase in the quantity of B-receptors under the influence of reserpine--one can conclude that a decrease in the level of catecholamines in the blood causes increased sensitivity of the adenylate cyclase system in the heart. The increase in sensitivity of adenylate cyclase to catecholamines and the increase in the quantity of B-receptors can be achieved in in vitro experiments. Preincubation of membranes with alprenolol for 20 minutes at 37° C, with subse- quent washing of the antagonist, leads to the development of effects similar to those which are observed after injection of rabbits with reserpine. The degree of activation of adenylate cyclase by guanyl nucleotides decreases and the ac- tivating effect of isoproterenol increases (figure 10). In parallel with this, the quantity of B-adrenergic receptors in the membrane increases (figure 11). Preincubation of the membranes with guanyl nucleotides without the hormone causes no increased sensitivity of the adenylate cyclase system. However, 10F © No reserpine pretreatment reserpine pretreatment 05 © — < control alprenolol bound/alprenolol free alprenolol bound (pmol/mg protein) 0 5 10 15 20 % 0.5 1.0 3H-alprenolol (nM) alprenolol bound (pmol/mg protein) FIGURE 9. Specific binding of 3H-alprenolol with heart membranes of control rabbits (0-0) and rabbits treated with reserpine (0-0). 209 Nn 3 g pmol cyclic AMP/mg protein/min 8 I Rm 50 i — HIE IE BLUES 0 - - preincubation preincubation with buffer with alprenolol FIGURE 10. Activity of adenylate cyclase of rabbit hearts after preincubation of membranes in buffer solution with and without 10~> M alprenolol (alp). Membranes were preincubated for 20 minutes at 37° C, and then washed four times with buffer solution. (J = Without guanosine triphosphate (GTP). B= With 10-4 M GTP. iso = isoproterenol. 1.2 £ + 2 10 ° a o 08 E _ 2 ost $ Cc —- o o 3S 8 fe 04+ o + ° 3 3 a2 2 g o2f |s| |B + 12 |Z a gl [8] || lel |& |B a © O QO [G] O 0 1 2 3 4 5 6 FIGURE 11. Effect of preincubation on the maximum quantity of binding sites of 3H-alprenolol in rabbit heart membranes. Membranes were preincubated in a buffer solution for 20 minutes at 37° C. 1. Without additives. 2. With addi- tion of 107° M alprenolol. 3. With 10=%4 M guanosine triphosphate (GTP). 4. With 10=2 M alprenolol + 104 M GTP. 5. With 104 M guanylyl imidodiphosphate [Gpp(NH)p]. 6. With 10=3 M alprenolol + 10=4 M Gpp(NH)p. After preincubation, membranes were then washed four times with the buffer solution and binding of 3H-alprenolol was determined. 210 guanyl nucleotides can develop the effect of increased sensitivity, occurring under the influence of alprenolol (figure 11). This indicates that the GTP- binding protein plays an important role in desensitization and resensitization of the adenylate cyclase system to catecholamines. Changes in the sensitivity of tissue (fat cells, liver, erythrocytes, and reproductive organs) to the effects of insulin, catecholamines, growth hormone, etc., have been described in the literature (13-15). In all cases, the effect developed as a result of long-term (4-24 hours) action of the hormones on the tissue. One distinguishing feature of all desensitized tissues is a reduced number of receptors, along with weak activation of adenylate cyclase by the hormone. The authors note that the activity of basal and NaF-stimulated adenylate cyclase in desensitized tissues remains unchanged. We demonstrated that the adenylate cyclase system of the heart is subject to desensitization. This process is reversible, both in vivo (under the influ- ence of reserpine) and in vitro (as a result of preincubation of membranes with alprenolol or alprenolol + guanyl nucleotides). Consequently, the decrease in the number of receptors occurring upon desensitization of the tissue can be explained by their masking rather than by their destruction, while the loss of sensitivity of adenylate cyclase to the hormone is related to some structural- functional restructuring of the enzyme-receptor complex. An important role in this process is played by the GTP-binding protein, since in the desensitized tissue, GTP and Gpp(NH)p cause a more clearly expressed activation of adenylate cyclase (figures 6 through 8) and accelerate the development of resensitization (figure 11). One necessary condition for the development of resensitization is long- term action of the hormone on the tissue. The hormone antagonists block this process. Desensitization can be achieved both in vivo and in vitro (with long- term incubation of the membranes with the hormones) (13-15). Therefore, we might think that, upon long-term action of the hormone on the receptor, it would be converted to the activated state: R+HZR +« H~> R*H or R+ HZ RH > R* + H. If an activated receptor can interact with adenylate cyclase (AC), R*H + AC ZT R*H + AC or R* + AC Z R* « AC, then the desensitized state can refer to the state of the adenylate cyclase system in which activation of the enzyme can be achieved without addition of the hormone. According to Rodbell's model (7-8,10), transition of adenylate cyclase to the activated state requires interaction of the enzyme with the hormonal recep- tor and the GTP-binding protein. If in the desensitized tissue the adenylate cyclase is preactivated by the receptor, the activation of the enzyme with the 211 guanyl nucleotide should occur without the addition of the hormones. Actually, we see that the desensitized adenylate cyclase is activated by guanyl nucleo- tides to a greater extent than is resensitized adenylate cyclase (figures 6 through 8, and figure 10). The adenylate cyclase system which we studied consists of several protein components, and the nature of the interaction between these components depends essentially on other membrane structures. The true mechanisms of conjugation of B-receptors with adenylate cyclase--development of processes of desensitiza- tion and resensitization--may be more complex. Based on the data presented in this work, we conclude that the guanyl nucleotides are among the most important participants in these processes. 212 10. 11. 12. 13. REFERENCES Murad F, Chi Y-M, Rall TW, Sutherland EW: Adenyl cyclase. III. The effect of catecholamines and choline esters on the formation of adenosine 3',5'-phosphate by preparations from cardiac muscle and liver. J Biol Chem 237:1233-1238, 1962 Levitzki A, Atlas D, Steer ML: The binding characteristics and number of beta-adrenergic receptors on the turkey erythrocyte. Proc Natl Acad Sci USA 71:2773-2776, 1974 Aurbach GD, Fedak SA, Woodard CJ, Palmer JS, Hauser D, Troxler F: B- adrenergic receptor: stereospecific interaction of iodinated B-blocking agent with high affinity site. Science 186:1223-1224, 1974 Lefkowitz RJ, Mukherjee C, Coverstone M, Caron MG: Stereospecific 3H-alprenolol binding sites, beta-adrenergic receptor and adenylate cyclase. Biochem Biophys Res Commun 60:703-709, 1974 Lefkowitz RJ, Roth J, Pastan I: ACTH-receptor interaction in the adrenal: A model for the initial step in the action of hormones that stimulate adenyl cyclase. Ann NY Acad Sci 185:195-209, 1971 Lefkowitz RJ: Heterogeneity of adenylate cyclase-coupled B-adrenergic receptors. Biochem Pharmacol 24:583-590, 1975 Rodbell M, Birnbaumer L, Pohl SL, Krans HMJ: The glucagon-sensitive adenylate cyclase system in plasma membranes of rat liver. V. An obliga- tory role of guanyl nucleotides in glucagon action. J Biol Chem 246: 1877-1882, 1971 Rodbell M: On the mechanism of activation of fat cell adenylate cyclase by guanine nucleotides. An explanation for the biphasic inhibitory and stimulatory effects of the nucleotides and the role of hormones. J Biol Chem 250:5826-5834, 1975 Lefkowitz RJ, Williams LT: Catecholamine binding to the beta-adrenergic receptor. Proc Natl Acad Sci USA 74:515-519, 1977 Rodbell M, Lin MC, Salomon Y: Evidence for interdependent action of glucagon and nucleotides on the hepatic adenylate cyclase system. J Biol Chem 249:59-65, 1974 Cassel D, Selinger Z: Catecholamine-stimulated GTPase activity in turkey erythrocyte membranes. Biochim Biophys Acta 452:538-551, 1976 Pfeuffer T: GTP-binding proteins in membranes and the control of adenylate cyclase activity. J Biol Chem 252:7224-7234, 1977 Lesniak MA, Gorden P, Roth J, Gavin JR: Binding 1251 -human growth hormone to specific receptors in human cultured lymphocytes. J Biol Chem 249: 1661-1667, 1974 213 14. 15. Mickey J, Tate R, Lefkowitz RJ: Subsensitivity of adenylate cyclase and decreased beta-adrenergic receptor binding after chronic exposure to (-)-isoproterenol in vitro. J Biol Chem 250:5727-5729, 1975 Lefkowitz RJ, Limbird LE, Mukherjee C, Caron MG: The beta-adrenergic receptor and adenylate cyclase. Biochim Biophys Acta 457:1-39, 1976 214 PROSTAGLANDIN SYNTHESIS AND REGULATION OF VASCULAR TONE AND PLATELET FUNCTION Philip Needleman SUMMARY A major new pathway of arachidonate metabolism has been discovered in in- tact perfused hearts and in isolated human coronary blood vessels which resulted in the synthesis of a novel, potent, and labile coronary dilator substance. We chemically characterized the primary cardiac-coronary arachidonate metabolite as 6-keto-PGF1,. The labile dilator provides the coronaries with a natural physio- logical antagonist against platelet-derived thromboxane Ap (a potent labile con- strictor and aggregator substance). Indeed, coronary tone and possibly spasm in ischemic myocardial zones may be influenced markedly by the interplay between arachidonate metabolites produced in platelets and blood vessels. Vane et al., in collaboration with the Upjohn Company, identified a labile vascular metabolite and discovered that it inhibited platelet aggregation. This compound degrades to 6-keto-PGFjy,. We found that the cardiac and coronary metabolite did indeed inhibit platelet aggregation and shared all the properties of prostacyclin, or PGI. We have subsequently been able to demonstrate that cardiac PGI), synthe- sis is restricted to the coronary vasculature. Investigations were initiated into platelet - blood vessel arachidonate metabolism and interactions. Contrary to existing but largely undocumented hypotheses, we demonstrated that intact perfused vasculature does not convert prostaglandin (PG)-endoperoxides to PGI. In addition, endothelial cells are not obligatory for vascular PG synthesis, whereas blood vessels readily synthesize both PGI and PGE) from endogenous substrate. Furthermore, thromboxane Aj; (and not PG-endoperoxides) is the major arachidonate metabolite released during in vitro aggregation (platelet-platelet interaction) or adhesion (platelet-blood vessel interaction), which therefore precludes the possibility that blood vessels utilize platelet PGH2 for PGI) synthesis. However, by inhibiting platelet throm- boxane synthesis during adhesion, we facilitated vascular PGI production and thereby favored local dilation and disaggregation at the site of the thrombus. Most recently, we have characterized the unique biological properties of a new family of PG (i.e., trienes). We discovered that the precursor fatty acid eicosapentaenoic acid is readily incorporated in platelet phospholipids and is From the Department of Pharmacology, Washington University School of Medicine, St. Louis, Missouri. 215 released by aggregatory substances. However, this fatty acid proves to be an effective competitive antagonist with the natural substrate arachidonic acid for two key platelet enzymes (cyclooxygenase and lipoxygenase). Some of the released eicosapentaenoate is converted by human platelets to PGH3 and thromb- oxane Aj, both of which surprisingly elevate cyclic AMP levels and thereby inhibit aggregation and also cause feedback inhibition of platelet phospholipase. Thus, triene PG's are potential antithrombotic agents since their precursor fatty acid (which can readily be supplied in the diet), as well as the transformation products (PGH3, thromboxane Aj, and PGI3), are capable of interfering with aggre- gation of human platelets. DISCOVERY OF CARDIAC AND CORONARY PGIp BIOSYNTHESIS Effect of PG Synthesis on Coronary Resistance We found that a PGE-like substance (PLS, measured by superfusion bioassay techniques) appeared to be the endogenous mediator of coronary vasodilation in the isolated perfused rabbit heart produced by several potent vasoactive sub- stances including arachidonic acid, bradykinin, or angiotensin II (AII) (1,2) since: (a) the concentration of PLS in the coronary venous effluent was directly proportional to the concentration of the coronary vasodilator stimulus (brady- kinin, AII, or arachidonate); (b) the PLS release and coronary dilation produced by the agonists was correlated both temporally and quantitatively; and (c) aboli- tion of cardiac PLS biosynthesis by indomethacin also abolished the decrease in coronary resistance produced by these agonists. Paradoxically, we observed that PGE2 at levels equal to or greater than the PG present in the cardiac effluent did not produce coronary dilation (2). One possible explanation was that the coronary dilation produced by arachidonate is due to the unstable endoperoxide intermediates, PGGy or PGHp, rather than to PGEp. However, no endoperoxide was detected in the coronary effluent from a heart perfused with arachidonate (2). A second possibility was that the existence of a novel PG causes the decrease in coronary resistance (which eventually proved to be the case). A third possible explanation was that the site of PG biosynthesis is at or very near the site of action, in this case, in the coronary vascular smooth muscle. Paradoxical Coronary Endogenous Synthesis of a Coronary Dilating Substance From Arachidonate and PG Endoperoxides We observed that isolated spiral strips of beef and human coronary artery were constricted by PGEp (3). Inhibition of PG-cyclooxygenase by aspirin, or indomethacin, led to a rapid and sustained increase in coronary tone suggesting continuous synthesis of endogenous vasodilator PG. Surprisingly, we discovered that, although the primary prostaglandin PGEp is a vasoconstrictor, administration of its precursor, arachidonic acid, caused relaxation of human as well as bovine and canine coronary arteries (3). This important experiment suggested that arachidonate was not converted by the coronary arteries to PGEy. However, the relaxation induced by arachidonate was inhibited by pretreatment with PG cyclo- oxygenase inhibitors. Furthermore, addition of indomethacin to strips pre- viously relaxed by arachidonate elicited contraction (3). 216 A candidate for the relaxing substance produced from arachidonate was the unstable endoperoxide PGG2 or PGHp. We administered PGH as a bolus injection across a superfusion cascade which contained a bovine coronary artery and a rabbit thoracic aorta strip. Surprisingly, we found that PGH)y produced a concentration-dependent relaxation of the bovine coronary artery strip but con- tracted the rabbit aorta (4). Comparable results were obtained with PGH3 (the PG intermediate between eicosapentaenoic acid and PGE3). In contrast, the PGHj (the PG intermediate between dihomo-y-linolenate and PGE]) which lacks the 5-6 bond caused a dose-dependent contraction of bovine coronary artery strips as well as porcine coronary and rabbit aorta (4). Thus, PGHyp and PGH3 caused a dose-dependent relaxation of bovine coronary artery strips, whereas PGHj pro- duced coronary contraction. The primary PG's exert effects directionally oppo- site to those of their endoperoxide precursors. Thus, PGE) and PGEj contracted and PGE] relaxed bovine coronary arteries. Although bovine coronary arteries do not possess thromboxane synthetase activity, they were profoundly constricted by exogenous thromboxane Ap (4). Relaxation of human and bovine coronary arteries induced by arachidonate and PG-endoperoxide suggested at least two possibilities. The vessel may possess a cyclooxygenase but lack PGH2 to PGEp isomerase. Thus, in the absence of enzy- matic generation of primary PG, the dilator endoperoxides PGHy and PGG2 may accumulate. On the other hand, the bovine coronary artery may be producing a novel PG-like substance. In fact, subsequent findings clearly indicate that the coronaries, numerous blood vessels, and other tissues possess an enzyme that catalyzes generation of a labile, potent, and coronary dilating substance from PGHjp. Chemical Characterization of the Novel Pathway Induction of coronary relaxation by arachidonate and PGHy suggested that these substances may be converted to a PG other than PGE. Therefore, we studied the metabolism of l4C-arachidonate by bovine coronary artery strips. Incubation of the 1-l4C-arachidonate-labeled bovine coronary arteries for 4 to 6 hours re- sulted in gradual release of a PLS into the medium (5). Acid-1lipid extraction of the medium followed by thin-layer chromatography showed the presence of a major radioactive product which comigrated with the standard PGE in chromatog- raphy solvent system C (chloroform:methanol:acetic acid:water--90:8:1:0.8) (5). When the extract of the medium bathing a labeled coronary artery strip was chromatographed in a different solvent system (benzene:dioxane:acetic acid-- 60:30:3), the "PGE" peak separated into two peaks, one with chromatographic mobility identical to PGEp and the other with chromatographic mobility similar to PGFpy (5). Lack of a radioactive peak comigrating with PGF2q in chroma- tography system C suggested to us that coronary arteries produce a substance different from PGEp or PGFj,. Alkali treatment (0.5 N KOH in methanol) of the total "PGE" peak isolated from system C resulted in only partial conversion of PGFy, (5). The chemical and chromatographic properties of the unknown substance are similar to those of the major PG produced by the rat fundus homogenate which has been identified by Pace-Asciak and Wolfe (6) as 6-keto-PGFy,. 217 Arachidonic acid infused through an isolated perfused rabbit heart produces coronary vasodilation and concomitant appearance in the effluent of a PG-like substance (2,3). The products of arachidonate metabolism were studied by pre- labeling the cardiac phospholipids with 1-l4c-arachidonate by the method we developed previously (7). The major product formed by the heart was not PGE but was, in fact, the novel PG (8). This substance, like the coronary compound described above, shared all the chemical and chromatographic properties of 6-keto- PGF1o. It has been suggested that the coronary dilation produced by arachidonic acid or hormone stimulation with bradykinin or angiotensin in the rabbit heart was due to PGE2 (1,2,9). The observation that the major product synthesized by both the isolated perfused heart and the coronary artery strips is not PGE) suggests that a novel PG is responsible for the coronary dilation which occurs with arachidonate or bradykinin administration to an intact heart. Biological Characterization of the Unique PG Coronary Vasodilator (9,10) In subsequent studies we found that addition of PGH) caused immediate con- traction of the rabbit thoracic aorta strip and a slightly delayed relaxation of the dog (or bovine) coronary artery. The same amount of PGH2 was then incubated for 2 minutes at room temperature with bovine coronary artery microsomes. The PGHy activity (aorta contraction) had disappeared but an immediate and profound coronary relaxation was produced. The radioactive product obtained from long- term (15 minutes) incubations of 14c-PGH) with blood vessel microsomes was a mixture of 6-keto-PGFj, and PGE. Thus, a labile intermediate between PGH) and 6-keto-PGF1, seems to be responsible for the coronary relaxation induced by either arachidonic acid or by PGGp and PGHp. Incubation of PGH] with bovine coronary artery microsomes did not result in disappearance of the PGHj-induced aortic contraction nor an alteration in the contractile effect on the coronary assay tissue. Thus, the conversion of endoperoxides to the labile relaxing sub- stance apparently requires the presence of a 5-6 double bond. This bond is present in PGHy and PGH3 (which we later proved was converted to PGI3, i.e., Al7-PGI) which dilate the bovine coronary but is absent in PGHj. We obtained analogous results in the isolated perfused rabbit heart. l4c-arachidonic acid was converted by the heart into 6-keto-PGFj, and PGE2, whereas 14c-dihomo-y- linolenic acid (the precursor of PGHj) was converted only to PGE} (8). In independent investigations by Vane et al., blood vessel microsomes from aorta, mesenteric, coeliac, and coronary arteries incubated with PGH2 produced a labile substance found to be a potent inhibitor of platelet aggregation (10) and a relaxant of some vascular smooth muscle. The stable, inactive metabolite of this pathway was found to be 6-keto-PGFi,. The structure of the intermediate was found by Vane et al., in conjunction with the Upjohn Company, to be a bicyclic-PG, designated prostacyclin (11) (now abbreviated as PGI). Release of the Labile Vasodilator Substance From the Isolated Perfused Rabbit Heart (12-14) Based on radiochemical demonstration of 6-keto-PGFj, production by the isolated rabbit heart, we attempted to demonstrate the presence of the labile 218 intermediate, PGI2, in the cardiac venous effluent following hormone stimulation. Intracardiac injection of the potent vasoconstrictor AII caused a decrease in coronary resistance in the perfused heart (as did bradykinin) and the appearance in the cardiac effluent of a substance which relaxed bovine coronary arteries. An aliquot of the cardiac effluent after peptide stimulation inhibited arachi- donate, PGHp, collagen, and thrombin-induced aggregation of washed human platelet suspensions. Administration of AII to perfused hearts pretreated with indo- methacin increased vascular resistance. In addition, no vasoactive PG-like sub- stance was present in the effluent. In previous experiments with isolated hearts, selected assay tissues were employed including chick rectum, rat stomach, and rat colon to detect effluent PGE, since PGEp contracts them. However, it is now clear that the isolated perfused rabbit heart releases a mixture of PGE2 and PGI). It is now apparent why exogenous PGEj did not account for the decrease in cardiac vascular resistance produced by arachidonic acid or by hormone stimula- tion since the intrinsic coronary dilator substance generated from arachidonic acid or endoperoxides is the labile intermediate (PGI) between PGH7 and 6-keto- PGF1y5. The data we obtained demonstrating that the isolated perfused rabbit heart is capable of synthesizing PGI include (5,8,12-14): (a) radiochemical identification of the stable end product 6-keto-PGFjy in the venous effluents of the heart exposed to hormone or ischemia; (b) biological identification of the labile product in the venous effluent released by arachidonate, angiotensin, bradykinin, or ischemia which causes relaxation of bovine coronary artery assay tissue and which inhibits platelet aggregation; and (c) confirmation that arachidonic acid (20:4) but not dihomo-y-linolenic acid (C20:3) serves as the precursor for the coronary dilator, platelet inhibitory substance. We attempted to selectively abolish PGI) biosynthesis by infusing 15- hydroperoxy arachidonic acid through the heart of kidney. Unfortunately, this agent did not appear to act on the intact organs (perhaps because of limited penetration or rapid enzymatic reduction to 15-hydroxy arachidonic acid), but the agent could block in vitro enzymatic PGI) formation from PGHy by bovine aorta microsomes and bovine coronary artery. Finally, administration of PGIy mimicked the decrease in coronary perfusion pressure produced by exogenous or endogenous arachidonic acid. BLOOD VESSEL - PLATELET INTERACTIONS We initially discovered that blood vessels possess the capacity for intrin- sic synthesis of vasodilator PG, thereby providing considerable potential for the local regulation of vascular tone (15). Prostacyclin, or its stable metab- olite 6-keto-PGFj,, has been demonstrated as the primary vascular PG produced. Since PGI2 has been discovered to be a potent in vitro inhibitor of platelet aggregation (10), it thereby became a prime candidate for the local vascular regulation of thrombotic events. Isolated blood vessel preparations rather inefficiently (less than 5 percent) convert exogenous arachidonate (via the enzyme cyclooxygenase) into PG-endoperoxides. However, the PGI) synthetase in vascular tissue quantitatively (more than 80 percent) metabolizes exogenous or endogenous PG-endoperoxides into PGIjp. 219 On the other hand, platelets possess a very active cyclooxygenase which readily cyclizes arachidonate into endoperoxide which is enzymatically converted into the potent aggregator, blood vessel constrictor thromboxane Ap. Vane et al. (16) performed mixing experiments and found that blood vessel segments treated with indomethacin (reversible cyclooxygenase inhibitor) and stirred with platelet-rich plasma prevented aggregation. They therefore proposed that endo- peroxides are released from platelets during aggregation and are used by the blood vessels to synthesize PGIp. Such an exchange of products simultaneously bypasses the inefficient vascular cyclooxygenase and deprives the platelets of substrate for thromboxane synthesis. One implication of such an interaction is that thrombosis and vascular tone may be mediated or modulated by the local balance of the ratio of vascular PGI; to platelet thromboxane. Understanding of platelet and blood vessel arachidonate metabolism and interactions has therefore become the focus of considerable research. We recently demonstrated that exogenous [l4C]-arachidonate added to intact human platelets in the presence of blood vessel microsomes (source of PGI) syn- thetase) resulted only in the production of [14C]-thromboxane Bp (17). PG- endoperoxides were released from the intact platelets only when thromboxane synthesis was inhibited by imidazole (which we discovered to be an in vitro thromboxane synthetase antagonist) (18), at which time the blood vessel micro- somes converted the released endoperoxide into PGI. Furthermore, incubation of unstimulated or thrombin-activated human platelets, which contain [l4c]- arachidonate-labeled phospholipids, with aspirin-treated intact rabbit aorta immediately resulted in adhesion (17). The only cyclized arachidonate product of this platelet - blood vessel adhesion reaction was labeled thromboxane, while no labeled 6-keto-PGFj, was detectable. However, thrombin stimulation of imidazole-inhibited platelets resulted in the release of platelet-derived, labeled PG-endoperoxides which were converted to labeled PGI by the vascular PGIy synthetase. Comparable results were obtained by Baenziger et al. (19) who incubated cultured human arterial smooth muscle and venous endothelial cells with human platelets in the presence of arachidonate and measured PGI) production by bioassay (i.e., inhibition of [l4C]-serotonin release from platelets). Sig- nificant PGI) synthesis from platelet-derived PG-endoperoxides by cultured cells pretreated with aspirin was observed only when platelet thromboxane synthesis was inhibited. These results mitigate against the attractive notion that blood vessels derive endoperoxide from platelets for subsequent conversion to PGI) and thereby are protected from deposition of platelets on vessel walls. Fortunately, these experiments have unmasked a potential pharmacological strategy (that we propose to pursue) which may permit therapeutic manipulation (e.g., in thrombosis, coronary, and cerebral vasospasm, and perhaps athero- sclerosis) of platelet - blood vessel interaction to locally improve the ratio of PGI to thromboxane. Thus, development of a thromboxane synthetase inhibitor effective in vivo could facilitate the release of platelet endoperoxide to vascular PGI2 synthetase. Since vascular injury initiates platelet adhesion and aggregation, a thromboxane synthetase inhibitor has the potential to selectively generate PGI2 at the critical site of interaction. Furthermore, it was initially suggested that vascular endothelial cells were the primary vascular source of antithrombotic PGI, and endothelial loss during injury therefore favored local 220 thrombosis (20). However, subendothelial vascular smooth muscle readily pro- duces PGI, under conditions which favor thrombosis (21). If blood vessels in vivo with damaged endothelium are stimulated to synthesize PGI, it is difficult to reconcile that this substance is a primary deterrent of throm- bosis. DISCOVERY AND FUNCTION OF THE TRIENE - PG PATHWAY We previously demonstrated (22) that the fatty acid 5,8,11,14,17-eicosapen- taenoic acid (C20:5) was converted by sheep seminal vesicle cyclooxygenase into a labile contractile substance which was a mixture of PG-endoperoxides PGG3 and PGH3 (22). The 3-series endoperoxides were then enzymatically converted by platelet microsomes into a potent and labile vasoconstrictor that was presumed to be thromboxane A3 (22,23). In addition, application of purified PGHj (pro- duced from arachidonic acid) or PGH3 to isolated spiral strips of bovine coronary artery caused a transient relaxation, whereas PGHj (produced by 8,11,l4-eicosa- trienoic acid) contracted the coronary strip (13,14). As indicated above, the primary product generated by bovine coronary arteries or by isolated perfused rabbit hearts from l4C-arachidonic acid was 6-keto-PGF1,, the stable endproduct formed from PGI, whereas ldc-eicosatrienoic acid was only converted to PGEj. These results indicated that the AS double bond of PG-endoperoxides is required for PGI) synthesis. Thus, PGHy is the precursor of PGI) and its stable aqueous end product 6-keto-PGFj,, and PGH3 which is an active coronary relaxant was presumably converted to PGI3 and ultimately degraded to its presumed end prod- uct Al7-6-keto-PGF1g. However, no direct chemical or biological proof of this latter pathway has been reported. Comparative study of the action of metabolites of arachidonate and eico- sapentaenoate on platelets in platelet-rich plasma produced unexpected results. Arachidonate, PGH2, and thromboxane Aj were potent aggregators of human platelets, but in sharp contrast, the eicosapentaenoic acid, PGH3, or thromboxane A3 did not cause platelet aggregation (22,23). Recently, this latter observation was given potential physiological perspective by the finding that Eskimos who have a tendency to bleed have elevated eicosapentaenoate and depressed arachidonate levels in their blood lipid fraction (24). These authors suggested that endo- genous PGI3 synthesis by vasculature from eicosapentaenoate contributed to the tendency to bleed. However, no direct evidence of PGI3 synthesis was presented, such as isolation and chromatographic identification of products, or abolition of the synthesis of the antithrombotic substance with a PGIp-synthetase inhib- itor such as 1l5-hydroperoxy-arachidonic acid. In addition, there is no evidence to indicate whether the Eskimos' bleeding disorder is due to a coagulation defect or to a platelet defect. We employed l4c_eicosapentaenoate and documented, biologically and chem- ically, the synthesis of PGI3 and its inactive metabolite Al7-6-keto-PGFj1y, and the synthesis of thromboxane Aj and its metabolite thromboxane B3. Furthermore, we analyzed the unique actions of the triene PG on human platelet-rich plasma and demonstrated an intrinsic platelet mechanism whereby PGH3 or thromboxane Aj inhibits aggregation by pro-aggregatory molecules (25). Thus, we found that the fatty acid, 5,8,11,14,17-eicosapentaenoic acid (C20:5) was converted by sheep 221 seminal vesicle cyclooxygenase into a mixture of PG-endoperoxides, PGH3 and PGG3 (25). Platelets enzymatically converted PGH3 into thromboxane A3. The conventional arachidonate metabolites, PGHy and thromboxane Aj, both aggregate human platelet-rich plasma in vitro. In contrast, PGH3 and thromboxane Aj do not. Addition of exogenous PGH3 and thromboxane A3 increases platelet cyclic AMP in platelet-rich plasma and thereby: (a) inhibits aggregation by all other agonists (i.e., ADP, thrombin, collagen), (b) blocks the ADP-induced release reaction, and (c) suppresses platelet phospholipase Ap activity. PGIg (A17-PGIy), synthesized from PGH3 by blood vessel enzyme, and PGI) exert similar effects. Both compounds are potent vascular relaxants that also inhibit aggregation in human platelet-rich plasma and increase platelet adenylate cyclase activity. Radioactive eicosapentaenoate and arachidonate are readily and comparably acylated into platelet phospholipids; in addition, stimulation of prelabeled platelets with thrombin releases comparable amounts of eicosapen- taenoate and arachidonate, respectively. Although, eicosapentaenoate is a relatively poor substrate for platelet cyclooxygenase, it appears to have a high affinity and thereby inhibits arachidonate conversion by platelet cyclo- oxygenase and lipoxygenase. It is, therefore, possible that the triene PG's are potentially antithrombotic agents because their precursor fatty acid, as well as their transformation products, PGH3 and thromboxane Aj (possibly mediated by PGD3), and PGI3 are capable of interfering with aggregation of platelets in platelet-rich plasma. Furthermore, recent evidence which we are pursuing sug- gests that another trienme product, PGD3, may in fact be the active antithrombotic member of this family causing the inhibitions of platelet aggregation. This latter compound offers the potential of an antithrombotic PG which, unlike PGI, does not cause profound vasodilation. This compound will be the focus of some of our continuing studies. 222 10. 11. 12. REFERENCES Needleman P, Key SL, Denny SE, Isakson PC, Marshall GR: The mechanism and modification of bradykinin induced coronary vasodilation. Proc Natl Acad Sci USA 72:2060-2063, 1975 Needleman P, Marshall GR, Sobel BE: Myocardial hormone interactions: Synthesis, inactivation and coronary vasomotor effects of prostaglandins, angiotensin, and bradykinin. Circ Res 37:802-808, 1975 Kulkarni PS, Roberts R, Needleman P: Paradoxical endogenous synthesis of a coronary dilating substance from arachidonate. Prostaglandins 12:337- 353, 1976 Needleman P, Raz A, Kulkarni PS, Wyche A, Denny S: Prostaglandin I syn- thesis and vascular effects in isolated coronary artery. In Mechanisms of Vasodilation, edited by PM Vanhoutte, I Leusen. Basel, Kargar, 1978, pp 122-128 Raz A, Isakson PC, Minkes MS, Needleman P: Characterization of a novel metabolic pathway of arachidonate in coronary arteries which generates a potent endogenous coronary vasodilator. J Biol Chem 252:1123-1126, 1977 Pace-Asciak C, Wolfe LS: A novel prostaglandin derivative formed from arachidonic acid by rat stomach homogenates. Biochemistry 10:3657-3664, 1971 Isakson PC, Raz A, Needleman P: Selective incorporation of l4c-arachidonic acid into the phospholipids of intact tissues and subsequent metabolism to l4c-prostaglandins. Prostaglandins 12:739-748, 1976 Isakson PC, Raz A, Denny SE, Puré E, Needleman P: The major product of arachidonic acid metabolism in rabbit heart. Proc Natl Acad Sci USA 74:101-105, 1977 Needleman P: The synthesis and function of prostaglandins in the heart. Fed Proc 35:2376-2381, 1976 Moncada SR, Gryglewski R, Bunting S, Vane JR: An enzyme isolated from arteries transforms prostaglandin endoperoxides to an unstable substance that inhibits platelet aggregation. Nature 263:633-635, 1976 Johnson RA, Morton DR, Kinner JH, Gorman RR, McGuire JC, Sun FF, Whittaker N, Bunting S, Salmon J, Moncada S, Vane JR: The chemical structure of prostaglandin X (prostacyclin). Prostaglandins 12:915-928, 1976 Needleman P, Bronson SD, Wyche A, Nicolau KC, Sivakoff M: Cardiac and renal prostaglandin Ip: Biosynthesis and biological effects. J Clin Invest 61:839-849, 1977 223 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, Needleman P, Raz A, Kulkarni PS, Pure E, Wyche A, Denny SE, Isakson P: Biological and chemical characterization of a unique endogenous vasodila- tor prostaglandin produced in isolated coronary artery and in intact perfused heart. In Biochemical Aspects of Prostaglandins and Thromboxanes, Proceedings of the 1976 Intra-Science Research Foundation Symposium, Decem- ber 1-3, Santa Monica, California, edited by N Kharasch, J Fried. New York, Academic Press, 1977 Needleman P, Kulkarni PS, Raz A: Coronary tone modulation: Formation and actions of prostaglandins, endoperoxides, and thromboxanes. Science 195:409-412, 1977 Needleman P, Marshall GR, Douglas JR: Prostaglandin release from vascula- ture induced by angiotensin II. Eur J Pharmacol 66:316-319, 1973 Bunting SR, Gryglewski R, Moncada S, Vane JR: Arterial walls generate from prostaglandin endoperoxides a substance (prostaglandin X) which relaxes strips of mesenteric and coeleaic arteries and inhibits platelet aggregation. Prostaglandins 12:897-915, 1976 Needleman P, Wyche A, Raz A: Platelet and blood vessel arachidonate metabolism. J Clin Invest 63:345-349, 1979 Needleman P, Raz A, Ferrendelli JA, Minkes M: Application of imidazole as a selective inhibitor of thromboxane synthetase in human platelets. Proc Natl Acad Sci USA 74:1716-1720, 1977 Baenziger NL, Becherer PR, Majerius PW: PGI2 production in cultured human arterial smooth muscle cells, venous endothelial cells and skin fibro- blasts. Cell 16:967-974, 1979 Moncada S, Herman AG, Higgs EA, Vane JR: Differential formation of pros- tacyclin by layers of the arterial wall. An explanation for the anti- thrombotic properties of vascular endothelium. Thromb Res 11:323-344, 1977 Puré E, Needleman P: Effect of endothelial damage on prostaglandin syn- thesis by isolated perfused rabbit mesenteric vasculature. J Cardiovasc Pharmacol 1:299-309, 1979 Needleman P, Minkes M, Raz A: Thromboxanes: Selective biosynthesis and distinct biological properties. Science 193:163-165, 1976 Raz A, Minkes MS, Needleman P: Endoperoxides and thromboxanes: Structural determinants for platelet aggregation and vasoconstriction. Biochem Biophys Acta 488:305-311, 1977 Dyerberg J, Bang HO, Stoffersen E, Moncada S, Vane JR: Eicosapentaenoic acid and prevention of thrombosis and atherosclerosis. Lancet II:117-119, 1978 224 25. Needleman P, Raz A, Minkes MS, Ferrendelli JA, Sprecher H: Triene prostaglandins: Prostacyclin and thromboxane biosynthesis and unique biological properties. 1979 Proc Natl Acad Sci USA 76:944-948, 225 PROSTAGLANDINS AND CYCLIC NUCLEOTIDES AS POSSIBLE REGULATORS OF HEART ADAPTATION TO ACUTE AND CHRONIC PRESSURE OVERLOAD V. D. Pomoinetsky, N. G. Geling, A. A. Nekrasova, Ts. R. Orlova, N. M. Cherpachenko, S. A. Kudrjashov, and D. S. Benevolensky INTRODUCTION In recent years, it has been shown that prostaglandins (PG's) not only play a role in the local regulation of myocardial blood flow (1-5), but also have a positive inotropic and chronotropic effect on cardiac muscle (1,6-8). The increase in the contractile force of the myocardium observed in this case, as in the case of inotropic agents, is related to an increase in the concentration of cyclic 3',5'-AMP (9-11) resulting from activation of cardiac adenylate cyclase (9,11-13). Data on the synthesis of PG in the heart itself (3,14) indicate that PG's probably are physiologic regulators of heart function. Synthesis of PG in the myocardium itself may also have a direct modulating effect on the con- tractile activity of the heart and on coronary circulation (16), controlling the release of neurotransmitters at the presynaptic level (16,17). In cardiac pathology modeled in vivo and in vitro, disruptions may occur in the synthesis and release of PG from the myocardium (3,18-21). Biosynthesis and release of PG from the myocardium of perfused rabbit hearts increase under various stimuli (mechanical irritation of the left ventricle with a rubber bal- loon, hypoxia, and anoxia) and decrease during fibrillation of the heart (18-20). These authors also note that an increase in the release of PG into the perfusate is accompanied by a simultaneous increase in contractile activity of the heart. Administration of PG (Ey, Ep, Fp,) during arrhythmias caused by various agents normalizes heart rhythm and function (6,22-25). Limas et al. (21) found a direct relationship between an increase in the synthesis of PG in the myocar- dium during the fifth minute of coarctation of the ascending arc of the aorta in rats and adenylate cyclase activity. Data from the literature (1-21) indicate that PG's have a direct relation- ship to the regulation of biochemical processes which lie at the basis of heart adaptation to external influences. This effect of PG on myocardial metabolism is apparently exerted through the cyclic nucleotides (cAMP and cGMP). Based on these findings, we set the following objectives for our study: From the All-Union Cardiology Research Center, USSR Academy of Medical Sci- ences, Moscow, USSR. 227 1. To study the change in content of PGE and PGFp, in the left ventricle of rabbit heart in acute cardiac overload, caused by measured constriction of the ascending arc of the aorta, and to determine whether there is any correla- tion between the contractile capacity of the heart and its endogenous level of PG. 2. To study the nature of the interaction between PG and the cyclic nucleo- tides (cAMP and cGMP) in adaptation of the rat heart to acute and chronic pres- sure overload. MATERIALS AND METHODS The first part of the study was performed on 53 male chinchilla rabbits weighing 2.5-3.0 kg. The animals were anesthetized with sodium pentobarbital (50-60 mg/kg body weight) and heparinized. The experiment was performed with open thorax and artificial pulmonary ventilation. Systolic and diastolic pres- sures in the left ventricle were recorded by an Elema (U.S.) electromanometer. The first derivative of intraventricular pressure (dP/dt) was recorded, and the contractile index (dP/dt/P) was calculated. The parameters were recorded on a mingograph at paper speeds of 3 and 300 mm/sec. It is important to emphasize that the pressure overload was strictly stan- dardized for all animals. This was achieved by performing coarctation of the ascending branch of the aorta so that the ratio of systolic pressure developed by the left ventricle (LVSP) to its maximum pressure (Ppgx) determined with the aorta completely compressed was comparable in different animals with differ- ent values of initial pressure. Ppgx was determined for each animal with full compression of the aorta for one to two cardiac contractions. After stabiliza- tion of the hemodynamic indices in the experimental animals, a screw clamp was used to constrict the ascending portion of the arc of the aorta in a measured manner so that the LVSP was increased to 80 percent of Ppyx for the animal in question. Throughout the entire time of observation (0-60 minutes), the hemo- dynamic indices of the left ventricle were recorded. In the control group of animals, the same experimental conditions were observed, but without constric- tion of the aorta. After various periods of time from the beginning of coarctation of the aor- ta, the left ventricle of the heart was rapidly removed, frozen in Wollenberger forceps, and preliminarily chilled in liquid nitrogen. Tissue content of PG was then determined. In the second part of the study, 80 male Wistar rats weighing 140-160 g were used. Compensatory hyperfunction and hypertrophy of the myocardium were evoked by twofold to threefold coarctation of the abdominal aorta using metal spring. Hypertrophy of the cardiac muscle was estimated from the increase in relative weight of the heart after creation of experimental aortic stenosis. The control group consisted of 24 rats subjected to surgery but without coarctation. Concentration of PG (E and Foo) in the heart was determined at various time intervals after induction of stenosis (i.e., after 15, 30, and 60 minutes 228 in the rabbits, and after 5, 15, and 40 minutes, 1.2-30.5 hours, and 1, 2, 3, 14, and 22 days in the rats). For each time interval, 5 to 10 animals were tested. The left ventricle was clamped in chilled Wollenberger forceps (26) and placed in liquid nitrogen. Samples of myocardial tissue powder that were frozen in liquid nitrogen (200-300 mg) were carefully homogenized in 4 ml of a methanol- water mixture (2:1, v/v), to which H3-PG (Ap, E1, and Fy) were added, 1,000 dpm each, in order to allow a count of the final output of PG in the process of ex- traction and chromatography on silicic acid. The homogenates were acidified with formic acid to pH 3.5, and then the subsequent stages of PG extraction with chloroform were performed by the method of Auletta et al. (27). Separation of the final extraction of PG into PGA, PGE, and PGF), was per- formed by chromatography of the specimens in small glass columns (D = 0.5 cm), filled with 0.5 g silica gel (Bio-Sil A, Bio-Rad Laboratory, 100-200 mesh). The evaporated fractions of PG were dissolved in methanol, and portions were taken for a count of the radioactivity on a Mark III scintillation counter. The percent yield of PG from the tissue was calculated from the internal stan- dard that was added initially. The other portion of the fraction was left for radioimmunoassay. The content of prostaglandins was determined in the column eluates using the radioimmunoassay procedure developed by Clinical Assay Inc. as modified by Andersen (28). The remaining procedures of PG analysis were performed as suggested by the firm, with the exception of the final stage. The protein complex formed was dissolved in 500 pl of NCS (Nuclear Chicago) solubilizer, and the radioactivity was counted in a 5 ml toluene scintillator (29). The cyclic nucleotides in the heart tissue were also determined by means of a kit of reagents and instructions produced by Amersham (England). RESULTS AND DISCUSSION Content of PG in the Myocardium in Acute Pressure Overload and Their Possible Relationship to Heart Function Analysis of the physiological indices throughout the entire time of observa- tion revealed two types of hemodynamic reactions among a randomly selected group of rabbits undergoing the identical degree of heart pressure overload. Some 50 percent of the animals adapted to the overload. In this group of animals, LVSP increased sharply in response to coarctation and remained practically unchanged throughout the entire experiment. The contraction index for this '"adapted' group of animals is shown by the upper curve in figure 1. In group II (50 percent of the animals), LVSP dropped after 15 minutes of aortic stenosis. This decrease amounted to more than 30 percent of the initial value of LVSP. The contraction index in this group of animals is shown by the lower curve in figure 1. We called this second group of animals "poorly adapted." In most of the animals of this group, cardiac insufficiency developed in response to acute pressure overload, if the observation was continued for several hours. During the first 229 140 ¢ © ¥ AS = sess" a / — 100 ®———————-— IF —— —— © OO 1 oO 1 60 11 1 1 1 AU 1 1 0 5 15 30 60 time after stenosis ( min) FIGURE 1. Change in contractility index (dP/dt/P) at various times after stenosis in adapted (e) and poorly adapted _(o) rabbits. Changes are given in percent of initial control value. M + SE. 5 minutes of the experiment, various types of transitory arrhythmias were ob- served in 60 to 70 percent of the poorly adapted animals. Changes in hemodynamic indices during the first few minutes after coarc- tation of the aorta indicate that the reserve capacities of the myocardium are "switched on," that is, directed toward maintenance of systemic hemodynamics. Our results show that the compensatory capabilities of various animals are not the same. It can be assumed that the increase in LVSP in animals adapting to the load is achieved by switching between one of two autoregulation mechanisms: either the Frank-Starling mechanism or changes in the inotropic states of the myocardium. In the group II animals, LVSP is governed only by the first mecha- nism, i.e., by the significant overfilling of the heart (which is indicated by the high final diastolic pressure in the left ventricle), and is not rein- forced by an increase in the contraction of the myocardium (the index of con- traction decreases). This method of providing a high LVSP seems less adaptive. In the group I animals, the rate of contraction and relaxation of the left ventricle increases with pressure loading, whereas in the group II animals, these indices decrease. 230 Figures 2 and 3 show the tissue content of PG, determined in quick-frozen specimens of the left ventricles of hearts from the adapted and poorly adapted animals. The ordinate shows the ratio of tissue PG content in the experimental animals to the same index in the control animals, at various time intervals. About five animals were used to produce each point on the diagram. Figure 2 shows the changes in PGE content in the left ventricle of the heart of adapted rabbits at various times after the beginning of coarctation of the aorta. As can be seen from the figure, the content of PGE in this group of animals increased by a factor of 5 to 10 as a result of aortic coarctation. In the poorly adapted animals, however, the content of PGE in the myocardial tissue is only slightly greater than the control level. These results show that the rise in LVSP and the contraction indices in adapted animals are accompanied by a simultaneous increase in the content of PGE in the left ventricle. On the other hand, inability of the heart to maintain high LVSP with pressure overload in this group of poorly adapted animals is accompanied by an inability to in- crease the endogenous PGE level in the left ventricle. It is important to note that these results are similar to those obtained for the other type of PG, PGFp,, shown in figure 3. The ordinate of this 15 + (5) (5) Co o \ 0 15 30 45 60 time after stenosis ( min ) FIGURE 2. Change in prostaglandin E (PGE) content in the left ventricle of hearts at various times after stenosis of the ascending branch of the aorta in adapted (e) and poorly adapted (o) rabbits. C4/C, = Ratio of tissue PG content in experimental animals to tissue PG content in control animals. M + SE. 231 time after stenosis (min) FIGURE 3. Change in prostaglandin Fp, (PGFp,) content in the left ventricle at various times after stenosis of the ascending branch of the aorta in adapted (e) and poorly adapted (o) rabbits. Cx/Co = Ratio of PGF, content in the tissue of experimental animals to PGF), content in the tissue of con- trol animals. M * SE. figure shows the ratio of the content of PGF), in the myocardium of the left ventricle of the experimental animals to its content in the control animals; the abscissa shows the time from the beginning of coarctation. Once more, in the group of animals adapted to acute pressure overload, the tissue content of PGF), increases and, as in the case of PGE, the content of PGFp, in the left ventricle of the poorly adapted animals differs little from the content in the control animals. The significant rise in the content of endogenous PG in the left ventricle of the adapted animals following coarctation of the aorta, and the absence of any such reaction in the poorly adapted animals, indicate that the synthesis of PG may be important in the physiological reaction of the heart to overload. Effect of Intraventricular Infusion of Indomethacin on PG Content, Myocardial Enzyme Systems, and Heart Function With Acute Pressure Overload If our assumption is correct, inhibition of PG synthesis in the hearts of animals adapted after coarctation should change the hemodynamic indices of this group of animals, making them less adapted with the same degree of loading. 232 To test this hypothesis, we used indomethacin, a well-known PG synthesis inhibitor, administered into the left ventricular cavity of adapted animals 15 minutes after coarctation of the aorta. Throughout the entire period of observation, we continuously recorded the hemodynamic indices of the heart. After 15 minutes, the left ventricle of the heart was removed for determination of PG tissue content. The control animals in these experiments received the same solution in the left ventricular cavity, but without indomethacin. As can be seen from figure 4A, intraventricular infusion of indomethacin significantly inhibits the synthesis of PGE and PGF), in the myocardium of the left ventricle of rabbits subjected to acute pressure overload of the heart. Thus, synthesis of PGE was inhibited (blocked) by 60 percent, and synthesis of PGFy, was inhibited by 80 percent. The effect of infusion of indomethacin on the hemodynamic indices of the heart (LVSP and the contractility index) is shown in figure 4B. We can see from this figure that, with 15 minutes of administra- tion of indomethacin, LVSP drops by 60 percent in the control animals. A sta- tistically significant decrease (by 20 percent) in the contractility indices is observed in the groups of animals receiving indomethacin. Thus, it follows from these experiments that blockage of PG synthesis in the hearts of animals adapted to overload leads to a change in the hemodynamic indices of the heart characteristic for the second group of animals, i.e., the poorly adapted group. With the same experimental design, we also studied the histochemical ac- tivity of certain enzymes involved in the Krebs cycle (succinate dehydrogenase, malate dehydrogenase), and in fat (B-hydroxybutyrate dehydrogenase) and carbo- hydrate metabolism (lactate dehydrogenase, glucose-6-phosphate dehydrogenase, a-glycerophosphate dehydrogenase, and phosphorylase), as well as one of the main enzymes responsible for breaking down PG (prostaglandin dehydrogenase). It should be emphasized that these enzymes play an important role in the metabolic processes in cardiac muscle by assuring its normal functioning. In studying this spectrum of enzymes, we did not detect any statistically signifi- cant differences in their activity in either the experimental or control groups of animals following 15 minutes of infusion of indomethacin into the left ven- tricle. However, some data in the literature indicate that longer (over 30 min- utes) or rapid administration of small doses of indomethacin may lead to a decrease in the activity of certain enzymes in the organs and tissues (30). Analysis of our results leads to the following conclusions: 1. Brief infusion of indomethacin (not over 15 minutes) apparently does not influence the activity of most enzymes in the heart (except PG synthetase). 2. With this experimental methodology, indomethacin, in and of itself, does not lead to a change in the metabolism of the heart, which could cause a decrease in the contractile activity of the myocardium. 3. The results of these studies once more support our hypothesis that a decrease in PG synthesis in the heart may lead to a change in the heart's con- tractile activity and may therefore be one cause of failure to adapt to pressure overload. 233 19.0 As) B 490 8 S 43h (5) ro ~ 4 T 3 3 460 © oO 96 $ gL 9 (6) 430 7 4.9 0.0 PGE PGF, 10 ® = 300° 7% Jeo 2 = Ee) (@] ~ © ? “= = = + aD c 200 qo 3 - 6) S, Qo (6) o a T S 0 3 2 100 2 00 = LVSP dP/dt/P FIGURE 4. Effect of intraventricular infusion of indomethacin on prostaglandin (PG) content in tissue of the left ventricle, and hemodynamic response in adapted rabbits. A: Relative changes in PG content. Cx/Co = Ratio of tissue PG content in experimental animals to tissue PG content in control animals. (J= Increase in tissue PG content in control animals within 30 minutes after beginning of coarctation. d= Corresponding figures for the group treated with indomethacin. B: Relative changes in left ventricular systolic pressure (LVSP) and the contractility index. (OJ= Increase in values in the control group of animals 30 minutes after beginning of coarctation. {d= Corresponding changes in the animals treated with indomethacin. M * SE. 234 Content of PG and Cyclic Nucleotides in the Heart With Chronic Pressure Overload Coarctation of the abdominal aorta in rats caused a clear and statistically significant increase in heart weight (figure 5). The animals were divided into three groups according to the rate of development and degree of cardiac hyper- trophy. Group I rats showed clear hypertrophy and were called rapidly adapting. Rats in this group were distinguished from the other animals by their more ac- tive behavior and absence of visible signs of cardiac insufficiency. Group II rats showed moderate hypertrophy and were called slowly adapting. The animals were not very active and showed clear signs of cardiac insufficiency (dyspnea, ascites). Group III rats showed slight hypertrophy of the myocardium and rapidly developing cardiac decompensation. These animals were called nonadapting. op A —" 130+ | — Z——— | 100d 70. . left ventricular weight (%) Oo -_ N+ w ay H N N N Oo T HOH PGE | PGF,,, o { \ < ol \ 0 2 3 14 22 time after stenosis (days) FIGURE 5. Change in relative heart weight (A) and ratio of prostaglandin E/Fy, (PGE/PGF2,) content (B) in the myocardium of rats at various times after steno- sis in rapidly adapting (e), slowly adapting (0), and nonadapting (Vv) animals. M = SE. 235 All three groups of animals were distinguished by the varying degree of changes in PG content in the myocardium in response to acute and chronic cardiac pressure overload. Figure 6 shows the dynamics of PGE and PGF), content, respec- tively, for each group of animals. Group I animals, with the most rapid development of hypertrophy and, con- sequently, heart adaptation to overload, were characterized by high PG content in the heart (figure 6). These data agree with the results described above. It is well known that this hypertrophy is based on stimulation of nucleic acid and protein synthesis, which is observed in the early period after induction of cardiac overload (15). Based on their own experimental data, Limas et al. (21) assumed that there should be a relationship between the activity of PG — 24 A DO j & - °l o— 4 ~~ Z of Pe 2 No Pe AP _ A = 2al B PGF,(ng/g tissue) @ —- @ Oo “N TT EO Hl / x 0 15min 1h 25h 5h 1d 3d 14d 22d time after stenosis FIGURE 6. Changes in prostaglandin E (PGE) content (A) and prostaglandin Fj, (PGF9,) content (B) in heart tissue at various times after aortic stenosis in rapidly adapting (e), slowly adapting (0), and nonadapting (A) animals. M # SE. V = Rats subjected to surgery but without coarctation. 236 synthetase in the heart and the synthesis of protein. They found a direct corre- lation between PG content in the heart of rats after the fifth minute of coarcta- tion of the ascending branch and the activity of the enzyme adenylate cyclase. These investigators, for the first time, supported the hypothesis that one of the initial stimuli starting the chain of biochemical reactions leading to pro- tein synthesis is early activation of PG synthetase. However, this activation of PG synthetase can also be observed with various other stimuli, particularly nervous and hormonal stimuli. The end result of these stimuli is rapid synthesis of both PG groups (E and Fj,) in the myocardium. An increase in PG content was found in group I. The literature contains data on the varying influence of PGE and PGFy, on the synthesis of nucleic acids and protein: For PGE, a stimulating effect has been found; for PGFj,, a blocking effect has been observed (31,32). Based on these findings, we assumed that the development of hypertrophy may be related to changes in the ratio of levels of PGE to PGFjy in the heart. Comparison of the dynamics of this ratio with the dynamics of the relative increase in heart weight indicated that there is some interrelationship between these two vari- ables (figures 5A and 5B). The greatest increase in the PGE/PFGy,, ratio in the direction of increasing PGE corresponds in time to a rapid increase in heart mass. The main question, however, remains unresolved: How do PG's influence pro- tein synthesis? Most probably, this occurs by the effect of PG on the cAMP system (35). Based on his experiments, Kuehl (33) has set forth the hypothesis that PGE and PGFp, act on cell metabolism through various mediators (cAMP and cGMP, respectively). Since one of the objectives of our study was to determine the possible relationship of PG to the cyclic nucleotides in heart adaptation to acute and chronic pressure overload, the nature of this relationship was analyzed for animals in group I, where these characteristics were more sharply expressed than for the animals in groups II or III. Figure 7 shows the relationship between the content of cAMP and PGE, and cGMP and PGFj,. Analysis of these data shows that the dynamics of PGE content in rat hearts coincides with the dynamics of cAMP content. This once more confirms the hypothesis of Kuehl (33) that PGE acts through cAMP. An inverse dependence was found for PGFy, and cGMP in heart adaptation to pressure overload. This seemingly paradoxical nature in the relationship between cGMP and PGFj, is difficult to explain at present. Future experiments will help to clarify this question. In comparing the changes in cAMP and cGMP content in figure 7, we see that these indices mirror each other, thus con- firming the hypothesis of Goldberg that cAMP and cGMP play opposite roles in the regulation of cell metabolism (34,35). 237 @® 24+ yd 1 60073 _ 3 fo) CY “o rs 0 7 — Let o 1400 © - o— | or o ND J \ 2 ~~ z nC a 8t 1200 , 2 oO a oO { — — } 1 F A 0 0 eB O p 0 \ 8 > o 0 n 2 { Ne {140 + L oO z * ~~ o * 0° = c \ a ~35 T ° / 12° 4 un NG N A o ° = o 0 15min 1h 2.5h 5h 1d 3d 14d 22d time after stenosis FIGURE 7. A: Interrelationship between prostaglandin E (PGE) content (e) and B: Inter- cAMP content (0) in the heart during adaptation to pressure overload. relationship between prostaglandin Fp, (PGFp,) content (o) and cGMP content (e) in the heart during adaptation to pressure overload. Possible Role of PG in Heart Adaptation to Pressure Overload Analysis of the data in the literature and the results of our own experi- ments allow us to formulate the following hypothesis about the role of PG in the adaptation of the heart to external influences. This proposed role is shown in figure 8. It follows from our model that PG's activate adenylate cyclase in the heart, and it has now been shown that this activation is independent of the action of catecholamines (the effect appears even in the presence of propranolol), start- ing the entire cascade of biochemical reactions directly related to Ca trans- port in the myocardium (36). At present, in vitro observations have been made 238 P 4 hormone ty, adenylate cyclase phospholipase cyclic 3,5-AMP free fatty acids & protein kinase prostaglandins ¢ regulation of Ca" transport FIGURE 8. Proposed mechanism of action of prostaglandins in adaptation of the heart to pressure overload. (37) that show stimulation of cAMP-dependent protein kinase and phosphorylase activity of rat heart by PGE, independent of the effect of catecholamines. These facts, together with others demonstrated previously, lead to the conclu- sion that PG's apparently influence all stages of regulation of Catt transport in the heart. The diagram in figure 8 shows one possibility of the effect of PG's on Catt transport as an iontophore (38). This hypothesis has not been en- tirely proven and, in our opinion, is not very probable. In sum, with the regulation of metabolic processes supporting the contractile activity of the myocardium, PG's probably serve as an intermediate link between physiologically active substances (catecholamines, etc.) and the intracellular regulatory sys- tem (adenylate cyclase, or cAMP). 239 10. 11. 12. 13. 14. 15. REFERENCES Hollenberg M, Walker RS, McCormick DP: Cardiovascular responses to intra- coronary infusion of prostaglandins Ej, Fj, and Fp, (1). Arch Int Pharmacodyn Ther 174:66-73, 1968 Nakano J: Effects of prostaglandins Ej, Aj, and Fy, on the coronary and peripheral circulations. Proc Soc Exp Biol Med 127:1160-1163, 1968 Needleman P: The synthesis and function of prostaglandins in the heart. Fed Proc 35:2376-2381, 1976 Kalsner S: Endogenous prostaglandin release contributes directly to coronary artery tone. Can J Physiol Pharmacol 53:560-565, 1975 Zayat AF, Hefnawi F, Amir YE, Bayoomi WK: Histophysiological studies of prostaglandin Fj, on isolated organs. I. Effect of prostaglandin F2q on the heart. Prostaglandins 13:131-142, 1977 Forster W, Mentz P: Effects of prostaglandin E-1, prostaglandin E-2 and prostaglandin F-2-alpha on isolated normal and damaged heart preparations. Adv Biosci 9:379-384, 1973 Schror K, Forster W: Interactions between isoproterenol and prostaglandin Eo in the dog heart in situ. Pol J Pharmacol Pharm 26:143-149, 1974 Su JY, Higgins CB, Friedman WF: Chronotropic and inotropic effects of prostaglandins Ej, A], and Fp, on isolated mammalian cardiac tissue. Proc Soc Exp Biol Med 143:1227-1230, 1973 Kaumann AJ, Birnbaumer L: Prostaglandin Eq action on sinus pacemaker and adenylyl cyclase in kitten myocardium. Nature 251:515-517, 1974 Moura AM, Simpkins H: Cyclic AMP levels in cultured myocardial cells under the influence of chronotropic and inotropic agents. J Mol Cell Cardiol 7:71-77, 1975 Sobel BE, Robison AK: Activation of guinea pig myocardial adenylcyclase by prostaglandins (abstr). Circulation 40 (suppl III):III-189, 1969 Levey GS, Klein I: Solubilized myocardial adenylate cyclase: Activation by prostaglandins. Life Sci 13:41-46, 1973 Levey GS: Phospholipids, adenylate cyclase, and the heart. J Mol Cell Cardiol 4:283-285, 1972 Piper P, Vane J: The release of prostaglandins from lung and other tissues. Ann NY Acad Sci 180:363-385, 1971 Schreiber SS, Klein IL, Oratz M, Rothschild MA: Effect of acute overload on protein synthesis in cardiac muscle microsomes. Am J Physiol 213: 1552-1555, 1967 240 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. 26. 27. 28. 29. 30. Gudbjarnason S: Prostaglandins and polyunsaturated fatty acids in heart muscle (editorial). J Mol Cell Cardiol 7:443-449, 1975 Wennmalm A: Studies on mechanisms controlling the secretion of neurotrans- mitters in the rabbit heart. Acta Physiol Scand [Suppl] 365:1-36, 1971 Block AJ, Poole S, Vane JR: Modification of basal release of prostaglan- dins from rabbit isolated hearts. Prostaglandins 7:473-486, 1974 Wennmalm A, Pham-Huu-Chanh, Junstad S: Hypoxia causes prostaglandin release from perfused rabbit hearts. Acta Physiol Scand 91:133-135, 1974 Needleman P, Key SL, Isakson PC, Kulkarni PS: Relationship between oxygen tension, coronary vasodilation and prostaglandin biosynthesis in the isolated rabbit heart. Prostaglandins 9:123-134, 1975 Limas CJ, Ragan D, Freis ED: Effect of acute cardiac overload on intra- myocardial cyclic 3',5'-AMP: Relation to prostaglandin synthesis. Proc Soc Exp Biol Med 147:103-105, 1974 Mest HJ, Blass KE, Forster W: [On the antiarrhythmic action of prostaglan- dins Aj, Ej, A2, Ep and Fj, on the strophanthin arrhythmia model of the cat] (Ger). Acta Biol Med Ger 35:63-67, 1976 Forster W: Prostaglandins and prostaglandin precursors as endogenous anti- arrhythmic principles of the heart. Acta Biol Med Ger 35:1101-1112, 1976 Mann D, Meyer HG, FOrster W: Preliminary clinical experience with the antiarrhythmic effect of PGFy,. Prostaglandins 3:905-912, 1973 Vergroesen AJ, De Boer J: The effects of prostaglandins on the hypodynamic frog heart compared with those of fatty acids, epinephrine and adenosine phosphates. J Am Oil Chem Soc 48:94A, 1971 Wollenberger A, Ristau O, Schoffa G: [A simple technic for extremely rapid freezing of large pieces of tissue] (Ger). Pfluegers Arch Ges Physiol 270:399-412, 1960 Auletta FJ, Zusman RM, Caldwell BV: Development and standardization of radioimmunoassays for prostaglandins E, F, and A. Clin Chem 20:1580- 1587, 1974 Andersen NH: Dehydration of prostaglandins: study by spectroscopic method. J Lipid Res 10:320-325, 1969 Florini JR, Dankberg FL: Changes in ribonucleic acid and protein synthesis during induced cardiac hypertrophy. Biochemistry 10:530-535, 1971 Roman RJ, Kauker ML, Terragno NA, Wong PY-K: Inhibition of renal prosta- glandin synthesis and metabolism by indomethacin in rats. Proc Soc Exp Biol Med 159:165-170, 1978 241 31. 32. 33. 34. 35. 36. 37. 38. Lupulescu A: Effect of prostaglandins on protein, RNA, DNA and collagen synthesis in experimental wounds. Prostaglandins 10:573-579, 1975 Ragni MV, Preuss HC: PGE; stimulation of 3H-thymidine incorporation into renal slice DNA. Clin Res 19:545, 1971 Kuehl FA: Prostaglandin cyclic nucleotide interaction in mammalian tissues. In Advances in Prostaglandin Research. Prostaglandins: Chemical and Bio- chemical Aspects, edited by SMM Karim. Lancaster, MTR Press Limited, 1976, pp 192-225 Goldberg ND, Haddox MK, Nicol SE, et al.: Biologic regulation through opposing influences of cyclic GMP and cyclic AMP: The Yin Yang hypothesis. In Adv Cyclic Nucleotide Res, vol 5, edited by GI Drummond, P Greengard, GA Robinson. New York, Raven Press, 1975, pp 307-351 Johuson M, Ramwell PW: Prostaglandin interaction with membrane components. Intra Sci Chem Rept 88:93-104, 1974 Silver MJ, Smith JB: Prostaglandins as intracellular messengers. Life Sci 16:1635-1648, 1975 Keely SL, Park CR: Effects of epinephrine and PGE] on heart cAMP dependent protein kinase, phosphorylase, and contraction (abstr). Fed Proc 36:283, 1977 Carsten ME, Miller J: Effects of prostaglandins and oxytocin on calcium release from a uterine microsomal fraction. J Biol Chem 252(5):1576-1581, 1977 242 GLUCOCORTICOIDS IN THE HORMONAL REGULATION OF CARDIAC METABOLISM Yu. M. Seleznev, S. M. Danilov, N. G. Volkova, G. V. Kolpakova, and A. V. Martynov Based on studies of the interaction of glucocorticoids with cytoplasmic and nuclear fractions of heart tissue, we have suggested that the heart is the target organ for glucocorticoid hormones (1-4). In the present paper, we sum- marize the experimental results to date which confirm this concept. The heart contains three glucocorticoid-binding fractions (binding factors I, II, and III). Some of the properties of binding factors I and II were pre- sented in 1977 at the Third Joint Symposium on Myocardial Metabolism (1). A number of the properties of binding factor I indicate that it is analogous to the corticosteroid-binding blood serum globulin--transcortin. Attempts to determine the location of this fraction in heart tissue have shown that at least most of it is present as an intercellular pool of transcortin, which finds its way into the cytoplasmic fraction upon homogenization of the tissue. The presence of transcortin in the intercellular space of perfused isolated hearts has decreased intracellular accumulation of naturally encountered glucocorti- coids, but not the synthetic glucocorticoid dexamethasone, which is not capable of forming a complex with transcortin. A second glucocorticoid-binding protein, which is detected in cytosol preparations, is binding factor II which, in contrast to transcortin, can bind with dexamethasone with great affinity, while the degree of its affinity for the other steroids correlates with the glucocorticoid activity of these steroids. This protein has been localized within the cells. Binding factor III is yet another glucocorticoid-binding protein which we have recently found in preparations of cytosol from rat heart (3). This fraction is found only after careful washing of the heart to remove extracellu- lar transcortin because of its significantly lower concentration in ordinary preparations of cytosol in comparison to preparations contaminated with trans- cortin. Binding factor III, like transcortin, forms complexes with the naturally encountered glucocorticoids but does not bind with dexamethasone. Treatment with sulfhydryl blockers inactivates it. From the All-Union Cardiology Research Center, USSR Academy of Medical Sci- ences, Moscow, USSR. 243 Table 1 presents the physical parameters of the binding proteins, calcu- lated on the basis of gel chromatography and sedimentation analysis of 31- glucocorticoid-protein complexes of rat myocardial cytosol. The molecular weights are given, considering deviation of the forms of the macromolecules from the globular. As shown in the table, binding factors I, II, and III differ in size. In cell-free systems, and particularly when there are no glucocorticoids present, binding factor II is extremely labile. Its binding activity is lost upon long-term storage at 0-4° C, short-term storage at over 25-30° C, or at high ionic strength. The formation of a complex with glucocorticoid increases its stability. Nevertheless, in a buffer system containing a relatively low concentration of the sulfhydryl group protector dithiothreitol (buffer 1, figure 1), at 0-4° C in the presence of the hormone, significant inactivation of binding factor II occurs within 18-24 hours. Equilibrium in the system consisting of free steroid and bound steroid occurs after approximately 18 hours of incubation (during this time, the values of the association constant [Kg] reach their maximum and constant value, whereas the highest values of concen- tration of binding locations [N] are noted by the fourth hour of incubation. TABLE 1. The Physical Parameters of the Binding Proteins Binding Binding Binding Parameters factor I factor II | factor III Stokes radius, A at low ionic strength 36.5 53.0 31.5 at high ionic strength 36.5 39.5 31.5 Coefficient of sedimentation, S at low ionic strength 4.0 7.0 4.0 at high ionic strength 4.0 4.0 4.0 Friction ratio at low ionic strength 1.24 1.32 1.13 at high ionic strength 1.24 1.31 1.13 Molecular weight at low ionic strength 63,500 161,400 54,800 at high ionic strength 63,500 68,700 54,800 Stokes radii were calculated as in (5), based on the results of gel chromatog- raphy of 3H-glucocorticoid-cytosol complexes on ultragel ACA-34. The sedimenta- tion coefficients were calculated as in (6), based on the results of ultra- centrifugation of complexes in a sucrose density gradient (7). The friction ratio and molecular weights were calculated as in (8). 244 10 410.5 £ J © a - o : los E db 5 2 » ° » E 3 a x =z 2 I" 6 12 18 24 ours FIGURE 1. Specific binding of 3H-dexamethasone by binding factor II of heart cytosol in adrenalectomized rats as a function of incubation time at 4° C. The concentration of protein in the cytosol was 5.5 mg/ml. The method of separation of cytosol and determination of binding are described in detail in (2). The cytosol was extracted using buffer I: Tris 25 mM, KCl 12.5 mM, sucrose 0.25 M, EDTA 0.5 mM, dithiothreitol 1.5 mM, pH 7.5. Each point on the graph is the arithmetic mean of three parallel determinations. Representative results are given from one of two individual experiments. Use of the buffer system, containing a high concentration of dithiothreitol (buffer 2, figure 2), greatly reduces the rate of inactivation of binding factor II, even when there is no glucocorticoid present (figure 2). The data indicate that the most correct value of the association content of binding factor II with dexamethasone is 1+109 M, rather than 108 M, as was reported earlier (1,2). Under analogous conditions, equilibrium in the system of free corticosterone- transcortin-bound corticosterone occurs in a few minutes, and the binding is independent of the presence of protectors or sulfhydryl group blockers (2). Presented below are the results of experiments in a cell-free system. These results help us to approach an understanding of the role of glucocorticoid- binding proteins in the transmission of glucocorticoid signals into the cell nuclei of the heart. Binding factor II, under certain conditions, can greatly increase the accumulation of 3H-glucocorticoids in the nuclei. At 0-4° C, com- plexes of glucocorticoids with this binding protein are in the preactive state. Activation of complexes by brief heating at 20° C causes a significant increase in the accumulation of hormones in the nuclear fraction (2). At physiological concentrations of hormones (about 10-8 M), under analogous conditions, other cytosol-binding proteins do not have this property. Furthermore, the free form of the hormones is practically not accumulated by the nuclei. As is shown in figure 3, other factors involved in the activation of the complex of glucocorticoids with binding factor II may be the treatment of cyto- sol with 0.4 M KCl or 10 mM theophylline. The same properties are characteris- tic for the transfer of glucocorticoids into the nuclei of the liver by the liver cytoplasmic receptor for glucocorticoid hormones (8). Thus, binding 245 2 1007 2 3 8 e 3 50+ | QQ i LU w I I FIGURE 2. Specific binding of 3H-dexamethasone by heart cytosol in adrenalec- tomized rats in various buffer systems. Binding was evaluated after 4 hours incubation of dexamethasone with cytosol. The first estimate of binding was performed immediately after extraction of cytosol. The second estimate was performed after 16 hours of storage of cytosol at 4° C without steroid. = Cytosol separated in buffer I. (J = Cytosol separated in buffer II. Composi- tion of buffer II: Tris 50 mM, KCl 50 mM, sucrose 0.25 M, EDTA 1.5 mM, dithio- threitol 10 mM, pH 7.5. The columns in the figure represent the arithmetic mean values of three parallel determinations. Representative results are given from one of two individual experiments. factor II obviously is the cytoplasmic receptor of glucocorticoids in heart tissue cells. Figure 4 shows data from experiments on the accumulation of 38- dexamethasone-receptor complex from the heart by nuclei extracted from the liver, and accumulation of 3H-dexamethasone-receptor complex of the liver by nuclei from heart cells. These experiments demonstrate that the two types of steroid receptor complexes are capable of accumulating steroids in both nuclei from the liver and nuclei from the heart. Unfortunately, in these experiments the concentration of steroid-receptor complex of the liver was significantly higher than the heart complex; therefore, it is more correct to only qualitatively estimate the accumulation of the com- plexes of the various nuclear fractions. However, since nuclear accumulation increased with an increase in the concentration of the same steroid-receptor complex in the system (see the two lower curves of figure 4), we can assume that the process is determined to a great extent by the quantity of the complex 246 n o Oo 3H-dexamethasone nuclear o oO accumulation, % il Ir FIGURE 3. Accumulation of SH-dexamethasone in nuclei after activation of 3H- dexamethasone-cytosol complexes by various factors. Extraction of 3H- dexamethasone-cytosol complexes and nuclei, plus determination of nuclear accumulation, were performed as described earlier (2). In all experiments, incubation of 3H-dexamethasone-cytosol complexes with nuclei was performed for 20 minutes at 4° C. A: 1. Unactivated complex. 2. Complex activated by heat- ing for 20 minutes at 20° C. 3. Complex activated by 10 mM theophylline. B: In these experiments, the initial concentration of cytosol for preparation of 3H-dexamethasone-cytosol complexes of protein was double the ordinary con- centration (2). Before incubation with nuclei, 3H-dexamethasone~-cytosol com- plexes were diluted with four volumes of buffer, and crystalline KCl was added to a final concentration of 0.1 M. 1. Unactivated complex. 2. Complex acti- vated by heating for 20 minutes at 20° C. 3. Complex activated with 0.4 M KCl before dilution. in the system, rather than by tissue (cellular) specificity. The total number of nuclear acceptors in each system was significantly greater than the quantity of the complex necessary for saturation of the nuclei. In a cell-free system, nuclear accumulation of a naturally encountered glucocorticoid--corticosterone--depends on the ratio of binding proteins present in the preparations (9). An increase in the fraction of corticosterone bound not with transcortin, but rather with cytoplasmic proteins (binding factors II and III), increases the absorption of the hormone by the nuclei. This shows that transcortin, under natural conditions, limits the access of the hormone or cytoreceptor. At the same time, the accumulation of corticosterone from transcortin-free cytosol does not reach the level of accumulation of dexametha- sone at concentrations of hormones close to the physiological level of corti- costerone in the cells. One can assume that this difference results from the presence of binding factor III, since this factor does not bind dexamethasone, but does bind corticosterone and should, therefore, limit the access of naturally encountered glucocorticoid to the receptor. An alternative explanation may be the different affinity of corticosterone- and dexamethasone-receptor complexes for the nuclei. The properties of corticosteroid-binding proteins found in the heart indi- cate that binding factor II is the only receptor which has a unique capability for transmission of the glucocorticoid signal to the nuclei at physiological concentrations of the hormones. One very interesting fact was that at higher 247 350 + 250 150 . -3 nuclear accumulation, pmoles (107%) 50 50 150 DNA, ug/probe FIGURE 4. Accumulation of 3H-dexamethasone-cytosol complexes by nuclear frac- tions extracted from the heart and liver. 1,11,2,21 = Cytosol from heart, nuclei from liver. 3,31 = Cytosol from liver, nuclei from heart. 111, 211 311 = Buffer solution containing 3H-dexamethasone, nuclei from corresponding tissue. 1,111,2,211 3,311 = Incubation at 20° C for 20 minutes. 11,2131 = Incubation at 0° C for 20 minutes. The concentration of 3H-dexamethasone-cytosol complexes is as follows: 1. 0.24 pmol/mg protein. 2. 0.33 pmol/mg protein. 3. 0.86 pmol/mg protein. The concentration of protein in the cytosol is 10-12 mg/ml. concentrations (higher than 5.108 M) the transcortin-bound and free form of the hormones take on the ability to be accumulated by the nuclear fraction in a cell-free system (9). However, in this case, the distribution of the hor- mone accumulated by the nuclei differs greatly (table 2). If the nuclei were incubated with 3H-glucocorticoid-receptor complex, almost all of the hormone included in the nuclei is extracted by two treatments: Triton X-100 (to extract the external nuclear membrane) and 0.4 M KCl (to ex- tract the dissolved chromatin fraction). The intranuclear fraction contained both the free and the bound form of the hormone. The bound form included a fraction corresponding approximately in size to the high salt form of the steroid-receptor complex (see table 1). The nuclei, incubated with transcortin- bound and free forms of 3H-corticosterone, contained a rather large quantity of unextractable radioactive material. Most of the extractable 3H-corticosterone was accounted for by the fraction of exterior nuclear membrane. These experi- ments confirm that the glucocorticoids can penetrate into the nuclei in a com- plex with the receptor. 248 6%¢ TABLE 2. Distribution of 3H-Glucocorticoid in Fractions of Nuclei (Percent From Incorporated Substance) Intranuclear Fraction (0.4 M KC1) Form and Concentration Outer Nuclear of Hormone Incubated Membrane Fraction Nonextractable With Nuclei (Triton X-100) Free Bound Fraction Cytosol-bound 3H-dexamethasone concentration ~ 2.7+10-9M 50 20 22 8 Cytosol-bound 3H-corticosterone concentration ~ 1.5+10-8M 60 14 12 14 Transcortin-bound 3H-corticosterone concentration ~ 10-8M concentration ~ 10-7M Free 3H-corticosterone concentration ~ 10~8M concentration ~ 10~/M Free JH-dexamethasone concentration ~ 10~8M concentration ~ 10~/M not found 50 not found 35 not found 2 not found 8 not found 10 not found not found 2 not found 1 not found not found 40 not found 54 not found Procedure for extraction of 3H-hormone incubated with nuclei with 10 percent Triton X-100 and 0.4 M KCl is described in (9). The quantity of 3H-hormone in nonextractable fraction is calculated by the difference between values of radioactivity of the nuclei before extraction and total radioactivity in Triton X-100 and KCl extracts. The Triton extracts of nuclei incubated with free SH-dexamethasone con- tained practically none of the radioactive label. It is possible that the outer surface of the nuclei has a certain quantity of binding factor III (and also transcortin) which facilitates accumulation of corticosterone, but not dexamethasone, and that this fraction is extractable with Triton. Obviously, the receptor is also capable, under certain conditions, of facilitating the accumulation of the hormone on the outer surface of the nuclear membranes. If we extrapolate these results with nuclear membranes to membranes in general, we can assume that transcortin in the intercellular space facilitates accumula- tion of naturally occurring glucocorticoids on the outside of the cytoplasmic membranes, while the intracellular glucocorticoid-binding proteins facilitate accumulation on the membranes of the subcellular structures. Additional experi- ments are required to confirm this assumption. In addition to the process suggested above, there is accumulation of the free form of the steroid and it is difficult to extract nuclear material bound with it within the nucleus, if the concentration of hormone in the cell exceeds 510-8 M. It is appropriate to ask whether this last mechanism is the cause of the toxic effect of glucocorticoids on the cell through nonspecific induction of transcription and subsequent synthesis of excess quantities of the enzymes. A change in activity of a number of enzymes has been noted in the heart after long-term administration of glucocorticoids (10). This mechanism doubtless differs from the specific activation of the genetic apparatus of the cell, which is caused by the interaction of steroid-receptor complexes with intranuclear acceptors. It also apparently does not relate to the rapid protective effect of high doses of glucocorticoids on the cell, which is related to direct stabi- lization of cell and lysosome membranes by hormones (11,12). On the left side of figure 5, we have attempted to show the mechanism of transfer of glucocorticoids into the cell nuclei in the heart. The intercellu- lar pool of transcortin limits the entry of naturally occurring glucocorticoids into the cells. The glucocorticoid-transcortin complexes partially accumulate on the outer surface of the cytoplasmic membranes. The dotted line shows the hypothetical path of glucocorticoid-transcortin complexes, if the transcortin can penetrate into the cells. The free form of hormones within the cells is distributed between binding factor II (receptor) and other binding proteins. The activated steroid-receptor complex relays the glucocorticoid signal to the intracellular acceptor or acceptors, where specific activation of the cell's genetic apparatus should occur. A portion of the steroid-protein complexes may be accumulated on the external membranes of the subcellular structures. At pharmacological concentrations, the free form of the hormones can also accumu- late within the nuclei, binding with as yet unidentified sections. At present, we know little of the processes illustrated on the right side of figure 5. We do not know whether the mechanism of transport of the steroid- receptor complex from the cytoplasm to the nuclei, demonstrated in vitro using preparations of cytosol and cell nuclei from the heart, reflects the regulation of activity of the genetic apparatus of heart tissue cells by glucocorticoids. Which proteins are synthesized de novo as a direct result of changes in tran- scription by glucocorticoids? How do these proteins (particularly enzymes) in- fluence cell metabolism? Judging from our preliminary studies, as well as from data in the literature (13,14), it seems improbable that changes in the 250 " Nucleus ’ oo [FE Glycogen, G+ Bindll GBindIl | _ Glucose / metabolism ? _lectrolytes activation metabolism AS‘ R= (RS) : atadholmines + N pre-active active induced degradation complex complex protein J 0 ld Se other effects ~ " S 8 o CELL FIGURE 5. Mechanism of transfer of glucocorticoid signal. Tr = Transcortin. G = Glucocorticoid. R = Cytoplasmic receptor. Bind. III = Binding factor III. concentration of these hormones could directly cause dramatic changes in overall protein synthesis. The effect of these hormones obviously involves changes in the synthesis of certain enzymes which perform important specialized functions, but the quantity of the enzymes in relationship to the total mass of protein synthesized is very small and can be detected only by special methods. In the liver, a classic target organ for glucocorticoids, one of the best studied functions of these hormones is activation of processes related to gluco- neogenesis, for example, the induction of transaminase synthesis, particularly tyrosine aminotransferase (TAT) (15,16). When glucocorticoids are administered repeatedly, maximum increase in TAT activity is observed with a characteristic lag period of 4 to 6 hours. The increase in activity results from synthesis of specific messenger RNA for TAT and subsequent synthesis of this enzyme as a re- sult of translation of induced mRNA. According to our data (figure 6), under analogous conditions, the changes in TAT activity in the heart are significantly less than in the liver, at least in mature animals. Induction of transaminase synthesis by glucocorticoids in the heart has no essential significance in the regulation of cardiac metabolism. In the heart, the appearance of permissive glucocorticoid effects is typical. The essence of permissive effects is that the glucocorticoids can, in some way which has not been established, change the sensitivity of the cells to other hormones and physiologically active substances. For cardiologists, the capability of glucocorticoids to potentiate the effects of catecholamines on the heart is of great interest. The potentiating effect of glucocorticoids is of tremendous physiological significance, since the effects of catecholamines occur in the presence of certain concentrations of glucocorticoids in the tis- sue, and both types of hormonal agents are necessary components of stress. 251 0.5} activity of TAT, umoles/min/mg prot. o rar heart liver FIGURE 6. Effect of hydrocortisone on tyrosine aminotransferase (TAT) activity in rat heart and liver. The rats used were Wistar males weighing 220-250 g. Each received 5 mg of hydrocortisone per 100 g body weight, intraperitoneally, 6 hours before sacrificing. The control animals received saline solution. TAT activity was determined in supernatant fluid after centrifugation of tissue homogenates at 3,500 g for 20 minutes according to the method of Levin (17). Each group of animals included 12-15 rats. (= Experimental animals. = Control animals. Fleckenstein (18) showed that preliminary, long-term administration of glucocorticoids to rats causes a sharp increase in the capability of the myocar- dium to increase the accumulation of calcium ions in response to the administra- tion of high doses of a synthetic analog of the catecholamines--isoproterenol. In our laboratory, we found that heart tissue of adrenalectomized rats completely loses the ability to increase the uptake of 45ca during administration of high doses of epinephrine. Kinetics of the change in sensitivity to epinephrine after adrenalectomy indicate that the effects of the hormones could hardly be a result only of their direct interaction with the cell membranes (figure 7). Although a sharp decrease in the concentration of glucocorticoids in the blood is observed 3 hours after removal of the adrenal glands, a stable lack of sensitivity of the myocardium to epinephrine is seen only on the third to fifth day after operation. Admin- istration of glucocorticoids can slow the loss of sensitivity in adrenalecto- mized rats (not shown in the figure), but their effect does not appear for many hours. However, administration of glucocorticoids intravenously results in their appearance in heart tissue within a few minutes. Obviously, some slowly developing processes are ultimately responsible for the changes in sensitivity to catecholamines. It can be assumed that these changes are localized in the cell membrane, but are dependent on intracellular processes including de novo synthesis of certain proteins that are capable of directly or indirectly influencing the function of the membranes. Based on current understanding, phosphorylation and dephosphorylation (20-22) are highly 252 se . 5 s 2 3 £ 8 3 £ g o B $ 2 2 3 8 8 t 3 r _— 3 5 3128 ¢ 06 18 30 48 72 120 hours after adrenalectomy FIGURE 7. Changes in accumulation of 43Ca in the myocardium (stimulated by epinephrine) and corticosterone concentration in blood serum after adrenalectomy. Six hours before sacrificing, the rats simultaneously received 10 uC of 43ca per 100 g of body weight intraperitoneally and 0.4 mg of epinephrine bitartrate per 100 g of body weight subcutaneously. Accumulation of 45Ca in heart and serum was determined after solubilization of specimens in formic acid. The ratio of accumulation of 45Ca in 1 g of tissue to accumulation in 1 ml of serum of intact rats which did not receive epinephrine was taken as 100 percent. The levels of corticosterone in serum were determined as in (19). Each group of animals in- cluded 5 to 14 rats. significant in the regulation of the passage of ions through biological membranes. It is possible, therefore, that glucocorticoids regulate the synthesis of en- zymes which take part in the phosphorylation of membrane proteins. However, this mechanism is still a hypothesis. There is a great deal of indirect data in the literature which indicate that the cell membrane is the place where the effects of glucocorticoids are manifested. When there are additional loads on the organism, the effect of glucocorticoids on Nat, K*-ATPase in the plasma membranes of the myocardium appears (23). In the urinary bladder of the frog, mineralocorticoids apparently participate in the synthesis of protein which is a component of the Nat, xt pump of the cell membranes (24). Nat, Kt-ATPase, as is known, also regulates the exchange of calcium ions. The point of application of the effect of gluco- corticoids may be the activity of phospholipase A), which releases arachidonic acid from the membrane phospholipids. Arachidonic acid is a substrate which limits prostaglandin synthesis. Prostaglandins, in turn, are effective regula- tors of calcium ion accumulation in the cells. The well-known anti-inflammatory effect of glucocorticoids results, it is thought, from the inhibition of prosta- glandin synthesis through regulation of the activity of phospholipase Ap (25- 29). Interestingly, synthetic and naturally occurring glucocorticoids can be 253 placed in the same sequence in terms of their capability to cause anti- inflammatory effects and their affinity for glucocorticoid receptors, including the glucocorticoid receptor of the heart. These facts indicate that cell membranes are probably one of the most im- portant sites in the cell for changes caused by glucocorticoids. At least some of the membrane effects of these hormones should result from intracellular pro- cesses regulated by the glucocorticoid receptor. Briefly discussed below is the interrelationship of processes mediated by changes in the level of cyclic AMP and the effects of glucocorticoids. Some investigators believe (30) that steroid hormones regulate the synthesis of spe- cific proteins participating in phosphorylation processes activated by cAMP in the corresponding target organs. These specific proteins have not yet been identified. In the heart, no significant changes in the level of the cyclic nucleotide have been detected under the influence of glucocorticoids. However, glucocorti- coids potentiate effects mediated by cAMP, such as activation of glycogen phos- phorylase by epinephrine. Miller et al. (31) believe that this is not related to a change in the synthesis of cAMP-dependent protein kinase, phosphorylase kinase, or phosphorylase b, but rather to the effect of intracellular accumula- tion of calcium ions which are necessary for activation of phosphorylase kinase. Our studies indicate that stable disruptions in the activation of phosphory- lase b in the heart by inephrine occurs between the third and fifth day fol- lowing removal of the adrenal glands, i.e., within the same time frame when heart tissue loses its capability to increase accumulation of calcium ions in response to epinephrine (figures 7 and 8). Thus, the changes in activation of phosphor- ylase may, in principle, follow changes in calcium ion accumulation. Interest- ingly, during the first day after adrenalectomy, the basal level of phosphorylase a in the heart is significantly increased and administration of exogenous epi- nephrine, with perfusion of an isolated heart, does not cause a further increase in enzyme activity. The elevated level of phosphorylase a during this time period results from trauma, and not from the decrease in the concentration of corticosteroids since phosphorylase a is also elevated in animals which undergo surgery without adrenalectomy (not shown in figure). By the second or third day after the sham operation or true adrenalectomy, the basal level of phosphorylase a drops to normal. Beginning at this time, one observes a stable loss of sensi- tivity to epinephrine, resulting from a deficiency of corticosteroids. Interesting data have been published on the possible relationship between the cytoplasmic receptor for glucocorticoids and cAMP-dependent processes. Cake and Litwack (33) first demonstrated the capability of methylxanthine (aminophyl- lin, theophyllin, caffeine), a phosphodiesterase inhibitor, to activate the inactive form of the glucocorticoid receptor of the liver. The stronger the inhibitory effect of methylxanthine, the more strongly activated was the cyto- plasmic receptor. Other phosphodiesterase inhibitors, nonmethylxanthines, were not capable of activating the receptor. We confirmed the capability of methyl- xanthines to activate the glucocorticoid receptor in the heart (see figure 3). Surprising data were reported by Litwack et al. (34) on the similarity between the cytoplasmic receptor for glucocorticoids in the liver and the regulatory 254 intact. + epinephrine 200 — 100 — intact. phosphorylase a, % 1 A 2 24 48 72 120 hours after adrenalectomy FIGURE 8. Effect of epinephrine on phosphorylase a activity of isolated rat heart at various times after adrenalectomy. Glycogen-phosphorylase activity was determined by the method of Stalmans and Hers (32) in supernatant fluid after centrifugation of rat heart homogenate at 3,500 g. The heart, isolated by the method of Langendorf, was preliminarily perfused with Krebs-Henseleit solution, with or without epinephrine (0.5¢10=7 M). Each group of animals included 5 to 15 rats. The ratio of phosphorylase a activity to total phosphorylase activity in intact rats was taken as 100 percent. @ plus epinephrine; O minus epinephrine. e subunit of cAMP-dependent protein kinase. However, we still do not know whether the cytoplasmic receptor for glucocorticoids is the regulatory subunit of protein kinase. In conclusion, we emphasize that the presence of a highly specific mecha- nism for transfer of the glucocorticoid signal in the heart indicates the pos- sibility of regulating metabolic processes in this organ by glucocorticoids through transcription and de novo synthesis of enzyme proteins. Regulation of the transfer of naturally occurring glucocorticoids to the cell nuclei involves both extracellular transcortin and intracellular proteins--the receptor and binding factor III. The latter apparently acts as an intracellular limitation for the access of hormones to the receptor. An important site of action for glucocorticoids on the cell should be the cellular (and possibly intracellular) membranes. This does not exclude the possibility of direct action of the gluco- corticoids on the membranes, rather than through transcription and protein syn- thesis. The capability of binding proteins to facilitate the accumulation of glucocorticoids on the surface of the membranes indicates that the hormone effects which result from direct interaction with the membranes probably depend on the affinity of the steroid molecules for the binding proteins. At pharma- cological concentrations of the hormones, both direct interaction of glucocorti- coids with the membranes and changes related to the primary synthesis of enzyme protein may occur. Accumulation on the surfaces of membranes is believed by a number of re- searchers to be the cause of the protective-stabilizing effect of glucocorti- coids on the cells. This effect is manifested very rapidly. On the other 255 hand, effects which are dependent on specific or nonspecific increases in transcription and enzyme synthesis characteristically have a rather long lag period. This factor indicates that the protective and toxic effect of high doses of glucocorticoids should be separated in time. 256 10. 11. 12. REFERENCES Smirnov VN, Seleznev YuM, Danilov SM, Preobrazhensky SN, Volkova NG, Kolpakova GV, Kuznetsova LA: Hormones and the heart: glucocorticoid receptors. In Proceedings of the Third Joint Symposium on Myocardial Metabolism, Williamsburg, Virginia, May 9-11, 1977. Washington, D.C. U.S. Department of Health, Education, and Welfare, Public Health Service, National Institutes of Health, DHEW Publication No. (NIH) 78-1457, pp 1-23 Seleznev YuM, Danilov SM, Preobrazhensky SN, Volkova NG, Kusnetsova LA, Smirnov VN, Volchek AG: Glucocorticoid-binding proteins of rat heart cytosol and their possible role in the transfer of glucocorticoids into nuclei. J Mol Cell Cardiol 10(10):877-891, 1978 Seleznev YuM, Danilov SM, Smirnov VN: Separation of three glucocorticoid- binding fractions from cytosol of rat heart. J Steroid Biochem 10(2):215- 220, 1979 Seleznev YuM, Smirnov VN: Glucocorticoid-binding proteins and cytoreceptor in the heart (abstr). 12th FEBS Meeting, Dresden, German Democratic Republic. Abstract 1355, 1978 Laurent TC, Killander J: A theory of gel-filtration and its experimental verification. J Chromatog 14:317-330, 1964 Martin RG, Ames BN: A method for determining the sedimentation behavior of enzymes. J Biol Chem 236:1372-1379, 1961 Seleznev YuM, Danilov SM, Volkova NG, Medvedeva LA: [Activation of gluco- corticoid cytoreceptors in rat heart] (Rus). Biokhimiia 44:245-251, 1979 Siegel LM, Monty KI: Determination of molecular weights and frictional ratios of proteins in impure systems by use of gel-filtration and density gradient centrifugation. Application to crude preparations of sulfite and hydroxylamine reductases. Biochim Biophys Acta 112:346-362, 1966 Seleznev YuM, Volkova NG, Medvedeva LA, Danilov SM, Smirnov VN: Uptake and distribution of glucocorticoids in cell nuclei of rat heart in cell-free system. Dependence on concentrations and forms of added hormones. J Mol Cell Cardiol 11:289-302, 1979 Sviryakin VT, Kosmach PI: [The effect of insulin on the reactivity of the myocardium to hydrocortisone] (Rus). Probl Endokrinol 22(4):58-60, 1976 Libby P, Maroko PR, Bloor CM, Sobel BE, Braunwald E: Reduction of experi- mental myocardial infarct size by corticosteroid administration. J Clin Invest 52(3):599-607, 1973 Okuda M, Lefer AM: Myocardial uptake and metabolism of [3H] methylprednis- olone in isolated cat hearts. J Mol Cell Cardiol 9:989-1001, 1977 257 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Bullock GR, Carter EE, Elliott P, Peters RF, Simpson Sh, White AM: Rela- tive changes in the function of muscle ribosomes and mitochondria during the early phase of steroid-induced catabolism. Biochem J 127:881-892, 1972 Rannels SR, Rannels DE, Pegg AE, Jefferson LS: Differential effects of glucocorticoids on initiation of protein synthesis in skeletal muscle and heart. Fed Proc 36(3):585, 1977 Baxter JD: Glucocorticoid hormone action. Pharmacol Ther [B] 2(3):605- 659, 1976 Gelehrter TD: Enzyme induction (third of three parts). N Engl J Med 294(12) :646-651, 1976 Levin FB: [A simple spectrophotometric method for determination of tyro- sine aminotransferase activity] (Rus). Vopr Med Khim 15(3):315-317, 1969 Fleckenstein A: Metabolic factors in the development of myocardial necrosis and microinfarcts. Triangle 14(1):27-36, 1975 Volchek AG: [Determination of microquantities of corticosterone in plasma of rats by the saturation analysis method] (Rus). Biol Nauki 16(10):124-129, 1973 Greengard P: Possible role for cyclic nucleotides and phosphorylated membrane proteins in postsynaptic actions of neurotransmitters. Nature 260:101-108, 1976 Kirchberger MA, Raffo A: Phosphoprotein phosphatase-catalyzed dephosphor- ylation of the 22,000-dalton phosphoprotein of cardiac sarcoplasmic reticu- lum. In Recent Advances in Studies on Cardiac Structure and Metabolism. Heart Function and Metabolism, vol 11, edited by T Kobayashi, T Sano, and N Dhalla. Baltimore, University Park Press, 1978, pp 285-291 Sperelakis N, Schneider JA: A metabolic control mechanism for calcium ion influx that may protect the ventricular myocardial cell. Am J Cardiol 37:1079-1085, 1976 K6rge PK: [Myocardial sodium-potassium pump and its adrenocortical regula- tion as factors limiting cardiac adaptation to severe physical exertion] (Rus). Kardiologiia 16(9):15-21, 1976 Edelman IS: Mechanism of action of steroid hormones. J Steroid Biochem 6:147-159, 1975 Hong SL, Levine L: Inhibition of arachidonic acid release from cells as the biochemical action of anti-inflammatory corticosteroids. Proc Natl Acad Sci USA 73(5):1730-1734, 1976 Grodzinska L, Panczenko B, Gryglewski R: Inhibition of release of prosta- glandin E-like material by non-steroid and steroid anti-inflammatory drugs. Acta Biol Med Germ 35:1099-1100, 1976 258 27. 28. 29. 30. 31. 32. 33. 34. Gryglewski RJ: Prostaglandins and the mechanism of drug action in circu- latory system. Acta Biol Med Germ 35(8-9):1097-1098, 1976 Gryglewski RJ, Panczenko B, Korbut R, Grodzinska L, Ocetkiewicz A: Cortico- steroids inhibit prostaglandin release from perfused mesenteric blood vessels of rabbit and from perfused lungs of sensitized guinea pig. Prostaglandins 10(2) :343-355, 1975 Foster SJ, Perkins JP: Glucocorticoids increase the responsiveness of cells in culture to prostaglandin Ej. Proc Natl Acad Sci USA 74(11):4816- 4820, 1977 Liu AY-C, Greengard P: Regulation by steroid hormones of phosphorylation of specific protein common to several target organs. Proc Natl Acad Sci USA 2:568-572, 1976 Miller TB, Exton JH, Park CR: A block in epinephrine-induced glycogenolysis in hearts from adrenalectomized rats. J Biol Chem 246:3672-3678, 1971 Stalmans W, Hers HG: The stimulation of liver phosphorylase b by AMP, fluoride and sulfate. Eur J Biochem 54:341-350, 1975 Cake MH, Litwack G: Effects of theophilline on the activation and nuclear translocation of the hepatic glucocorticoid receptor at low temperature. Biochem Biophys Res Comm 66:828-835, 1975 Litwak G, Filler R, Cake MH: [Glucocorticoid receptor in the liver and its possible interrelation with cAMP] (Rus). Vestn Akad Med Nauk SSSR 3: 42-46, 1977 259 STUDY OF THE INTERACTION OF CARDIOSTEROIDS WITH Na', K'- ADENOSINE TRIPHOSPHATASE BY EPR* R. I. Zhdanov, N. M. Mirsalikhova, and Yu. Sh. Moshkovsky SUMMARY 19-(1-0xy-2,2,5-5~-tetramethyl-3-iminomethylpyrrolidine) derivative of strophantidine (SL-Str) was synthesized as a spin probe for studying the inter- action of cardiac glycosides with Na't, Kt-adenosine triphosphatase (ATPase). It was found that SL-Str has the same inhibitory capacity as strophantidine (Kinhib 5.80°10-7 and 6.25-10~7, respectively). The EPR spectra of SL-Str, in combination with ATPase in the presence of various concentrations of Na™t, Kt, and Mgt, show no significant limitations of the rotational mobility of the radical, which may indicate that the cardiosteroid interacts with the enzyme with the reverse side of the cyclopentaphenanthrene skeleton. The values of the CIC constant ay = 16 6e of the complex agrees with polar circling for the nitroxyl fragment, close to the polarity of water. The parameters of binding of SL-Str with ATPase (Kpound = 2.0 * 0.5105) are sensitive to the concentra- tion of Kt or strophantidine ions. Binding of SL-Str deteriorates with an in- crease in the concentration of Kt and strophantidine ions. Finding an increase in the affinity of SL-Str for model lecithin membranes in the presence of cho- lesterol, it is assumed that cardiac glycosides probably interact with Nat, k*+- ATPase by the formation of complexes with cholesterol. INTRODUCTION Based on current understanding, the positive inotropic effect of cardiac glycosides is related to the inhibition of Nat and Kt ion transport through the cell membrane and the parallel inhibition of Nat, Kt-ATPase (1). It has been suggested that inhibition of Nat, Kt-ATPases from cardiac muscle is the initial mechanism or receptor of the positive inotropic effect of cardiosteroids, the so-called digitalis receptor (2). Only the radioisotope method ([3H]- ouabain) has been used to study this process (1). A report has also been pub- lished on the photoaffinity derivatives of ouabain (3). *EPR = Electron paramagnetic resonance From the Institute for Biological Testing of Chemical Compounds, Moscow, USSR, and from the Institute of Biochemistry, Academy of Sciences of the Uzbek SSR, Tashkent, USSR. 261 Although studies using spin-labeled (4) and hydrophobic spin probes of the conformation and structure of membrane preparations of Nat, Kt-ATPase (4,5) are known, use of spin-labeled inhibitors, particularly cardiosteroids, to study this enzyme is not found in the literature. In general, study of Na¥t, Kt-ATPase preparations by the spin-label method is very difficult due to the presence of other proteins, while the use of spin-labeled substrate derivatives, e.g., spin- labeled ATP, requires highly concentrated preparations of ATPase. We suggest the use of highly active spin probe, spin-labeled derivative of the Na¥t, Kt- ATPase inhibitor, cardiosteroid strophantidine, for studying the sites of its interaction with Nat, Kt-ATPase. Since the steroid framework with the lactone ring should address the nitroxyl radical directly to the site of interaction of the cardiosteroids with ATPase, we can hope to obtain information on the param- eters and mechanism of this process and, possibly, the mode of action of cardiac glycosides. MATERIALS AND METHODS For chemical synthesis, nitroxyl I was obtained as described by Hsia and Piette (6). Synthesis of the spin-labeled strophantidine derivative (II) is shown in figure 1. For production of spin-labeled strophantidine, a solution of strophantidine (0.5 mmol) and nitroxyl radical with a primary amino group (0.75 mmol) in 4 ml of a mixture of DMPA and benzene (1:1) was boiled for 14 hours with reflux cooling and a Dean-Stark trap. The solvents were then evap- orated in a vacuum, and the spin-labeled strophantidine derivative was separated from the oily residue after chromatography on silica gel (acetone eluant) (Legal test positive). 19-(1-0xyl-2,2,5,5-tetramethyl-3-iminomethylpyrrolidine)-strophantidine (II) was produced with the following characteristics: dark yellow crystals, yield 28 percent, m.p. 120-125° C, Rf = 0.65 (acetone-benzene, 3:1). IR spec- trum (KBr), cm~l: 3350-3280. vy 0-H: 2865, 2815 (m) y > CHp, > CH; 1770, 1735 y C=0 (lactone); 1665 y C=N (imine); 1620 y C=C (lactone). EPR spectrum (ethanol-water): ay = 15.5 e; 5.5 * 0.31023 spin/mol. Found: C 64.63, H 8.90, N 3.61; C32H4gN)0g2H20 m/e 557. Calculated: C 64.73, H 9.0, N 4.72; m.w. 593.8. Three different enzyme preparations of Nat, KT-ATPase were used in the study. These were a microsomal ox brain fraction containing Nat, KT-ATPase obtained by the method of Skou et al. (7), a preparation of ATPase from pig 0 0 I CH + OH N Ie 0 HO OH FIGURE 1. Synthesis of spin-labeled strophantidine derivative. 262 cardiac muscle, extracted as described by Akera et al. (8), and a purified preparation of Na, K-ATPase from a microsomal fraction of the medullary layer of pig kidney, obtained by centrifuging in a zonal rotor according to Jorgensen (9). This last preparation contained 90 percent protein Nat, K'-ATPase (7.9 mg protein per ml). All ATPase specimens were free of nonspecific Mgtt-ATPase, stored, and refrigerated in 25 mM Tris-HCl buffer with 1 mM EDTA, pH 7.5. The ATPase activity was determined by the increase in inorganic phosphate per mg of protein per hour. A solution of the ATPase preparation was diluted with an equal volume of standard buffer solution containing 30 mM Tris-HCl, 130 mM NaCl, 20 mM KCl, 4 mM MgCl,, and 1 mM EDTA, pH37 = 7.5. A standard buffer (440 pl) was added to the aliquots consisting of 10 pl of the enzyme solution, while the same quantity of a standard buffer containing strophantidine or various concentrations of the spin probe were added to other 10 pl aliquots. The samples were placed in a constant temperature bath at 37° C, and the reaction was begun by adding 50 pl of a 10 mM solution of ATP, pH 7.5. After 10 minutes of incubation at 37° C, the reaction was stopped by adding 0.7 ml of a 5 percent solution of HClO, and reagents were added to the samples (1.2 ml) to determine inorganic phosphate by the 'cold" modified method of Del'sal (10). The level of Nat, Kt-ATPase activity was calculated from the difference in en- zyme activities as determined in the incubation medium with the full complement of ions (Mgtt, Nat, Kt) and upon addition of 0.1 M strophantidine to this medium. The initial specific activity, e.g., for the preparation from the medullary layer of the pig kidney, was 600 umol P;/mg protein per hour. Protein content was determined by the method of Lowry, using serum albumin as a standard for the calibration curve. Liposomes were prepared from egg leci- thin by sonic irradiation (operating frequency 22 kHz) of an aqueous dispersion containing 20 mg/ml lecithin (0.3 percent oxidation). Liposomes with cholesterol contained 5 mg/ml egg lecithin and 2.5 mg/ml cholesterol. Measurement of activity was conducted using the EPR method. The EPR spectra were recorded on a Varian E-9 radiospectrometer: At 9.4 GHz, P = 5 mW, Happ = 1; Tconst = 0.3 sec; T = 8 min at t = 37° C. Concentrations of the spin probe of 10-6 to 6.5+10-5 were created by successive addition of aliquots of 2 ul to 100 ul ATPase in a buffer solution containing 20 mM Tris-HCl, 130 mM NaCl, 5 M KCl, 4 mM MgClp, 1 M EDTA, and 1 mM ATP (the concentration of ATPase was 10-5 M/liter). The concentration of spin probe II not bound with the enzyme was determined from the calibration curve of the variation of intensity of 147 of the compo- nents of the EPR spin probe spectrum with its concentration in buffer solution without ATPase. The concentration of spin probe II bound with the enzyme was calculated from the difference of the total concentration of II and the concen- tration of the probe bound to the enzyme. The correlation time was calculated by the formula: T, = 6.65 AH 1 (VI1/1I_ - 1)-10~10 sec Cc 263 where AH) is the width of the low-field component of the EPR spectrum and AH_7 is the amplitude of the low-field and high-field components of the EPR spectrum (11). RESULTS AND DISCUSSION All of these preliminary experiments to test the biological activity of strophantidine and its spin-labeled derivatives were performed on microsomal preparations of Nat, Kt-ATPase from pig cardiac muscle, from ox cerebral cortex cells, and from the medullary layer of pig kidney (preparation contained 90 per- cent Nat, Kt-ATPase protein). Table 1 presents the concentrations of 50-percent inhibition (Isp) of these three preparations of Na't, Kt-ATPase by strophantidine. We found similar values for Nat, K'-ATPase of specimens from the different sources, and they had identical antigenic properties. Therefore, subsequent experiments were performed using the purified enzyme from the medullary layer of the kidney. Selection of this preparation for studies by the spin probe method was favored because experiments using EPR and spin probes are usually performed with enzyme preparations containing some tens of mg of protein per ml of solution. Table 2 shows that the parameters of inhibition of Nat, Kt-ATPase with strophantidine and its spin-labeled derivative are similar. The kinetic param- eters of interaction of strophantidine with Na™t, Kt-ATPase change little when a radical with the aminomethyl group (II) is added at Cjg9. This may indicate that the steroid framework interacts with the digitalis receptor with its reverse side "b". Therefore, introduction of the radical on a spacer arm TABLE 1. Inhibition of Various Preparations of Na, kK -ATPase by Strophantidine (Isq x 107 M/1) Nat, Kt-ATPase From Ox Brain | ATPase From Pig Kidney | ATPase From Pig Heart 6.50 6.25 6.50 TABLE 2. Inhibition of Nat, K'-ATPase by Strophantidine and Its Spin-labeled Derivative (II) 7 % Inhibition at Concentration Kinhip * 107M of 2.104 M/1 Strophantidine 6.25 94 Derivative (II) 5.80 98 264 at CjgHE causes immobilization of the radical and alteration of the inhibitor properties. The form of the lines on the EPR spectrum and the value of the constant of the superfine interaction of the spin probe in complex with the enzyme agree with this assumption. Figure 2 shows the EPR spectra of spin-labeled strophantidine (II) in a buffer solution (A) and in the presence of SL-Str and ATPase (B). These EPR spectra are practically identical, indicating that there is no significant limi- tation of the rotational mobility of the nitroxyl fragment of the radical (tc = 6-101 sec). 10 A/m —_— FIGURE 2. The EPR spectrum of spin-labeled strophantidine derivative in a buffer solution (Tris-HCl) (A) and in the presence of Nat, Kt-ATPase (10-5 M/ liter) (B). 265 The value of the CTC constant for the spin probe II, in complex with ATPase, ay, equals 16 e, indicating that the polarity of the surroundings of the nitroxyl group in (II) is similar to the polarity in water. The results of titration of ATPase with the spin probe (IVb) are consistent with the existence of one type of binding site for cardiosteroids with a binding constant Kpound = 2.0 * 0.5+105 (on the assumption that the molecular weight of Nat, Kt-ATPase is 300,000). Under the experimental conditions (aqueous solution, 37° C), the spin probe (II) is stable. In a control experiment, water was evaporated from an aqueous solution (10-3 M/liter) (IVb) after holding at 37° C for 6 hours. Thin-layer chromatography on silica gel (L 100-160, CSSR) showed one spot with Rg = 0.65 (acetone-benzene, 3:1), as in (II) before the experiment. The combination of the enzyme plus the spin probe (IVb) was sensitive to Kt ions and strophantidine (table 3). Addition of an excess of Kt ions caused a greater increase in the amplitude of the EPR signal of the unbound radical (i.e., a greater degree of dissociation of the enzyme-spin probe complex) than did the addition of an excess of strophantidine. Thus, nitroxyl (II) retained the properties of the strophantidine cardiosteroid and is, therefore, a highly specific spin probe for Nat, KT-ATPase. Since Na™, Kt-ATPase is a protein component of biological membranes, the question arises as to the nature of binding of cardiosteroids with this enzyme. Since there are three possibilities for binding with the protein, the lipid portion, and with both, we attempted to use experiments with spin-labeled strophantidine, specifically the interaction with model membranes, in order to select one of these possibilities. Lack of anisotropy of rotation and the value of the CTC constant indicated that cardiosteroids do not bind with the lipid portion. Upon titration of lecithin vesicles with the spin probe (II) (figure 3), the correlation time in- creased slightly and did not depend critically on concentration, also indicating a weak interaction of SL-Str with phospholipids. TABLE 3. Effect of Kh Ions and Strophantidine on the Intensity of the EPR Spectrum (Relative Units) of the Complex of Nitroxyl (II) and ATPase Enzyme + (II) (2.1075 M/1) - ea kt (40 mM) + Strophantidine Intensity 13 28 - EPR spectrum, mm (I_y) 11 - 16 266 1 0 [st-str] 0 M FIGURE 3. Variation of time of rotational correlation of spin-labeled strophan- tidine derivative as a function of its concentration in buffer solution (----), in the presence of lecithin liposomes (-A-), and in the presence of cholesterol- lecithin liposomes (-0-). However, upon titration of the spin probe with mixed lecithin-cholesterol vesicles (see experimental portion), we observed an interesting effect: As the concentration of SL-Str increased, the correlation time 1, of the spin-labeled cardiosteroid also increased, i.e., in the presence of cholesterol, the affinity for the model membranes increases. At low concentrations, the affinity for these liposomes is probably low. Thus, the inhibitory capacity of strophantidine results from its interaction with the active centers of the protein, and not with the phospholipids. And, cholesterol apparently increases this interaction. 267 10. 11. REFERENCES Wallick ET, Lindenmeyer GE, Lane LK, Allen JC, Pitts JR, Schwartz A: Recent advances ‘in cardiac glycoside - Nat, Kt-ATPase interaction. Fed Proc 36:2214-2218, 1977 Repke K, Portius HJ: [On the identity of the ion-pump ATPase in the cell membrane of the myocardium with a digitalis-receptor enzyme] (Ger). Experientia 19(7):452-458, 1963 Forbush B III, Kaplan JH, Hoffman JF: Characterization of a new photo- affinity derivative of ouabain: labeling of the large polypeptide and a proteolipid component of Na, K-ATPase. Biochemistry 17:3667-3676, 1978 Boldyrev A, Remge E, Smirnova I, Tabak: Na, K-ATPase: The role of state of lipids and Mg ions in activity regulation. FEBS Lett 80:303-307, 1977 Raikhman LM, Moshkovskii YuSh: [Ion-dependent conformational transitions in membrane preparations of transport ATPase] (Rus). Mol Biol 8(5):612- 617, 1974 Hsia JC, Piette LM: Spin-labeling as a general method in studying antibody active site. Arch Biochem Biophys 129:296-307, 1969 Skou JC: Preparation from mammalian brain and kidney of the enzyme system involved in active transport of Nat and K*. Biochim Biophys Acta 58:314- 325, 1962 Akera T, Larsen FS, Brody TM: The effect of ouabain on sodium- and potassium-activated adenosine triphosphatase from the hearts of several mammalian species. J Pharmacol Exp Ther 170:17-26, 1969 Jorgensen PL: Purification and characterization of (Nat + Kt) -ATPase III. Purification from the outer medulla of mammalian kidney after selec- tive removal of membrane components by sodium dodecylsulphate. Biochim Biophys Acta 356:36-52, 1974 Panusz HT, Graczyk G, Wilmanska D, Skarzynski J: Analysis of orthophosphate- pyrophosphate mixtures resulting from weak pyrophosphatase activities. Anal Biochem 35:494-504, 1970 Buchachenko AL, Vasserman AM: [Signs of constants of ultrafine interaction and spin densities in azotoxy radicals] (Rus). Zh Strukt Khim 8:27-32, 1967 268 SOME FEATURES OF THE INHIBITION OF Nat, K'-ATPase IN HEART MUSCLE BY CARDIOTONIC GLYCOSIDES N. M. Mirsalikhova, N. Sh. Palyants, and N. K. Abubakirov Current understanding of the mechanism of action of cardioactive glycosides establishes beyond doubt that their effect on heart muscle involves interaction with Nat, Kt-ATPase. This interaction does not occur with the entire enzyme, the molecular weight of which, depending on the methods of determination, has been calculated to be 250,000-1,000,000 daltons (1), but rather with only the active site or digitalis receptor. The reaction between cardiac glycosides and the receptor involves the entire molecule of the cardiosteroid. However, before the discovery of the biochemical basis of the effect of cardiotonic glycosides, it was learned that a key role in the positive inotropic effect is played by an unsaturated, five-member lactone ring in the cardenolides and a six-member ring in the bufadienolides. The un- saturated lactones, oriented in a definite way with respect to the plane of the steroid nucleus, give the molecule the specificity of action for heart muscle not present in other steroid compounds. After discovery of the parallelism existing between Nat, Kt-ATPase inhibition and the positive inotropic effect, the primary attention of researchers was therefore directed toward determining the role of the lactone group. Based on the assumption that the interaction of cardiosteroids with the en- zyme occurs through the lactone ring, significant work was done on restructuring and modification of this essential element in the structure of the cardiac agly- cons. A steroid framework and hydroxyl group at C-3 has been assumed to be nec- essary in the remaining portion of the molecule. The significance of the sugar component in the mechanism of action of car- diosteroids is not clear. The fact that the activity of cardiac glycosides decreases or increases as a function of the structure and length of the monosac- charide links seems natural. However, it has recently been discovered that sugar not only changes the cardiotonic effect inherent in aglycons, but also, in some cases, may eliminate the effect, even if all of the structural elements of the steroid portion are retained. For example, a compound of digitoxigenin and strophanthidin with glucose differs little in biological activity from From the Institute of Biochemistry and from the Institute of Chemistry of Plant Substances, Academy of Sciences of the Uzbek SSR, Tashkent, USSR. 269 ouabain and other natural glycosides. If glucuronic acid is introduced into the molecule instead of glucose, this glycoside has approximately one-tenth of the inotropic effect of ouabain (2). The replacement of alcohol function with a carboxyl function in the sugar portion leads to a large reduction in activity. We synthesized cardenolides mono- and bisglycosides with additional mono- saccharide groups in the C-5 and C-19 positions (3,4) (figure 1). The new compounds were tested for their inhibition of Nat, Kt-ATPase under approximately the same conditions as had been used earlier (5). To confirm the parallelism between inhibition and the cardiotonic effect in synthesized glycosides, we simultaneously determined their lethal doses for cats. To determine the con- centration of glycosides causing 50-percent inhibition, we used the microsomal fraction of pig heart muscle. Our findings are detailed in table 1. When the relationship between structure and action is compared, the most surprising finding is the lack of cardiotonic activity in 38, 19-0-bis a-L- rhamnopyranosyl-strophanthidol (III). The results of our experiments with this glycoside agree with the observation made earlier (6) that 19-0-a-L-rhamnopyrano- syl-strophanthidol (II) has no cardiotonic effect. Its isomer 3-o-L-rhamno- pyranosyl-strophanthidol (convallotoxol) has rather high biological activity (0.099 mg/kg of body weight of the cat). What is the reason for the loss of activity of strophanthidol connected with L-rhamnose through the hydroxyl at C-19? The most probable causes are two: steric hindrance and a change in hydrophobicity. Steric hindrance. Biological activity of cardiosteroids is based on their very strict complementarity with the receptor portion of the enzyme, not Q — CH 0—CH im \ 0 2 2 H,C OH HO OF Hood on "0 Il. 19-Rh, I- HO Ou La epyanosy I. Convallotoxol (inactive) lll. 3,19-Bisrhamnopyranosyl- (strongly active) strophanthidol (inactive) Q : : He m HC OH OH OH 0 HO 0 HO, 0 Ho 6 OH 0 @) 0 H Gy Ho Hi Ho 0 MO OH 6 IV. Convallotoxin V. 6-Isoconvallotoxin HO OH (strongly active) (moderately active) VI. 3,5-Bisrhamnopyranosyl- strophanthidin (moderately active) FIGURE 1. Natural and synthetic strophanthidol and strophanthidin glycosides with rhamnose. 270 TABLE 1. Dependence of the Biological Activity on the Structures of Strophanthidin and Its Derivatives Concentration Causing 50-Percent Inhibition Lethal Dose for Aglycon and Glycoside Compounds (pmol/1) Cats (mg/kg) Strophanthidol 0.60 0.700 Strophanthidin 0.21 0.324 Strophanthidylic acid 53.0 2.649 3=Rhamnopyranosyl-strophanthidol (I) (convallotoxol) (6) Not determined 0.099 19-Rhamnopyranosyl-strophanthidol (11) (6) Inactive Inactive 3,19-Bisrhamnopyranosyl- strophanthidol (III) (4) Inactive Inactive 3=-Rhamnopyranosyl-strophanthidine (IV) (convallotoxin) 0.12 0.079 5-Rhamnopyranosyl-strophanthidin (V) (5-isoconvallotoxin) (3) 5.20 0.620 3,5-Bisrhamnopyranosyl- strophanthidin (VI) (3) 0.47 0.172 permitting the slightest deviation. The angular groups at C-10 and C-13 are of no little significance, projecting as they do beyond the plane of the cardio- steroid ring. All natural cardiosteroids have a methyl group at C-13. The chem- ical nature of the angular group at C-10 varies. It may be a methyl group (e.g., digitoxigenin, periplogenin), a primary alcohol (strophanthidol, ouabagenin), or an aldehyde group (strophanthidin, pachygenin). In nature, we find aglycons with a carboxyl group, but it has not been proven that they are native. Approximate calculation of the internuclear distances in the sequence in which the volume of oxygen-containing functions at C-10 increases, while the remaining elements of the molecular structure remain unchanged (periplogenin- strophanthidin-strophanthidol-strophanthidylic acid-19-a-L-rhamnopyranosyl- strophanthidol) shows that the critical dimensions, assuring high cardiotonic effect, are characteristic of strophanthidin and strophanthidol. Furthermore, even a slight increase in the dimensions of the angular group leads to a serious disruption in the complementarity between the receptor and the cardiosteroid 271 molecule. Whereas strophanthidylic acid inhibits Nat, KT -ATPase to some extent (5), 19-rhamnosyl-(II) and 3,19-bisrhamnosyl-strophanthidol (III) do not have this capacity at all. Change in hydrophobicity. Angular methyl groups give steroid compounds a certain hydrophobicity. It cannot be excluded that the active site of the ATPase grabs the molecules of cardiosteroids by their nonpolar groups, extract- ing them from aqueous solution. Due to the hydrophobic interaction, the methyl groups help the cardiosteroids to leave the polar medium and go over to the receptor portion of the enzyme. The methyl radical is among the electron-donor groups and usually increases the basicity of the molecule. Consequently, the second possible cause for lack of activity in 19-rhamnosyl- strophanthidol and 3,19-bisrhamnosyl-strophanthidol is increased hydrophilicity of the molecular fraction at C-19, preventing direct contact between the cardiac glycoside and the receptor portion of the enzyme (figure 2). Another group of compounds was also tested for cardiotonic activity and inhibition effect: convallotoxin (IV), 5B-0-oa-L-rhamnopyranosyl-strophanthidin (5-isoconvallotoxin, V), and 3B,5B-0-bis-a-L-rhamnopyranosyl-strophanthidin (VI). FIGURE 2. Schematic models of adsorption section of receptor center of Na™t, Kt- ATPase. Molecule of convallotoxin (A) is completely complementary with receptor. Strophanthidylic acid (B) has low activity, since the carboxyl group has little contact with the enzyme slot for the angular group. In 3,5-bisrhamnopyranosyl- strophanthidin (C), the sugar group at C-5 and the aldehyde group at C-19 are located above the plane of the steroid ring. This compound disrupts the steric form of the enzyme slot and therefore has less activity than convallotoxin. In 3,19-bisrhamnopyranosyl-strophanthidol (D), the sugar group is large and hydro- philic. Contact with the surface of the receptor is impossible, and this com- pound is inactive. 272 The distinguishing feature of the last two compounds is the presence of a sugar molecule at C-5. The introduction of L-rhamnose not through the primary alcohol hydroxyl at C-19, but rather through the tertiary at C-5, also influenced the activity of the cardiosteroids, but not in the extreme form of 3,19-dirhamnosyl- strophanthidol (III). Compounds V and VI have an inhibiting effect on Nat, Kt- ATPase, although to a lesser extent than convallotoxin (see table 1). Why do cardiac glycosides of differing structures have differing, and at times greatly differing, activities? The specificity of the action of cardiosteroids is closely related to the specificity of the structure of the receptor portion of Nat, Kt-ATPase. The strict orientation of peptide chains prevents distribution of polar and nonpolar groups; therefore, only molecules which contact the active site sufficiently closely, not blocked by dimensions, steric placement of individual rings and functional groups of the steroid series, degree of polarization, lipophilicity, etc., can interact with the enzyme. In our opinion, the active site of Nat, KT-ATPase consists of two sec— tions which differ in terms of function. One section, which we will call the adsorption section, binds the cardiosteroids to the receptor, and concentrates and orients the molecules relative to the second, or catalytic section, where groups which interact directly with the lactone ring are concentrated. It is most probable that adsorption occurs by means of angular groups and particularly by means of their hydrophobic interaction with the corresponding sections of the receptor. Unstable intermediate compounds may develop for a short period of time. It is possible that when cardiosteroids are bound with the active site, some additional energy is released, stimulating the work of another section of the enzyme responsible for the reaction with the participation of the lactone group. The more easily the molecule of the cardiosteroid penetrates into the enzyme ''slot,'" the more firmly it is bound with the adsorption section, and the more energetic the interaction is upon extraction, in which the lactone ring plays the primary role. Conversely, if full complementarity is not achieved at the adsorption section, the effect of biological action will be reduced to a greater extent with the greater deviations from the optimal position. The two-center model explains why cardiac glycosides which have five- or six-membered lactone rings, but differ from each other in the structural ele- ments of the steroid or sugar component, have different cardiotonic effects. In particular, the cardenolide bisglycosides with unaltered butenolide ring, but with sugar components in unusual positions (at C-19 or C-5), are less active than the glycosides used in medicine because, due to steric hindrances or re- duced forces of hydrophobic interaction, they fit poorly into the enzyme slot, disrupting its usual steric form (see figure 2). A definite alteration of polar and nonpolar groups is characteristic for the receptor section of Na™t, KT-ATPase. Therefore, if the steroid portion of the molecule of cardiac glycosides binds to an adsorption section by means of hydrophobic interaction, the sugar portion should interact with the digitalis receptor by means of the polar groups. This position is separated in some way from the point of contact of the steroid por- tion and is more suited for binding of the sugar molecule. The absence of a 273 sugar component has a negative influence on the inotropic effect--aglycons, as a rule, are less active than glycosides. In regard to the mechanism of action on the catalytic section of the recep- tor, Repke (7) set forth as early as 1963 the hypothesis that the lactone ring is bound to the active site by a hydrogen bond, the strength of which is pro- portional to the fractional negative charge on the carbonyl group of the lactone. A great deal of experimental material has been accumulated on this subject (1, 8), but we do not dwell on this in this paper. In sum, we suggest a two-center model of the digitalis receptor to explain the varying activity of different cardiac glycosides in the process of inhibi- tion of transport Nat, Kt-ATPase in heart muscle. For the experimental portion of this work, see reference 9. 274 REFERENCES Thomas R, Boutagy J, Gelbart A: Synthesis and biological activity of semisynthetic digitalis analogs. J Pharm Sci 63:1649-1683, 1974 Petersen R, Flasch H, Heinz N: [Demonstration and properties of some glucuronides and sulfates of cardenolides and cardenolide-glycosides] (Ger). Arzneim Forsch 27:642-649, 1977 Palyants NSh, Gorovits MB, Abubakirov NK: [Glycosylation of cardenolides. IV. 5-Alpha-L-rhamnoside and 3,5-bisglycosides of strophanthidin] (Rus). Khim Prir Soedin No.5:765-772, 1976 Palyants NSh, Abubakirov NK: [Glycosylation of cardenolides. V. Periplo- genin fucoside and strophanthidol dirhamnoside] (Rus). Khim Prir Soedin No.1l:125-126, 1977 Mirsalikhova NM, Umarova FT, Palyants NSh, Abubakirov NK: [Inhibition of transport (sodium-potassium ion)-dependent ATPase by cardenolides of the strophanthidin group and study of the structure-activity relation] (Rus). Khim Prir Soedin No.2:188-194, 1976 Chernobai VT: [Partial synthesis of the cardiac glycosides. Strophanthidol (3,19)-0~-L dirhamnoside] (Rus). Zh Obshch Khim 34:1690-1691, 1964 Repke K: Effect of digitalis on membrane adenosine triphosphatase of cardiac muscle. In Proceedings of the International Pharmacological Meeting, 2nd, Prague, 1963. Drugs and enzymes, vol 4, edited by BB Brodie and JR Gillette. New York, McMillan, 1965, pp 65-87 Repke KRH: [Biochemical basis of the development of new types of cardiac stimulants of the digitalis type] (Ger). Pharmazie 27:693-701, 1972 Mirsalikhova NM, Palyants NSh, Abubakirov NK: [Characteristics of the inhibition of transport (sodium-potassium ion)-dependent ATPase by cardenolide bisglycosides] (Rus). Khim Prir Soedin No.1:95-102, 1978 275 re ACTIVATION OF LIPID PEROXIDATION AS THE DECISIVE LINK IN THE PATHOGENESIS OF STRESS DAMAGE TO THE HEART, AND PREVENTION OF STRESS AND HYPOXIC DAMAGE BY THE ANTIOXIDANT IONOL F. Z. Meerson, L. Yu. Golubeva, V. E. Kagan, M. V. Shimkovich, and A. A. Ugolev INTRODUCTION The important role of emotional stress in the etiology of primary circula- tory diseases is no longer in doubt. However, the molecular mechanism by which the high concentrations of catecholamines and glucocorticoids that are present in stress damage the heart remains largely unknown. This lack of knowledge greatly limits our ability to prevent stress damage chemically. One approach to this problem is based on the observation that high doses of catecholamines (1) and the oxidation products of catecholamines--adrenochromes (2)--activate peroxidation of cardiac muscle lipids, while the lipid hydroperoxides formed in the process of peroxidation can damage the sarcoplasmic reticulum membrane and sarcolemma of muscle cells in vitro (3). Based on these facts, we hypothe- sized that activation of lipid peroxidation, induced by an excess of catechol- amines, was the decisive link in the pathogenesis of stress damage to the heart, and that stress damage could be prevented by antioxidants strong enough to block peroxidation. In order to test the above hypothesis, emotional-pain stress (EPS) was reproduced in rats in the form of alarm neurosis according to Desiderato's well- known method (4). The most important feature of this stress effect is that the animals, during 6 hours of tense expectation of an electrically induced pain shock, actually receive these shocks at random intervals. The result is ac- tivation of the adrenergic and hypophyseal-adrenal systems, reaching their maximum during the stress period and continuing for at least 4 to 5 days after stress is stopped. The increase in concentration of catecholamines and gluco- corticoids in the blood regularly leads to the development of ulcers in the gastric mucosa (5) and disorders in the function and metabolism of the heart (6). From the Institute of Normal and Pathological Physiology, USSR Academy of Medical Sciences, and from the Department of Biology, Moscow State University, Moscow, USSR. 277 RELATIONSHIP OF LIPID PEROXIDATION TO STRESS DAMAGE TO THE HEART In the first stage of the study, we investigated the effect of this severe EPS on peroxidation of lipids (POL), the status of the lysosomes, and the struc- ture and function of the myocardium. We determined the primary molecular prod- , ucts (polyene lipid hydroperoxides) and the endproducts of POL (fluorescent Schiff bases) in the heart, skeletal muscle, and brain of rats undergoing stress. Lipids were separated from these organs using the method of Folch (7). Accumula- tion of hydroperoxides in the polyene lipids was estimated from the ultraviolet absorption spectra of lipid solutions in methanol-hexane (5:1) characteristic for diene conjugates, assuming the molar extraction coefficient at Apgx = 232 nm to be 2.1+104 mm~1 cm~1 (8). The spectra were recorded on a Shimadzu spectro- photometer. The endproducts of POL--the products of interaction of short- chain dialdehydes with aminophospholipids--were recorded on the basis of the fluorescence spectra of lipid solutions in chloroform with a maximum of excita- tion of fluorescence at 360 nm and a maximum of emission in the area of 420-440 nm on an Aminco-Bowman spectrofluorimeter. Figure 1A shows the typical ultraviolet absorption spectra, and figure 1B shows the emission fluorescence spectra of lipid solutions taken from the cardiac muscles of control and EPS-exposed animals. Figure 1A shows that the lipids taken from the hearts of animals subjected to EPS have the absorption spectrum typical for polyene lipid hydroperoxides, with maxima at 230-235 nm and 270-280 nm, which are practically absent in the controls. We see further that the in- tensity of fluorescence of the Schiff bases, which are the endproducts of POL, is significantly higher for the lipids of animals exposed to EPS than in the controls (figure 1B). The shaded areas in figure 1 reflect the accumulation of intermediate and endproducts of POL with EPS. Table 1 demonstrates that, in cardiac muscle, the content of hydroperoxides increases under the influence of EPS by a factor of 3, while the intensity of fluorescence of the endproducts of POL increases by a factor of 5. For the skeletal muscles the corresponding quantities are 2.0 and 2.7, respectively, and A B Conjugated dienes Schiff’s bases 34 > @ 1.0 3 c Q 8 82 3 $ 2 0.5 Er Oo \ uw \ \ o 200 300 nm 400 500 nm FIGURE 1. Effect of emotional-pain stress (EPS) on ultraviolet absorption spec- tra (A) and emission spectra (B) of lipids extracted from heart muscle. =--- = Control animals. —— = Animals which experienced EPS. 278 TABLE 1. Accumulation of Products of Lipid Peroxidation During Emotional-Pain Stress (EPS) Group of Skeletal Index Animals Myocardium Muscle Brain Lipid hydro- Control 0.35 + 0.05 0.40 + 0.05 0.15 = 0.05 peroxides n = 10 (optic den- sity units) EPS 0.90 + 0.05 0.80 + 0.10 0.25 + 0.05 n=11 Significance of differences p < 0.001 p < 0.01 p < 0.01 Fluorescence Control 19.2 + 2.9 13.0 + 1.7 3.3 + 0.4 of Schiff n = 10 bases (rela- tive fluores- EPS 95.5 + 20.8 35.3 + 20.8 20.0 + 1.5 cence units) n= 11 for the brain, the quantities are 1.7 and 2.1. Thus, after EPS, the process of POL is activated in the organism, and is manifested more in the heart than in the other organs. Since lipid peroxides are capable of disrupting the membrane structure of cells, further studies were directed, first of all, toward examination of the status of the lysosomes, the labilization of which under the influence of lipid peroxides could represent the next step in the pathogenesis of stress damage to heart muscle. Second, studies were directed toward measurement of enzyme ac- tivity in blood plasma, since the release of enzymes from cells into the blood is one possible direct result of damage to the cell membranes. Figure 2 shows that EPS leads to a decrease by 25 to 40 percent in the activity of acid cathepsins in the lysosome plus mitochondrial fraction, with a simultaneous increase by 45 to 70 percent in the activity of these enzymes in the supernatant fraction. This change indicates with some probability that EPS causes labilization of lysosome membranes and liberation of proteolytic enzymes from them. The data in the diagram indicating that EPS doubles the activity of cathepsins in blood plasma also agree with this concept of pathogenesis. Figure 3 shows that, 2 hours after EPS, the activity of aspartate trans- aminase, alanine transaminase, lactate dehydrogenase, and malate dehydrogenase in blood plasma was increased by approximately a factor of 2. Thus, after stress, in addition to activation of POL, and possibly under the influence of the products of POL, labilization of lysosomal membranes and liberation of lysosomal proteolytic enzymes into the cytoplasm and blood plasma occur. 279 Myocardium Plasma — % 7 I 7 Control Zam % 1 1 T 1 ¥ Zz iz EPS 29 | 22 33148 05110 os 1 gs Fraction of Suprasedimen- lisosomes tal fraction .and . mitochondria FIGURE 2. Effect of emotional-pain stress (EPS) on the activity of cathepsins in heart muscle and plasma (ug tyrosine/mg protein/h). VX, 4783 45] o DH 0.9 © 36 N N No Se eI aaa. AST ALT LDG M oO G Control , EPS FIGURE 3. Effect of emotional-pain stress (EPS) on the activity of plasma enzymes (standard units/ml). Activity of lactate dehydrogenase (LDG) is given in pm NADH/min/ml). AST = Aspartate transaminase. ALT = Alanine transaminase. MDG = Malate dehydrogenase. 280 Simultaneously, more extensive damage to cell membranes develops, resulting in marked release of enzymes into the blood. As we continued the studies, it was important to determine which morpho- logical changes in heart muscle result from such damage to the membranes. To do this, a series of topographic sections of heart muscle 5-7 pm thick was stained with picrofuchsin according to the method of Van Gieson, and colloidal iron according to the method of Hale and Selye. The Perls reaction was con- ducted with subsequent H&E stain. The PAS and Brashe reactions were measured. Phase contrast microscopy was used to estimate the status of the contractile apparatus of heart muscle cells and the nature of changes in the myofibrils (9). The most important fact established in these morphological studies was that, even in the first few hours after stress, the muscle cells of the heart develop focal changes of contracture, and these changes progress to reach a maximum, i.e., full contracture of the muscle cells, after 39-45 hours. In some in- stances, the contracture is accompanied by clearly manifest necrobiotic changes that lead to the death of muscle cells, which are subsequently resorbed, forming a fibroblastic granule. In other instances, the contracture reverses and the cell structures are restored. Figure 4 shows the myocardium 45 hours after EPS, i.e., in the phase of maximum morphological changes. Photograph A shows the results of phase contrast microscopy and demonstrates the deep contracture with merging of disks A of the myofibrils, forming in some fibers continuous anisotropic conglomerates. Photo- graph B, with less enlargement, shows a large number of such contracture-altered muscle fibers with significant anisotropy. Photograph C shows the results of Perls reaction and groups of necrobiotically altered muscle fibers with positive reaction to iron. Finally, photograph D shows the formation of a cellular infiltrate around the necrotized muscle fibers. At the focus of necrosis, we see pyknotically altered nuclei, and disruption and fragmentation of muscle fibers. In evaluating the fact that stress causes focal contracture changes in heart muscle cells, we should keep in mind that the concentration of ATP in the myocardium of animals following EPS was not different from the controls, either at rest or under maximum heart load. Consequently, it is difficult to explain the focal contracture of the muscle cells by hypoxia or other shortage of energy-rich phosphorus compounds. The most probable cause of this type of con- tracture is a disorder in membrane transport of calcium, since active removal of this cation from the myofibrils for storage in the sarcolemma and sarco- plasmic reticulum is the basis for normal relaxation, while focal contracture after stress is nothing other than a focal disorder of relaxation. Thus, it is possible that, in stress, lipid peroxides and proteolytic lysosomal enzymes damage the membrane apparatus which transports calcium and, as a result, focal contracture damage of the myocardium arises. In order to evaluate this hypothe- sis, we must analyze data characterizing the effect of stress on contraction and relaxation of heart muscle. To evaluate the effect of EPS on myocardial contractile function, a study was performed on papillary muscle using Sonnenblick's method (10), and on an 281 X UOT3IedTJTudeW ‘uoT3oeal sSTIg ‘uoTjeuRTdX® I0J 3X9] °9§ 2 ‘67 Xx uorjledryTudew ‘IYSTT pazZTIIRIOd = ¢ "67 X uoTlIEOTITUSEW ‘UTEIS FRH = A °9TT */6T X uorledIITudRU ‘3Yy3TT pe9zTIiB[Og = V °Ssai1ls ured-TeUOTjIOWD JO sSInOoy Gh I931Je wnTpiedoku jex ol a8eweq ‘H TINOIJ 282 entire working heart. These experiments were conducted with the myocardium re- moved from the regulatory effect of the organism. Consequently, the disruptions in contractile function that were detected could be caused only by damage to the structure and metabolism remaining after stress. It was found that papillary muscles of animals which had experienced stress manifested a 40 percent reduction in the rate and amplitude of contraction, an equal decrease in the rate of relaxation, a decrease in the positive inotropic effect of high frequency stimulation, and a significant decrease in adrenoreac- tivity. At the same time, the maximum tension developed by the papillary muscle under isometric conditions, related to maximum ATP and oxygen consumption, did not change significantly with stress (11). This indicates that the reduced rate and amplitude of contractions following stress is independent of the energy deficit, and instead may be caused by the disruption of membrane transport of calcium in myocardial cells. Data obtained in a study of the isolated working heart as described by Neely (12) support this assumption. The important fact in these experiments was that the reaction of the animal heart following stress to a change in calcium concentration in the perfusate was greatly increased. As shown in the lower portion of figure 5, when the concen- tration of calcium in the perfusate was decreased from 2.5 to 1.25 mM, the maxi- mum reduction in the stroke volume of isolated hearts from control animals was Stroke volume, % 75mM Catt 150 - | —— EPS 1 ———- Control 7 / 100 1X < S50 1 1.25 mM Cat+ T T 3 5 10 15 20 min FIGURE 5. Dynamics of reaction of the isolated working heart of control animals (=---) and animals which experienced emotional-pain stress (EPS) ( ) to change in the calcium ion concentration in the perfusate. 283 less than 40 percent, while the hearts of animals which had undergone stress decreased their stroke volume to 0, i.e., by 100 percent. The upper portion of the figure shows that, with a subsequent increase in the concentration of cal- cium from 1.25 to 7.5 mM, the rise in stroke volume is 20 to 40 percent in the control animals, and 80 percent in the animals which had undergone stress. These changes are expressed in the first 3 to 5 minutes after a change in the concentration of calcium; subsequently, heart function in both groups gradually returns to the initial level due to adaptation of the ion transport mechanisms. The curves in figure 6 trace the same phenomenon with respect to aortic pressure. The upper portion of the figure shows that, 3 minutes after a de- crease in the concentration of calcium from 2.5 to 1.25 mM, systolic pressure in the control animals drops by approximately 20 mm Hg. The lower portion of the figure shows that the heart of an animal which has undergone stress responds to the same decrease in concentration of calcium with a tremendous drop in sys- tolic pressure. Thus, it is beyond doubt that, after stress, changes arise in the heart muscle of animals which greatly increase the dependence of contractile function of the heart on the concentration of calcium in the perfusate. In evaluating this fact, we must keep in mind the data of comparative physiology which indicate that the reaction of muscle to a change in the exter- nal concentration of calcium depends on the capacity of the transport mechanisms of this cation. Thus, in skeletal muscle, the well-developed system of the sar- coplasmic reticulum supports a position such that almost all the calcium that leaves the cisternae of the sarcoplasmic reticulum during the action potential and causes the contraction of myofibrils subsequently returns to the tubules. With this 100-percent recirculation of calcium, the skeletal muscle can contract 25mM Cat 1.25mM Ca** mm Hg 100 Control dh 0 mm Hg 100 EPS A Ae | sec FIGURE 6. Effect of a decrease in calcium ion concentration in the perfusate on aortic pressure in an isolated working heart in control animals (upper curve) and in animals which experienced emotional-pain stress (EPS) (lower curve). 284 in a calcium-free solution for several hours. In mammalian heart muscle cells, the calcium transport mechanisms localized in the sarcoplasmic reticulum and sarcolemma are not as strong; recirculation is not 100 percent, and, in a calcium-free solution, the heart stops after performing a few dozen contractions. In the frog heart the sarcoplasmic reticulum is poorly developed, and after calcium is eliminated from the solution, the heart stops instantly (13). These data indicate that, in animals undergoing stress, the increasing de- pendence of hearts on the external concentration of calcium results from damage to the membrane calcium transport mechanisms in myocardial cells--due to a de- crease in the capacity of the sarcolemmal and sarcoplasmic membranes to bind and accumulate calcium. This agrees with our finding that the activation of POL under stress is the decisive link in stress damage to heart muscle, since hydroperoxides can damage the membranes of lysosomes and the membranes of sys- tems responsible for calcium transport, resulting in the development of focal contractures and damage to the contractile function of the entire heart. PREVENTION OF STRESS DAMAGE WITH IONOL Obviously, the most direct means of checking this finding is to block the activation of POL before stress develops with a strong powerful antioxidant, thus eliminating the accumulation of hydroperoxides, and to observe whether damage to the heart muscle and disorders in the contractile function of the heart will develop under these conditions. For this second stage of the study, we utilized the strong, nontoxic antioxidant Ionol, which effectively suppresses POL (14), but which has never previously been used in cardiology research. Ionol was administered intraperitoneally in a dose of 120 mg/kg each day for 3 days before stress was induced. The data provided in table 2 show that the great increase in myocardial content of hydroperoxides and Schiff bases ob- served after stress is prevented to a great extent by the administration of Ionol. Ionol did not significantly affect peroxidation in the heart muscle of control animals. Studies of the papillary muscle and working of the entire heart showed that blockage of peroxidation with Ionol prevents labilization of lysosomes, development of focal contracture damage in the myocardium, and dis- orders in contractile function in animals exposed to stress. In order to finally resolve this question about the preventive effect of Ionol blockage of peroxidation on stress damage to the heart, experiments were also conducted on an isolated isovolumic heart, using the method of Fallen et al. (15). In these experiments, we recorded contractile function and took specimens of perfusate which had passed through the coronary bed, following the method of Bergmeyer (9,14), to determine the activity of creatine phosphokinase (CPK). As is known, release of CPK from the myocardium is one of the most reliable criterion for indicating damage to heart muscle cells. In hypoxia, CPK regularly increases in proportion to the decrease in ATP concentration in the myocardium (16). Therefore, the following hypoxic test was carried out. The oxygen content of the perfusion solution, as a percentage of saturation, was decreased over a period of 20 minutes from 96 percent to 20 percent. Dur- ing this hypoxic test, reduction in the contractile function of the heart devel- oped regularly, and release of CPK from the myocardium increased significantly. 285 TABLE 2. Effect of Ionol on Activation of Lipid Peroxidation in Heart Muscle With Emotional-Pain Stress Accumulation of Lipid Hydroperoxides Accord- ing to Ultraviolet Spectroscopy (nmol/mg lipids) Intensity of Fluorescence of Schiff Bases Series (relative units) 1. Control 16.7 + 1.4 10.5 + 3.5 n=2>5 2. Stress 47.6 £ 2.4 27.6 + 4.2 n=17 Pi» < 0.001 Pio < 0.01 3. Ionol 16.7 = 0.5 10.6 + 2.2 n=17 4. Ionol + stress 28.6 + 2.4 13.2 + 1.8 n=7 Poy © 0.001 Py_y © 0.01 Table 3 presents data on the release of CPK from isolated rat heart into the perfusate. The table shows that the hearts from control animals release a great deal of CPK into the perfusate 35 minutes after the beginning of perfu- sion, while after 95 minutes, the quantity of CPK released is decreased to less than half. The kinetics of CPK release have been demonstrated earlier by other researchers (17) and are determined by the fact that, as the heart is excised and connected to the apparatus, the myocardium is damaged by hypoxia, after which and during long-term functioning of the heart with good oxygenation, damage due to this factor no longer occurs. The hypoxic test conducted under these conditions leads once more to significant release of CPK. For the hearts of animals which had tolerated stress, the kinetics of CPK release remain the same as in the controls, but the absolute value of CPK activity released was increased in all stages by 1.5 times or more, i.e., stress decreases the resis- tance of the heart to hypoxic damage arising during excision or resulting from special hypoxic testing. Ionol, administered to intact animals, decreased the release of CPK in comparison to the controls at the beginning of perfusion by almost a factor of 3; during long-term perfusion with sufficient oxygenation, by more than a factor of 2; and during the hypoxic test, by a factor of 3. In other words, preliminary administration of Ionol was proven to increase the resistance of the heart to hypoxic damage. It also follows from table 3 that the hearts of animals which received Ionol before stress released approximately the same amount of CPK in all stages of the experiment as the hearts of intact animals which received Ionol. Thus, 286 TABLE 3. Effect of Hypoxia and Ionol on Creatine Phosphokinase (CPK) Activity in the Perfusate of Isolated Rat Hearts CPK Activity, IU/liter Series 35 Min Perfusion 95 Min Perfusion 20 Min Hypoxia 1. Control 46.86 + 6.7 19.65 + 2.15 31.64 += 3.78 n=38 2. Stress 86.55 + 10.52 29.56 + 1.89 47.26 + 5.25 n=17 Pio < 0.01 p < 0.01 p < 0.05 3. Ionol 16.11 + 4.31 8.80 += 2.59 10.02 = 3.62 n= 38 Py_3 < 0.01 p < 0.01 p < 0.01 4. TIonol + stress 18.44 + 4.85 10.57 = 1.27 18.70 + 1.96 n=17 Pig < 0.01 p < 0.01 p < 0.02 P3_y4 < 0.001 p < 0.001 p < 0.01 Ionol prevents both the damaging effect of hypoxia, and the potentiating effect of stress on this damage. Data on the contractile function of hearts contracting in the isovolumetric mode, i.e., the same hearts for which CPK release was determined, are presented in figure 7. The figure shows that, with equal contractile frequency, syn- chronized by electrical stimulation, the pressure developed by the hearts of animals which had experienced stress was lower than that of the controls by approximately 1/3 (p < 0.05). In animals which had experienced stress following administration of Ionol, this defect in contractile function was absent. In in- tact animals which had received Ionol, the pressure developed also did not differ from the controls. The curves in figure 8 reflect the reaction of the isovolumetric heart to the hypoxic test. The left portion of the figure shows that, after hypoxia develops, systolic pressure drops sharply; for the hearts of animals undergoing stress, this decrease occurs from a lower base level as indicated by the dotted curve which is lower than the other three curves. Beginning at the 10th minute of hypoxia, systolic pressure becomes very low and is practically identical (8-9 mm Hg) for all the animals in the group. Restoration of contractile func- tion after elimination of hypoxia occurs differently for animals in different groups. The contractile function of the hearts of animals which have experi- enced stress recovers most slowly. One minute after elimination of hypoxia, systolic pressure is only 25 mm Hg, whereas, in the control animals, it has reached 33-34 mm Hg, and in the intact animals which received Ionol it is about 50 mm Hg. These differences in response to hypoxia are also maintained for longer periods; 20 minutes after cessation of hypoxia, the pressure in the 287 mm Hg ] , Control 140 A - / Steececicnctiannnss _ lonol o NT J lonol+EPS 7 120 41 / ~~ ~~" T~ EPS | 7 / J 100 - ~ / {7 H 80 J eart rate 30 200 270 340 4i0 (beats/min) FIGURE 7. Effect of emotional-pain stress (EPS) and Ionol on the contractile force of the left ventricle of isolated rat heart with pacing of contractions. LVP mm Hg 100 reeenne lonol * —.— lONoOl+EPS — Control -—— EPS 0 | r ~T 1F — v —f 0 5 10 20 25 40 min FIGURE 8. Disorders in contractile function of isolated rat heart caused by emotional-pain stress (EPS) and hypoxia, and the prevention of these disorders by Ionol. LVP = Left ventricular pressure. 288 hearts of animals who experienced stress is 58 percent of the initial pressure developed before hypoxia. For the controls which experienced stress and received Ionol, the pressure is 74 to 75 percent of the initial level; and, finally, for the hearts of intact animals which had received Ionol, the pressure is about 90 percent of the initial level. Thus, stress potentiates hypoxic damage to the heart, based on data obtained from the liberation of enzymes, and simultaneously delays posthypoxic recovery of contractile function of the heart. Preliminary administration of Ionol com- pletely eliminates the potentiating effect of stress on posthypoxic damage to contractile function. Finally, administration of Ionol to intact animals com- pletely protects the heart from hypoxic damage, as measured by the release of CPK, and simultaneously greatly stimulates recovery of posthypoxic contractile function. CONCLUSIONS The data which we have presented lead to two main conclusions. First, administration of nontoxic doses of the antioxidant Ionol to animals completely prevents activation of lipid peroxidation, which develops regularly in the myo- cardium under stress. At the same time, stress damage to the heart, estimated by CPK release from myocardial cells, is eliminated, as are stress disorders in contractile function. Consequently, activation of lipid peroxidation is a necessary and decisive link in the pathogenesis of stress damage to the heart. Second, administration of Ionol prevents hypoxic damage to the heart, as estimated by CPK release from the myocardium, and significantly increases the rate and degree of posthypoxic recovery of contractile function of the heart upon reoxygenation. These results, together with data from other investigators concerning the positive effect of antioxidants in cases of ischemic heart damage (13,17), in- dicate that the clinical and physiological significance of preventive and thera- peutic use of antioxidants in cardiology is quite promising. 289 10. 11. REFERENCES Kogan AKh, Kudrin AN, Nikolaev SM: [The question of the role of free radical peroxidation of lipids in the mechanism of damage of the myocardium by adrenalin] (Rus). In [Materials of the Symposium on Free Radical Oxidation of Lipids in Normal and Pathologic Subjects]. Moscow, Nauka, 1976, p 71 Gudbjarnason S, Doell B, Oskarsdottir G, Hallgrimsson J: Modification of cardiac phospholipids and cathecholamine stress tolerance. In Proceedings of the International Symposium on Tocopherol, Oxygen, and Biomembranes, Yamanakako-Mura, Japan, 1977, edited by C de Duve and O Hayaishi. Amsterdam, New York, Elsevier, 1978, pp 297-310 Kagan VE, Churakova TD, Karagodin VP, Arkhipenko IuV, Bilenko MV: [Dis- ruption of the Catt transport enzyme system in sarcoplasmic reticulum mem-— branes upon exposure to phospholipid hydroperoxides and fatty acid hydroperoxides] (Rus). Biull Eksp Biol Med 87(2):145-149, 1979 Desiderato O, Mac Kinnon IR, Hisson H: Development of gastric ulcers in rats following termination. J Comp Physiol Psychol 87:208-214, 1974 Meerson FZ, Giber LM, Markovskaia GI, Radzievskii SA, Rozhitskaia II, Kogan AKh: [Prophylaxis of the disturbances of heart contractile function and ulcerous injury of stomach at the emotional stress by means of sodium oxybutyrate and vitamin E] (Rus). Dokl Akad Nauk SSSR 237(5):1230-1233, 1977 Meerson FZ, Pavlova VI, Yakushev VS, Kamilov FKh: [Disorders of energy metabolism in the myocardium in emotional-pain stress and prevention of these disorders by means of sodium gammaoxybutyrate] (Rus). Kardiologiia 18(3):52-59, 1978 Folch J, Lees M, Stanley GHS: A simple method for the isolation and pur- ification of total lipides from animal tissues. J Biol Chem 226:497-509, 1957 Bolland JL, Koch HP: The course of autooxidation reaction in polyisoprenes and allied compounds. Part 9. The primary thermal oxidation product of ethyl linoleate. J Chem Soc 1945, pp 445-447 Bergmeyer HU: Methoden der enzymatischen Analyse. Weinheim, Verlag Chemie, 1970 Sonnenblick EH: Force-velocity relations in mammalian heart muscle. Am J Physiol 202:931-939, 1962 Csallany AS, Ayaz KL: Quantitative determination of organic solvent soluble lipofuscin pigments in tissues. Lipids 11:412-417, 1976 290 12. 13. 14. 15. 16. 17. Neely JR, Libermeister H, Battersby EJ, Morgan HE: Effect of pressure development on oxygen consumption by isolated rat heart. Am J Physiol 212:804-814, 1967 Morad M: Contracture and catecholamines in mammalian myocardium. Science 166:505-506, 1962 Emanuel NM, Lipchina LP: [Leucosis in mice and some features of its development under the action of certain inhibitors of oxidizing chain pro- cesses] (Rus). Dokl Akad Nauk SSSR 121(1):141-144, 1958 Fallen ET, Elliott W, Gorlin R: Apparatus for study of ventricular function and metabolism in the isolated perfused rat heart. J Appl Physiol 22:836- 839, 1967 Spiekerman FG, Gerhard MM, Nordbeck H: Myocardial high energy phosphates and enzyme release during anaerobiosis in the isolated perfused rat heart. In Abstracts. International Symposium of International Study Group for Research in Cardiac Metabolism. Bruxelles, 1975, p 41 Miller DG, Mallov S: Quantitative determination of stress-induced myocar- dial damage in rats. Pharmacol Biochem Behav 7:139-145, 1977 291 RELATIONSHIP OF THE GENERATION AND DETOXIFICATION PROCESSES OF LIPID PEROXIDES IN THE NORMAL AND HYPERTROPHIC HEART A. A. Kubatiev and S. V. Andreyev INTRODUCTION According to today's concepts, peroxidation of lipids (POL) plays an im- portant role in cell damage mechanisms (1-4). The capability of cell membranes to be destroyed by the action of products of POL results from the specific nature of their chemical structure, particularly the presence of phospholipids (2,5). As complex, polyfunctional compounds, phospholipids play an active role in supporting the supermolecular organization of biological membranes and in regulating their metabolic activity. In recent years, it has been shown that, due to the chemical weakness of phospholipids, they are easily subject to oxidative destruction by molecular oxygen with loss of hydrophobicity of the lipid matrix and subsequent penetration of free-radical initiators of auto-oxidation into the inner layers of the mem- brane (5). This is facilitated equally by two factors: either expulsion of fatty acid acyls from the phospholipids, with the related formation of lysophos- pholipids, or direct breakdown of the hydrophobic barrier of the membrane by polar peroxide groups, also located in the phospholipid environment. With their high reactivity, lipid peroxides cause disintegration of intracellular structures, and oxidation and polymerization of low-molecular intracellular components and macromolecules (3). Under ordinary physiological conditions, this process is strictly limited, but various external and pathological effects, involving damage of the membrane structures, lead to intensification of POL (6,7). One of the results of chemical modification of the hydrophobic layer of phospholipid membranes is inhibition of membrane-bound enzyme complexes and an increase in their nonspecific permeability to ions and nonelectrolytes. This ability has been demonstrated recently in experiments with fragments of the sarcoplasmic reticulum (SR) of rabbit skeletal muscles. These studies showed that accumulation of toxic products of POL causes a large increase in the permeability of the SR membranes for calcium ions, separation of calcium ion transport and hydrolysis of ATP, as well as inhibition of Ca-dependent ATPase (8). From the Institute of Morphology, USSR Academy of Medical Sciences, Moscow, USSR. 293 If one assumes that the same changes in the transport function of the SR cause disruption of the contractile capacity of the muscle tissue (9), a natural question arises: Are these disruptions related to the action of products of POL? An indirect answer to this question can be found in the works of Omaye et al. (10), and Zalkin et al. (11), who have studied processes of POL during genetically mediated Duchenne type muscular dystrophy, as well as dystrophy evoked by E-avitaminosis. It has been found that the generation of lipid per- oxides in SR phospholipids in dystrophic muscles actually results from qualita- tive and quantitative changes in phospholipid composition, and disruptions in the structuring and ordering of carbohydrate chains in the hydrophobic lipid layer of the membrane. In recent years, reports have appeared on the relationship of POL to the function of heart muscle (11,12-16). According to Kogan (12), administration of adrenalin to animals causes a large increase in the level of lipid peroxides in the myocardium, and the development of myocardial necrosis. Analogous data were obtained by Meerson (13) in modeling stress damage to the myocardium. The author assumes that the elevated level of lipid peroxides in the myocardium during stress reflects an increase in their oxidation. The fact that products of POL are produced in the damaged myocardium is of significant interest, in and of itself. However, it is not clear how this generation occurs. We know that the intracellular concentration of lipoperoxides at any given time is determined by the balance between two processes--generation of the products of POL and detoxification of these products. The latter may occur by either an enzymatic or a nonenzymatic mechanism (3). In enzymatic breakdown of hydroperoxides, occurring by nucleophilic substitution for oxygen, homogeneous molecular products are formed which are isomers of hydrogen peroxide and which are decomposed by the enzymes glutathione peroxidase and catalase (5). At the same time, the enzyme superoxide dismutase causes a blockage in the forma- tion of the superoxide anion radical. In the nonenzymatic mechanism of decompo- sition of hydroperoxides, which is less favorable for the organism, large numbers of free radicals are formed which interact with each other and which may attack functionally important targets and induce new chains of radical oxidation with very toxic properties (5). Clearly, the mechanisms of genera- tion of lipid peroxides in the organism are complex and cannot be evaluated objectively without parallel study of the processes involved in their detoxifi- cation. The objectives of the present study were, first, to study the relationship of the generation and detoxification processes of lipid peroxides in the normal heart and, second, to determine the specific contribution of each of these pro- cesses to oxidative destruction of the myocardium in hypertrophy. The struc- tures used in the study were microsomes and mitochondria from the heart, in which the processes of POL occur particularly intensively in comparison to other subcellular fractions (3). We studied the activity of the two most fre- quently investigated systems of POL--NADPH-dependent (NDP) and ascorbate- dependent--as well as the level of diene conjugation, the total antioxidant capacity of the myocardium, and the activity of the antioxidant enzyme system, including superoxide dismutase, glutathione peroxidase, and catalase. 294 MATERIALS AND METHODS Experiments were performed on chinchilla rabbits weighing 3-3.5 kg, main- tained in a vivarium under normal diet conditions. In order to reproduce heart muscle hypertrophy, we used an original method of pulmonary vessel thrombosis, produced by a combination of acting on the coagulation capacity of the blood and immune hemostasis (17). Intact rabbits were given microdoses of thrombin in combination with atro- pine, and, at the height of the increase in coagulation potential of the blood and the decrease of its fibrinolytic properties, all rabbits were given specific antipulmonary immunoglobins intravenously. Three hours after administration of the substances, diffuse damage to the microcirculatory bed of the lungs was ob- served, with subsequent growth of thrombic masses over the next 3-4 days into the medium and large branches of the pulmonary arteries. On the fifth to seventh day, thrombosis involved as much as 70 to 80 percent of the entire bed of the pulmo- nary arteries, and the animals died with symptoms of right heart insufficiency. This model had the advantage that, in addition to high reproducibility, it was nontraumatic and allowed all stages of the compensatory adaptation of the right heart myocardium to be observed, beginning with hyperfunction and terminat- ing with development of insufficiency. The degree of thrombosis was evaluated in percent by the method of Boruach et al. (18). The degree of right heart hypertrophy was judged on the basis of functional (indications of right heart hemodynamics and phase structure of the systole of the right ventricle) and bio- chemical (content of nucleic acids) criteria, as well as on the increase in mass of the right heart (in percent) during the course of thrombotic occlusion of the pulmonary vessels. All biochemical studies were performed both on homogenates of heart muscle and on the subcellular structures--microsomes and mitochondria. In order to produce the subcellular fractions of the rabbit heart, the animals were sacri- ficed by hexenal narcosis (3.0 ml of 1 percent solution intravenously). The hearts were perfused with cooled 0.14 M NaCl solution, carefully cleansed of blood, fat, and connective tissue elements, and the heart tissue was homogenized with nine volumes of 0.25 M sucrose, 0.01 M Tris-HCl, pH 7.4, and 0.2 mM EDTA in a glass Potter-Elveheim homogenizer. The nuclei, myofibrils, and undamaged cells were removed by centrifugation at 900 g for 10 minutes. The mitochondria were sedimented by centrifugation of the supernatant fluid in a Beckman refrigerated ultracentrifuge (U.S.) at 8,000 g for 10 minutes. The sediment produced was washed twice in the homogenization medium, then resuspended and used as the mitochondrial fraction. Recentrifugation of the mitochondrial supernatant at 105,000 g for 1 hour in the same medium precipitated the microsomal fraction. All of the procedures involved in separation of the subcellular fractions after the rabbits were sacrificed were performed at 2-4° C. Purity of the prep- arations of microsomes and mitochondria produced was tested by marker enzymes (1). Peroxidation of membrane phospholipids was evaluated by the kinetics of accumulation of malonic dialdehyde (MDA) during aerobic incubation (3,15). In studies of the NADPH-dependent POL in microsomes and mitochondria of heart muscle, the incubation mixture, 1 ml in volume, contained 1 mM ATP, 0.1 295 mM FeCl3, 50 mM Tris-maleic buffer, pH 6.75, 0.5-1.0 mg microsomal protein, and the NADPH generating system, containing 0.5 mM NADP, 10 mM glucose-6-phosphate, and 1.0 pug glucose-6-phosphate dehydrogenase. The NADPH generating system was not added to the control sample. Incubation was performed at 37° C for 20 minutes with constant shaking. The reaction was stopped by addition of 1 ml of 30 percent TCA and the sediment was removed by centrifugation at 5,000 rpm for 10 minutes. Of the supernatant produced, 100 pl were placed into special microscopic test tubes, to which 20 pl of 0.6 N HCl and 80 pl of 0.12 M thiobarbituric acid were added. The entire mixture was placed in a water bath at 100° C for 10 minutes. The rate of color formation was measured on a Beckman spectrophotometer (U.S.) at 535 nm. The reaction rate was expressed in mmol MDA*mg~lemin-1. In studies of ascorbate-dependent POL, the incubation mixture included 1 mM ATP, 0.1 mM FeCly, 50 mM Tris-maleic buffer, pH 6.75, 0.5-1.0 mg of protein preparation, and 0.8 mM ascorbic acid. The reaction was conducted as above. The control sample included all components of the mixture except the ascorbic acid. The level of diene conjugation was determined by the method of Placer (19), and antioxidant activity was determined by the method of Glevind (20). Superoxide dismutase (SOD) activity was studied by a modification of the method of Misra et al. (21), based on the principle of inhibition of auto- oxidation of adrenalin. Standard determinations were performed in a thermo- stated cuvette at 25° C in 3 ml of 0.05 M Na carbonate buffer, pH 10.2, containing 10=%4 EDTA. The reaction mixture contained 3-104 M of L-adrenalin and 10~5 M of adrenochrome. The reaction was initiated by introduction of 0.4 ml of acid solution of adrenalin, pH 2.5, and the initial rate of oxidation of the adrena- lin was recorded, equal to 0.025 min—l, and based on the formation of adreno- chrome on a Gilford spectrophotometer at 480 nm. The unit of SOD activity was the quantity required to inhibit the initial rate of auto-oxidation of adrenalin by 50 percent under the conditions described above (21). Glutathione peroxidase (GP) activity was studied at 25° C by a modifica- tion of the method of Paglia et al. (22), and catalase activity was determined by the method of Bergmeyer (23). Protein content was determined by the method of Lowry (24), and the concentration of nucleic acids was estimated by the method of Spirin (25). RESULTS AND DISCUSSION Typical kinetic curves of reoxidation of polyene lipids in a homogenate, and in mitochondria and microsomes from rabbit heart and liver, are presented in figure 1. As is shown, normal functioning of the heart, in contrast to the liver, characteristically manifests a very low stable level of POL. The period of induction of POL for the myocardium is almost double, and the level of diene conjugation and MDA is much lower. In subcellular fractions of the myocardium, the rate of POL is slightly higher than in the homogenate, which is explained apparently by localization of the electron transfer chains in these fractions (1). Comparison of the intensity of POL in various subcellular fractions shows 296 £ 2 ° a oo 0.282 + 0.02 £ 100% 2 Bo 0 @ = 33 © 3g E oS c SE 3% ag s t min FIGURE 1. Typical kinetic curves of NADPH-dependent (NDP) and ascorbate- dependent lipid peroxidation and the level of diene conjugates in the tissue of the intact rabbit. 1 and 2 = NDP in the heart microsomes (J) and in the homogenate (MB). 3 and 4 = Ascorbate-dependent peroxidation in the heart mito- chondria (A) and in the homogenate (A). 5 = NDP in the hepatic microsomes (0). 6 = Ascorbate-dependent peroxidation in the hepatic mitochondria (O). 8 = The level of diene conjugates (DC). MDA = Malonic dialdehyde. that NDP predominates in the microsomes, and ascorbate-dependent peroxidation predominates in the mitochondria. In almost all cases, reduced capacity of the intact heart muscle for genera- tion of lipid peroxides was combined with a high level of fat- and water-soluble antioxidants, as well as a high level of antioxidant enzymatic system activity (figure 2). In some cases, SOD activity, and particularly GP activity, was so great that it was practically the same as activity in the liver tissue which, as is known, is very rich in these enzymes. Since the heart is distinguished by a high level of oxidative metabolism, there is every reason to believe that the increased level of antioxidant enzyme activity in the intact myocardium is a reflection of an important and general biological mechanism to protect the myocardium from constant attack by products of POL. This assumption is seen clearly in experiments on thrombosed animals. With 10 percent occlusion of the pulmonary arteries and sufficient balance in the hemodynamic activity of the right heart, and no signs of hypertrophy (figure 3), the level of lipid peroxides in the mitochondria and microsomes is essen- tially unchanged (table 1). There is also a slight increase in the antioxidant activity of the myocardium. The activity of SOD in the mitochondria increased by 38 percent of the initial level, GP activity increased by approximately 20 percent, and catalase activity increased by 32 percent. In the microsomes, the 297 40 7 8 120 gz a hi 2 = 30 s 6 {15 = Ho 2 2 ag — > I o = ee E £3 5 o a N, 20 ¢ as 4 1 10 4 Oo a oe E 2 gd 3 = 5a o 2g 2 g 57 = 10 ag 2 5 b - oO b [oN Oo HEART LIVER HEART LIVER FIGURE 2. The level of endogenous antioxidants (A) and the activity of the antioxidant enzymes (B) in the myocardium and liver of intact rabbits. [= Mitochondria. EMM = Microsoma. I = Superoxide dismutase (SOD). 2 = Glutathione peroxidase (GP). 3 = Catalase (CA). a = Fat-soluble. b = Water-soluble anti- oxidants (AO). 100 90 + 80 0 2 70 £ 60 S £ 50 c 40 S 3 30 20 10 1 1 1 1 1 1 1 1 1 1 / 0 7 0 10 20 30 40 50 60 70 80 90 100 400 Percent hypertrophy FIGURE 3. Dynamics of the weight increase in the right ventricle of the heart in developing hypertrophy. 298 66¢ TABLE 1. Lipid Peroxides and Antioxidative Activity of the Enzymatic System in the Subcellular Structures of the Right Ventricle of the Heart in Hypertrophy Parameters NDP ADP DC AO(F-S) SOD* GP CA 1 -1 a 1 mmol NADPH/min MDA x ng ~ x min MDA x mg =~ x min E,,,0m/100 mg mEq/g u/mg prot mg DNA mmol/min/mg DNA Series mit mic mit mic mit mic mit mic mit mic mit mic mit mic Intact rabbits 1.8 = 2.5 + 2.0 + 1.4 + 0.289 0.312 18 + 26 * 3.4 + 4.6 x | 4.09 + 4.4 2 15.3 18.4 0.06 0.09 0.05 0.02 + 0.04 + 0.03 1.6 2.3 0.6 0.5 0.62 0.56 + 4.1 t+ 3.9 Percent of | No + 8% + 7% + 127 + 10% + 9% + 67% + 247 + 17% 38% + 53% | + 20% + 46% + 32% + 58% hypertrophy | hypertrophy | > 0.05 > 0.05 > 0.05 > 0.05 > 0.05 > 0.5 < 0.05 < 0.05 0.05 < 0.01| < 0.05 < 0.01 | < 0.05 < 0.01 20-25 + 197 + 287% + 10% + 15% + 267% + 41% - 10% - 147% 56% + 62% | + 767 + 93% + 59% + 80% percent < 0.05 < 0.05 > 0.05 > 0.05 < 0.05 < 0.05 [> 0.05 > 0.05 0.01 < 0.01] < 0.05 < 0.001 < 0.05 < 0.01 60-70 + 128% + 164% + 156% + 99% + 150% + 90% - 30% - 447% 147% + 10% | + 18% + 20% + 5% + 8% percent < 0.01 < 0.05 < 0.001 < 0.01 < 0.05 < 0.05 | < 0.01 < 0.05 0.05 > 0.05| < 0.05 < 0.05| > 0.01 > 0.5 Over 100 + 260% + 192% + 210% + 220% + 175% + 126% | - 39% - 50% 30% - 35% | - 26% - 30% - 44% - 60% percent < 0.001 < 0.001 < 0.001 < 0.001 |< 0.01 < 0.001 < 0.01 < 0.01 0.05 < 0.05] < 0.05 < 0.05] < 0.01 < 0.01 NDP = NADPH (sodium diphosphate)-dependent peroxidation ADP = Ascorbate-dependent peroxidation DC = Diene conjugates AO(F-S) = Antioxidant (Fat-soluble) SOD = Superoxide dismutase GP = Glutathione peroxidase CA = Catalase MDA = Malonic dialdehyde mit = Mitochondria mic = Microsome *See "MATERIALS AND METHODS" and reference 21 intensity of increase in antioxidant enzyme activity was more significant--53, 46, and 58 percent, respectively. The data show that compensated loads on the right ventricle, caused by a moderate increase in pulmonary vascular resistance, lead to activation of the antioxidant capacity of the myocardium, in spite of the fact that there are no clear signs of hemodynamic disorder and that contractile activity is preserved. It can be assumed that activation of mechanisms which facilitate detoxification of lipid peroxides is related, on the one hand, to acute myocardial hypoxemia resulting from mechanical blockage of a portion of the pulmonary circulation, and insufficient oxygenation of the blood arriving at the right ventricle, and, on the other hand, by increasing load on the right heart. As thrombosis progresses with occlusion of approximately 30 percent of the pulmonary vascular bed, the hemodynamic load on the right ventricle increases still more and its contractile capacity is more than tripled in order to over- come the resistance to blood flow. During this period, an increase in intraven- tricular systolic pressure is noted, as well as an increase in pressure at the end of diastole, and an increase in the length of the period of expulsion. The degree of myocardial hypertrophy was approximately 20 to 25 percent (figure 3). In addition to the appearance of indications of right heart overload, an increase was observed in POL in subcellular structures of the myocardium (table 1). The level of NDP in the myocardial microsomes was increased by an average of 28 percent, and in the mitochondria by 19 percent. There was a lesser, al- though significant, increase in the level of ascorbate-dependent peroxidation: in the microsomes by 15 percent, and in the mitochondria by 10 percent. There was also a clear increase in the content of diene conjugates in the mitochon- dria and microsomes. This indicates that the early periods of compensatory myocardial hypertrophy are accompanied by an increase in the accumulation of not only the final, but also the initial, products of free-radical auto- oxidation of lipids. Antioxidant capacity of the myocardium reached its maximum in animals with 30 percent thrombosis of the pulmonary vessels. The increase in GP and catalase activity in the microsomes was particularly high (93 and 80 percent, respec- tively). In the mitochondria, the increase in the activity of these two enzymes was 76 and 59 percent, respectively. The SOD activity increased by 62 percent in the microsomes and by 56 per- cent in the mitochondria. All of this indicated that, in spite of the increase in POL processes, the cells of the myocardium were sufficiently supplied with protective antioxidant enzymes during the initial stages of hypertrophy. These enzymes break down the lipid peroxides by the mechanism most favorable for the cells, without causing the formation of free radicals capable of inducing new chains of radical oxidation. They, therefore, can be considered '"antiradical" cell enzymes (22). The greatest degree of thrombotic occlusion of the pulmonary vessels (50 to 60 percent of the total bed of the pulmonary arteries) was accompanied by hypertrophy of the right heart up to 70 percent of its initial weight and was 300 associated by a further increase in the intensity of POL and a reduction in the antioxidant capacity of the myocardium. The NDP activity in the microsomes increased by 164 percent; ascorbate- dependent peroxidation activity increased by 99 percent. In the mitochondria, activation of NDP and ascorbate-dependent peroxidation reached 128 and 156 per- cent, respectively. The level of diene conjugates in the microsomes increased by 90 percent and in the mitochondria by more than 150 percent. This abundant flooding of the cellular structures with products of POL was possible due to the significant reduction in their antioxidant potential. Table 1 shows that the intensity of antioxidant activity of fat-soluble microsomal lipids decreased by 44 percent, while that of mitochondrial lipids decreased by 30 percent. Parallel study of the antioxidant enzyme system showed that, in contrast to the period of moderate hypertrophy of the right heart, its activity in this case was greatly reduced, almost to the initial level. This indicates that, in cases of severe hypertrophy, caused by significant disruption in the expulsion of blood from the right ventricle, the generation of lipid peroxides is so in- tense that it cannot be compensated for by enzymatic protective mechanisms. The imbalance in the formation and effective breakdown of POL creates con- ditions which favor free-radical auto-oxidation of phospholipids in the hyper- trophic myocardium, thus decreasing the heart's contractile capacity. This is clearly manifested as a slowing in the rate of contraction and weakening of the right ventricle, a decrease in the contractile index (by almost 30 percent), a decrease in the intensity of functioning of the structures, and a lengthening in the period of expulsion. In the final stage of thrombosis, when the area of occlusion of the pulmo- nary vessels reached as much as 70 to 80 percent and higher, and the degree of hypertrophy of the right heart was as much as 350 percent of the initial mass, the functional capacity of the heart was practically exhausted and, in addition to a critical drop in the contractile capacity of the ventricle, the level of lipid peroxides reached its maximum value. The highest content of MDA, reflect- ing the formation of the end products of POL, was recorded in the microsomes incubated with NADPH. Ascorbic acid increased the production of MDA in the microsomal fraction by almost 220 percent. In the mitochondria, the quantita- tive content of MDA was increased by 260 and 210 percent under the influence of NADPH and ascorbate, respectively. When subcellular structures were incubated in a medium containing ascorbic acid, the rate of accumulation of primary products of POL was almost equally high, using diene conjugation as the index. Introduction of NADPH to the reac- tion mixture was accompanied by a still greater increase in diene conjugates. Parallel study of the antioxidant enzyme system showed that the generation of lipid peroxides in hypertrophy occurred with significant blockage of SOD, the activity of which was only 35 percent of the initial level, and of GP, the activity of which was decreased by almost 70 percent in comparison with the normal level. Catalase activity was decreased by 40 percent. 301 The data show that the level of lipid peroxides in the hypertrophic myo- cardium increases due to a decrease in the concentration of endogenous antioxi- dants and due to a decrease in the activity of antioxidant enzymes. In contrast to the hypertrophic heart, the cells of the normal myocardium are reliably pro- tected from excess accumulation of products of POL. This is achieved by the presence of effective intracellular mechanisms for inactivation of lipid per- oxides in these cells, which begin operating during the formation of active forms of oxygen, and also during the period of their increased generation. The high level of activity and mobility of the antioxidant protective system of the heart allows it to be considered a unique buffer system which assures stability and preservation of intracellular hemostasis. Disruption of the buffer properties of this system results in myocardial cells that can no longer cope with the oxidative, free-radical attack, and products of oxidative degradation of lipids begin to accumulate in them. This mechanism is apparently the basis for genera- tion of lipid peroxides in the development of compensatory cardiac hyperfunction and hypertrophy. These experiments have shown that, when the activity of antioxidant enzymes is high, the cells of the myocardium, with their compensatory hyperfunction, cope well with the increased load. However, as the buffer properties of the antioxidant system are suppressed, the intensity of POL increases and, due to the destabilization of the intracellular membranes, the capacity of the myocar- dium to withstand the increasing resistance to blood flow decreases rapidly. The contractile capacity of the myocardium gradually decreases and cardiac in- sufficiency develops. The fact that the products of POL not only damage the cellular membranes, but also cause disruptions in oxidative phosphorylation and swelling of the mitochondria (26), damaging the hypertrophic muscle still further, is also important. 302 10. 11. 12. 13. REFERENCES Archakov AI: [Microsomal Oxidation] (Rus). Moscow, Nauka, 1975 Burlakova E: [Study of Physical and Chemical Properties of Lipids in Certain Pathological States] (Rus). Moscow, Akad Nauk SSSR, 1970 Vladimirov IuA, Archakov AI: [Peroxidation of Lipids in Biological Mem- branes] (Rus). Moscow, Nauka, 1972 Chernukh AM: [Inflammation] (Rus). Moscow, Meditsina, 1979 Kozlov Iu, Danilov V, Kagan VE, Sitkovskii MV: [Free Radical Oxidation of Lipids in Biological Membranes] (Rus). Moscow, Moscow State Univer- sity, 1972 Kubatiev AA, Andreyev SV: [Lipid peroxides and thrombosis] (Rus). Biull Eksp Biol Med 5:414-417, 1979 Hochstein P, Ernster L: Microsomal peroxidation of lipids and its possible role in cellular injury. In Ciba Foundation Symposium on Cellular Injury. London, 1964, pp 123-130 Arkhipenko YuV, Bilenko MV, Dobrina SK, Kagan VE, Kozlov YuP, Shelenkova LN: [Ischemic damage to the sarcoplasmic reticulum of skeletal muscles; the role of lipid peroxidation] (Rus). Biull Eksp Biol Med 83(6):683- 686, 1977 Chazov EI, Smirnov VN, Aliev MK, Saks VA, Rozenstraukh LV, Levitsky DO, Undrovinas AI: [Molecular mechanisms of cardiac insufficiency in myo- cardial ischaemia] (Rus). Kardiologiia 16(4):5-13, 1976 Omaye S, Tappel AL: Glutathione peroxidase, glutathione reductase and thiobarbituric acid-reactive products in muscles of chickens and mice with genetic muscular dystrophy. Life Sci 15(1):137-145, 1974 Zalkin H, Tappel A, Caldwell K, Shibko S, Desai J, Holliday J: Increased lysosomal enzymes in muscular dystrophy of vitamin E-deficient rabbits. J Biol Chem 237(8):2678-2682, 1962 Kogan AKh: [Correlation between the intensity of free-radical peroxidation of lipids and the degree of myocardial damage in coronary-occlusion in- farct] (Rus). In [News in the Diagnosis and Treatment of Cardiovascular Disease]. Moscow, Meditsina, 1976, pp 75-77 Meerson FZ, Giber LM, Markovskaya GI, Radziyevskiy SA, Rozhitskaya II, Kogan AKh: [Prophylaxis of the disturbances of heart contractile function and ulcerous injury of stomach at the emotional stress by means of sodium oxybutyrate and vitamin E] (Rus). Dokl Akad Nauk SSSR 237(5):1230-1233, 1977 303 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. George C, Balasubramaniam A, Cherian G: Lipid peroxidation in rat heart: effect of feeding coconut and sunflower seed oil. J Mol Cell Cardiol 10 (suppl 1):26, 1978 Takeshige K, Minekami S: Reduced nicotinamide adenine dinucleotide phosphate-dependent lipid peroxidation by beef heart submitochondrial particles. J Biochem 77:1067-1073, 1975 Meyers CE, McGuire WP, Liss RH, Ina I, Grotzinger K, Young R: Adriamycin: the role of lipid peroxidation in cardiac toxicity and tumor response. Science 197(4299):165-167, 1977 Andreyev SV, Kubatiev AA: [Method of modeling thrombosis of the pulmonary vessels] (Rus). Biull izobretenii otkrytii 12:144, 1977 Boruach KD, Chakravati RN, Wahi PL: Evaluation of xantinol nicotinate (complamine) in experimental pulmonary fibrin thromboembolism. Indian J Med Res 62:923-929, 1974 Placer ZA, Cushman LL, Johnson BC: Estimation of product of lipid peroxi- dation (malonyl dialdehyde) in biochemical systems. Anal Biochem 16:359- 365, 1966 Glevind I: Antioxidants in animal tissue. Acta Chem Scand 17(6) :1635- 1640, 1963 Misra HP, Irwin F: The role of superoxide anion in the autoxidation of epinephrine and a simple assay for superoxide dismutase. J Biol Chem 247(10):3170-3175, 1972 Paglia D, Valentine W: Studies on the quantitative and qualitative characterization of erythrocyte glutathione peroxidase. J Lab Clin Med 70:158-169, 1967 Bergmeyer HU: Methods of Enzymatic Analysis. New York, Academic Press, 1965, pp 837-886 Lowry OH, Rosenbrough NS, Farr AL, Randall T: Protein measurement with folin phenol reagent. J Biol Chem 193:265-275, 1951 Spirin A: [Spectrophotometric measurement of nucleic acids] (Rus). Biokhimiia 23:656-662, 1958 Kondrashova MN, Mironova GD: [The necessity of oxygen for phosphorylation of ADP under conditions of cyanide block] (Rus). Biokhimiia 36:864-866, 1971 304 PART III PROTEIN BIOSYNTHESIS AND PATHOLOGY OF THE HEART BRANCHED-CHAIN AMINO ACIDS AND THE REGULATION OF PROTEIN TURNOVER IN RAT HEART Balvin Chua, Daniel L. Siehl, Ellen O. Fuller, and Howard E. Morgan SUMMARY Leucine, but not the other branched-chain amino acids, accelerated protein synthesis and inhibited protein degradation in hearts that were perfused as Langendorff preparations with buffer containing glucose and normal plasma levels of other amino acids. In working hearts, only inhibition of degradation was observed following leucine addition. Products of the metabolism of all three branched-chain amino acids inhibited proteolysis and stimulated protein synthe- sis. In some, but not all instances, inhibition of degradation and acceleration of synthesis were accompanied by an increase in intracellular leucine. When insulin was present, proteolysis was reduced by 40 percent, but leucine had no effect. Rates of branched-chain amino acid uptake varied widely, depending upon the perfusate concentration and the availability of other substrates. At a leucine concentration of 2 mM, metabolism of the amino acid could account for 36 percent of oxygen consumption in hearts that were perfused with buffer con- taining glucose. On the other hand, uptake of leucine could account for less than 3 percent of oxygen consumption when hearts were supplied physiological mixtures of substrates and normal plasma levels of leucine. These experiments indicate that leucine and a variety of substrates that are oxidized in the citric acid cycle are involved in regulation of protein turnover in heart muscle. INTRODUCTION Protein turnover in heart and skeletal muscle is affected by provision of amino acids, particularly the branched-chain compounds. As a result of stimula- tion of peptide-chain initiation, protein synthesis was accelerated in rat hearts perfused with buffer that contained 5 times the normal plasma levels of amino acids (1,2). This effect was found to depend upon increased availability of the branched-chain amino acids (3). Since the perfused rat heart can oxidize branched-chain amino acids (4), and since various noncarbohydrate substrates such as lactate, acetoacetate, B-OH-butyrate, palmitate, and oleate, enhanced From the Department of Physiology, The Milton S. Hershey Medical Center, The Pennsylvania State University, Hershey, Pennsylvania. This work was supported by grants HL-20388, HL-18258, HL-07223, and NU-05238 from the National Institutes of Health. 305 protein synthesis to a similar extent (3), the acceleration of protein synthesis was suggested to be related to their oxidation. Branched-chain amino acids also were found to accelerate protein synthesis and to restrain protein degradation in rat diaphragm (5-7). The effects were attributed to either leucine or a mixture of isoleucine and valine (5) or to be specific for leucine (6,7). In diaphragm, addition of metabolites of the branched-chain amino acids had no inhibitory effect (7). A mixture of the three branched-chain amino acids or leucine alone also has been found to increase protein synthesis and inhibit pro- teolysis in skeletal muscles that were contained in the perfused hemicorpus preparation (8). The possibility that these effects were dependent upon pro- vision of substrates for energy metabolism was rendered unlikely by the finding that addition of glucose or palmitate did not mimic these effects. In order to assess the role of branched-chain amino acids in regulation of protein synthesis and degradation, hearts were perfused with buffer that con- tained glucose or mixtures of glucose and branched-chain amino acids or their metabolites. The experiments were designed to determine (a) whether leucine was unique among the branched-chain amino acids in stimulating protein synthesis and inhibiting protein degradation, (b) whether this effect could be simulated by metabolites of branched-chain amino acids, and (c) whether leucine affected protein synthesis and degradation during in vitro cardiac work. EXPERIMENTAL PROCEDURES Perfusion Apparatus The apparatus for working hearts that was described earlier (9,10) was modified to reduce the circulating volume and to facilitate oxygenation of albumin-containing buffer (figure 1). The apparatus consisted of six major parts: the cannula assembly, aortic compliance chamber, combined oxygenator and left atrial reservoir, heart chamber, buffer reservoir and filter, and peri- staltic pump. The cannula assembly for mounting of the aorta and left atrium and the aortic compliance chamber were as described earlier (10). Outflow from the compliance chamber was returned to the buffer reservoir via a length of Tygon tubing with a hypodermic needle inserted into its distal end. A needle with an internal diameter of 0.58 mm (20 gauge) provided sufficient resistance to result in systolic and diastolic aortic pressures of approximately 145 mm Hg and 75 mm Hg; use of a needle with an internal diameter of 0.33 mm (23 gauge) produced pressures of approximately 160/100 mm Hg. The oxygenator-left atrial reservoir was adapted from an apparatus described by Hems et al. (11) for liver perfusion. Perfusate entered the top of the oxy- genator and spread as a film over the entire surface of a series of glass bulbs. The lower end of the oxygenator terminated in a tapered glass joint that was fitted with a Teflon stopper that contained a piece of stainless steel tubing (2.8 mm, internal diameter). A short length of Tygon tubing (3.2 mm, internal diameter) connected this tubing to the atrial inflow cannula. A side-arm on the lower end of the oxygenator served as an overflow that returned excess buffer to the reservoir. The height of the overflow above the level of the left atrium determined left atrial filling pressure (14 mm Hg). A gas mixture (95 percent 02, 5 percent CO2) that was saturated with water vapor entered the oxygenator 306 GAS INLET OXYGENATOR- ATRIAL RESERVOIR — 10 cm COMPLIANCE CHAMBER INLET FOR PRELIMINARY PERFUSION NEEDLE CANNULA ASSEMBLY HEART CHAMBER —— 0 PUMP BUFFER RESERVOIR —— FILTER FIGURE 1. Modified apparatus for perfusion of working rat hearts. The compo- nents are described in the text. Oxygen tension of buffer leaving the oxygenator averaged 623 + 10 mm Hg and 603 + 21 mm Hg (four observations each), when 20- and 23-gauge needles were located in the aortic outflow tract, respectively. In these same hearts, oxygen tension of the coronary effluent that was obtained from the pulmonary artery averaged 210 * 9 mm Hg and 284 * 14 mm Hg in the pres- ence of the 20- and 23-gauge needles, respectively. When a 20-gauge needle supplied the outflow resistance, aortic pressures, heart rate, aortic output, and coronary flow of hearts that were perfused with buffer containing 15 mM glucose as oxidizable substrate were stable for 90 minutes. Systolic pressures were maintained at approximately 145 mm Hg for the first 90 minutes and were still at 90 percent of this value after 120 minutes. Diastolic pressures averaged approximately 77 mm Hg and were well maintained throughout the 2-hour period. Aortic output and coronary flow averaged approximately 42 ml/min and 307 via an upper side-arm and was carried into the buffer reservoir through the overflow side-arm. The cannula assembly fitted into the heart chamber which was seated in the top of the buffer reservoir. In addition, the heart chamber had a Y-shaped side-arm that received excess buffer from the oxygenator and out- flow from the compliance chamber. The buffer reservoir was a condenser (20 cm in length) with a coarse sintered glass filter inserted into the lower end. The perfusate was recirculated by a peristaltic pump (Cole Parmer Co., Model 7545; pump head, Model 7014) whose out- put was maintained at 120 ml/min. Temperature of water circulating through the jacketed portions of the apparatus was maintained at 37° C. The minimum circu- lating volume was 40 ml in contrast to 65 ml in the earlier apparatus (10); albumin-containing buffer could be oxygenated without formation of foam. Use of the apparatus was as described earlier (10). Langendorff preparations were perfused with a hydrostatic pressure of 60 mm Hg, as described earlier (10). Perfusion Method A modified Krebs-Henseleit bicarbonate buffer, pH 7.4, was used in all experiments. Final concentrations of buffer components were: 117 mM NaCl, 4.7 mM KCl, 3.0 mM CaClp, 1.2 mM KH2PO4, 1.2 mM MgSO4, 0.5 mM EDTA, 24.7 mM NaHCO3, 0.2 percent bovine serum albumin, 15 mM glucose, and normal plasma levels of amino acids (1). Before addition to the buffer, the amino acids and metabolites were dissolved in saline, and the pH was adjusted to 7.4. In some experiments, the exact volume of buffer that was recirculated was determined by isotope dilution using D-[3H]sorbitol. In experiments in which rates of pro- tein degradation were assessed, recirculation of buffer containing 0.01 mM phenylalanine, various amino acid mixtures, and 0.02 mM cycloheximide followed a preliminary perfusion of 10 minutes. When rates of protein synthesis were measured, (14C]phenylalanine (550 dpm/nmol) was added at a concentration of 0.4 mM. At this concentration, the specific activities of phenylalanyl-tRNA and phenylalanine in the perfusate were equal (12). At the end of perfusion, hearts were cut from the cannula into beakers of 0.15 M NaCl (2° C), opened, blotted on filter paper, and dropped into preweighed tubes containing 5 percent tri- chloroacetic acid. Hearts were removed from Sprague-Dawley rats (250-300 g) that were fasted overnight and anesthetized with nembutal. The heart was mounted on the cannula assembly, as described earlier (10). Aortic and atrial pressures of working (FIGURE 1 continued): 27 ml/min, respectively. When a 23-gauge needle supplied the outflow resistance, mechanical performance was not as stable as with a 20- gauge needle. Systolic pressures averaged approximately 160 mm Hg for the first 60 minutes, but by the end of 2 hours of perfusion, systolic pressures were simi- lar whether a 20- or 23-gauge needle offered the outflow resistance. Diastolic pressures were approximately 100 mm Hg when a 23-gauge needle was used. Higher resistance in the aortic outflow tract resulted in redistribution of cardiac out- put from aortic to coronary flow. Aortic flow (approximately 10 ml/min) was stable during the 2-hour period, but coronary flow fell from about 35 ml/min to 25 ml/min over the 2-hour period. 308 hearts were recorded on light sensitive paper (Honeywell Visicorder, Model 1508C). Aortic flow of working hearts represented outflow from the compliance chamber, and coronary flow was estimated by measuring fluid dripping from the heart. In some instances, the pulmonary artery was cannulated and a portion of the outflow was passed through a water-jacketed lucite chamber (37° C) that housed a Clark- type oxygen electrode. Estimations of Phenylalanine Incorporation and Uptake and Intracellular Concentrations of Amino Acids Incorporation of [l4Cc]phenylalanine into whole heart protein was estimated as described earlier (1). Duplicate samples of 4-8 mg of protein were solubilized in 1 ml of tissue solubilizer (NCS, Amersham/Searle) at 50° C for 1 hour, and 10 ml of scintillation fluid (Formula 949, New England Nuclear) were added. Radioactivity was estimated using a Beckman LS-350 liquid scintillation counter; [14C]toluene (New England Nuclear) was used to correct for quenching. When uptake of branched-chain amino acid was measured, buffers were recircu- lated through the heart for 1 hour. These buffers contained phenylalanine (0.01 mM), normal plasma levels of other amino acids, and either (a) 15 mM glucose and 0.2 percent bovine serum albumin, (b) 2 mM lactate, 10 mM glucose, 400 punits insulin/ml, and 0.2 percent bovine serum albumin, or (c) 1.5 mM palmitate, 10 mM DL-B-hydroxybutyrate, 8 mM glucose, and 4 percent bovine serum albumin. Branched- chain amino acid levels were measured with a Model 119CL Beckman amino acid analyzer. Leucylalanine was added as an internal standard. When leucine uptake was measured, buffer that contained glucose (15 mM), phenylalanine (0.01 mM), leucine (0.5, 1, 2, or 5 mM), 0.02 mM cycloheximide, and normal plasma levels of other amino acids (except for isoleucine and valine), was recirculated through the heart for 2 hours. Leucine concentrations were measured either with the amino acid analyzer or by tRNA binding assay (13,14). Uptake values were calcu- lated from disappearance of amino acids from the perfusate. Results are ex- pressed as a function of the average amino acid concentration during the uptake period. Intracellular amino acid concentrations, nmol/ml, were calculated using the formula: [Intracellular amino acid] = Heart amino acid, nmol/g - sorbitol space, ml/g °* perfusate amino acid, nmol/ml . Total water, ml/g - sorbitol space, ml/g Content of leucine in the heart was measured in trichloroacetic acid extracts by use of either the tRNA binding assay or the amino acid analyzer. Trichloroacetic acid extracts of heart were treated with ion retardation resin (AG11A8, Bio-Rad Laboratories; 0.25 g resin/ml extract) before assay of leucine concentration. Sorbitol space was determined as described earlier using D-[1-H3]sorbitol (1). Total water content was taken to be 0.8 ml/g (1). 309 Perfusion of Hearts for Measurement of Conversion of [l4C]Amino Acids to l4co) When production of l4co, from radioactive leucine or isoleucine was mea- sured, hearts were perfused without recirculation of buffer for 30 minutes. Glucose (15 mM), phenylalanine (0.01 mM), and normal plasma levels of other amino acids, except for the branched-chain compounds, were added to the buffer. Wash-through perfusion was continued with the same buffer that contained 2 - 10-5 M cycloheximide and either radioactive leucine or isoleucine. Coronary effluent was collected as l-minute samples from the pulmonary artery into tubes to which 3 ml of paraffin oil had been added. An aliquot of the coronary efflu- ent (1 ml) was injected into flasks that contained 0.2 ml of 5 N H9SO4 and a hanging center well to which folded filter paper, wet with 1 M methylbenzethonium hydroxide (0.2 ml), had been added. After shaking for 1 hour at 25° C, the center well was dropped into a vial containing scintillation fluid (Formula 949, New England Nuclear), and radioactivity was measured. Rates of l4co, production were calculated using the specific activities of leucine or isoleucine in the perfusate. Statistical Analysis Significance of difference between means was established by student's t- test. Values are expressed per gram of wet heart. RESULTS Uptake, Intracellular Concentrations of Branched-Chain Amino Acids, and Production of 14CO2 The first series of experiments was designed to establish conditions for measurement of rates of conversion of [U-l4C]leucine and [U-l4C]isoleucine to l4co,, to measure rates of leucine uptake and oxidation over a range of leucine concentrations in the perfusate, and to evaluate the contribution of leucine as a substrate for oxidative metabolism. It was found that, after 6 minutes of ex- posure to [l4C]leucine, production of 14c09 reached a plateau value that was used, along with the specific activity of leucine in the perfusate, to calculate the rate of conversion to 14co,. To validate the use of the specific activity of perfusate leucine in calculating rates, specific activities of leucine in the perfusate and intracellular water were measured. After 8 minutes of exposure to [l4C]leucine, specific activities of leucine in the intracellular water and perfusate were calculated from measurements of content and specific activity of leucine in perchloric acid extracts of heart and perfusate, sorbitol space, and total water content (12). In heart extracts, 77 percent and 56 percent of radioactivity applied to columns of the amino acid analyzer were in the leucine peak at perfusate concentrations of 0.1 and 0.5 mM, respectively. Specific activities of leucine were calculated using radio- activity in the leucine peak. At a leucine concentration in the perfusate of 0.1 mM, the specific activity of intracellular leucine was 73 = 3 percent (four observations) of the specific activity of leucine in the perfusate. 310 In earlier experiments, Buse et al. (4) found that specific activity of intracellular leucine was approximately 60 percent of leucine in the perfusate at this concentration. At a perfusate concentration of 0.5 mM, these values were equal. A similar relationship between specific activities of intracellular and extracellular phenylalanine was described earlier in the perfusate rat heart (12). These findings indicate that the rate of conversion to COy is underestimated by about 25 percent at a leucine concentration of 0.1 mM by use of the specific activity of perfusate as compared to intracellular leucine, but that the values are unaffected when the concentration in the perfusate is 0.5 mM or greater. Uptake of leucine from the perfusate increased linearly as the concentration was raised to 3.7 mM (figure 2). At a concentration of 1.0 mM, the rate of up- take was 137 nmol/g * min. In comparison, the rate of conversion of [u-l4c) leucine to CO» was 118 * 13 nmol/g + min, indicating that 85 percent of the leucine that was taken up was decarboxylated and converted to intermediates in the tricarboxylic acid cycle. At a leucine concentration of 2 mM, the rate of conversion of [U-1l4C]leucine to 14co, (207 nmol/g + min) was 72 percent of the rate of uptake. Conversion of [U-1l4C]isoleucine to 14C09 occurred at a rate that was about 40 percent of the rate observed with leucine. Intracellular leucine concentrations increased as perfusate leucine was raised from 0.04 to 2.5 mM (figure 2B). At the higher concentrations, the increase in intracellular level was less proportionately than the increase in extracellular concentration. Although the heart was able to convert [u-14C]leucine to l4co, at a rapid rate when the perfusate concentration was high and only glucose was provided as oxidizable substrate, the rate of uptake was much lower when normal plasma levels of branched-chain amino acids were added and the mixture of substrates normally oxidized by the heart were present (table 1). In either Langendorff preparations or working hearts, leucine uptake was approximately 35 nmol/g - min in hearts perfused with buffer containing only glucose. Uptake was reduced still further in hearts provided mixtures of either lactate, glucose, and insulin or palmitate, B-hydroxybutyrate, and glucose. In the presence of either sub- strate mixture, cardiac work increased the rate of leucine uptake. Uptake of isoleucine and valine was lower than leucine and affected in a similar way by addition of the substrate mixtures. Intracellular concentrations of leucine and isoleucine were increased by addition of any of the substrate mixtures and, in all but one instance, were decreased by cardiac work. In Langendorff preparations, intracellular valine was not increased by addition of the substrate mixtures, but was decreased by cardiac work. These findings indicate that the extracellular supply of branched-chain amino acids made only a negligible contribution to oxidative metabolism in hearts supplied with physio- logical mixtures of substrates. Cardiac work appeared to increase the rate of oxidation as indicated by increased uptake and decreased intracellular concen- trations. Effect of Amino Acids on Protein Degradation Rates of protein degradation were measured by following the release of phenylalanine from hearts that were perfused with buffer containing cycloheximide. At all rates of degradation that were investigated, release of this amino acid 311 600 400} £ LEUCINE UPTAKE N p- J 14 3 u-c'*CILEUCINE g 200F _*5 CONVERSION To 4co, c ~~ u-c'4C1ISOLEUCI NE .0 CONVERSION TO 14co, UPTAKE OR CONVERSION TO '4c02, 0 . 0 1000 ~ L B w z L O > wu - [vd — £ 500 NN 28 w oc [&] < x » - | | | 0 . | BRANCHED-CHAIN AMINO ACID, mM FIGURE 2. Effect of concentration of [U-l4C]leucine and [U-l4C]isoleucine in the perfusate on uptake, intracellular concentration, and conversion to laco,. Hearts were perfused for measurement of leucine uptake and conversion to l4co, as described in "Experimental Procedures.' Between 6 and 10 minutes of perfusion with buffer containing either 0.1 mM [U-l4C]leucine or [U-l4C]isoleucine, coronary effluent was collected from the pulmonary artery and analyzed for lédco, (panel A). Before perfusion was switched to buffer containing the next higher concentra- tion of leucine or isoleucine, the heart was perfused for 15 minutes with buffer to which no branched-chain amino acid was added. The total period of perfusion for each heart was 115 minutes. Three or four hearts were perfused for measure- ment of conversion of [U-l4C]leucine or [U-l4C]isoleucine to 14C0y, respectively. In each case, the buffer contained 145,200 dpm/ml of perfusate. Five hearts were perfused for measurement of leucine uptake at each leucine concentration. The relationship between leucine uptake and concentration in the perfusate was described by the following formula: y = 149x - 12.2. Values represent the mean + S.E. 312 TABLE 1. Effect of Substrate Supply on Uptake and Intracellular Levels of Branched-Chain Amino Acids Langendorff Preparation Working Hearts Branched-Chain Amino Acid in Intracellular Intracellular Perfusate, Uptake, Concentration, Uptake, Concentration, nmol/ml nmol/g/min nmol/ml nmol/g/min nmol/ml 15 mM glucose Leucine, 183 31 +3 71 + 11 40 + 7 ND? Isoleucine, 118 21 + 2 15 + 7 23 + 4 ND Valine, 204 20 + 1 86 + 13 ND ND? 2 mM lactate, 10 mM glucose, 400 uU insulin/ml Leucine, 183 8.9 + 1.1° 154 + oP 15 + 13D 89 + 80 Isoleucine, 118 3.7 * 0.5° 64 * 7° 6.0 + 0.5%P 25 t 53P Valine, 204 2.2 + 0.4 58 + 16 ND ND? 1.5 mM palmitate, 10 mM DL-B-hydroxybutyrate, 8 mM glucose Leucine, 183 14 + 3° 193 + 4° 264 + 1%P 125 + 16%°P Isoleucine, 118 7 + 2° 102 + 7° 11 + 1° 68 + 527° Valine, 204 9 + 4° 74 + 8 12 + 1 46 + 33D Hearts were perfused for 1 hour with buffer that contained 0.01 mM phenylalanine, normal plasma levels of other amino acids, and the substrate mixture as indi- cated. Working hearts were perfused with a 23-gauge needle in the aortic out- flow tract. Values represent the mean + S.E. of four to six hearts. op < 0.05 versus Langendorff preparation supplied the same substrate p < 0.05 versus the same preparation supplied 15 mM glucose ND = None detected 313 was linear for 2 hours. Since intracellular phenylalanine did not accumulate under these circumstances (15), appearance of phenylalanine in the perfusate was a valid index of protein degradation. In hearts that were perfused with buffer that contained only glucose and phenylalanine, protein degradation proceeded at the fastest rate (figure 3). This rate was decreased 25 percent by addition of either five times normal plasma levels of all amino acids or five times normal plasma levels of the branched-chain compounds along with normal plasma levels of other amino acids. When the branched-chain amino acids were not included in the mixture of amino acids at five times the normal plasma levels, the rate of degradation was identical to that observed in hearts perfused with buffer to which only glucose and phenylalanine were added. These findings suggest that 400 Ix ALL Ix BR-CH 5x OTHERS ow oO oO T & QB BK & 030, XX 35 QS & 5 QQ > QQ CQ XX 3 0 x0 5x BR-CHAIN Ix OTHERS 5x ALL 200 100 - PROTEIN DEGRADATION, nmol PHE/g/h 0 : AMINO ACID MIXTURE FIGURE 3. Effect of amino acids on protein degradation. Following 10 minutes of preliminary perfusion with buffer that contained 15 mM glucose, 0.01 mM phenyl- alanine (PHE), and 0.02 mM cycloheximide, hearts were perfused for 2 hours with the same buffer to which various amino acid mixtures were added. Values repre- sent the mean * S.E. of 6 to 10 hearts. BR-CH = Branched-chain amino acids. 314 addition of branched-chain amino acids could inhibit protein degradation. Pre- vious studies indicated that these compounds stimulated protein synthesis (3). Effect of Branched-Chain Amino Acids and Their Metabolites on Protein Degradation In hearts perfused in the absence of insulin, leucine (1 or 2 mM) inhibited protein degradation by approximately 40 percent (figures 4 and 5). Similar ef- fects could not be demonstrated by addition of either isoleucine or valine (1 or 2 mM) to the perfusate (figures 6 and 7). Concentrations of leucine greater than 1 mM (approximately 5 times the normal plasma levels) maximally inhibited the degradative rate in the absence of the hormone. On the other hand, leucine was ineffective in the presence of insulin (figure 4). Insulin significantly reduced the rate of degradation at all leucine concentrations. 300 Ww an : CONTROL I 200 ££ —— oo + INSULIN I — J Oo < E 100} > 2 wl I a. (0) 1 1 1 1 1 0 0.5 1.0 1.5 2.0 LEUCINE(mM) IN THE PERFUSATE FIGURE 4. Effect of leucine and insulin on protein degradation. Hearts were perfused with buffer that contained 15 mM glucose, 0.01 mM phenylalanine, normal plasma levels of amino acids, except for the branched-chain compounds, and 0.02 mM cycloheximide. Leucine and insulin (25 mU/ml) were added, as indicated. Five to 15 hearts were perfused under each condition. Values represent the mean * S.E. 315 300 — 600 200 400 jw /|owu 3ANION3T H¥VINTI3IOVHLNI 100 200 PROTEIN DEGRADATION nmol PHE/g /h Oo LL] ImM NO ADDITIONS Ba LEUCINE, ImM ISOVALERATE, FIGURE 5. Effect of leucine and its metabolites on proteolysis and intracellular leucine. Hearts were perfused as described in figure 4. Intracellular leucine was measured as outlined in "Experimental Procedures." Values represent the mean * S.E. of 6 to 10 hearts. PHE = Phenylalanine. * = p < 0.05 versus no addition. Since the heart oxidized branched-chain amino acids, metabolites of these compounds were added to the perfusate to explore whether leucine played a unique role in regulating protein degradation (figures 5-8). When added at the same concentrations as leucine, the transamination products of leucine and valine, a-ketoisocaproate (figure 5) and o-ketoisovalerate (figure 7) inhibited proteo- lysis. Addition of a-ketoisocaproate increased intracellular leucine concentra- tion. Similar effects were observed with isovalerate (figure 5) and isobutyrate (figure 7), the decarboxylation products of leucine and valine. Tiglic acid, a compound whose CoA derivative is an intermediate in isoleucine oxidation, in- hibited protein degradation when added at a concentration of 10 mM (figure 6). Products of oxidation of branched-chain amino aeids, acetoacetate, acetate, and propionate also inhibited proteolysis, but did not affect intracellular leucine (figure 8). These metabolites are utilized rapidly by heart muscle (16) and were added in concentrations sufficient to support total oxygen consumption over a period of 2 hours. These results indicate that leucine is unique among the branched-chain amino acids in inhibiting proteolysis. However, metabolites of branched-chain amino acids could mimic the leucine effect. Some of these metabo- lites did not increase intracellular leucine. 316 300 -600 200 | - 400 jw / |[owu 3NION3T "YVINTI3OVHLNI 100 4 200 PROTEIN DEGRADATION nmol PHE /g /h | mM ImM I0 mM NO ADDITIONS ISOLEUCINE, 2 mM TIGLIC ACID, FIGURE 6. Effect of isoleucine and its metabolites on proteolysis and intra- cellular leucine. Hearts were perfused as described in figure 4. Values repre- sent the mean + S.E. of 6 to 10 hearts. PHE = Phenylalanine. * =p < 0.05 versus no addition. Effect of Branched-Chain Amino Acids and Their Metabolites on Protein Synthesis In earlier experiments, addition of five times the normal plasma levels of amino acids accelerated protein synthesis; the same effect could be obtained by addition of five times the normal plasma levels of branched-chain amino acids (3). The present experiments demonstrated that leucine, but not isoleucine or valine, stimulated protein synthesis (figure 9). Metabolites of the branched- chain amino acids including isobutyrate, acetoacetate, acetate, and propionate accelerated protein synthesis. Of this group, only acetoacetate increased intracellular leucine. These data confirm earlier findings that oxidizable noncarbohydrate substrates, such as fatty acids, ketone bodies, pyruvate, lac- tate, and acetate, maintained protein synthesis (3). Leucine is unique among the branched-chain amino acids in having this effect. Effect of Leucine on Protein Turnover in Working Hearts As noted above, leucine accelerated protein synthesis and inhibited proteo- lysis in Langendorff preparations provided glucose as an oxidizable substrate. Cardiac work accelerated protein synthesis and inhibited protein degradation 317 300k 600 2 20 o < S 37 = ze N as 2001+ 400 3 § AN C7 Zu o a o 2 —- _— w 2 loo} 200 z 5c oc a 2 0 0 NO ADDITIONS VALINE, | mM 2mM o-KETOISO VALERATE, Im ISOBUTYRATE, ba ImM 10 mM FIGURE 7. Effect of valine and its metabolites on proteolysis and intracellular leucine. Hearts were perfused as described in figure 4. Values represent the mean + S.E. of 6 to 10 hearts. PHE = Phenylalanine. * = p < 0.05 versus no addition. in the absence of leucine (table 2). Addition of leucine to the perfusate of working hearts had no further effect on protein synthesis, but significantly reduced the rate of degradation. DISCUSSION Leucine decreased the negative nitrogen balance that was characteristic of both Langendorff preparations and working hearts that were perfused with buffer containing glucose and normal plasma levels of amino acids (table 2). However, the rate of protein synthesis was only 60 percent of the rate of degra- dation in the working preparation that was supplied leucine. None of the other amino acids, including isoleucine and valine, affected synthesis or degradation in the perfused rat heart (figures 3, and 5-7). In comparison to the effect of insulin, addition of leucine had only a modest effect on either synthesis or degradation. On the other hand, the combined effect of leucine addition and cardiac work on proteolysis was as large as the effect of insulin (table 2 and figure 4). The mechanisms of the effects of leucine on protein synthesis and 318 -1 600 200 <1 400 jw /|owu ANION3T ¥VINTT3OVYHLNI nmol PHE/g/h 100 1 200 PROTEIN DEGRADATION o i] NO ADDITIONS ACETOACETATE}: 10 mM 20 mM PROPIONATE, I0mM ACETATE, FIGURE 8. Effect of metabolites of branched-chain amino acids on proteolysis and intracellular leucine. Hearts were perfused as described in figure 4. Values represent the mean + S.E. of 6 to 11 hearts. PHE = Phenylalanine. * = p < 0.05 versus no addition. degradation are incompletely understood. The stimulatory effect on protein synthesis appeared to involve acceleration of peptide chain initiation (2,3), while the inhibition of protein degradation by leucine was associated with an increase in lysosomal latency (16). Leucine was utilized rapidly by the perfused rat heart (4,17) (figure 2). At a concentration in the perfusate of 1 mM, approximately 85 percent of the leucine that was taken up was decarboxylated and further metabolized. If the leucine that was metabolized was converted quantitatively to acetyl CoA and oxidized in the citric acid cycle, metabolism of the amino acid would account for 36 percent of myocardial oxygen consumption (4.3 umol/g heart + min) (10) when the concentration of leucine in the perfusate was raised to 2 mM. These data should not be interpreted to indicate that leucine is a major oxidative substrate in vivo. As shown in table 1, leucine uptake was markedly inhibited in the presence of physiological mixtures of substrates and would account for only 1.4 to 3.7 percent of oxygen consumption in Langendorff and working prep- arations. Instead, these data indicate that leucine oxidation is closely 319 100 200 T 50 |w/owu‘INION3T YY INTIIOVHLNI » Ei PROTEIN SYNTHESIS, nmol PHE /g/ h EF] ° EE = 8 4 4 b a Z Y¥ 3s EZ OF ofs Sf 05% S 8 J gE Zo 3° w3E wo &o eS 4 £ a” a = {2 gv £7 FIGURE 9. Effect of branched-chain amino acids and their metabolites on protein synthesis. Hearts were perfused for 2 hours with buffer that contained 15 mM glucose, 0.4 mM [l4C]phenylalanine (PHE), and normal plasma levels of other amino acids, except for the branched-chain compounds. Additions were as indi- cated. Values represent the mean of 6 to 12 observations. * = p < 0.05 versus no addition. controlled and depends upon leucine concentration, the presence of other oxi- dizable substrates, and the level of ventricular pressure development. Uptake and conversion of isoleucine and valine to CO) occurred at a rate that was approximately 40 percent of that observed with leucine (table 1 and figure 2). These compounds are thought to share the same transaminase (18) and perhaps the same branched-chain o-ketoacid dehydrogenase. In this regard, early studies (19) indicated that more than one dehydrogenase might be involved, but recent experiments aimed at purification and characterization of the dehydrog- enase from rat liver and muscle have indicated that a single enzyme is present (20-22). In contrast to the relative rates of uptake of leucine and valine in the perfused heart, a-ketoisovaleric acid was decarboxylated by heart mitochon- dria at a rate that was two to three times faster than that of a-ketoisocaproic acid (23). These results could be explained by greater accumulation of a pro- duct of valine metabolism, such as isobutryl-CoA, than of the decarboxylation product of leucine, isovaleryl-CoA. The latter compound competitively inhibited the dehydrogenase from liver (21). Different steady-state levels of intermedi- ates in the pathways of isoleucine and valine metabolism, as compared to leucine metabolism, would not be unexpected because the former compounds enter the 320 citric acid cycle as succinyl CoA and contribute to the carbon pool. Net oxida- tion of isoleucine and valine is achieved only if cycle intermediates are con- verted to pyruvate and then to acetyl-CoA (24,25). On the other hand, leucine is converted directly to acetyl CoA and oxidized to COp. Additional studies are needed to relate rates of metabolism of branched-chain amino acids to steady- state levels of intermediates in the degradative pathway. Neither uptake, oxidation, nor intracellular accumulation of leucine corre- lated with the effect of the amino acid on protein degradation (figures 2 and 4). Uptake, oxidation, and intracellular accumulation increased progressively as the concentration in the perfusate was raised to 2 mM and above. In contrast, the maximal inhibitory effect of leucine on proteolysis was achieved when the leucine concentration in the perfusate reached 1 mM. Similarly, the increase in synthesis and decrease in degradation that resulted from provision of acetate and proprionate, for example, was not associated with an increase in concentra- tion of intracellular leucine (figures 8 and 9). These findings suggest that factors other than intracellular concentration and oxidation of leucine, such as tissue levels of metabolites, may be involved in inhibition of protein breakdown. Metabolites of leucine that could be re- lated to acceleration of protein synthesis and inhibition of protein degradation include (a) intermediates involved in protein synthesis, or (b) metabolites whose intracellular concentration changed as a result of leucine transamination, TABLE 2. Effect of Leucine on Protein Turnover in Working Rat Hearts Leucine Protein Synthesis Protein Degradation Preparation 1 mM nmol phenylalanine/g/h Langendorff 0 82 + 4 272 = 9 + 102 + 22 235 + 12° b b Working 0 112 + 4 243 + 7 + 107 + 2 172 + 7%P Hearts were perfused for 2 hours with buffer that contained 15 mM glucose and normal plasma levels of amino acids, except leucine and phenylalanine. Leucine was added as indicated. When protein synthesis was measured, [1%4C]phenylalanine was 0.4 mM; when protein degradation was measured, phenylalanine was omitted. Working hearts were perfused with a 20-gauge needle in the aortic outflow tract. Four to eight hearts were perfused under each condition. Values represent the mean * S.E. a . . bP < 0.05 versus no leucine, same preparation p < 0.05 versus Langendorff preparation supplied the same concentration of leucine 321 decarboxylation, and oxidation. Alternatively, because amino acids are com- partmented in heart (12), an increase in leucine concentration in a particular compartment of the cell could be responsible for the inhibitory effect (7). The regulatory mechanism can be identified only when measurements of metabolites have been carried out in hearts provided leucine and other noncarbohydrate substrates and when effects of potential regulators have been explored in cell-free systems that carry out peptide chain initiation and proteolysis. 322 10. 11. 12. 13. 14. REFERENCES Morgan HE, Earl DCN, Broadus A, Wolpert EB, Giger KE, Jefferson LS: Regu- lation of protein synthesis in heart muscle. I. Effect of amino acid levels on protein synthesis. J Biol Chem 246:2152, 1971 Morgan HE, Jefferson LS, Wolpert EB, Rannels DE: Regulation of protein synthesis in heart muscle. II. Effects of amino acids and insulin on ribosomal aggregation. J Biol Chem 246:2163, 1971 Rannels DE, Hjalmarson AC, Morgan HE: Effects of non-carbohydrate sub- strates on protein synthesis in muscle. Am J Physiol 226:528, 1974 Buse MG, Biggers JF, Friderici KH, Buse JF: Oxidation of branched-chain amino acids by isolated hearts and diaphragm of the rat. J Biol Chem 247:8085, 1972 Fulks RM, Li JB, Goldberg AL: Effects of insulin, glucose, and amino acids on protein turnover in diaphragm. J Biol Chem 250:290, 1975 Buse MG, Reid SS: Leucine: A possible regulator of protein turnover in muscle. J Clin Invest 56:1250, 1975 Buse MG, Weigand DA: Studies concerning the specificity of the effect of leucine on turnover of proteins in muscles of control and diabetic rats. Biochim Biophys Acta 475:81, 1977 Li JB, Jefferson LS: Influence of amino acid availability on protein turnover in perfused skeletal muscle. Biochim Biophys Acta 544:351, 1978 Morgan HE, Neely JR, Wood RE, Liebecq C, Liebermeister H, Park CR: Factors affecting the transport of glucose in heart muscle and erythrocytes. Fed Proc 24:1040, 1965 Neely JR, Liebermeister H, Battersby EJ, Morgan HE: Effects of pressure development on oxygen consumption by the isolated rat heart. Am J Physiol 212:804, 1967 Hems R, Ross BD, Berry MN, Krebs HA: Gluconeogenesis in perfused rat liver. Biochem J 101:284, 1966 McKee EE, Cheung JY, Rannels DE, Morgan HE: Measurement of the rate of protein synthesis and compartmentation of heart phenylalanine. J Biol Chem 253:1030, 1978 Rubin IB, Goldstein G: An ultrasensitive isotope dilution method for the determination of L-amino acids. Anal Biochem 33:244, 1970 Chua B, Kao R, Rannels DE, Morgan HE: Hormonal and metabolic control of proteolysis. Biochem Soc Symp 43:1, 1978 323 15. 16. 17. 18. 19. 20. 21. 22. 23. 24, 25. Chua B, Kao RL, Rannels DE, Morgan HE: Inhibition of protein degradation by anoxia and ischemia in perfused rat heart. J Biol Chem 254:6617, 1979 Chua B, Siehl DL, Morgan HE: Effect of leucine and metabolites of branched-chain amino acids on protein turnover in heart. J Biol Chem, in press Clarke EW: A simplified heart-oxygenator preparation suitable for isotope experiments: With some observations on the metabolism of acetate, pyruvate and amino acids. J Physiol 136:380, 1957 Ichihara A, Noda C, Ogawa K: Control of leucine metabolism with special reference to branched-chain amino acid transaminase. Adv Enzyme Regul 11:155, 1973 Connelly JL, Danner DJ, Bowden JA: Branched-chain o-keto acid metabolism. J Biol Chem 243:1198, 1968 Danner DJ, Lemmon SK, Elsas LJ II: Substrate specificity and stabilization by thiamine pyrophosphate of rat liver branched-chain a-ketoacid dehydrog- enase. Biochem Med 19:27, 1978 Parker PJ, Randle PJ: Partial purification and properties of branched- chain 2-oxoacid dehydrogenase of ox liver. Biochem J 171:751, 1978 Odessey R, Goldberg AL: Leucine degradation in cell-free extracts of skeletal muscle. Biochem J 178:475, 1979 Bremer J, Davis EJ: The effect of acylcarnitines on the oxidation of branched-chain a-keto acids in mitochondria. Biochim Biophys Acta 528: 269, 1978 Goldstein L, Newsholme EA: The formation of alanine from amino acids in diaphragm muscle of the rat. Biochem J 154:555, 1976 Chang TW, Goldberg AL: The metabolic fates of amino acids and the forma- tion of glutamine in skeletal muscle. J Biol Chem 253:3685, 1978 324 CHANGE IN ANTIGENIC PROPERTIES OF RAT MYOCARDIAL CHROMATIN UPON ADAPTATION TO HYPOXIA N. A. Fedorova, I. V. Muzurov, and L. B. Nesterchuk INTRODUCTION Adaptation to hypoxia characteristically involves an increase in the work performance of the respiratory and circulatory systems, as well as an increase in the capacity of cells to utilize oxygen. The latter primarily results from an increase in the biogenesis of mitochondria (1,2). Since only 10 percent of the proteins in mitochondria are coded by mitochondrial DNA, the synthesis and assembly of component parts of mitochondria are, obviously, complex and multi- stage processes under the dual control of chromosomal and mitochondrial genes (3). Consequently, one can assume that an increase in the biogenesis of mito- chondria with adaptation to hypoxia will be accompanied by increased transcrip- tion of the nuclear genome. Comparative studies of the antigenic properties of chromatin are of inter- est because changes in the matrix activity of chromatin result from modification of its protein components and/or variations in their spatial organization. We undertook this type of study using immunochemical methods which enabled us to determine not only changes in the composition of the proteins but also in the three-dimensional organization of their antigenic determinants. MATERIALS AND METHODS Adaptation to periodic hypoxia was achieved by placing rats in a pressure chamber for 6 hours each day for 30 days at a simulated altitude of 7 km. Chromatin was extracted from the nuclei of the cardiac muscle according to Widnell et al. (4). Antiserum was obtained through immunization of chinchilla rabbits. Chromatin preparations were administered intramuscularly with Freund adju- vant (1:1). Reimmunization was performed after 45-60 days. Approximately 40 mg of cardiac muscle chromatin protein was administered during the entire cycle of immunization. In some experiments, we used the gamma globulin fraction of From the Laboratory of Protein Chemistry, Leningrad State University, Leningrad, USSR. 325 immune serum (5). Comparison of cardiac muscle chromatin proteins of animals adapted to hypoxia and control animals was made using a double immunodiffusion micrometer in agar—agarose gel according to the method of Ouchterlony as modi- fied by Gusev and Tsvetkov (6). RESULTS AND DISCUSSION When the double immunodiffusion reaction was performed, clear differences were found between the control animals and those adapted to hypoxia in the spectra of the precipitation lines formed by the antiserum to the total cardiac muscle chromatin. Adaptation to hypoxia involved enrichment of the spectrum of precipitation lines by two lines. One of these lines was located in the immediate vicinity of the well into which the antiserum was placed, while the other formed a wall near the well filled with antigen (figure 1). The location of the additional precipitation lines indicates the appearance of antigenic determinants in the chromatin of the adapted animals, differing sig- nificantly in rate of diffusion in the gel and, consequently, having different molecular weights. Because we used antiserum to the chromatin of the control animals, we can assume that quantitative differences were found in the immuno- genic properties of the chromatin of both the control animals and the animals adapted to hypoxia. In other words, during adaptation to hypoxia, an addi- tional quantity of antigenic determinants appears in the cardiac muscle chro- matin; some of these determinants were opened and had immunogenic properties in the cardiac muscle chromatin of the control animals. FIGURE 1. The lines of precipitation formed by the antiserum to rat heart chromatin. As = Antiserum. Cha = Chromatin of the animals adapted to hypoxia. Che = Chromatin of the control animals. 326 Analogous results were obtained earlier in comparing the antigenic proper- ties of chromatin in rat brain and liver, following adaptation to periodic hypoxia and in control animals (7). It can be assumed that, upon adaptation to hypoxia, the chromatin of tissues, particularly those sensitive to oxygen deficiencies, undergoes significant conformational restructuring leading to changes in the degree of shielding of the antigenic groups. Data obtained in in vitro experiments also indicate that oxygen shortage causes a change in protein composition, structure, and functional activity of the nuclear genome. In these studies, a decrease in the partial pressure of oxygen led to formation of puffs in several sections of the polytene chromo- somes of Drosophila, indicating a local increase in transcription (8,9). When the oxygen content was reduced to 5 percent, an increase by a factor of more than 3 was noted in the number of polyamines, with an increase in histone acetylation in a culture of chick embryo heart ventricle cells (10). At present no data are available to explain the molecular mechanisms of action of oxygen on the amount of polyamines and histone acetylation. However, accumulation of polyamines and a sharp increase in the degree of histone acety- lation indicate a possible reason for restructuring of the nuclear genome and an increase in its transcription under conditions of hypoxia. The change which we found in the spectrum of surface-located antigenic determinants of cardiac muscle chromatin apparently results from modification of certain proteins and changes in the spatial organization of chromatin. Such conformational restructuring of genetic material is probably one of the neces- sary elements occurring at the cellular level in adaptation to hypoxia. Tran- scription apparently increases in the areas of the genome which control synthesis of the enzymes of oxidative phosphorylation and the cytochromes. CONCLUSIONS Further studies of changes in the antigenic properties of chromatin under extreme influences, and adaptation to these changes, should be combined with identification of the proteins and their complexes resmpnsible for the forma- tion of the spectrum of "open" antigenic determinants of chromatin, which de- fine its immune properties. Until now, significant differences in the chromatin protein spectra of identical rat tissue had been noted in in vivo experiments only in such processes as regeneration of the liver. In this study, we found changes which were related neither to changes in the type of proliferation nor to the process of cell dif- ferentiation. These changes can only reflect a certain functional restructuring of the metabolism. 327 10. REFERENCES Meerson FZ, Pomoinitskii VD, Yampolskaia BA: [The role of biogenesis of mitochondria in adaptation of the organism to high altitude hypoxia] (Rus). Dokl Akad Nauk SSSR 203:973-976, 1972 Meerson FZ: [Heart Adaptation to Heavy Load and Heart Insufficiency] (Rus). Moscow, Nauka, 1975 Neyfakh SA: [Mechanisms of cellular integration in mitochondrial biogenesis] (Rus). In [Genetic Functions of Cytoplasmic Organoids], edited by SA Neyfakh. Leningrad, Nauka, 1974, pp 58-70 Widnell CC, Hamilton IA, Tata JR: The isolation of enzymically active nuclei from the rat heart and uterus. J Cell Biol 32:766-770, 1967 Levy HB, Sober HA: A simple chromatographic method for preparation of gamma globulin. Proc Soc Exper Biol Med 103:250-252, 1960 Gusev AI, Tsvetkov VS: [Techniques of performing microprecipitation in agar] (Rus). Lab Delo 2:43-45, 1961 Ashmarin IP, Konarev VG, Sidorova VV, Simanovskii LN, Fedorova NA: [Change in protein composition of rat brain and liver chromatin during functional reconstruction of cells. The effect of adaptation to hypoxia] (Rus). Dokl Akad Nauk SSSR 228:222-224, 1976 Leenders HJ, Berendes HD: The effect of changes in the respiratory metabolism upon genome activity in drosophila. 1. The induction of gene activity. Chromosoma 37:433-444, 1972 Koninkx JG, Leenders HJ, Birt LM: A correlation between newly induced gene activity and enhancement of mitochondrial enzyme activity in the salivary glands. Exp Cell Res 92:275-282, 1975 Clo C, Orlandini , Guarnieri C, Caldarera CM: Role of oxygen on growth rate and gene activity in cultured chick-embryo heart cells. Biochem J 154:253-256, 1976 328 STUDY OF NITROGEN METABOLISM IN THE CARDIAC MUSCLE USING THE ISOTOPE 19N 0. I. Pisarenko, A. V. Artemov, and V. N. Smirnov INTRODUCTION This work is dedicated to the study of one aspect of nitrogen metabolism-- the mechanisms of bonding of free ammonia in the heart muscle. In the works of a number of authors using experimental models (1,2) and studies on patients suffering from various heart diseases (3-5), the formation of ammonia has been reported on contraction of the heart muscle, with expulsion into the blood of the coronary sinus. Studies performed on patients have shown that the pro- cesses of formation and bonding of ammonia in the heart are quite labile, that ammonia content in the blood of the coronary sinus may vary widely, and that both absorption and liberation of this compound by the heart muscle are ob- served (3,5-7). Regulation and testing of the level of free ammonia in the cardiac muscle, involving intensive aerobic metabolism, is of particular im- portance since this substance has strong cytotoxic properties (8). Study of the coronary arteriovenous difference of the nitrogenous compo- nents of the blood, performed in animal models (3,9,10) and in patients with various cardiac pathologies, has shown that only a slight portion of the ammo- nia is liberated into the bloodstream in free form. The ammonia formed upon contraction of the cardiac muscle can be bonded by 2-oxoglutarate in the glu- tamate dehydrogenase reaction (9). Another method of detoxication of ammonia from the heart is the formation of glutamine (9,11,12). There are reports of the possibility of removal of ammonia in the form of transamination products of glutamic acid--alanine and aspartic acid (12,13), as well as by synthesis of urea, which has a tendency to increase with myocardial infarction (12,4). In connection with this, there is interest in studying the degree of partici- pation of each of these mechanisms in the neutralization of ammonia in the heart, and in determining the role of synthesis of urea as an additional means of elimination of ammonia in the intact heart and in the heart under experi- mental myocardial infarction, using the isotope 15, From the Laboratory of Myocardial Metabolism, All-Union Cardiology Research Center, USSR Academy of Medical Sciences, Moscow, USSR. 329 MATERIALS AND METHODS Reagents and Isotopes We used lithium citrate, lithium oxide hydrate, sodium acetate, picric acid, ninhydrin, and hydrochloric acid (from Merck, FRG); l-ornithine, 1- glutamine, l-arginine, l-leucine, pyruvate, and isoproterenol produced by Nutritional Biochemicals Co. (USA); and ammonium chloride (from the USSR). The solutions were prepared in ammonia-free water (Zerolit demineralite, England). Ammonium acetate made with 15 (95.8 atom percent) was obtained from VEB Berlin-Chemie (GDR). Ion exchange resins were obtained from the following firms: M-72 from Beckman (USA), Amberlit CG-120 III, 400 mesh from Serva (FRG), Dowex 2 x 8, 200 mesh from Bio-Rad (USA). Perfusion of Isolated Rat Heart With Ammonium Salts and 1-Amino Acids All experiments on the metabolism of 15N-ammonia and synthesis of urea were performed on isolated rat heart models in order to eliminate the influ- ence of enzyme systems in the blood and other organs on the metabolism of the nitrogenous compounds. Male Wistar rats weighing 200-230 g were used. The animals were anesthetized with nembutal, and the hearts were removed and per- fused by Langendorf's method using oxygenated (95 percent 05:5 percent C09) Krebs-Henseleit bicarbonate buffer containing heparin (1 mg/ml) for 3 minutes (without recirculation) at 60 mm Hg. After the blood was washed from the heart, recirculation perfusion with 25 ml buffer was performed. To study the synthesis of amino acids and urea from ammonia, ammonium acetate (95.8 atom percent 15N) was added to the buffer to concentrations of 0.1, 1.0, 1.6, 2.2, 2.8, and 3.4 mM. Recirculation perfusion with the 15N- ammonium acetate was performed for 30 minutes. The heart was then perfused for 3 minutes with a buffer not containing the label, frozen in liquid nitro- gen, and carefully ground to powder. To evaluate the influence of the various substrates on the synthesis of urea, recirculation perfusion was performed with a bicarbonate buffer contain- ing ammonium chloride (10 mM), l-ornithine (10 mM), l-arginine (10 mM), 1- glutamine (10 mM), l-leucine (5 mM), and pyruvate (5 mM). The heart tissue was treated by the method described above. Perfusion of hearts with experimental myocardial infarction with 15N- ammonium acetate was performed under the same conditions as those used for intact hearts. Myocardial necrosis was evoked in rats in vivo by three in- traperitoneal injections of isoproterenol (25 mg/kg body weight) at intervals of 24 hours. Its presence in the subendocardial layers of the myocardium was confirmed morphologically. Three hours after the third injection of isoproterenol, the hearts were extracted from the anesthetized animals for perfusion. 330 Preparation of Tissue and Perfusate for Determination of Content of Nitrogenous Metabolites Extraction of amino acids and urea from the heart tissue was performed with perchloric acid. Six percent HC104 (5 mg/g tissue) was added to the tis- sue powder frozen in liquid nitrogen. After 5 minutes of thawing at room tem- perature, the protein was separated by centrifugation at 5,000 rpm for 10 minutes at 0° C. The supernatant was adjusted to pH 7.0 with 2 N KOH and placed in ice for 30 minutes. The KC1l0, precipitate was removed by centrifu- gation at 5,000 rpm for 15 minutes at 0° C. The supernatant was evaporated in a vacuum at 40° C to the minimal volume, excluding precipitation of salts, and then acidified to pH 2.2 with 2 N HCI. The proteins in the perfusate were precipitated with 3 percent HC1lO4 (5 ml/ml perfusate). All subsequent operations were analogous to those performed above. Concentrations of nitrogen exchange metabolites in the protein-free tissue extracts and perfusates were determined by means of a Beckman M-121 amino acid analyzer. Analysis was performed by a single-column method with a lithium citrate buffer (14). Norleucine was the internal standard. Preparation of Specimens of 15N-Amino Acids and 15N-Urea for Isotope Analysis After perfusion of the hearts with 15N-ammonium acetate, the 15N-amino acid and urea were separated from the protein-free tissue extracts. The spec- imens of extract and perfusates obtained from 5 g of wet tissue were applied to the preparative column (2.5 x 60 cm) of an amino acid analyzer filled with Amberlite CG-120 III, 400 mesh resin. Elution was performed with a pH 2.80 lithium citrate buffer at 260 ml/hr at 40° C. The amino acid and urea frac- tions were purified to remove buffer components by means of ion exchange resins (15). The eluates were evaporated under a vacuum at 40° C. The dry residues containing amino acids and urea were burned in a CHN analyzer in the presence of NiO in a current of helium. The products of com- plete oxidation (Np, COp, H20) were separated on the chromatographic column of a CHN analyzer filled with Polysorb-1l. The nitrogen was trapped with 50 mg of CaA zeolite in a glass ampule cooled with liquid nitrogen. The ampules were preliminarily degassed in a current of helium at 300° C. The quantity of ni- trogen in each specimen was determined from the area of the peak on the chro- matogram. The isotope analysis required 100 ug of nitrogen in the sample. The content of 15N in the amino acids and urea was determined on an MI- 1305 mass spectrometer. Statistical Processing The significance of the differences between data obtained in the experi- ments on intact and infarcted hearts was evaluated using Students t-test. 331 Kinetic analysis of the effect of l-amino acids on the synthesis of urea was performed using a special computer program on the MIR-1 computer. RESULTS Enrichment With 15N-Amino Acids and Urea Upon Perfusion of Hearts With 15N-Ammonium Acetate Before performing experiments on the transfer of 15N from l15N-ammonium acetate to amino acids and urea in rat cardiac muscle, it was necessary to establish the optimal concentration of the label in the perfusate and the op- timal duration of perfusion. The concentration of 15N-ammonium acetate had to satisfy two criteria: It had to be high enough to allow measurement of the level of inclusion in the nitrogenous compounds, but not so high as to lead to metabolic disorders in the myocardium due to hyperammonemia. This condition was found to correspond to a concentration of 15N-ammonium acetate of 2.2 mM, approximately 1.2 times the tissue concentration of ammonia. The ammonia con- centration gradient in the heart-perfusate system should facilitate diffusion of 15N-ammonium acetate into the myocardial cells, thus assisting in the inclu- sion of 15N in the compounds studied. As the studies of Davidson and Sonnen- blick have shown, perfusion of the intact rat heart with ammonium chloride of practically the same concentration (2.06 mM) leads to a significant increase in the ammonia content in heart tissue (16). As the concentration of the label is decreased to 1.0 mM, enrichment of the ammonia, aspartic acid, and alanine become comparable to the experimental error in measurement of 15x. Preliminary experiments showed that 30 minutes of perfusion with 2.2 mM of l5N-ammonium acetate was sufficient to record the incorporation into the amino acids and ammonia significantly higher than the measurement error (* 0.005-0.010 atom percent). Longer perfusion (60 minutes) resulted in a de- crease in the uptake of 15N. One reason for this might be the dilution of the label with unlabeled ammonia formed upon operation of the heart and accumulat- ing in the perfusate. After 60 minutes of perfusion of the heart with 2.2 mM 15N-ammonium acetate, the quantity of ammonia in the perfusate had increased to an average of 2.3 mM (p < 0.05). To illustrate these aspects of the inclusion of 15N in the nitrogenous compounds, table 1 presents data on the enrichment of urea with 15N, obtained under various experimental conditions. Perfusion of the hearts with 0.1 mM 15N-ammonium acetate (i.e., five times excess of ammonia in comparison to its physiologic content in the blood plasma) did not lead to inclusion of 15N in the urea. 15N was found in the urea when 1.0 mM 15N-ammonia was administered to the heart for 30 minutes. However, in this case the increase was very close to the experimental error. After perfusion of the heart for 1 hour with the same concentration of l5N-ammonium acetate, the inclusion of 15N into the urea decreased to less than half. It became possible to measure the content of 15N when the concentration of labeled ammonia was increased by 100-150 times in comparison to its concentration in the blood plasma, i.e., in the range of 1.6-3.4 mM. The enrichment of urea with 15N increased with increasing concen- tration of the label in the perfusate. Higher levels of inclusion of 15N were recorded after 30 minutes of perfusion. 332 TABLE 1. Inclusion of 15N in Urea of the Intact Rat Heart (Atom Percent Excess 15N)* Upon Perfusion With 15N-Ammonium Acetate (95.8 Atom Percent) Concentration of 15N-ammonium acetate in perfusate (mM) Perfusion time (min) | 0.1 1.0 1.6 2.8 3.4 30 0 0.011 + 0.008 [0.062 + 0.005 [0.098 + 0.006 | 0.122 + 0.005 60 0 0.006 + 0.005 [0.042 + 0.006 - 0.095 + 0.007 *Mean + error of mean, n = 3-6. In subsequent experiments, the synthesis of 15N-amino acids and 15N-urea was studied by perfusion of the hearts with 2.2 mM l5N-ammonium acetate for 30 minutes. To explain the variation in inclusion of 15N in the urea of the infarcted heart, experiments were performed with 15N-ammonium acetate with concentrations of 1.6, 2.8, and 3.4 mM. Further increases in the concentra- tion of the label were considered undesirable due to the possibility of cyto- toxic effect by the ammonia. Nonenzymatic Enrichment of Nitrogenous Compounds With 15x Incorporation into urea (as in the other compounds studied) can result not only from enzymatic enrichment, but also from contamination with 15N- ammonia during isolation from the biologic material. In order to evaluate the contamination of amino acids and ammonia during their separation on an ion exchange column, experiments were performed with an artificial mixture of unlabeled compounds and 15N-ammonium acetate. The quantity of unlabeled amino acids and urea in the mixture corresponded to their content in 5 g of rat heart tissue. The 19N-ammonium acetate (95.8 atom percent) was taken in the quantity necessary for perfusion of five hearts, considering 4.0 mM of the solution. The mixture was applied to the preparative column of the amino acid analyzer. The subsequent stages of nitrogen separation were unchanged. The enrichment of urea with 19N under these conditions was 0.005 atom percent ex- cess 15, and the amino acids averaged 0.010-0.012 atom percent excess 15N, which was only twice the experimental error. Incubation of this same mixture in Krebs-Henseleit bicarbonate buffer for 60 minutes at 40° C with subsequent separation on the column of the amino acid analyzer showed that isotope exchange between 15N-ammonium acetate and urea amounts to not over 0.003, and of amino acids 0.005-0.008 atom percent excess 15N. Thus, enrichment of urea with 15N due to nonenzymatic processes in the experiment on perfusion of 15N- ammonium acetate at a concentration of 1.6-3.4 mM should not exceed an average of 9 percent of the values indicated in table 1. 333 We note that the enrichment of urea with 15N was the lowest achieved for any of the compounds studied. Inclusion of 15N in Nitrogenous Compounds of Intact and Infarcted Hearts Data on the inclusion of 19N in amino acids and urea in the intact rat heart after 30 minutes perfusion with 2.2 mM l5N-ammonium acetate are pre- sented in table 2. The transfer from l15N-ammonium acetate was observed in all of the compounds studied. The highest inclusion of 15N was noted in glutamic acid and glutamine. Aspartic acid and alanine were enriched to a lesser extent. The differences between the isotope composition of nitrogen in amino acids were relatively slight. The urea was enriched with 15N significantly less than the amino acids. In 30 minutes of recirculation perfusion with 2.2 mM l5N-ammonium acetate, equilibrium was established in the heart-perfusate system, characterized by equality of the concentrations of 15N in the compounds in the myocardial tis- sue and the medium. Special measurement showed that the 15N enrichment of the same substance liberated from the perfusion medium and the heart tissue was the same. This factor allowed us to separate urea for isotope analysis from the perfusates (due to its significant accumulation in them as a result of diffu- sion from the heart during perfusion), while taking amino acids from the heart muscle extracts. Since we know that the metabolism of ammonia upon myocardial infarction undergoes significant changes, we were interested to see what would be the changes in the synthesis of 15N-amino acids and 15N-urea in the infarcted rat TABLE 2. Inclusion of 15N in Amino Acids and Urea of Intact and Infarcted Rat Hearts (Atom Percent Excess 19N)* After 30 Minutes Perfusion With 2.2 mM 15N-Ammonium Acetate (95.8 Atom Percent) Compound Intact heart Infarcted Heart Aspartic acid 0.40 =£ 0.01 0.21 + 0.01 Alanine 0.316 + 0.007 0.181 + 0.007 Glutamic acid 0.560 + 0.007 0.46 + 0.02 Glutamine 0.481 + 0.008 0.37 + 0.01 Urea 0.078 + 0.006 0.126 + 0.006 “Mean + error of mean, n = 3-5. 334 heart muscle. The results on inclusion of 19N in the nitrogenous compounds of the infarcted heart were compared with data on the 15N enrichment of these compounds in the intact heart in table 2. Experimental myocardial infarction led to a decrease in the inclusion of 15N in all amino acids. The decrease of inclusion of 15N in aspartic acid and alanine (by 48 and 42 percent, respec-— tively, p < 0.02) was particularly great. The decrease in concentration of 15N in these amino acids corresponds to sharp depression of alanine- and aspartamino- transferase, observed upon damage to the myocardium caused by isoproterenol (17). The decrease in the content of 15N in the transamination products of glutamic acid and glutamine (by 23 percent, p < 0.02) may also be a result of the slight enrichment of the glutamic acid itself with 15N. If we consider that the trans- fer of 15N from the labeled ammonia to the amino group of glutamate occurs ba- sically upon reducing amination of 2-oxoglutarate, the decrease in enrichment of this amino acid (by 20 percent, p < 0.05) must indicate a decrease in the activity of glutamate dehydrogenase in the infarcted heart. In contrast to the amino acids, the formation of l5N-urea in the infarcted heart increases. The increase in the inclusion of 15N in urea is 0.048 + 0.006 atom percent excess L5N, corresponding to 62 percent (p < 0.02). The synthesis of 15N-urea in the heart with experimental myocardial infarc- tion was studied with various concentrations of 15N-ammonium acetate (figure 1). These experiments showed that the degree of enrichment of 15N-urea increases with an increase in the content of l5N-ammonia in the perfusion medium. For all concentrations of l5N-ammonium acetate, the value of 15N enrichment of urea in the infarcted heart was greater than in the intact heart (table 1). Content of Amino Acids and Urea in Heart-Perfusate System After Perfusion With 15N-Ammonium Acetate To determine the quantitative relationships between nitrogenous compounds synthesized upon perfusion of the heart with 15N-ammonium acetate, their con- tent in the heart tissue and perfusate was assessed. The need to consider the total content of nitrogenous compounds in the heart-perfusate system resulted from their diffusion from the myocardial cells into the medium during perfu- sion. Data on the content of amino acids and urea in the heart-perfusate sys- tem after perfusion of intact and infarcted hearts using labeled ammonia are presented in table 3. In systems of intact and infarcted hearts, the content of the metabolites of nitrogen metabolism basically agreed with their concentration in nonperfused rat hearts washed clear of blood. Significant changes in the relationship of metabolites after perfusion of the hearts with 2.2 mM of ammonium acetate were not detected. As in rat heart tissue, the highest content was noted for gluta- mine and glutamic acid. These were followed, in order of decreasing concentra- tion, by alanine, urea, and aspartic acid. In the system consisting of the infarcted heart plus perfusate, our atten- tion is drawn by the significant loss of glutamine-—the main carrier of ammonia in the myocardium (1.03 + 0.18 umol/g tissue, p < 0.02) and the increase in the content of urea (by 0.15 * 0.06 umol/g tissue, p < 0.05) in comparison to the 335 T I | Ra § - 0.15 / ® - 01S . / n — Yd 4 S$ I / ~ 7 . = 00 — / —0I0 LN rt 7 | 3 4 o / = i a [&) / 5 / 005 / 0.05 Be / i , n = oS : 7 ° | [— 0.05). The concentrations of ala- nine in the tissue and perfusate of the intact and infarcted hearts coincide. The total content of nitrogenous compounds taking part in bonding of ammonia in the tissue and perfusate of the infarcted heart is lower than in the intact heart-perfusate system. The content of urea in the tissue and perfusate of the infarcted heart after its perfusion with Krebs-Henseleit buffer without 15N-ammonium acetate was higher than in the system of the intact heart (1.39 * 0.07 and 1.24 * 0.05 umol/g tissue, respectively). The increase in the content of urea in the in- farcted heart-perfusate system was also retained when the hearts were loaded with various concentrations of 15N-ammonium acetate (figure 2). The linear nature of the increase in urea content with increasing concentration of ammo- nia corresponded to a variation in the enrichment of this compound with 15N as a function of the concentration of the label (figure 1). 336 TABLE 3. Total Content of Amino Acids and Urea (umol/g Wet Tissue) in Rat Heart Tissue and Perfusate After 30 Minutes Perfusion With 2.2 mM 15N-Ammonium Acetate (95.8 Atom Percent) Compound Control* Intact Heart®* Infarcted Heart*¥* Aspartic acid 0.62 + 0.08 0.76 + 0.10 0.98 + 0.12 Alanine 1.22 + 0.16 1.92 + 0.14 1.91 + 0.15 Glutamic acid 3.08 + 0.20 3.75 + 0.24 4.22 + 0.28 Glutamine 3.68 + 0.21 4.93 + 0.12 3.90 + 0.18 Urea 1.00 + 0.08 1.44 + 0.05 1.59 + 0.06 *Content of amino acid and urea in intact unperfused rat heart washed of blood with 25 ml Krebs-Henseleit buffer; n = 6. *#%n = 3-5. Values presented are mean * error of mean. Content of 19N in Amino Acids and Urea of the Intact and Infarcted Heart Based on the data on 19N enrichment of amino acids and urea and their content in heart tissue and perfusion medium, we determined the total quan- tity of 15N which had entered each of the compounds during perfusion with 15N- ammonium acetate. The content of 15N in the compounds (in pg-atom 15N/g tis- sue) was calculated by the equation: _ meatom¥% 19N 100 where m is the content of the compound in the myocardial tissue and perfusate in pymol/g tissue, n is the number of nitrogen atoms in the molecule of the compound, atom? 15N is the atom percent excess 15N in the compound. Figure 3 shows the results of calculating the content of 15N in the compounds studied after perfusion of the heart with 2.2 mM 15N-ammonium acetate. The total quantity of 15N transferred from the 15N-ammonium acetate to the amino acids and urea of the intact heart is 0.080 + 0.003 pg-atom 1IN/g of tissue. Of this quantity, about 60 percent is accounted for by glutamine: 0.047 + 0.002 pg-atom 15N/g tissue. 15N-ammonia is utilized to a lesser extent by glutamic acid: 0.021 + 0.001 pg-atom 15N/g tissue, corresponding to about 26 percent of the 15N neutralized by the substances studied. The remaining portion of 19N is distributed among alanine (7 percent), aspartic acid (4 percent), and urea (3 percent). The total absorption of 15N by amino acids 337 7] pr L7 -1.7 — = 1.6 mn rd -1,6 7 w _ - n ~~ 15 _ _ 1.5 2 - < = 1.4 1 oe 7 rad -1.4 7 7 \ pd 7 = 1,3 ad - 1.3 a= o-~ =D 1.2 1.2 CHCO01oNHy, mM FIGURE 2. Variation in total content of urea in tissue and perfusate of intact (a) and infarcted (b) rat hearts as a function of concentration of l19N-ammonium acetate administered. Hearts were perfused 30 minutes with 15N-ammonium acetate at a concentration 1.6, 2.2, 2.8, and 3.4 mM and Krebs-Henseleit buffer without 15N-ammonium acetate. Each point is a mean value produced in a series of three to four experiments. and urea in the intact heart is about 0.1 percent of the 15N administered to the perfusate in the form of 19N-ammonium acetate. When an infarcted heart is perfused, the transfer of 15N to these nitroge- nous compounds is significantly decreased to 0.058 + 0.002 ug-atom 15N/g tissue (p < 0.02), which corresponds to a consumption of 0.073 percent of the 5N con- tained in the label compound. The quantity of 15N transferred to glutamine in the infarcted heart decreases by 40 percent, to 0.028 * 0.001 pg-atom 15N/g tissue (p < 0.02), whereas in the intact heart this substance neutralizes most of the 15N ammonia--about 50 percent. The absorption of 15n by amino acids in the infarcted heart is less than in the intact heart: aspartic acid--by 37 percent (p < 0.05), alanine--by 43 percent (p < 0.05). Experimental infarction had no influence on the content of 15N in glutamic acid (p > 0.05). The con- tent of 19N in urea in the infarcted heart increased by 80 percent (p < 0.02). However, this increase (0.002 ug-atom 15n/g tissue) was an order of magnitude less than the decrease in 15N (0.021 ug-atom 15N/g tissue) due to the decrease in inclusion of this isotope in the amino acids. 338 [CJ NORMAL 80 INFARCTION ™ 2 _ <| 3 60 6 [72] =n wo |r . — — [+ 5 “© 41 P= © | Bn n = 20 - 2 TOTAL GLN GLU ALA ASP UREA FIGURE 3. Content of 15N in amino acids and urea, and total content of 15N in these compounds of intact and infarcted rat heart after 30 minutes of perfusion with 2.2 mM ammonium acetate - 15N (98.5 atom percent). n = 3-4. GLN = Gluta- mine. GLU = Glutamic. ALA = Alanine. ASP = Aspartic acid. The quantity of 15N neutralized by urea increases with increasing concen- tration of 15N-ammonium acetate in the perfusate (figure 4). For all concen- trations of labeled ammonia, the content of 15N in the urea of the infarcted heart is somewhat higher than in the intact heart. The data produced confirm the possibility of bonding ammonia in the heart muscle by the reaction of urea synthesis, although most of the ammonia is absorbed in the formation of glutamine. The predominance of this process over the remaining mechanisms of neutralization of ammonia does not decrease the significance of urea synthesis for the myocardium. In a heart with an experimental infarct, the quantity of ammonia removal by this way increases almost by a factor of 2 upon suppression of the formation of amino acids and glutamine, which bond ammonia. Formation of Urea Upon Perfusion of an Isolated Heart With 1-Amino Acids and Ammonium Chloride Synthesis of 15N-urea in a rat heart from l5N-ammonia does not eliminate the possibility of formation of this compound from other precursors. In exper- iments to study the influence of other nitrogenous compounds on the synthesis of urea, the nitrogen donors used were l-arginine, l-ornithine, l-glutamine, and ammonium chloride, usually used for activation of this process in the liver (18,19). The effect of l-leucine and pyruvate was also evaluated. 339 I I 1 T T be 5.0 ad ~-50 7 / 7 S / x 40 0 7 a a —4.,0 0 [+ / 7 —_ / 7 = = / 7 Slo 7 - + J 7 30 7 _-® — 30 = 7 7 oo ° ." © 20 7 —20 Bod CHC00 NH, mM FIGURE 4. Influence of concentration of 1ON-ammonium acetate in perfusate on the content of 12N in urea of intact (a) and infarcted (b) rat heart. The values correspond to the total content of 19N in the urea of heart tissue and perfusate. n = 3-5. To check the quantity of urea synthesized upon perfusion of the heart by these compounds, it was necessary to determine the initial content of this substance in the heart tissue before perfusion and the change in the content of urea caused by perfusion with Krebs-Henseleit bicarbonate buffer containing no additives. The content of urea in the Krebs-Henseleit bicarbonate buffer washed from the heart was 1.00 * 0.08 umol/g of wet tissue (n = 20). The nature of the change in the content of urea in the rat heart tissue and in the heart upon perfusion with Krebs-Henseleit buffer is shown in figure 5. During the course of perfusion as a result of diffusion of urea from the myocardial cells, a decrease in its content in the tissues and accumulation in the perfusate occurs. The concentration of urea found in the tissue with the passage of time was 0.19 * 0.04 umol/g of wet tissue and, as experiments involving addition of l-amino acids to the perfusate showed, was practically independent of the composition of the medium. The concentration of urea in the perfusate was several times higher than in the tissue. Perfusion with 340 - \ ~~ - 0.6 ¢ = » = 140 oc 2 do. ] PERFUSION, MIN. FIGURE 5. Change in content of urea in rat heart tissue (a) and perfusate (b) upon perfusion with Krebs-Henseleit buffer, n = 6-8. A = Content of urea in rat heart washed of blood with 25 ml Krebs-Henseleit buffer, 1.00 + 0.08 umol/g moist tissue, n = 20. Krebs-Henseleit buffer resulted in an increase in the substance with the pas- sage of time in the heart-perfusate system. The total quantity of urea in the perfusate and tissue after 40 minutes was significantly greater than the con- tent of urea in the washed heart (1.26 + 0.10 umol/g wet tissue, p < 0.05). The influence of substrates added to the perfusate on synthesis of urea is shown in figure 6. Perfusion with Krebs-Henseleit buffer containing the precursors of urea-—ammonium chloride and l-arginine--significantly stimulated the formation of urea in the heart. After 40 minutes of perfusion with these substances, the total content of urea in the heart tissue and medium increased by 1.02 + 0.09 and 0.92 + 0.08 pumol/g wet tissue, respectively (p < 0.02). A lesser effect was discovered for l-ornithine 0.63 * 0.09 umol/g wet tissue, Pp < 0.05). The activating effect of l-glutamine appeared only in the late stages of perfusion and was weaker than the precursors listed: In 40 minutes of perfusion the content of urea in the heart-perfusate system increased by 0.38 + 0.10 umol/g of wet tissue. When l-leucine and pyruvate were introduced to the perfusion medium, inhibition of urea synthesis was observed (p < 0.05). 341 = 104 A _ I Fio{ B 7 = | o” | $ , L [&>] 7 ~~ < i J Lo 7 L - ; - = h / FA ad + =, 9 PE « = / 2 - - 7 - o /© 6 id Mo os ly 105. 1 L — / / o / wn 4 ry Lo ,/ __@--6 | = ’ / / ° 77 ° = i / J Lo] o/ 7 | = / | f 7 72} {1/0 © __-—-—- Foq 7 - , = . < 8 J - / 5 o 17 8g °° . RTA 0 i = 0 > : T T 4 T > T 0 O00" T T T 10 20 30 40 10 20 30 40 PERFUSION, MIN. PERFUSION, MIN. FIGURE 6. Influence of various substrates on the synthesis of urea upon per- fusion of isolated rat heart. Hearts were perfused with Krebs-Henseleit buffer without additives (1), buffer containing l-ornithine at 10 mM (2), NH4Cl 10 mM (3), pyruvate 5 mM (4), l-leucine 5 mM (5), l-glutamine 10 mM (6), and l-arginine 10 mM (7). The quantity of synthesized urea was determined from the difference between the total content in the heart tissue and medium and the initial content in the washed heart (1.00 * 0.08 umol/g wet tissue). n = 6-8. To provide a quantitative description of the effectiveness of the nitrogen donors, we used the following kinetic model of urea synthesis: K K IK 2 Precursor Ky Precursor* —% 4, Urea* + Urea (® > (PX) oe) (M) perfusate Ky | perfusate | heart | We assumed that these donors resulted only from processes occurring in the heart. The exchange between the medium and the tissue was achieved by diffu- sion of the nitrogen donor into the myocardial cells (P XL; P*) and reverse diffusion of the synthesized urea into the medium (M* Xs y M). It was also assumed that the stage of diffusion of the nitrogen donor and urea occurs more rapidly than the transformations occurring in the heart tissue: Kj, Kgs >> Kp, K3, K4. Under the experimental conditions [P] >> [P*], therefore, we considered that active transport of amino acids from the medium into the myocardial cells is a practically irreversible process. In this case, the rate of formation of urea was: WM = K4[X]. (equation 1) 342 Considering that the rate of formation of the intermediate complex X and the rate of its conversion are approximately equal (stable state), dX Fra 0, (equation 2) and also assuming that [P*] = [X] + [M¥*], (equation 3) . dX . we obtain ax Ko ([X] + [M*]) - (Ky + K3) [X] = 0, (equation 4) K,) [M* ] 1 ER ——————————————————— %* i from which [X] K; + K3 - Ko K[M#*] (equation 5) K where K = — K3 Substituting equation 5 into equation 1, we finally obtain Wy = Kegg[M*] (equation 6) where Koff = K,+K is the effective value of the rate constants or, after integration, A . 1n Be +1) = Kegs® T (equation 7) 0 where A = [M*] - [M5] is the increase in urea. Comparison of the experimental data (points) and calculated data (straight lines) in coordinates log (A/[M*] + 1) versus perfusion time (figure 7) indi- cates that the kinetic model suggested satisfactorily describes the synthesis of urea in the heart upon perfusion with Krebs-Henseleit buffer without addi- tives, and also with l-glutamine and 1l-leucine throughout the entire time of the experiment. For the remaining nitrogen donors, the comparison is satis- factory only for the period of perfusion, which does not exceed 12 minutes. For longer perfusion, disagreement was observed between the calculated and experimental data. The values of the constants Keff presented in table 4 indicate that the most active nitrogen donors are l-arginine and ammonium chloride. These re- sults agree well with the data of a number of researchers obtained from studies of isolated rat livers (18-20). Introduction of l-ornithine and glutamine to the perfusion medium results in lower values of constant Keff. The inhibiting capacity of l-leucine is confirmed by the lower value of Kgff. 343 2 025 / o / / 4 0.20 A 0.5 / / rd o A = alo / / 7 __.o ] | / ~ _— : 005 oy // pd _— . Ce +1) A [M ] log ( PERFUSION, MIN. FIGURE 7. Variation of log (a/ [ME] + 1) as a function of perfusion time for various media: (1) l-arginine 10 mM; (2) NH4Cl 10 mM; (3) l-ornithine 10 mM; (4) 1-glutamine 10 mM; (5) Krebs-Henseleit buffer; and (6) l-leucine 5 mM. DISCUSSION The experimental data presented above allow us to analyze some of the specifics of the metabolism of ammonia in heart muscle. The great enrichment of glutamine with 15N indicates the presence of glutamine synthetase in the myocardium, the existence of which has been doubted by some researchers due to the difficulties involved in isolating it (21). The inclusion of 15N in alanine and aspartic acid confirms the assumption that the corresponding amino- transferases participate in bonding of ammonia (12,13). 344 TABLE 4. Influence of Composition of Medium on Effective Rate Constant of Formation of Urea Upon Perfusion of Isolated Rat Heart No.* Perfusion medium K¥es x 102, min-1 1 1-Arginine, 10 mM 2.88 2 NH,C1, 10 mM 1.98 3 1-Ornithine, 10 mM 1.62 4 1-Glutamine, 10 mM 0.62 5 Krebs-Henseleit buffer 0.03 6 1-Leucine, 5 mM 0.02 *Corresponds to calculated lines of figure 7. Experimental myocardial infarction leads to significant changes in the metabolism of ammonia due to a decrease in the activity of transamination and amination enzymes in the cardiac muscle (17,22). This is manifested as a decrease in the bonding of 15N ammonia by aspartic acid, alanine, and glu- tamine (figure 3). It is significant that in infarcted hearts, synthesis of urea is activated, although the contribution of the glutamate dehydrogenase and glutamine synthetase reactions to the detoxication of ammonia remains most significant. The linear relationship between the concentration of 15N-ammonium acetate and perfusate and the content of 15N in urea (figure 4) confirms the conclusions drawn earlier (4) that urea synthesis by the heart muscle is a response to an increase in the level of ammonia in the muscle. This result agrees well with the increase in arginase activity in the necrotic zone of heart muscle observed upon examination of human biopsy materials (4). We should note that the increase in urea synthesis in the infarcted heart occurs with blockage of the basic mechanism of bonding of ammonia--reduction in glutamine synthesis (figure 3). Thus, this process can be considered a com- pensatory restructuring of the nitrogen metabolism of the heart in response to disruption of the normal mechanisms of elimination of ammonia. These data agree with the results obtained from the synthesis of urea in experimental myocardial infarction in dogs in vivo (10). In these experiments, a decrease was discovered in the rate of urea synthesis in the liver as well as a de- crease in the level of urea in the blood, while its content in the heart re- mained unchanged, accompanied by an increase in the level of urea in the blood of the coronary sinus. This observation not only confirms the capability of the myocardium in mammals to form urea, but also shows that activation of its synthesis is a compensatory phenomenon in cardiac pathology. 345 For urea, more than for the amino acids studied, there is disagreement be- tween its synthesis as a result of perfusion with labeled nitrogen (figure 4) and the increase in its total content in the system of the intact (or infarcted) heart plus perfusate (figure 2). For example, upon perfusion of an infarcted heart with 1.6 mM 15N-ammonium acetate, the quantity of 15N-urea synthesized is 0.003 pumol/g of tissue (figure 4), while its absolute increase in the system consisting of the heart plus the perfusate is 0.011 * 0.005 umol/g of tissue (figure 2). The absolute increase in urea exceeds the quantity of the labeled compound synthesized at all concentrations of 15N-ammonium acetate. This fact forces us to consider the reason for this disagreement under the experimental conditions and to select noncontradictory statements to explain it. When the heart is perfused with increasing 15N-ammonium acetate concentrations, the content of urea may increase as a result of (a) synthesis from the 15N-ammonia introduced (to a slight extent), (b) formation of arginine with increased pro- teolysis under the influence of ammonia and subsequent splitting with arginase, or (c) weakening of deamination of urea (23). Further experiments will allow us to evaluate the contribution of these processes more definitely. In addition to ammonium salts, synthesis of urea is strengthened upon introduction of l-arginine, l-ornithine, and l-glutamine to the perfusate. Of these amino acids, the most effective is the closest precursor of urea--arginine (figure 6, table 4). The possibility of urea synthesis in rat cardiac muscle from the arginine administered was determined in experiments on the transfer of 15N from 15N-arginine to urea (4). The increased synthesis of urea under the influence of arginine can be related not only to the formation of the com- pound under the influence of arginase, but also to the stimulating capacity of the ornithine formed from it, since the latter compound also activated the syn- thesis of urea, although to a lesser extent (table 4). 1-Glutamine had less influence on the formation of urea than equimolar quantities of l-arginine and l-ornithine (table 4). The glutaminase activity in the heart is not high enough to support its effective supply with ammonia for the synthesis of carbamoyl phosphate, in spite of the fact that glutamine easily penetrates through the cell membranes (20). The increase in urea ob- served in the late stages of perfusion with glutamine (figure 6) may occur due to the ammonia formed upon partial hydrolysis of the amide group of glutamine in the perfusion medium. As we know, accumulation of ammonia in cardiac tissue, leading to inhibi- tion of oxidative metabolism, is a pathogenic factor in the development of myo- cardial damage (9,12,24). Therefore, activation of urea synthesis in cardiac muscle by l-amino acids can be looked upon as a metabolic defense against its toxic effect. Therefore, it is possible that the restoration of contractile capacity of the ischemic myocardium under the influence of l-arginine and 1- glutamic acid (25) is also related to the participation of these amino acids in ammonia detoxication. Of the nitrogenous compounds studied, l-leucine and pyruvate had an in- hibiting effect on the synthesis of urea. The effect of l-leucine can be explained by its stimulation of glutamate dehydrogenase in the direction of 346 formation of glutamic acid and suppression of the deaminating activity of this enzyme (26). Since glutamate dehydrogenase is responsible for the formation of ammonia for its participation in the carbamoyl phosphate synthetase reaction and, consequently, one of the two atoms of nitrogen in the urea molecule, the reason for weakening of the synthesis of urea when l-leucine is present becomes understandable. It is quite probable that in the heart, as in the liver, leu- cine inhibits the synthesis of citrulline from ornithine and thus blocks the formation of urea (27). The influence of pyruvate on the synthesis of urea in the liver is related to the high activity of ornithine pyruvate transferase: Pyruvate leads to com- plete consumption of ornithine and disrupts the synthesis of urea (28). It is known that in the heart, as in the liver, transamination of ornithine to pyru- vate occurs (29). Since the level of ornithine in the myocardium is much lower than the concentration of other metabolites (4), its intracellular concentra- tion may be the determining factor in the effectiveness of the ornithine cycle and, consequently, the synthesis of urea. Pyruvate is also capable of blocking the transformation of citrulline to arginine--a stage which limits the speed of the ornithine cycle, as has been shown in liver sections (30,31). These considerations probably remain valid for cardiac muscle, since arginine suc- cinase and other enzymes in the ornithine cycle have been shown to be active in the myocardium (4,32), although it is possible that the paths leading to the formation of urea in the myocardium are not identical to the cycle of urea in the liver. CONCLUSIONS The results obtained on the synthesis of amino acids and urea in cardiac muscle allow us to draw the following conclusions: 1. The ammonia formed upon contraction of cardiac muscle, as well as that entering the heart with the bloodstream, is bonded by amino acids and the urea of the myocardium. The primary quantity is neutralized by glutamine and glutamic acid. Ammonia is utilized to a lesser extent by alanine and aspartic acid, and in reactions leading to the formation of urea. 2. The significance of urea synthesis as an additional mechanism for regulating the level of ammonia in heart tissue increases with myocardial infarction. Activation of this process by amino acids provides a possibility for correction of ammonia metabolic disorders in various cardiac pathophysio- logic states. 347 10. 11. 12. 13. 14. REFERENCES Clark AJ, Gaddie R, Stewart CP: Metabolism of the isolated heart of the frog. J Physiol (London) 72:443-466, 1931 Feinberg H, Alma M: Ammonia production in the isolated working rabbit heart. Am J Physiol 200:238-242, 1961 Chazov EI, Smirnov VN, Mazaev AV, et al.: [Coronary arteriovenous dif- ference in the content of some nitrogen metabolites in man] (Rus). Kardiologiia 12(No. 10):11-19, 1972 Smirnov VN, Asafov GB, Cherpachenko NM, et al.: Ammonia neutralization and urea synthesis in cardiac muscle. Circ Res 35 (suppl 3):58-73, 1974 Brodan V, Fabidn J, Pechar K, et al.: [Metabolism of ammonia in the ischemic myocardium at rest and during pacing] (Cze). Cas Lek Cesk 114(31-32):988-991, 1975 Bessman AN, Evans JM: Blood ammonia in congestive heart failure. Am Heart J 50:715-719, 1955 Hani H: Blood ammonia levels in patients with chronic congestive heart failure. Noukogude Eesti Tervishoid (Soviet Estonian Health) 6:495-496, 1973 Katunuma N, Okada M, Nishii Y: Regulation of the urea cycle and TCA cycle by ammonia. Adv Enzyme Regul 4:317-336, 1966 Kobayashi T: Myocardial amide-nitrogen metabolism with special reference to ammonia metabolism; studies by the use of the coronary sinus catheteri- zation technique. Jpn Circ J 31(3):33-47, 1967 Fetisova TV: [Some nitrogen-containing components of the heart, of the blood in the coronary sinus and the aorta in experimental myocardial infarction] (Rus). Kardiologiia 16(No. 6):89-93, 1976 Kato T: Myocardial amide-nitrogen metabolism with special reference to ammonia metabolism. Jpn Circ J 32:1401-1416, 1968 Watanabe T, Yamazaki N, Aoyama S: Significance of myocardial ammonia metabolism in the failing heart. Israel J Med Sci 5(4):496-500, 1969 Carlsten A, Hallgren B, Jagenburg R, Svanborg A, Werks L: Myocardial metabolism of glucose, lactic acid, amino acids, and fatty acids in healthy human individuals at rest and at different work loads. Verh Dtsch Ges Kreislaufforsch 27:195-196, 1961 Kedenburg CP: A lithium buffer system for accelerated single-column amino acid analysis in physiological fluids. Anal Biochem 40:35-42, 1971 348 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. Chain EB, Chiozzotto M, Pocchiari F, et al.: Participation of the ammo- nium ion in the transformation of glucose to amino acids in brain tissue. Proc R Soc Lond [Biol] 152(B):290-297, 1960 Davidson S, Sonnenblick EH: Glutamine production by the isolated perfused rat heart during ammonium chloride perfusion. Cardiovasc Res 9:295-301, 1975 Yakushev VS, Lifshits RI, Slobodin VB: [Effect of l-aspartic acid in isoproterenol necrosis of the myocardium] (Rus). Kardiologiia 12(No. 12): 49-55, 1972 Saheki T, Katunuma N: Analysis of regulatory factors for urea synthesis by isolated perfused rat liver. 1. Urea synthesis with ammonia and glu- tamine as nitrogen sources. J Biochem 77:659-669, 1975 Kamin H, Handler P: The metabolism of parenterally administered amino acids. II. Urea synthesis. J Biol Chem 188:193-205, 1951 Chamalaun RAFM, Tager JM: Nitrogen metabolism in the perfused rat liver. Biochem Biophys Acta 222:119-134, 1970 Igbal K, Ottaway JH: Glutamine synthetase in muscle and kidney. Biochem J 119:145-156, 1970 Fetisova TV, Frolkis RA: [Biochemistry of Myocardial Infarction] (Rus). Kiev, 1976 Gorkin VZ, Akopyan ZhI, Verevkina IV, et al.: [Deamination of some bio- genic amines and other nitrogenous compounds by liver mitochondria] (Rus). Biokhimiia 35:140-151, 1970 Watanabe T: Significance of ammonia in myocardial metabolism. Jpn Circ J 32(suppl):1811-1814, 1968 Rau EE, Shine KI, Gervais A, Douglas AM, Amos EC: Enhanced mechanical recovery of anoxic and ischemic myocardium by amino acid perfusion. Am J Physiol 236:H873-H879, 1979 Mendes-Mourao J, McGivan JD, Chappell JB: The effects of l-leucine on the synthesis of urea, glutamine and glutamate by isolated rat liver cells. Biochem J 146:457-464, 1975 McGivan JD, Bradford NM, Crompton M, Chappell JB: Effect of l-leucine on the nitrogen metabolism of isolated rat liver mitochondria. Biochem J 134:209-215, 1973 Metz R, Salter JM, Brunet G: Effect of pyruvate and other substrates on urea synthesis in rat liver cells. Metabolism 17:158-167, 1968 Krebs HA: The role of chemical equilibria in organ function. Adv Enzyme Regul 13:449-472, 1975 349 30. 31. 32. Gornall AG, Hunter A: The synthesis of urea in the liver with special reference to citrulline as an intermediary in the ornithine cycle. J Biol Chem 147:593-615, 1943 Ratner S: Urea synthesis and metabolism of arginine and citrulline. Adv Enzymol 15:319-387, 1954 Menne F, Kossmann KT, Lange K: [On the origin of urea in the myocardium] (Ger). Hoppe-Seyler's Z Physiol Chem 332:314-315, 1963 350 HIGH ENERGY PHOSPHATE AND CELL VOLUME CONTROL IN LETHAL ISCHEMIC INJURY Robert B. Jennings, Hal K. Hawkins, James E. Lowe, Mary L. Hill, and Keith A. Reimer SUMMARY Brief periods of severe myocardial ischemia induced by coronary occlusion cause reversible cell injury. However, after longer periods of ischemia (40 minutes), the injury becomes irreversible in most myocytes in that necrosis occurs even if blood flow is reestablished. Ultrastructural and metabolic fea- tures of this transition to irreversibility have been studied in the severely ischemic, subendocardial region of dog myocardium following 15 minutes to 4 hours of coronary occlusion. High energy phosphate depletion occurred rapidly. By 40 minutes, nearly all of the original creatine phosphate, more than 90 percent of the adenosine triphosphate (ATP), and 65 percent of the total adenine nucleo- tide pool were gone. At this time, slices of this tissue when incubated in vitro swelled markedly and were incapable of maintaining ion gradients. Inadequate supplies of ATP may be the cause of failure to maintain volume. However, elec- tron microscopy revealed that the injured cells also showed breaks in the plasma- lemma of the sarcolemma. These breaks were associated with other ultrastructural features of irreversibility and with an increased inulin diffusible space (IDS) in the incubated slices. Because there was little interstitial edema on electron microscopy, this increased IDS is functional evidence that the sarcolemma had become permeable to large molecules. These studies demonstrate that marked ATP depletion and sarcolemmal defects are two early features of irreversible myo- cardial ischemic injury. The cause of the sarcolemmal defects is unknown, but it may be due to the depletion of ATP below levels essential for the maintenance of membrane structure. INTRODUCTION Brief episodes of severe ischemia, produced by proximal occlusion of a major coronary artery in dogs, cause reversible injury to left ventricular myo- cytes. The injury is known to be reversible because elimination of the ischemic state by restoration of coronary arterial flow prevents cell death (1,2). However, From the Department of Pathology, Duke University Medical Center, Durham, North Carolina. This work was supported in part by grants HL-23138 and HL-17670 from the National Institutes of Health. 351 the injury becomes irreversible in the subendocardial region of severe ischemia if the cells are exposed to 40 or more minutes of severe ischemia. At this time, restoration of arterial flow is followed by striking cell swelling, al- tered ion gradients, myofibrillar contraction bands, mitochondrial catt accumu- lation, and focal disruption of the sarcolemma (3-5). The proximate cause(s) of the transition to irreversibility are still un- known. However, defects in cell volume regulation observed at or about the time ischemic cells pass the "point of no return" suggest that alterations in sarcolemmal permeability or structure might be an early feature of the irrevers- ible state (6). Unfortunately, sarcolemmal permeability cannot be assessed in the ischemic zone in vivo because markers of transsarcolemmal fluxes cannot be perfused into areas of markedly depressed arterial flow. For this reason, we have developed an in vitro method of assessing the capacity of tissues in- jured in vivo to maintain volume and to exclude molecules, such as inulin, which normally remain confined to this extracellular space (5-7). We also have examined the ultrastructure of the severely ischemic tissue before and after incubation in order to relate any functional changes observed to structural evidence of irreversible injury. Since volume regulation is an active process requiring ATP to fuel the Nat, Kt-ATPase of the cell membrane, we have also assessed the effect of severe in vivo ischemia on the ~PO4 and total adenine nucleotides (ZAd) of the tissue (1). The results of the studies presented in this paper show that there is a rapid and marked loss of “PO4 in the zone of severe ischemia. Creatine phos- phate (CP) is depleted by 75 percent within 60 seconds of ischemia (8-10) and ATP progressively decreases thereafter (1,8-11). The decrease in ATP is accom— panied by degradation of all adenine nucleotides. At the time when irreversibility occurs, ATP has been depleted to less than 10 percent of control and cell volume no longer can be maintained in incubated slices. At the same time, the IDS in- creases, suggesting that the sarcolemma has become permeable to large molecules, and structural defects can be observed in the plasmalemma by electron-microscopic evaluation. The pathogenesis of the membrane defects and their relation to the concomitant loss of high energy phosphate depletion is unknown. However, both the high energy phosphate depletion and membrane defects are early features of irreversible injury and may be causally related to ischemic cell death. MATERIALS AND METHODS Experiment Design Ischemic tissue was obtained from the posterior papillary muscle (PP) of the left ventricle of dogs following 15, 30, 40, 60, 120, or 240 minutes of ischemia induced by occlusion of the circumflex artery. This ischemic tissue was compared to nonischemic tissue from the anterior papillary muscle (AP) of the same animals. Special precautions (see below) were taken to identify the zone of severe ischemia and to include only severely ischemic tissue (tissue with collateral blood flows of less than 15 percent of control) (12,13). 352 Table 1 lists the parameters evaluated. After each period of ischemia, tissue metabolites, electrolytes, TTW (total tissue water), and ultrastructure were assessed. Cell volume regulation and sarcolemmal permeability were assessed in slices of control and injured tissue incubated in vitro at 37° C in oxygenated media. The various methods used are briefly summarized in the following paragraphs. Operative Technique Thirty-five adult, healthy mongrel dogs were used. They were anesthetized with intravenous sodium pentobarbital, intubated, and ventilated with a Harvard Model 607 respirator pump. Ventilation was maintained at a rate of 200 ml/min/kg of room air containing sufficient Oj; to maintain the PO, PCO, and pH at physio- logical levels. The left chest was opened through a small incision in the fourth intercostal space. The pericardium was opened and the heart was suspended in a pericardial cradle. After isolating the circumflex artery 7-10 mm from its origin at the aorta, it was ligated with a silk suture (14). Electrocardiograms were recorded on lead II with a Brush model 440 recorder. All animals included in this study developed cyanosis in the circumflex bed and elevation of ST segments in leads II, III, and aVF. After the desired interval of ischemia had passed, thioflavine S was injected (see below) and the heart was excised quickly and rapidly cooled to about 5° C in ice-cold isotonic KCl. TABLE 1. Parameters Evaluated in Severe Ischemia Induced In Vivo in the Left Ventricle of Dogs Papillary Muscle Anterior Papillary (After 15, 30, 40, 60, 120, 240 Minutes Muscle (Control) of Ischemia) 1. Ultrastructure 2. TTW, Nat, Kt, Mgt 3. ATP, ADP, AMP, CP, and Lactate 4, Slice function in vitro 37° C for 60 minutes (a) Ultrastructure (b) TTW, Nat, kt, Mgtt (c) ATP (d) Synthesis of CP (e) IDS TTW = Total tissue water. ATP = Adenosine triphosphate. ADP = Adenosine di- phosphate. AMP = Adenosine monophosphate. CP = (Creatine phosphate. IDS = Inulin diffusible space. 353 Identification of the Zone of Severe Ischemia With Thioflavine S In order to identify the zone of severe ischemia, 10 ml of a freshly prepared 4 percent solution of thioflavine S dye (Pfaltz & Bauer, Inc., Flushing, New York) was injected one to two circulation times (10-15 seconds) prior to excision of the heart (15). This dye binds to endothelium of vessels exposed to a flow of more than 20 percent of control and fluoresces brightly in ultraviolet light. The zone of severe ischemia was identified as the subendocardial region of non- fluorescence and was visualized by illuminating the tissue with ultraviolet light in a darkened room. Fluorescent and nonfluorescent tissue was identified most easily if a yellow filter was used to view the heart (15). If the boundary of the ischemic region extended into the tip of the AP, the ischemic part of the AP also was identified by nonfluorescence and was excluded from the control tissue sample. After 60, 120, and 240 minutes of ischemia, severely ischemic and irre- versibly injured tissue of the PP was readily identified by its pale color and thioflavine S was not necessary to identify it (see figure 3 in reference 16). The severely ischemic PP and control AP tissue blocks were cut and placed in ice-cold Krebs-Ringer's-phosphate (KRP) (6). Thin slices of these tissue blocks were subsequently cut with a razor blade parallel to the long axis of the myofibers for ultrastructure examination, metabolite assays, and assessment of cell volume regulation. Metabolite Assays For metabolite assays, the slices were weighed quickly on a Cahn Model DTL microbalance and placed in ice-cold 3.6 percent perchloric acid (PCA). Weighing and transfer to the PCA required 10 to 15 seconds. After 60 to 180 minutes for extraction, the slices were homogenized with a Tri-R homogenizer and neutralized with K2C03 and KOH. The extracts were centrifuged to remove KC1l04 and the super- natant was frozen at -20° C. Samples were assayed within 2 to 3 weeks by enzymatic techniques for lactate (17), ATP (18), adenosine diphosphate (ADP) (19), adenosine monophosphate (AMP) (19), and glucose-6-phosphate (G6P) (18). Assessment of Cell Volume and Electrolytes Control slices of unincubated tissue were weighed quickly on a Cahn Model DTL electrobalance, placed directly into rinsed, ion-free scintillation vials (mini- vials, Corning), and dried at 105° C. The dry slices were reweighed and TTW was calculated in ml per 100 g dry tissue. Other slices were incubated at 37°C in oxygenated KRP solution for 60 minutes. The KRP was prepared fresh, pH 7.2 to 7.4, and contained the following ions (mM): Nat 151.6, Kt 4.81, Catt 1.29, Mg*t 1.20, S042- 1.24, PO, 15.63, and Cl~ 121.7. Trace quantities of 1l4c-hydroxymethyl-inulin were added to this incubating medium to determine the IDS. The medium was equilibrated with 100 percent oxygen prior to placing it into flasks, and the flasks were continuously gassed with 100 per- cent oxygen during incubation. All flasks were shaken 180 times per minute on a Dubnoff shaker. As many as six slices weighing 20 to 80 mg wet weight were in- cubated in 15 ml of media at 37° C in a 25-ml Erlenmeyer flask. After different 354 periods of incubation, these slices were rinsed by rapid dipping in 0.25 M sucrose (J. T. Baker), blotted quickly on Whatman filter paper, weighed, and dried as described above. Electrolytes were extracted from each dried slice in 5 ml of 0.75 N HNOj (1). The ions were determined in a 1 to 100 dilution of each HNOj3 extract pre- pared in ion-free glassware which had been shown to contain no detectable sodium. RbCl, 5,000 ppm, was added to each dilution and to the standard to equalize ionization between standards and unknowns. Electrolytes were measured against appropriately diluted standards in an IL 351 atomic absorption spectrophotometer interfaced with a Tektronix Model 31 computer calculator. Standard curves for Mgtt, Nat, and Kt were prepared from a standard solution containing these ions and PO4 in physiological concentrations (20). To calculate the IDS, the dry slices were rehydrated by adding one or two drops of deionized water to the vial and solubilized in Soluene 350. l4c ac- tivity was counted in a Hewlett-Packard liquid scintillation counter (1,7). Electron Microscopy For each of the four periods of ischemia, blocks from both the surface and center of at least two slices before and two slices after incubation of both AP and PP muscles were cut under 4 percent glutaraldehyde in 0.1 M cacodylate buffer. After 1 to 4 hours of fixation, the tissue was postosmicated, dehydrated in a graded ethanol series, rinsed in propylene oxide, and embedded in Epon 812. Thin sections were cut from an average of three blocks per slice, stained with toluidine blue, and examined by light microscopy. Ultrathin sections were cut from at least two representative blocks of each slice, stained with uranyl acetate/lead citrate, and examined in an Hitachi Model HS-8 electron microscope. RESULTS Changes Occurring in Severely Ischemic Tissue Gross changes. The subendocardial layer of the circumflex bed including PP was grossly normal until 40 or more minutes of ischemia had passed. However, in 3 of 6 dogs at 40 minutes, and in all hearts after 60 or more minutes of ischemia, tissue in the PP and the inner one-third of the circumflex bed was pale and stiff in comparison to nonischemic control tissue. The areas which were grossly pale were the same areas which were nonfluorescent under ultraviolet light. Thus, either parameter could be used to identify the severely ischemic, irreversibly injured myocardium. Ultrastructural changes. The ultrastructural features of myocytes in the center of freshly cut slices of nonischemic AP muscle or of PP muscle injured by various periods of ischemia have been described previously (16) and are reviewed here only briefly. The characteristic features of nonischemic myocardium are illustrated in figure 1. After 15 minutes of severe ischemia, myocytes of the PP muscle showed only minor differences from control tissue, including some margination of nuclear chromatin, occasional damaged mitochondria, and decreased amounts of glycogen (figure 2). Myofibrils were relaxed and showed I bands. 355 FIGURE 1. Nonischemic anterior papil- lary muscle. A: The nuclear chromatin (N) is evenly arranged. The mitochon- dria (M) are dense and contain occa- sional matrix granules. Glycogen is present but faintly stained as usual in tissue fixed in glutaraldehyde. (Magnification x 31,000) B: The sar- colemma is scalloped over two con- tracted myofibrils. The plasma membrane (P) and glycocalyx (G) are typical of control myocardium. (Magnification x 50,000) (Reprinted by permission of The American Journal of Pathology from reference 1) 356 FIGURE 2. Posterior papillary muscle after 15 minutes of ischemia. A: The significant alterations from control include margination of nuclear chroma- tin (N) and slight mitochondrial swelling. One markedly swollen mito- chondrion (M) is present. There ap- pears to be less glycogen than in control myocardium. (Magnification x 31,000) B: The sarcolemma is flattened over relaxed myofibrils. However, the plasma membrane and glycocalyx appear intact. (Magnifi- cation x 74,000) (Reprinted by per- mission of The American Journal of Pathology from reference 1) After 30 minutes to 4 hours of ischemia, PP myocytes showed progressively more extensive changes, with marked margination of nuclear chromatin, mito- chondrial swelling, and mitochondrial amorphous matrix densities (figures 3 and 5). The sarcolemma often was flattened. The basal lamina remained intact, but there were occasional tiny defects in the plasma membrane which became more prom- inent as the period of ischemia was prolonged (1). The effects of in vitro warm incubation for 60 minutes on slices injured by 60 minutes of in vivo ischemia are illustrated in figure 4. The most striking change was the development of subsarcolemmal blebs covered by focally disrupted sarcolemma. Remnants of the plasma membrane could be observed as circular pro- files under the basal lamina, which often remained intact. Cells with membrane defects also developed prominent myofibrillar contraction bands. In comparison to warm incubation, when injured slices were swollen by cold incubation at 0-1° C, the sarcolemma exhibited similar changes (figure 6), but the myofibrils did not develop contraction bands (1). Metabolite changes. Severe ischemia caused a rapid decrease in tissue CP to negligible levels by 30 minutes. The ATP content of ischemic and nonischemic tissue was measured in 21 dogs given 15-60 minutes of severe ischemia (see table 2 in reference 1). The ATP of the control nonischemic AP was similar at all time intervals studied, and the data have been pooled. The ischemic PP tissue exhibited a progressive loss of ATP to 35, 16, 9, and 7 percent of control levels after 15, 30, 40, and 60 minutes of ischemic injury, respectively. During severe ischemia, anaerobic glycolysis is the only significant source of new high energy phosphate; thus, the lactate content of the PP muscle was 400 percent greater than the AP after only 30 minutes of ischemia. Progressive loss of ATP in the ischemic tissue was associated with a de- crease in the total adenine nucleotide pool. The distribution of adenine nucleo- tides was measured in 12 dogs (figure 7). The control tissue contained 7.41 * 0.27 umole of IAd per gram wet weight of which 6.02 pymole was ATP, 1.23 umole was ADP, and 0.16 umole was AMP. After 15 minutes of ischemia, total nucleotides were reduced by 55 percent, due chiefly to the loss of 3.9 umoles of ATP without a significant reciprocal increase in either ADP or AMP. In fact, there was a small but significant decrease in ADP. By 30 or 40 minutes, there were further decreases in the total adenine nucleotide pool, again chiefly due to the loss of ATP. The ADP content remained depressed. The AMP content gradually in- creased, reaching a level of six times greater than control after 40 minutes of ischemia. However, the rise accounted for only a small fraction of the ATP lost. Thus, depletion of the total adenine nucleotide pool presumably was due to degra- dation of the nucleotides to adenosine, inosine, xanthine, and hypoxanthine (11, 22). The adenylate charge ratio (ACR)* of the nonischemic tissue was 0.90 but _ ATP + 1/2 ADP *ACR = 3TP + ADP + AMP" cellular adenine nucleotides which contain high energy phosphate. The ratio in normal cells is 0.85 to 0.9. If the adenine nucleotide pool were all ATP, the ratio would be 1.0; if the pool were all AMP, the ratio would be O. This ratio estimates the proportion of the total 357 FIGURE 3. Posterior papillary muscle after 40 minutes of ischemia. A: The significant alterations from control include the marked margination of nuclear chromatin (N) and mitochondrial swelling (M) with the appearance of a clear matrix which often contains amorphous matrix densities (arrow). No glycogen is detectable. (Magnifi- cation x 31,000) B: Several small breaks are visible in the plasma mem- brane although the glycocalyx appears continuous. Swollen mitochondria are present (M). The myofibrils are re- laxed and show I bands. (Magnifica- tion x 38,000) (Reprinted by permission of The American Journal of Pathology from reference 1) 358 FIGURE 4. slice after 40 minutes of in vivo ischemia and 60 minutes of incuba- Posterior papillary muscle tion. Slice incubation caused char- acteristic changes in myofibrillar architecture, including disappear- ance of the I bands and widening of the Z band zone. The mitochondria (M) are swollen and contain amor- phous matrix densities (A). The sarcolemma of a subsarcolemmal bleb (B) of an adjacent cell is disrupted. Circular profiles of plasma membrane (arrows) are present beneath the remaining glycocalyx. (Magnification x 31,000) (Reprinted by permission of The American Journal of Pathology from reference 1) FIGURE 5. after 40 minutes of ischemia of the Posterior papillary muscle same dog shown in figure 3. This thin section in high power view shows a structurally intact glycocalyx (G) with breaks in the trilaminar membrane of the plasmalemma beneath it (arrows) The circular profiles in the extra- cellular space are crosssections of collagen bundles. A swollen mito- chondrion is at M. (Magnification x 108,000) 359 FIGURE 6. Posterior papillary muscle slices after 60 minutes of in vivo ischemia and 60 minutes of incubation at 0-1° C. Note the bleb (B) of edema fluid beneath the sarcolemma. The plasma membrane of the sarcolemma of this bleb is broken into circular profiles. Note that the myofibrils of this irreversibly injured cell on the left side of the figure did not contract. (Magnification x 75,000) 7 AMP’ 4 Y ps 6 / 5 J w = = of ~N a. - < rr 3 ATP AMP, - 0 //) s (4 . 2 J 2 1 ATP 7 / « ATP / ATP CONTROL 15 30 40 MIN MIN MIN FIGURE 7. The distribution of adenine nucleotides in nonischemic (control) tissue and tissue after 15, 30, or 40 minutes of ischemia in vivo is shown. There was a progressive loss of adenosine triphosphate (ATP) from the ischemic tissue to 9 percent of control by 40 minutes. Although adenosine monophosphate (AMP) increased by 6 times during this period of ischemia, the total nucleotide content of the tissue decreased markedly because of further degradation of AMP to adenosine, inosine, xanthine, and hypoxanthine (11,12). The control levels of adenosine diphosphate (ADP) and AMP are higher than true levels in vivo be- cause of continued metabolism during the relatively slow sampling procedure necessitated by the desire to identify severely ischemic tissue for study. (Reprinted with permission of Bergman Verlag from reference 21) was reduced significantly to 0.76, 0.54, and 0.32 after 15, 30, and 40 minutes of ischemia, respectively (see table 2 in reference 1). Cell Volume Regulation and High Energy Phosphate Characteristics of control tissue. The general features of volume control in thin, free-hand slices of papillary muscle sampled at intervals during 60 minutes of incubation are shown in figure 8. Note that during the first 5 min- utes of incubation, there was a small decrease in Kt, a small increase in Nat, and a modest increase in TTIW. This initial electrolyte flux probably reflects loss of Kt from the cut cells on the edge of the slice and gain of Nat due both to cut cells and to a small increase in the interstitial space. Following these 360 40 35 w o — ~N w a BN 400 mMOL Na OR K OR ML Hy0/G DRY N o foe cee 20 .. 300 15 200 104 \S eer, cP, no 00s grr -— # ATP umols — bono pe 30 60 90 120 MINUTES FIGURE 8. The effect of incubation of thin nonischemic (control) slices of papillary muscle on slice electrolytes, water, and metabolites is shown. The slices all came from the same animal and were incubated in Krebs-Ringer's- phosphate at 37° C for up to 120 minutes. Individual slices were removed and analyzed after 5, 10, 15, 30, 60, and 120 minutes of incubation. The initial changes in electrolytes and H90 are presumed to be due to the effect of cold, because the temperature of the tissue from which the slices were prepared was between 1-5° C. The synthesis of creatine phosphate (CP) was largely complete at 30 minutes. Initial adenosine triphosphate (ATP) loss occurred, after which a stable ATP content was maintained. initial electrolyte changes, Nat, kt, and TTW were maintained constant in the control slices throughout 60 minutes of incubation. The ATP of the slices was initially reduced by more than 50 percent, perhaps again due to the presence of cut cells on the edge of the slice plus ATPase activity during preparation of the slice. Slice ATP content was maintained constant thereafter. Creatine phosphate also was initially low but gradually was resynthesized to about 75-80 percent of the values found in vivo. Maintenance of cell volume requires active metabolism. Thus, if nonischemic tissue slices are cooled to 0-1° C in oxygenated KRP to prevent enzyme activity, there is an enormous increase in TTW and Nat and a decrease in Kt. Resumption of metabolism by transferring the cold swollen slices to a 37° C media is accom- panied by rapid restoration of ionic gradients and restoration of volume toward normal. The kinetics of Kt reaccumulation were also studied by placing 86rbt in the medium as a tracer of K*. At the end of the period of cold incubation, 86gpt had equilibrated in the TTIW so that the slice-to-medium ratio was 1.0. Resump- tion of metabolism was associated with prompt accumulation of 86Rb* in the slices 361 (figure 9). This process was completely inhibited by ouabain, indicating that the accumulation of 86Rb*, and therefore of Kt, is dependent on the Nat, Kt- ATPase of the cell membrane. Thus, slices of nonischemic papillary muscles main- tain cell volume control for 3 or 4 hours through a process involving active metabolism and the Nat, Kt-ATPase of the cell membrane. In addition, the ultra- structure of nonischemic slices incubated at 37° C for 60 minutes, whether or not initially cold swollen, was similar to control tissue. Cell volume control in incubated slices of tissue injured by episodes of in vivo ischemia. Electrolyte and water content of unincubated PP tissue follow- ing ischemia for 15, 30, 40, or 60 minutes was indistinguishable from that of the control, unincubated AP. Even after 120 and 240 minutes of ischemia, TTW and Nat were increased only slightly and Kt was decreased only slightly from 40.3 + 1.8 in control to 36.5 * 1.5 in PP. However, the capacity of injured tissue to maintain volume during in vitro incubation was altered by all inter- vals of ischemia. Slices incubated after 15 minutes of ischemia in vivo showed a slight increase in water and Nat and a slight concomitant reduction in Kt and Mgt*t. These changes became progressively more marked with increased durations of ischemia. After 60 minutes or more of ischemia, the incubated slices were markedly swollen and the concentrations of Nat and Kt approached the concentra- tions in the media (figures 10 and 11). 2004 1504 s/M RATIO 86Rp 504 25 ae...s0UabaIN, oid” | warm 1 30 60 90 120 150 180 210 MINUTES OF INCUBATION DOG 2992 FIGURE 9. The effect of cold incubation on subsequent accumulation of 86Rb in thin slices of papillary muscles is shown. The 86Rb content/ml of slice water is compared to the medium 86Rb concentration (S/M ratio). The ratio is about 1.0 in nonmetabolizing slices at 0-1° C. Once the metabolism is turned on by transferring the slices to 37° C medium, the S/M ratio rises quickly as 86Rb and Kt (for which it is a marker) are accumulated. Note that 10-3 M ouabain, through its action on the Na, K-ATPase, completely inhibits this accumulation of 86Rb and presumably Kt. 362 SODIUM CONTENT OF SLICES OF ISCHEMIC MYOCARDIUM AFTER 60 MIN. OF 37° C — 80 Sodium, millimoles /I00 gm dry weight 20 | 1 Ll 1 1 1} —r 1 15 3040 60 90 120 240 Minutes of Ischemia in vivo FIGURE 10. The effect of varying periods of ischemia on the sodium content of slices incubated at 37° C for 60 minutes is shown. The 0- minute value is the sodium content of slices of nonischemic anterior papillary muscle of the same animals. Each interval of ischemia produced a significant increase in sodium which tended to parallel the increase in total tissue water shown in figure 12. POTASSIUM CONTENT OF SLICES OF ISCHEMIC MYOCARDIUM AFTER 60 MINUTES AT 37°C 301 Potassium, millimoles /I00 gm dry weight 1 1 1} J 0 1 1 1 {} 1 15 3040 60 90 120 240 Minutes of Ischemic in vivo FIGURE 11. The effect of varying periods of ischemia on the potassium content of slices incubated at 37° C for 60 minutes is shown. The 0- minute value is the potassium content of slices prepared from nonischemic anterior papillary muscle of the same animals. The decrease in potassium noted after each interval of ischemia is significantly different from con- trol. IDS of incubated slices of tissue injured by episodes of in vivo ischemia. The alteration in IDS of incubated slices of injured tissue is shown in figure 12. There was no change following less than 40 minutes of in vivo ischemia, but, at this time, the IDS became significantly increased. The IDS of the damaged tissue continued to increase to a maximum of 300 ml/100 g dry weight. However, the IDS never became equal to the TTW. The effect of 60 minutes of severe ischemia in vivo on TTW and IDS of incubated slices under several different conditions is shown in figure 13. Slices of PP incubated at 37° C showed the increased TTW and IDS just described. Equivalent degrees of swelling occurred in both nonischemic (control) and in- jured tissue incubated at 0° C, but the IDS of control tissue was not increased by cold swelling. During subsequent warm incubation, the control tissue restored volume and the IDS still remained unchanged. In contrast, the damaged tissue was unable to restore volume during warm incubation and the IDS continued to increase. ATP and CP changes in incubated slices following in vivo ischemia. The progressive loss of cell volume regulation in slices of ischemic tissue (figures 10-12) was associated with progressive depletion of ATP in vivo and with progres- sive inability to resynthesize CP during slice incubation. Although the injured tissue had low initial ATP levels, neither further loss nor resynthesis occurred during subsequent incubation (see figures 1 and 2 in reference 1). Some CP was synthesized in slices damaged by 15 and 30 minutes of in vivo ischemia, but, at 40 minutes, when some 70 to 80 percent of the myocytes were irreversibly injured (1), three of the five dogs showed no CP synthesis and, at 60 minutes, when nearly all the myocytes were irreversibly injured, no CP synthesis occurred during slice incubation. DISCUSSION These experiments show that total adenine nucleotides decrease quickly and markedly in severely ischemic myocardium (figure 7). By 15 minutes, tissue ATP is reduced by 65 percent and, by 40 minutes, less than 10 percent remains. This © = TOTAL TISSUE WATER OF SLICES OF ISCHEMIC MYOCARDIUM o« INULIN DIFFUSIBLE SPACE a o Oo T a Q [e} T 400 TTW, milliliters /1I00gm dry weight nN o o T FINN OOF EF DIFFUSIBLE i Lo L 1 L — 0 15 3040 6 90 120 240 Minutes of Ischemia in vivo FIGURE 12. The total tissue water (TTW) and inulin diffusible space of slices of muscle injured by varying intervals of in vivo ischemia are shown. The 0O- minute values are the results found in nonischemic control muscle slices of the same animals. The increase in the inulin diffusible space is significant at all intervals of ischemia from 40 minutes or greater. 364 60 MINUTES PERMANENT OCCLUSION 500 400} 300} 200} = © 100 37°60° 0%60° 0°60°37°60° OUABAIN 0°60°37°60° 600 500} 400} 300} mlH20/100 gm DRY WEIGHT © © 200+ 100} FIGURE 13. This bar diagram shows the tissue distribution of water in control anterior papillary (AP) muscle and injured posterior papillary (PP) muscle from four dogs. ICS = Intracellular space. IDS = Inulin diffusible space. The total height of each bar is equal to the total tissue water. Control muscle incubated at 0° C for 60 minutes developed a great increase in the ICS but no change in the IDS. Resumption of metabolism resulted in extrusion of much of the intra- cellular edema. Ouabain inhibited this process. The muscles damaged by 60 min- utes of ischemia in vivo also showed an increased ICS following cold incubation and extrusion of some of this edema by subsequent warm incubation. However, the IDS of injured slices was increased under all conditions compared to control slices. This result is considered to indicate that inulin has penetrated water from which it was previously excluded, namely, by passing through the sarcolemma into the ICS. The alternative explanation, i.e., that injured slices have in- creased interstitial edema, is not supported by the ultrastructural findings. Note that the IDS never becomes equal to the total tissue water, possibly be- cause inulin is still excluded from an intracellular compartment such as mito- chondria. loss of ATP is accompanied by a moderate decrease in ADP and a 600 percent rise in AMP. However, the increase in tissue AMP accounts for only a small fraction of the nucleotide lost. Although rapid loss of ATP during ischemic or anoxic injury has been observed previously (9-11), in general, the losses have not been as marked as those noted in our study. The greater ATP depletion observed in our study is undoubtedly the result of our improved sampling technique: Use of thioflavine S provides relatively pure samples of severely ischemic tissue with little contamination from normal or mildly ischemic tissue. The massive destruction of ZAd also has been observed previously (9-11,22,23), but has not been the subject of much comment. This decrease in myocardial IAd occurs in part because of degradation of AMP by the 5' nucleotidase of the cell membrane to adenosine which is further degraded to inosine, hypoxanthine, and xanthine (9,11,22-25). Virtually all of the adenine nucleotide loss can be accounted for by equivalent increases in the above metabolites in severely 365 ischemic PP (unpublished data). Adenine nucleotide resynthesis cannot occur while the tissue is ischemic and it occurs slowly by either the salvage or de novo pathways in oxygenated myocardial tissue (26-28). Thus, even a brief episode of reversible ischemia produces a striking and relatively persistent depletion of adenine nucleotides. Relationship of “PO, to Ultrastructural and Ion Gradient Changes Sudden onset of severe ischemia is associated within seconds with proton accumulation, onset of anaerobic glycolysis, decrease in cellular glycogen, accumulation of lactate and acylcarnitine, and loss of CP, ADP, and AMP (29-32). These changes progress with time and are well advanced at 15 minutes. Neverthe- less, the injury is still reversible at this time. The ultrastructure of the tissue including the sarcolemma remains intact, except for the loss of glycogen, relaxation of myofibrils, and other mild changes. Incubation of slices of PP injured by this period of ischemia results in a slight increase in TTW and a slight decrease in ion gradients. Thus, myocardial cells with a mean ATP of 2.12 umoles/g of wet tissue maintain intact structure and volume control. In addition, restoration of arterial flow in vivo is associated with resumption of contractile function and prevention of necrosis (2,4). Thus, myocardial cells will tolerate a 65 percent loss of ATP and resume contractile function (1) in vivo. Restoration of function does not appear to be dependent on net resyn- thesis of ATP because total tissue ATP content remains low during the first 24 hours of reflow (unpublished data). The 2.12 umoles of ATP/g wet weight represents an ATP concentration of 2.8 mmoles of ATP/1 of intracellular fluid (ICF). (This calculation is based on the assumptions that all ATP within the cells is available to maintain the structure and function of the myocyte, and that the tissue measured ATP is primarily myocyte ATP.) Because not all ischemic cells in the area under study enter the irreversible state simultaneously, it is not possible to calculate the precise intracellular ATP concentration associated with transition to irre- versibility. However, at 60 minutes, when virtually all myocytes in the zone of severe ischemia are dead, the intracellular ATP concentration is about 0.55 mmol/l ICF and, at 40 minutes, when 72 percent are dead, it is 0.65 mmol (table 2). Thus, ischemic cell death in this model is associated with levels of tissue ATP of less than 8 percent of control or with calculated intracellular ATP concentration of less than 0.7 mM. The 0.55 mmol mean ATP concentration in the ICF observed at 60 minutes approximates the estimated Kj of many intracellular enzymes which in- volve ATP (33). Thus, it is possible that ATP deficiency is the principal cause of the failure in ion gradient generation (34). However, the failure to maintain ion gradients at 40 and 60 minutes (figures 10 and 11) also occurs simultaneously with the appearance of ultrastructural changes characteristic of irreversibility (figure 3). Essentially, all gradients are lost at the time that most of the myocytes in the slices are lethally injured. Thus, one cannot distinguish an effect of ATP deficiency on gradients from the effect of plasmalemma defects shown in figures 3-5. The plasmalemmal defects probably permit a rapid backleak of ions and thereby cause a continuous loss of ion gradients, even if the Nat, Kt pump was fully operational. 366 TABLE 2. Calculation of Intracellular ATP* Minutes of ATP TTW - IDS = ICF mmol ATP In Vivo Ischemia mmol/100 g dry ml/100 g dry 1 ICF 0 2.8 345.5 83.7 261.8 10.6 15 1.06 364.8 83.7 281.1 3.8 40 0.19 370.0 83.7 286.3 0.65 60 0.19 425.5 83.7 341.8 0.55 *These data are derived from tables 2 and 3 of reference 1. The mean inulin diffusible space (IDS) of the incubated control tissue of the anterior papillary muscle of 21 dogs was 83.7 * 3.94 ml/100 g dry tissue. There were 21 dogs in the 0 time (control) group and five, five, and four dogs respectively at 15, 40, and 60 minutes. In calculation of the intracellular adenosine triphosphate (ATP) concentration, IDS was held constant at 83.7 ml/100 g dry tissue because the in- crease in IDS observed in slices containing cells irreversibly injured in vivo most likely is due to penetration of inulin into the intracellular fluid (ICF) rather than to a real increase in the extracellular fluid space of the tissue. This assumption is supported by the ultrastructural features of injury which in- clude little interstitial edema but massive cell swelling. Calculation of intra- cellular ATP concentration is based on the assumption that the ATP measured is primarily myocyte ATP. TTIW = Total tissue water. Sarcolemmal Changes Although the observed plasmalemmal breaks are quite striking, their presence must be interpreted with caution. For example, small defects could be artifacts caused by grazing cuts of pinocytotic vesicles. Thus, we require presence of a sharp cutoff with a trilaminar membrane structure on either side of the break, such as that shown in figure 5, before accepting an apparent break as a true change in sarcolemmal structure. The possibility that membrane breaks are induced by fixation of the tissue for electron microscopy also must be considered. How- ever, breaks in the plasmalemma of the sarcolemma always occur in cells which also exhibit other ultrastructural signs of irreversibility. Development of membrane breaks in irreversibly injured cells by swelling slices in the cold is further evidence that the sarcolemma of cells injured by ischemia in vivo is different from the sarcolemma of control tissue. Even though myocytes of PP and AP swell to an equivalent extent, plasmalemma breaks occur only in the ischemic cells while the plasmalemma remains intact in the control cells. Development of myofibrillar contraction bands and mitochondrial calcium accumulation in the irreversibly injured cells either following reperfusion in vivo or slice incuba- tion in vitro is further evidence that the sarcolemma of irreversibly injured cells is no longer capable of excluding a massive influx of calcium from the extracellular space. 367 Alterations in the IDS of slices injured by in vivo ischemia and incubated at 37° C in vitro also are quite striking. In classic physiological terms, inulin is confined to the extracellular compartment, and the difference between TTW and IDS should be the ICF volume. Using this classic assumption, analysis of the parallel increases in TTW and IDS with increasing duration of ischemia (figure 12) suggests that the ICF volume stays roughly constant and that the extracellular space increases. However, ultrastructural data clearly show that the swelling is principally intracellular. We conclude that inulin has penetrated the intracellular space from which it was previously excluded. In contrast, cold swollen control myocytes, although increasing their total water content, do not have an increase in IDS (figure 13). Thus, several lines of evidence indicate that irreversible ischemic injury is closely associated with defective membrane structure and function. The pathogenesis of this alteration remains unknown. Relation Between Adenine Nucleotide Depletion and Cell Death Results of the studies presented in this paper show a strong association between adenine nucleotide depletion and irreversible injury of myocytes. Whether or not this association is a cause-and-effect relationship has not been estab- lished. However, it may be that the absence of a reaction driven by ATP causes development of irreversibility. The tissue ATP concentration, after 40 to 60 minutes of ischemia, is at or below the Kp of many ATP-dependent reactions (33). If the remaining ATP is compartmentalized, local ATP concentrations, such as the sarcoplasmic ATP concentration, may be even lower than the 0.5 mM estimated to be present. Thus, at or around the time of transition to irreversibility, ATP is sufficiently low to affect the reaction rate of many intracellular enzymes. It is worth speculating that membrane defects develop when ATP is too low to permit a membrane phosphorylation reaction(s). On the other hand, there are a large number of reactions occurring simul- taneously in ischemic cells which are not dependent on ATP and which could also cause or contribute to the development of irreversibility. These include, for example, accumulation of acylcarnitine which may act as a detergent on the cell membrane and cause the observed breaks (32), or proton excess with consequent enzyme denaturation. Thus, the hypothesis that depletion of ATP or high energy phosphates may be a proximate cause of the development of irreversibility is viable but not proven. Furthermore, ATP depletion is not an indicator associated with irreversibility in all tissues (35,36). Whether or not ATP is causally related to irreversibility, low ATP and sarcolemmal alterations are early markers which indicate that myocardial cells injured by ischemia have passed the '"point of no return." 368 10. 11. 12. REFERENCES Jennings RB, Hawkins HK, Lowe JE, Hill ML, Klotman MS, Reimer KA: Relation between high energy phosphate and lethal injury in myocardial ischemia in the dog. Am J Pathol 92:187-214, 1978 Jennings RB, Sommers H, Smyth GA, Flack HA, Linn H: Myocardial necrosis induced by temporary occlusion of a coronary artery in the dog. Arch Pathol 70:68-78, 1960 Whalen DA, Hamilton DG, Ganote CE, Jennings RB: Effect of a transient period of ischemia on myocardial cells. I. Effects on cell volume regula- tion. Am J Pathol 74:381-398, 1974 Kloner RA, Ganote CE, Whalen D, Jennings RB: Effect of a transient period of ischemia on myocardial cells. II. Fine structure during the first few minutes of reflow. Am J Pathol 74:399-422, 1974 Jennings RB: Cell volume regulation in acute myocardial ischemic injury. Acta Med Scand [Suppl] 587:83-93, 1976 Ganote CE, Jennings RB, Hill ML, Grochowski EC: Experimental myocardial ischemic injury. II. Effect of in vivo ischemia on dog heart slice func- tion in vitro. J Mol Cell Cardiol 8:189-204, 1976 Grochowski EC, Ganote CE, Hill ML, Jennings RB: Experimental myocardial ischemic injury. I. A comparison of Stadie-Riggs and free-hand slicing techniques on tissue ultrastructure, water and electrolytes during in vitro incubation. J Mol Cell Cardiol 8:173-188, 1976 Braasch W, Gudbjarnason W, Puri PS, Ravens KG, Bing RJ: Early changes in energy metabolism in the myocardium following acute coronary artery occlu- sion in anesthetized dogs. Circ Res 23:429-438, 1963 Jones CE, Thomas JX, Parker JC, Parker RC: Acute changes in high energy phosphates; nucleotide derivatives and contractile force in ischaemic and non-ischaemic canine myocardium following coronary occlusion. Cardiovasc Res 10:276-282, 1976 Allison TB, Ramey CA, Holsinger JW Jr: Transmural gradients of left ven- tricular tissue metabolites after circumflex artery ligation in dogs. J Mol Cell Cardiol 9:837-852, 1977 Parker JC, Jones CE, Thomas J Jr: Effect of ischemia and infarction on regional content of adenine nucleotides and derivatives in canine left ven- tricle. Cardiology 61:279-288, 1976 Kloner RA, Reimer KA, Jennings RB: Distribution of collateral flow in acute myocardial ischemic injury--Effect of propranolol therapy. Cardiovasc Res 10:81-90, 1976 369 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Reimer KA, Jennings RB: The wavefront phenomenon of myocardial ischemic cell death. II. Lab Invest 40:633-644, 1979 Jennings RB, Wartman WB, Zudyk ZE: Production of an area of homogenous myocardial infarction in dog. Arch Pathol 63:580-585, 1957 Kloner RA, Ganote CE, Reimer KA, Jennings RB: Distribution of coronary arterial flow in acute myocardial ischemia. Arch Pathol 99:86-94, 1975 Jennings RB, Ganote CE: Structural changes in myocardium during acute ischemia. In Proceedings of the First Joint US-USSR Symposium on Myocardial Metabolism, Ponte Vedra, Florida, November 4-6, 1973. Circ Res 35 (suppl IIT):156-172, 1974 Lowry OH, Passoneau JR: A Flexible System of Enzymatic Analysis. New York, Academic Press, 1972, pp 194-201 Lamprecht W, Trautschold I: Determination of ATP with hexokinase and glucose-6-phosphate dehydrogenase. In Methods of Enzymatic Analysis, edited by HV Bergmeyer. New York, Academic Press, 1974, pp 2101-2110 Jawouk D, Gruber W, Bergmeyer HV: Adenosine-5'-diphosphate and adenosine- 5'-monophosphate. In Methods of Enzymatic Analysis, edited by HV Bergmeyer. New York, Academic Press, 1974, pp 2127-2131 Jennings RB, Moore CB, Shen AC, Herdson PB: Electrolytes of damaged myo- cardial mitochondria. Proc Soc Exp Biol Med 135:515-522, 1970 Jennings RB, Lowe JE, Hawkins HK, Reimer KA: Ultrastructural changes in acute myocardial ischemic injury. In Brain and Heart Infarct, II, edited by KJ Zilch. Munich, Bergman Verlag, in press, 1979 Thomas RA, Rubio R, Berne RM: Comparison of the adenine nucleotide metabolism of dog atrial and ventricular myocardium. J Mol Cell Cardiol 7:115-123, 1975 Imai S, Riley AL, Berne RM: Effect of ischemia on adenine nucleotides in cardiac and skeletal muscle. Circ Res 15:443-450, 1964 Gerlach E, Deuticke B, Dreisbach RH: Der Nucleotid-Abbau der Sauer stoffmangel and seine mogliche Bedeutung fur die coronarydurchbluting. Naturwissenschaften 50:228-229, 1963 Benson ES, Evans GT, Hallaway BE, Phibbs G, Freier EF: Myocardial creatine phosphate and nucleotides in anoxic arrest and recovery. Am J Physiol 201:687-693, 1961 Reibel DK, Rovetto MJ: Myocardial ATP synthesis and mechanical function following oxygen deficiency. Am J Physiol 234:H620-H624, 1978 370 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. Berne RM, Rubio R: Adenine nucleotide metabolism in the heart. Circ Res 25 (suppl IIT):109-118, 1974 Zimmer HG, Trendelenberg C, Kammermeier H, Gerlach E: De novo synthesis of myocardial adenine nucleotides in the rat. Acceleration during recovery from oxygen deficiency. Circ Res 32:635-642, 1973 Williamson JR: Glycolytic control mechanism. II. Kinetics of intermediate changes during the aerobic anoxic transition in perfused rat heart. J Biol Chem 241:5025-5036, 1966 Wollenberger A, Krause E: Metabolic characteristics of the acutely ischemic myocardium. Am J Cardiol 22:349, 1968 Jennings RB: Early phase of myocardial ischemic injury and infarction. Am J Cardiol 24:753-765, 1969 Idell-Wenger JA, Neely JR: Effects of ischemia on myocardial fatty acid oxidation. In Pathophysiology and Therapeutics of Myocardial Ischemia, edited by AM Lefer, G Kelliher, M Rovetto. New York, Spectrum Publication, 1977, pp 227-238 Decker K: Quantitative aspects of biochemical mechanisms leading to cell death. In Pathogenesis and Mechanisms of Liver Cell Necrosis, edited by D Klepper. Baltimore, University Park Press, 1975, pp 45-56 Hoffman JF: Tonic transport across the plasma membrane. Cell membranes. In Biochemistry, Cell Biology and Pathology, edited by G Wessman, R Claiborne. New York, New York Hospital Practice, 1975, pp 95-103 Kleihues P, Kobayashi K, Hassman K: Purine nucleotide metabolism in the cat brain after one hour of complete ischemia. J Neurochem 23:417-425, 1974 Kleihues P, Hassman K, Pegg AE, Kobayashi K, Zimmerman V: Resuscitation of the monkey brain after one hour of complete ischemia. III. Indications of metabolic recovery. Brain Res 95:61-73, 1975 371 COMPARATIVE STUDY OF CARDIOMYOCYTE MEMBRANE PERMEABILITY DEFECTS IN SEVERE MYOCARDIAL ISCHEMIA AND UPON EXPOSURE TO ISOPROTERENOL USING COLLOIDAL LANTHANUM V. G. Sharov, R. B. Jennings, H. K. Hawkins, Yu. M. Seleznev, and A. V. Martynov INTRODUCTION The idea that severely damaged myocytes exhibit changes in the permeability of their membranes early in the course of injury has been developing over the past 20 years. Alterations in permeability first were detected at the level of light microscopy, by Veress et al. (1), and by Kent (2,3), who showed that plasma proteins were detectable in the cytoplasm of necrotic muscle cells. Later, improvement of electron microscopic tracer techniques allowed disorders of the permeability of plasma membranes of injured cells to be detected much earlier (4-6). Hoffstein et al. (7) showed that cells with no ultrastructural signs of injury on the periphery of a 3.5-hour-old experimental myocardial in- farct were sufficiently damaged to allow the penetration of colloidal particles of lanthanum [La(OH)3] into their sarcoplasm. On the other hand, lanthanum did not precipitate as phosphate in markedly damaged cells in the center of the zone of ischemia, presumably because insufficient intracellular inorganic phosphate ion was available in these cells to allow precipitation. Rona et al. (8) showed that toxic doses of B-adrenergic catecholamines (isoproterenol and norepinephrine) also led to early changes in the permeability of the sarcolemma of the cardiomyocytes. They showed that peroxidase penetrated into the sarcoplasm of cells which had no visible ultrastructural alterations, as well as into damaged cells, while peroxidase remained extracellular in con- trol hearts. Because of the similarity of the ischemic- and catecholamine- mediated changes in sarcolemmal permeability in cells showing no other signs of damage, the possibility that the increased sarcolemmal permeability observed in cells on the periphery of the infarct was mediated by excess levels of B- adrenergic catecholamines was investigated. In the present work, we attempted to determine whether the nature of damage to permeability of the cardiomyocyte membranes varies with the agent or condi- tion causing the damage. Accordingly, we compared membrane permeability defects From the All-Union Cardiology Research Center, USSR Academy of Medical Sciences, Moscow, USSR, and from the Department of Pathology, Duke University Medical Center, Durham, North Carolina, USA. 373 in contractile cells of the heart in cases of catecholamine-mediated ''calcium" myocardial necrosis to changes in severe ischemia in experimental myocardial infarcts. MATERIALS AND METHODS The electron-microscopic tracer used was lanthanum. The toxicity of this metal can be avoided only in fixed tissue, but fixation limits its usefulness. Nevertheless, it was selected because it allows control tissue to be compared to experimental tissue and because it is the smallest available electron-micro- scopic marker of sarcolemmal permeability. The mean diameter of a colloidal particle of La(OH)3 is 20 A (9). 1It is much smaller than other tracers: The diameter of a ferritin particle is 100-200 A, of peroxidase--35-40 A, of cyto- chrome c--30 A (9-12). Comparative study of tissue subjected to various treat- ments allowed us to avoid artifacts related to the use of the toxic tracer, since these artifacts, with identical processing, would be expected to be simi- lar in control and experimental tissues. Our most important aim was to assess early changes in membrane integrity, and in this situation, colloidal lanthanum seems to be the best marker. Colloidal lanthanum was prepared from La(NO3)3 by a method suggested by Revel and Karnovsky (13). The final mixture used for fixation consisted of 1 percent freshly prepared La(OH)5 (pH = 7.8) and 4 percent glutaraldehyde buf- fered with sodium cacodylate (pH = 7.4). Acute myocardial ischemic injury was induced by proximal occlusion of the circumflex branch of the left coronary artery in the anesthetized open-chest dog (14). Occlusions of 15-minute durations were used to produce reversible ischemic injury while occlusions of 40-minute durations were used to induce irreversible injury. Severely ischemic tissue was identified in the posterior papillary muscle of the excised cooled heart by use of an intravenous injection of thioflavine S (TS) one circulation time prior to excision of the heart. Thin tissue slices (about 0.1-0.3 mm thick) were cut using a sharp razor from the posterior (damaged) and anterior (control) papillary muscles parallel to the long axis of the fibers (15). Some of the sections were immediately fixed, while the remaining sections were incubated for 30 minutes in an oxygenated Krebs-Ringer's solution at 37° C before fixation (15,16). Subcutaneous iso- proterenol (30 mg/kg of body weight) was used to induce necrotic myocyte foci containing calcium, hereafter referred to as ''calcium necrosis," in 150-200 g Wistar rats. The animals were sacrificed 1 and 6 hours after the injection, and the papillary muscles of the left ventricles of the experimental rats were used for study. The myocardium was fixed with lanthanum overnight with constant agitation, rapidly dehydrated in alcohol of increasing concentrations, and placed in Epon- 812 (17). Some of the material, after glutaraldehyde fixation with lanthanum, was postfixed in buffered osmic acid, not containing lanthanum (7). 374 RESULTS The results were similar with or without postfixation of the myocardium with osmic acid. After 15 minutes of acute ischemia (reversible injury), all the colloidal lanthanum, as in the control material, was still localized outside the sarcoplasm (figure 1). Colloidal particles penetrated into many of the car- diomyocytes after 40 minutes of acute ischemia, i.e., when many cells of the muscle were first biologically irreversibly injured (18). In many cardiomyo- cytes, lanthanum accumulated in the parts of the sarcoplasm over the myofibrils, leading to a picture of variable staining. Such variable staining of myo- fibrils occurred even within the limits of a single cell (figure 2). The layer beneath the outer membrane of the mitochondria was stained with lanthanum only in those locations where intensive accumulations of lanthanum could be seen over the myofibrils. No lanthanum precipitate was observed in the mitochondrial matrix and lumen of the cisternae of the sarcoplasmic reticulum. The effect of 30 minutes of incubation of thin slices of tissue injured in vivo by 15 minutes of ischemia did not change the picture of distribution of lanthanum in the myocardium. The damaged tissue was similar to the control tissue; all of the colloid remained outside the limits of the sarcoplasm (fig- ure 3). However, incubation of slices of myocardium subjected to ischemia for 40 minutes prior to testing led to the appearance of large electron-dense accumu- lations of colloidal lanthanum in the matrix of the mitochondria and lumen of the cisternae of the sarcoplasmic reticulum of some cells (figure 4). Colloidal lanthanum did not penetrate into the sarcoplasm of the cardiomyo- cytes of control rats. All La(OH)j3 could be seen as an electron-dense, homogene- ous or granular precipitate, localized over the surface of the sarcolemma or its derivates (figure 5). One hour after administration of isoproterenol, about half of the apparently unaltered or little-altered cardiomyocytes contained La(OH)3 in the sarcoplasm, located primarily around the outer membranes of the mitochondria and the exterior surface of the lipid drops (figure 6). Some 6 hours after injection of the isoproterenol, the sarcolemma of most of the appar- ently unaltered or little-altered contractile cells became permeable to colloidal lanthanum. The cardiomyocytes with signs of contracture of myofibrils all con- tained La(OH)3 in their sarcoplasm, and the colloid disappeared from the outer surface of the sarcolemma (figure 7). However, electron-dense precipitates were not found around the mitochondria of these cells after glutaraldehyde fixation without postfixation of the tissue with osmic acid. In order to eliminate the possibility of localization due to a histochemical reaction dependent on osmic acid, the electron-dense precipitates in these cells were investigated with energy dispersive X-ray (EDAX). Results of this analysis confirmed the presence of lanthanum in the precipitates (see inset, figure 7). One possibility is that the osmic acid merely helps to attach the precipitate of colloidal lanthanum to the cell after glutaraldehyde fixation alone and prevents it from being washed out during dehydration and embedding. In irreversibly al- tered cardiomyocytes containing contraction bands and torn sarcomeres, the electron-dense sediments of colloidal lanthanum, regardless of the method of fixation, appeared in the matrix of almost all mitochondria and in the lumina of some cisternae of the sarcoplasmic reticulum (figures 8 and 9). The appar- ently unaltered cells, in which the colloidal lanthanum was located around the 375 9L¢ FIGURE 1. Dog cardiomyocytes after 15 minutes of acute ischemia. outside the sarcolemma and primarily in the lumen of the intercalated disk (id). 15,000. & Ja. : . 9 5 “ vo os - “ ' ~~ 3 - er” Ft - . v A hy ’ : ~, Lvs on ES Cag : Sr -* Taye iy vw * - - 8 % . . os ao 3 Colloidal lanthanum is located Magnification x LLE FIGURE 2. Dog cardiomyocytes after 40 minutes of acute ischemia. Colloidal lanthanum penetrates through the sarcolemma, staining the myofibrils and membranes of mitochondria (M), but remains outside their matrix. L = Capillary lumen. Magnification x 10,000. 8L¢ FIGURE 3. A 5-2 apm Dog cardiomyocyte after 15 minutes of acute ischemia in vivo and an additional 30 minutes of incubation in vivo. Colloidal lanthanum remains outside the sarcoplasm, and primarily in the lumen of the intercalated disk (id). Magnification x 15,000. 6L¢ of the sarcoplasmic retic ulum (SPR) (see arrow). Magnification x 30,000. 08¢ FIGURE 5. Cardiomyocytes of control rats. Colloidal lanthanum is located outside the sarcoplasm in the lumen of the intercalated disk (id) and in intercellular slits. M = Mitochondria. L = Capillary lumen. Magnification x 25,000. 18¢ FIGURE 6. Rat heart cardiomyocytes 1 hour after administration of isoproterenol. Colloidal lanthanum is found in the lumen of the intercalated disk (id) and in the sarcoplasm of cells in intimate contact with external mitochondrial membrane (M) (see arrow) and at the boundary of lipid drops (L). Magnification x 25,000. 8¢ FIGURE 7. Rat heart cardiomyocyte showing signs of initial lysis of the myofibrils 6 hours after administration of isoproterenol. All mitochondria (M) are surrounded with the residue of colloidal lanthanum (see arrow); the colloid disappears from the boundary membrane of cells. M = Mitochondria. Magnification x 30,000. The inset shows an electron microscope EDAX analysis of electron-dense precipitate in intimate contact with the cell mitochondria shown in the figure. The recorded peaks correspond to lanthanum. €£8¢ FIGURE 8. Irreversibly altered cardiomyocyte of rat heart 6 hours after injection of isoproterenol. The mitochondrial matrix (M) and the lumen of the sarcoplasmic reticulum (SPR) show precipitated colloidal lanthanum (see arrow). Mf = Myofibrils. Magnification x 30,000. %8¢ FIGURE 9. Mitochondria from an irreversibly altered rat heart cell 6 hours after administration of isoproterenol. The precipitate of LaOH3 can be seen in the organelle matrix (see arrow). Magnification x 150,000. mitochondria, and the irreversibly altered cells, in which the matrix of the mitochondria was filled with large conglomerate particles of La(OH)3, were frequently located side by side (figure 10). DISCUSSION The results of the work show that, in both ischemic injury and catechol- amine toxicity, the permeability of the sarcolemma is disrupted before that of the other membranes of the cell. However, in ischemia, 20 A pores appear in the plasma membrane only in irreversibly damaged cardiomyocytes, whereas follow- ing administration of isoproterenol, the particles of colloidal lanthanum were observed in the sarcoplasm of unaltered or only slightly altered cells. The effect of isoproterenol on the myocyte is related initially to the con- trolled influx of external calcium into the cardiomyocytes (19). Toxic doses are associated with uncontrolled entry of Ca2t through an altered sarcolemma. Myocytes reversibly injured by 15 minutes of ischemia do not accumulate caZt when reperfused with arterial blood. However, cells injured by 40 minutes of ischemia accumulate extraordinary quantities of Cat when reperfused. The cast appears as calcium phosphate in mitochondria. The mitochondria load massively, presumably because of increased cat entry through a damaged sar- colemma (20). These data show that cells damaged by 40 minutes of ischemia allow penetration of the colloidal lanthanum into the sarcoplasm when they are incubated in vitro at the same time that uncontrolled calcium influx occurs in vivo when these damaged cells are reperfused for arterial blood. Thus, col- loidal lanthanum is a convenient electron-microscopic marker of defects in the plasma membrane of the cardiomyocytes sufficient to allow uncontrolled trans- membrane flow of calcium, regardless of the direct cause of its appearance or the extent of intracellular ultrastructural changes. Like Rona et al. (8), who used the peroxidase tracer, we used colloidal lanthanum and demonstrated severe disruption of the permeability of the outer and inner membranes of the mitochondria and the membranes of the sarcoplasmic reticulum after irreversible isoproterenol damage. However, acute ischemia of 40-minutes duration failed to result in the appearance of particles of colloidal lanthanum in the mitochondrial matrix or in the lumen of the sarcoplasmic retic- ulum. This confirms the results of Hoffstein et al. (7) and Hawkins et al. (17), who also showed that lanthanum does not enter mitochondria in cells injured by longer periods of ischemia. However, reincubation in vitro of myocardium that was subjected to irreversible ischemic damage in vivo leads to the appearance of La (OH) 3 in the mitochondrial matrix and in the lumen of the sarcoplasmic reticulum. Since Ca?t is available in the incubation media to support its intracellular accumulation in vitro and since it accumulates when the injured tissue is reperfused (21) in vivo, and since the inner membrane of the mito- chondria contains Cat under both these circumstances (21), the penetration of La3t into this space may be a CaZt-mediated effect. The results of this study suggest that early penetration of lanthanum into slightly altered cardiomyocytes located at the boundary of experimental myocar- dial infarction may be explained not by ischemic damage (7), but by the effect 385 98¢ FIGURE 10. Slightly altered and irreversibly altered rat heart cardiomyocytes 6 hours after injec- tion of isoproterenol. M = Mitochondria. L = Capillary lumen. The arrows show precipitated col- loidal lanthanum in a slightly altered cell outside the mitochondrial matrix and in an irreversibly altered cell in the mitochondrial matrix. Magnification x 25,000. of toxic doses of B-adrenergic catecholamines on the cell. On the other hand, it remains unclear why only a portion of the cells of the marginal zone of the infarct demonstrate altered permeability. Local microcirculatory deficiencies also could explain the observed changes. 387 10. 11. 12. 13. REFERENCES Veress B, Kerényi T, Hiittner I, Jellinek H: The phases of muscle necrosis. J Pathol Bact 92:511-517, 1966 Kent SP: Diffusion of plasma proteins into cells: a manifestation of cell injury in human myocardial ischemia. Am J Pathol 50:623-637, 1967 Kent SP: Diffusion of plasma proteins into cells: a manifestation of cell injury in rabbit skeletal muscle exposed to lecithinase C. Arch Pathol 88:407-412, 1969 Fahimi HD, Cotran RS: Permeability studies in heat induced injury of skeletal muscle using lanthanum as a fine structural tracer. Am J Pathol 62:143-157, 1971 Hiittner I, Rona G, Theodosis D, More RH: Ultrastructural studies on myo- cardial fibrin deposition in experimental hypertension. In Recent Advances in Studies on Cardiac Structure and Metabolism, vol 1, Myocardiology, edited by E Bajusz and G Rona. Baltimore, University Park Press, 1972, pp 376- 385 Rona G, Hiittner I, More RH: Fibrin as a natural tracer in cardiac muscle cell injury. In Present Status of Thrombosis. Its Pathophysiology, Diag- nosis and Treatment, edited by R Losito. Stuttgart and New York, Schattauer Verlag, 1973, pp 21-33 Hoffstein S, Gennaro DE, Fox AC, Hirsch J, Streuli F, Weissmann G: Col- loidal lanthanum as a marker for impaired plasma membrane permeability in ischemic dog myocardium. Am J Pathol 79:207-218, 1975 Rona G, Boutet M, Hiittner I: Membrane permeability alterations as mani- festations of early cardiac muscle cell injury. In Recent Advances in Studies on Cardiac Structure and Metabolism, vol 6, Pathophysiology and Morphology of Myocardial Cell Alteration, edited by A Fleckenstein and G Rona. Baltimore, University Park Press, 1975, pp 439-451 Hayat MA: Positive Staining for Electron Microscopy. New York, Van Nostrand Reinhold, 1975 Thoenes W, Langer KH, Paul N: [In vivo labeling of glycocalyx in rat nephrone using horseradish peroxidase] (Ger). Cytobiol 6:487-491, 1973 B&ck P: Adsorption of horseradish peroxidase to negatively charged groups. Acta Histochem 43:8-14, 1972 B8ck P: [Adsorption of horseradish peroxidase on ruthenium red-positive structures in unmyelinated nerves] (Ger). Acta Histochem 46:146-149, 1973 Revel JP, Karnovsky MJ: Hexagonal array of subunits in intercellular junctions of the mouse heart and liver. J Cell Biol 33(3):C7-Cl2, 1967 388 14. 15. 16. 17. 18. 19. 20. 21. Jennings RB, Sommers HM, Smyth GA, Flack HA, Linn H: Myocardial necrosis induced by temporary occlusion of a coronary artery in the dog. Arch Pathol 70:68, 1960 Grochowski EC, Ganote CE, Hill ML, Jennings RB: Experimental myocardial ischemic injury. I. A comparison of Stadie-Riggs and free-hand slicing techniques on tissue ultrastructure, water and electrolytes during in vitro incubation. J Mol Cell Cardiol 8:173-187, 1976 Ganote CE, Jennings RB, Hill ML, Grochowski EC: Experimental myocardial ischemic injury. II. Effect of in vivo ischemia on dog heart slice func- tion in vitro. J Mol Cell Cardiol 8:189-204, 1976 Hawkins HK, Sharov VG, Jennings RB: Cell membrane permeability defects demonstrated with colloidal lanthanum early in ischemic cell injury in dog heart. In Proceedings of the Ninth International Congress on Electron Microscopy, vol 2. Toronto, 1978, pp 306-307 Jennings RB, Ganote CE, Reimer KA: Ischemic tissue injury. Am J Pathol 81:179, 1975 Fleckenstein A, Janke J, D6ring HJ, Leder O: Key role of Ca in the produc- tion of noncoronarogenic myocardial necrosis. In Recent Advances in Studies on Cardiac Structure and Metabolism, vol 6, Pathophysiology and Morphology of Myocardial Cell Alterations, edited by A Fleckenstein and G Rona. Baltimore, University Park Press, 1975, pp 21-32 Jennings RB, Shen AC: Calcium in experimental myocardial ischemia. In Recent Advances in Studies on Cardiac Structure and Metabolism, vol 1, Myo- cardiology, edited by E Bajusz and G Rona. Baltimore, University Park Press, 1972, pp 639-655 Shen AC, Jennings RB: Kinetics of calcium accumulation in acute myocardial ischemic injury. Am J Pathol 67:441-452, 1972 389 BIOCHEMICAL EVALUATION OF DAMAGE TO THE HUMAN MYOCARDIUM WITH COMPLETE ISCHEMIA DURING OPEN-HEART SURGERY V. I. Burakovsky, D. B. Saprygin, and L. S. Kashtelian INTRODUCTION Open-heart surgery is frequently complicated by disorders in cardiac muscle function. It is thought that the main cause of damage to the myocardium in open- heart surgery is complete ischemia of the stopped heart (cardiac arrest), the duration of which varies from a few minutes to 2 or more hours. In addition to the ischemia, the myocardium is also affected by various factors related to anes- thesia, extracorporeal circulation, stress, and surgical trauma to the heart. Evaluation of the nature and degree of damage to the myocardium requires answers to the following questions: What specific metabolic and structural disorders of the myocardial cells arise under the above conditions? During what operative and postoperative periods do they appear? And, finally, is the damage reversible or does it result in death (necrosis) of the myocardial tissue? One of the earliest signs of damage to the contractile cells of the heart is a change in their membrane permeability (1,2); together with ischemia, these changes in the sarcolemma are considered the first symptom of irreversible damage (3). Direct determination of membrane defects of cardiac cells under clinical conditions is difficult, and therefore an indirect indicator of dis- orders in myocyte membrane permeability is used--that is, determination of an elevated level of intracellular myocardial enzymes in the circulating blood. This phenomenon--a sharp increase in creatine phosphokinase (CPK) and glutamic- oxaloacetic transaminase (GOT) activity--is widely used in diagnosing acute myocardial infarction. However, these enzyme tests are not very informative in evaluating damage to the heart during surgery, since the levels of CPK and GOT increase in the blood in all types of surgery (4,5). This increase is primarily due to entry of these enzymes into the bloodstream from the skeletal musculature where they are found in even greater quantities than in the myocardium. From the A. N. Bakulev Institute of Cardiovascular Surgery, USSR Academy of Medical Sciences, Moscow, USSR. 391 Therefore, in this study we evaluated the damage to the myocardium by determining the arteriovenous difference of CPK, GOT, and other enzymes, as well as myoglobin and lactate, in the heart. Furthermore, during and after heart surgery we determined the content of the serum isoenzyme CPK MB, which is specific for the myocardium. Determination of this enzyme is considered a sign of irreversible (necrotic) cardiac muscle damage (6). MATERIALS AND METHODS Clinical Studies were performed on 30 patients who were operated on in the Division of Coronary Surgery (group 1) and 30 patients who received prosthetic valves (group 2). (One patient in group 2 received an aortic prosthesis with a pros- thetic ascending segment of the aorta, implantation of the left coronary artery, and aortocoronary bypass of the right coronary artery.) The arteriovenous difference in the heart was studied in 24 patients of the first group and 22 patients of the second group. The first group of patients included two sub- groups: those operated on for the purpose of resectioning left ventricular aneurysms (10 persons), and 20 persons who received aortocoronary bypass of one, two, or three coronary arteries. The operations were performed with artificial circulation and interrupted coronary circulation (clamping of the aorta), the period of which varied from 5 minutes to 2.5 hours, under whole body hypothermia (30° C). In group 1, pro- tection of the myocardium during clamping of the aorta was achieved by external cooling of the heart (16-18° C). In group 2 protection was achieved by a com- bination of local external cooling of the heart (16-19° C) and administration into the root of the aorta or directly into the mouth of the coronary artery a cooled oxygenated cardioplegic solution containing a number of ingredients: ions of potassium, magnesium, sodium, and calcium; chlorides, glucose, mannitol, and novocaine. The osmolarity of the cardioplegic solution was 300 mosm, pH = 7.5. The solution was administered in several doses: immediately after clamp- ing of the aorta and then after every 30 minutes of ischemia, so that the tem- perature of the heart was maintained at 15-19° C. Constant electrocardiographic (ECG) testing was performed on all patients. Blood was taken to study the arteriovenous difference in the heart during two stages of surgery: immediately before clamping the aorta and after removal of the clamp from the aorta (coronary circulatory restart). Blood was taken from the right atrium, from the area of the coronary sinus with the vena cava completely clamped but the aorta free at the beginning of extracorporeal circu- lation (8-10 minutes) in the first stage, and with continuing artificial circu- lation in the second stage. Blood was simultaneously taken from the arteries. During surgery, the level of CPK MB was determined in the venous blood during the two stages mentioned above, and also before surgery at the moment of cannulation of the peripheral vessels (the beginning of anesthesia) and at the end of surgery (closure of the thorax). During the postoperative period, testing of CPK MB in the venous blood was performed after 2, 5, 8, 13, 18, and 30 hours following completion of surgery. 392 The studies also included a control group of patients with ischemic heart disease (IHD). This group consisted of 12 persons who received atrial stimula- ton of the heart with simultaneous study of CPK, CPK MB, and lactate in the arterial blood and blood from the coronary venous sinus of the heart. To take the blood samples and perform the atrial stimulation test, a bipolar electrode catheter was introduced througl. the left subclavian vein and placed in the lumen of the coronary sinus under fluoroscopic observation. The position of the electrode catheter in the lumen of the coronary sinus was constantly moni- tored by contrast radiography, ECG, and the pressure curve in the sinus. Stim- ulation was performed up to a cardiac contraction rate of 160-170/minute or until chest pains and ECG changes occurred. Biochemical Creatine phosphokinase was determined by a modification of an activated, kinetic method (7). GOT and glutamic-pyruvic transaminase (GPT) were determined by kinetic methods with activation by pyridoxal-5-phosphate (8). Enzyme activ- ity measurements were performed on a reaction rate analyzer (LKB, Sweden), with computer calculation of the kinetic curves (HP-9815A, Hewlett-Packard, USA) using standard programs supplied by the manufacturer. Each point corresponded to the mean value produced as a result of at least three repeated measurements. At the same time, constant quality control was performed by daily testing with commercial control sera (Boehringer, FRG). The CPK isoenzymes were determined electrophoretically on cellulose acetate films (Titan III, USA) with subsequent fluorodensitometry of specimens and computer calculation of densitograms (Helena, USA). As a control, periodic determination of the CPK isoenzyme spectrum was performed in control commercial sera produced by the same firm. An isoenzyme level of less than 3 percent was not taken into consideration. The cumulative activity of CPK MB, liberated into the blood from the heart over the entire period of investigation, was calculated by means of a modified Shell-Sobel equation (9). Lactate was determined by the standard enzyme method, and myoglobin was determined by the radioimmunologic method. Statistical processing of data on the arteriovenous difference was per- formed using the t-test for paired quantities. The remaining data were pro- cessed using the standard t-test and chi-square method. RESULTS Studies of CPK MB in the blood serum in all stages mentioned above showed that the isoenzyme was determined at only two or three points in almost all of the patients (95 percent). However, the dynamics of the change in the curves of CPK MB activity in the serum in some patients varied significantly, differing both in the maximum elevation of isoenzyme activity and in the time of appear- ance of peaks of CPK MB and the period of circulation in the blood. Analysis of the curves of change in CPK MB activity allowed the three most typical curve profiles to be determined, which are shown in figure 1. In most patients, the 393 CPK MB (IU/L) » o ° CPK MB(IU/L) 033 J 50 C S 40 & 30 2 20 x = 10 oO 0 L Vy operative T v v v vv 717 v v J period 0 2 4 6 8 1S 25 35 time after operation, hours FIGURE 1. Nature of changes in creatine phosphokinase isoenzyme (CPK MB) activity in patients following heart surgery over the entire period of obser- vation (three types of curves). A = Type 3 curves. B = Type 2 curves. C = Type 1 curves. curve type which we called type 1 is characterized by a slight elevation in CPK MB activity during surgery and a rapid drop within a few hours after operation (figure 1C). In another, smaller group of patients (table 1), the elevation of CPK MB activity was recorded later in comparison to the type 1 curve, and the maximum value reached was higher (figure 1B). We called these curves type 2 curves. Finally, there was a small group of eight persons (table 1) in whom the nature of the change in CPK MB activity differed sharply from the others, in that there was a significant, later rise in isoenzyme activity (figure 1A). A more detailed description of the various types of curves is shown in table 1, from which we see that the curves of the third type differ signifi- cantly in cumulative release, maximum values, and times of appearance of ele- vated activity from the curves of types 1 and 2. In order to determine the significance of this great excursion of CPK MB from the myocardium of patients after heart surgery, corresponding to type 3 curves, we compared the clinical and biochemical data. We found that out of 394 TABLE 1. Characteristics and Number of Patients for Each Type of Curve Based on Change in Activity of the Creatine Phosphokinase Isoenzyme (CPK MB) Time to Peak Peak CPK MB Number of Curve CPK MB CPK MB Release Patients Type (hours) (1IU/1) IU ml-1 kg (%) Type 1 2-4 40.3 + 3.2 <1 60.7 Type 2 6 - 8 94.2 + 10.0 10.02 + 1.7 25.3 Type 3 15 - 18 444.6 + 121.9 102.3 + 23.3 14.0 eight persons with altered CPK MB activity of type 3, six patients died with acute cardiac insufficiency. Autopsy revealed myocardial infarctions in five of the patients. In one of the victims, infarct damage to the heart was not found, but the time of clamping of the aorta (disruption of coronary blood flow) had been 148 minutes, and the postoperative period revealed definite signs of acute cardiac insufficiency; we therefore cannot exclude the possi- bility of diffuse damage to the myocardium. Of the two survivors in this group with type 3 curves, in one the ECG showed signs of focal damage to the left ventricle, and in the other cardiac activity was restored only after six defibrillations. From the results of this clinical-biochemical comparison we concluded that changes in CPK MB activity following a type 3 curve probably indicate the de- velopment of zones of irreversible (necrotic) damage in the heart. This con- clusion was also confirmed by the fact that none of the patients with a type 1 or 2 curve, either those surviving or those who died for reasons unrelated to cardiac insufficiency, showed indications of myocardial infarction. We further established that the type 3 CPK MB curves do not result from surgical trauma (figure 2). Type 3 curves were not found in any of the patients who experienced resection of a postinfarct aneurysm of the left ventricle, which is a very traumatic operation for the heart, but rather were most frequently seen (25 percent) in patients with aortocoronary bypass, which does not involve direct trauma to the cardiac muscle. In a smaller number of cases (10 percent), these type 3 curves appeared in patients with valve prostheses. On the other hand, type 1 curves, representing the minimum expulsion of CPK MB from the myocardium into the bloodstream (table 1), were also most frequently seen in the group of patients with resection of left ventricular aneurysms. However, type 3 curves were found in patients in groups 1 and 2 with aortal clamping times exceeding 50 minutes and 80 minutes, respectively (figure 3), but were not found in any patients with disruption of coronary circulation less than these time intervals. This dependence of type 3 curves on the time of clamping 395 30, ACB 2 20. o 8 a S 8 0 . = AR ol FIGURE 2. Frequency of distribution of patients (in percent) with type 3 curves as a function of the degree of surgical trauma. ACB = Patients with aortocoro- nary bypass. AR = Patients with resection of aneurysms of the left ventricle. VR = Patients with heart valve replacement. 40. n=15 32 30. £ Q n=15 8 20] © Te [0] € 10 | = n=15 n=15 0 L ee IT.< 50 IT.>50 IT<80 IT>80 Ischemic time (LT), min ACB + AR VR FIGURE 3. Frequency of occurrence of type 3 curves (in percent) among patients in groups 1 and 2 with different times of aortic clamping (total ischemia). ACB + AR = Patients with aortocoronary bypass and resection of left ventricular aneurysm. VR = Patients with heart valve replacement. 396 of the aorta indicates that the significant expulsion of CPK MB into the blood- stream results from irreversible ischemic damage to the myocardium. However, the reasons for finding CPK MB in the remaining patients with type 1 and 2 dynamics remain unexplained. These dynamics correspond, as shown in table 1, to a slight increase in CPK MB activity and cumulative liberation of CPK MB into the bloodstream. Since the peak isoenzyme activity in these patients occurred at the end of surgery or in the hours immediately following surgery, we were interested to determine whether there was an arteriovenous difference in cardiac isoenzyme at different stages during the surgery: before clamping of the aorta and immediately after removal of the clamp from the aorta. In addition to the enzymes, we studied the content of myoglobin and lactate in the arterial blood and the blood flowing from the heart in these stages (see "Materials and Methods'). The results of the studies showed that, in the stage preceding clamping of the aorta and heart surgery, there is "leakage" of intracellular enzymes, which is not accompanied by production of lactate (figure 4A). Simultaneously, as follows from figure 4, the myocardium loses myoglobin, whereas the content of alanine aminotransferase, a hepatospecific enzyme, is the same in the arte- rial blood and the venous blood flowing from the heart. The creatine phospho- kinase isoenzyme CPK MB is a myocardial-specific isoenzyme and was found in this stage in the coronary sinus blood in almost 50 percent of our patients. MYO A B p<0.0l 200. MYO p<0.01 CPK p<0.00! goT p<0.00! LAC GPT p<0.0I p>0.05 Enzyme activity or concentration, % oS o OL FIGURE 4. Arteriovenous difference in the heart from enzymes, myoglobin (MYO) and lactate (LAC) in two stages of operation. A = Stage before clamping of the aorta and beginning of surgery. B = Stage after removal of the clamp from the aorta and restart of coronary circulation. The activity of enzymes and concen- tration of components in arterial blood was taken as 100 percent. CPK = Creatine phosphokinase. GOT = Glutamic-oxaloacetic transaminase. GPT = Glutamic-pyruvic transaminase. 397 After the period of total ischemia (clamping of the aorta), an increase was noted in the arteriovenous ''gradient" in the heart for the enzymes and for myoglobin, which was accompanied by significant production of lactate (figure 4B). The frequency of CPK MB determination in the venous blood of the heart in this stage increased and reached 80 percent of the cases. We did not succeed in detecting losses of CPK, CPK MB, and GOT by the myo- cardium in patients in the control group subjected to the '"pacing' procedure, either before or after coupling of the rhythm. However, we did note consump- tion of lactate before atrial stimulation, and production of lactate afterward. DISCUSSION Testing of the level of enzymes contained in large quantities in the myo- cardium (CPK, CPK MB, GOT, etc.) in the blood is widely used in clinical prac- tice to establish irreversible, necrotic damage to the cardiac muscle (myocardial infarction). Of these, only CPK MB is absolutely specific for myocardial tissue, and therefore determination of this isoenzyme in the blood is considered a re- liable sign of irreversible myocardial damage (6) and myocardial infarction in patients (10,11). From this standpoint, we were surprised that a number of other researchers (9,12-15), as well as ourselves, observed almost 100 percent frequency of CPK MB appearance in the circulating blood in heart surgery cases using artificial circulation. This fact remained difficult to explain, since it required that we assume the presence of infarct damage to the myocardium in almost all heart surgery patients. Using serial studies of CPK MB during and after heart surgery, we found a significant difference in the variation of CPK MB activity, the maximum values of enzyme activity, and cumulative liberation of CPK MB in the bloodstream in certain patients (figure 1 and table 1). We found high CPK MB activity and significant liberation of CPK MB into the bloodstream accompanied by type 3 curves only in patients with confirmed myocardial infarction. Correlation be- tween duration of clamping of the aorta and type 3 curves and the simultaneous lack of correlation with surgical trauma (figures 2 and 3) led us to conclude that type 3 curves reflect necrosis of myocardial tissue related to ischemic (hypoxic) cardiac muscle damage. Of particular interest from our point of view are the results of studies of arteriovenous difference in the heart in various stages of surgery based on enzymes, myoglobin, and lactate (figure 4). As we found, even before the stage of clamping of the aorta and surgery, the myocardium loses CPK, GOT, and myo- globin, as well as CPK MB (figure 4A). It is important to emphasize that this "leakage" of intracellular components from the heart, which indicates disorders in permeability of the sarcolemmal membranes, is not accompanied by production of lactate, i.e., this leakage may occur without significant or serious meta- bolic disorders in the cardiac cells. In the second stage of our study (after removal of the clamp from the aorta), this liberation of enzymes and myoglobin by the myocardium increased 398 and was accompanied by significant production of lactate, i.e., disruption of membrane permeability increased and was combined with metabolic changes in the myocardial cells. We know that ischemic damage to cardiac cells, which reaches the irreversi- ble stage, is accompanied by sharp changes in the structural organization of the sarcolemmal membranes (3), which leads to significant losses of intracellular enzymes and other components from the myocardium into the bloodstream. This liberation of CPK MB, which indicates the death of myocardial cells, was found in the form of type 3 curves. However, disorders in permeability of the sar- colemmal membrane may arise following other actions on the heart cells, e.g., under the influence of catecholamines (2,16), without alteration of the cellular organelles (16). Probably, it is just this sort of thing which causes detection of CPK MB in the blood in iatrogenic heart trauma, without documented myocardial infarction (17). Therefore, one cannot exclude that the reason for the loss of enzymes and myoglobin by the myocardium which we discovered before clamping of the aorta is not hypoxia due to microcirculation disorders, etc., but rather the influence of catecholamines (norepinephrine) or a number of other hormones on membrane permeability, systemic or local liberation of which may be induced by manipulations of the vena cava and surgical stress. The later effect of total ischemia (aortic clamping) may worsen this damage and lead to the devel- opment of myocardial necrosis. The probable sequence of events in such "non- ischemic" damage to the heart differs from ischemia beginning with defects of the sarcolemmal membrane, leading to loss of important intracellular components, and as a result of this, metabolic changes. With ischemic damage, the flow of events is the reverse: metabolic disorders cause damage to the structural or- ganization of the membrane. The data which we obtained indicate that both of these types of damage to cardiac cells, either individually or in combination, probably occur during open-heart surgery with extracorporeal circulation. However, further clinical and experimental studies are required for final determination of the mechanisms of damage to the myocardium during heart surgery. The authors would like to express their deep gratitude to the following colleagues at the Institute of Cardiovascular Surgery: A. I. Malashenkov, M. E. Polonskaya, L. A. Migalina, and A. I. Belichkov for technical assistance in performing a number of studies and analysis of the clinical materials. 399 10. 11. 12. 13. REFERENCES Hoffstein S, Gennaro DE, Fox AC, et al.: Colloidal lanthanum as a marker for impaired plasma membrane permeability in ischemic dog myocardium. Am J Pathol 79:207-218, 1975 Rona G, Boutet M, Hiittner I: Membrane permeability alterations as mani- festation of early cardiac muscle cell injury. In Recent Advances in Studies on Cardiac Structure and Metabolism, vol 6, Pathophysiology and Morphology of Myocardial Cell Alterations, edited by A Fleckenstein and G Rona. Baltimore, University Park Press, 1975, pp 439-451 Ganote CE, Jennings RB, Hill ML, et al.: Experimental myocardial ischemic injury. II. Effect of in vivo ischemia on dog heart slice function in vitro. J Mol Cell Cardiol 8:189-204, 1976 Dixon SH Jr, Fuchs JC, Ebert PA: Changes in serum creatine phosphokinase activity following thoracic, cardiac, and abdominal operations. Arch Surg 103:66-68, 1971 Phornphutkul KS, Anuras S, Koff RS, et al.: Causes of increased plasma creatine kinase activity after surgery. Clin Chem 20(3):340-342, 1974 Ahmed SA, Williamson JR, Roberts R, et al.: The association of increased plasma MB CPK activity and irreversible ischemic myocardial injury in the dog. Circulation 54(2):187-193, 1976 Szasz G, Gruber W, Bernt E: Creatine kinase in serum: 1. Determination of optimum reaction conditions. Clin Chem 22(5):650-656, 1976 Bergmeyer HU, Bowers GN Jr, Horder M, et al.: Provisional recommendations on IFCC methods for the measurement of catalytic concentration of enzymes. Clin Chem 23(5):887-899, 1977 Delva E, Maillé JG, Solymoss BC, et al.: Evaluation of myocardial damage during coronary artery grafting with serial determinations of serum CPK MB isoenzyme. J Thorac Cardiovasc Surg 75:467-475, 1978 Roberts R, Sobel BE: Creatine kinase isoenzymes in the assessment of heart disease. Am Heart J 95:521-528, 1978 Klein MS, Shell WE, Sobel BE: Serum creatine phosphokinase (CPK) iso- enzymes after intramuscular injections, surgery, and myocardial infarction. Experimental and clinical studies. Cardiovasc Res 7:412-418, 1973 Klein MS, Coleman RE, Weldon CS, et al.: Concordance of electrocardio- graphic and scintigraphic criteria of myocardial injury after cardiac surgery. J Thorac Cardiovasc Surg 71(6):934-937, 1976 Gray RJ, Shell WE, Conklin C, et al.: Quantification of myocardial injury during coronary artery bypass graft. Circulation 58(3), part 2(suppl 1): 38-42, 1978 400 14. 15. 16. 17. Adappa MG, Jacobson LB, Hetzer R, et al.: Cold hyperkalemic cardiac arrest versus intermittent aortic cross-clamping and topical hypothermia for coronary bypass surgery. J Thorac Cardiovasc Surg 75(2):171-178, 1978 Weisel RD, Lipton IH, Lyall RN, et al.: Cardiac metabolism and performance following cold potassium cardioplegia. Circulation 58(3), part 2(suppl 1): 217-226, 1978 Waldenstrom AP, Hjalmarson AC, Thornell L: A possible role of noradrena- line in the development of myocardial infarction: an experimental study in the isolated rat heart. Am Heart J 95(1):43-51, 1978 Tonkin AM, Lester RM, Guthrow CE, et al.: Persistence of MB isoenzyme of creatine phosphokinase in the serum after minor iatrogenic cardiac trauma: absence of postmortem evidence of myocardial infarction. Circulation 51 (4):627, 1975 401 POSSIBILITY OF USING LIPOSOMES FOR TARGETING OF DRUGS IN THE TREATMENT OF CARDIOVASCULAR DISEASES V. P. Torchilin, V. R. Berdichevsky, Ban-An Khaw, V. M. Zemskov, E. Haber, V. N. Smirnov, and E. I. Chazov INTRODUCTION In recent years, artificial phospholipid vesicles, i.e., liposomes, fre- quently produced by mechanical dispersion of phospholipids in aqueous solution (1), have come to be looked upon as a promising means for transport of drugs in the organism (2,3). The drug, enclosed within a liposome, does not come in contact with blood components, causes no side effects, and does not undergo rapid biodegradation (4). Furthermore, liposomes, as drug carriers, are not only distinguished by a high degree of biocompatibility and total utilization of the phospholipids which they are made of in the organism, but they also rep- resent a unique possibility for delivering drugs into the cells since the lipo- somes can interact with the cells by fusion or endocytosis (2). Unfortunately, as many experiments have shown, after a very brief period of time following administration of liposomes into the circulation, they are rapidly cleared from the bloodstream by the cells in the reticuloendothelial system, primarily in the liver. This occurs regardless of method of administration, phospholipid composition, or charge or size of the liposomes (5-7). In order to improve the effectiveness of drug transport in liposomes, one can, in principle, take two approaches. First, the liposomes can be modified by some means to eliminate liver absorption. Attempts have been made, for example, to use liposomes made with phospholipid analogs which are not sub- jected to the effects of phospholipases (8,9). These attempts have, however, met with ambiguous results. It might also be promising to cover the surface of ordinary liposomes with some protein or protein aggregate which is not absorbed by the liver. A second approach would be to increase the effectiveness of transport by giving the liposomes which carry the drugs specific affinity for the target organ. This affinity might be achieved by using specific macro- molecular ligands bound to the outer surface of the liposomes. Antibodies against various chemical compounds which are specific components of the target organ are most frequently utilized as such specific ligands. From the All-Union Cardiology Research Center, USSR Academy of Medical Sci- ences, Moscow, USSR, and from Massachusetts General Hospital, Boston, Massa- chusetts, USA. 403 The idea of using antibodies for directed transport of drugs has been dis- cussed in the scientific literature for several years (10,11). Suggestions have been made for the construction of synthetic systems, including a biocompatible polymer carrier, to which molecules of the drug and a molecule of a specific ligand are covalently bound (11). It is not conclusive that the efficiency of specific macromolecular ligands will be maximal when they are used for binding with liposomes, each of which may include a significant quantity of the drug. Attempts have also been described for increasing the specificity of the inter- action of liposomes with various cells by covering the outer surface of the liposomal membranes with specific or nonspecific immunoglobins (12,13). How- ever, as was specifically noted in a recent review on the use of liposomes for protein transport, the problem of directed transport remains unsolved (14). This lack of success can be explained both by insufficient development of methods for binding specific ligands to the surface of the liposomes, leading directly to a significant loss in the activity of the ligand, and by an insuf- ficient number of experiments with individual antibody-antigen systems, rather than nonspecific immunoglobins. In the present work, we have attempted to find a new method of covalent coupling of proteins with the surface of liposomes and to determine the feasi- bility of utilizing antibodies specific for cardiac myosin and type 1 collagen for covalent coupling to liposomes. COVALENT BINDING OF PROTEINS WITH THE SURFACE OF LIPOSOMES THROUGH A "SPACER" For successful functioning of proteins bound to liposomes, it is generally considered that the surface of the liposomes must bind a sufficient quantity of protein, that the binding with the membrane must be firm, that the specific binding properties of the protein must remain unchanged, and that the integrity of the liposome which contains the drug or model compound must not be damaged in the process of binding. Traditional methods of binding proteins to liposomes by sonication in the process of liposome formation, or by adsorption onto the surface of liposomes al- ready formed usually do not satisfy these requirements fully. Obviously, a firm binding of the protein to liposomes by adsorption can only be achieved when there are hydrophobic interactions between the liposomal membrane and the nonpolar sec- tions of the protein molecules. Such hydrophobic interaction is achieved only in certain cases (15,16). The protein, incorporated into the membrane by cosoni- cation, may significantly decrease its specific binding capacity due to steric hindrances, evoked by the closeness of the membrane, or due to partial or com- plete denaturation under the influence of nonpolar components of the membrane. On the other hand, it is well known from studies of affinity chromatography and immobilization of enzymes that such undesirable phenomena can be avoided if the reactive group on the protein intended for binding is shifted from the sur- face of the matrix through coupling to an inert hydrocarbon 'spacer" (17,18). It is this principle of protein binding which we have used in our experiments (19,20). 404 For this purpose, phosphatidylethanolamine, which has a reactive group, was introduced into the lipid mixture (the details of the experiment are de- scribed in reference 20). The liposomes were activated by treatment of the surface with dialdehyde or diimidate; this allowed groups reactive with protein to be introduced and separated from the surface of the liposomes by the length of the chain of the bifunctional reagent used. It was demonstrated simultane- ously that, with this type of treatment, the radioactive marker previously en- closed within the liposomes did not leave them, i.e., the liposomes remained intact. The activated liposomes produced were bound to a model protein--the enzyme a-chymotrypsin--and both the quantity of bound enzyme and the preserva- tion of its specific activity and ability to be inhibited by a high-molecular protein inhibitor were determined (the enzyme - high molecular inhibitor inter- action can be looked upon as a model of the interaction between an antigen and an antibody). The results were compared with the same parameters of the enzyme bound to the surface of the liposomes by traditional methods. The results of the experiment are presented in table 1. The data indicated that significant quantities of protein can be bound to the liposomes by cosonication or covalent binding. Covalent binding through a "spacer" has two definite advantages: First, during cosonication, a portion of the total concentration of the enzyme enters the internal space of the liposomes (as demonstrated by the breakdown of liposomes with addition of Triton X-100 and subsequent determination of the degree of increase in enzymatic activity in the reaction solution), whereas, with covalent attachment, this does not occur at all. Second, during cosonication, some of the protein molecules (up to 50 TABLE 1. Properties of a-Chymotrypsin, Bound to Liposomes by Various Methods Ratio of Apparent Constants of Inhi- Percent Retention bition of Immo- Quantity of of Internal Marker bilized and Native Bound En- in Liposomes after o—~Chymotrypsin by zyme (mol/ Association of Pancreatic Trypsin Binding Method mol lipid) the Enzyme Inhibitor Adsorption 2.0-107° 70 1 Incorporation 4.7-107° 80 5 Covalent binding -5 through diimidate 4.9.10 100 1 Covalent binding 5 through dialdehyde 7.110 75 1 For experimental conditions see reference 20. 405 percent) lose their capability to interact with the protein inhibitor due to steric hindrances created by the closeness of the membrane. However, with cova- lent attachment through a ''spacer,'" over 80 percent of the molecules of the bound enzyme showed unaltered affinity for the protein inhibitor (figures 1 and 2). Thus, covalent binding of proteins with the surface of the liposomes through a "spacer," after previous introduction of reactive lipids into the liposomal membrane and treatment of the liposomes with bifunctional reagents, enables the binding of a sufficient quantity of macromolecular ligand with unchanged binding properties to the surface of the liposomal membrane. CONSERVATION OF SPECIFIC ACTIVITY OF BOUND ANTIBODIES The next stage in our work was to investigate the possibility of developing actual transport systems suitable for use in the treatment of cardiovascular disease. In a number of recent experiments, it has been shown that purified antibodies against myosin from the cardiac muscle or their (Fab'), fragments interact effectively with cardiac myosin both in vitro and in vivo, and can be successfully used for visualization of the necrotic zone in experimental myo- cardial infarction following administration of antibodies labeled with radioac- tive isotopes (21-23). One could assume that the same antibodies could be used for directed transport of drugs enclosed in liposomes directly to the damaged zone in myocardial infarction. FIGURE 1. Schematic picture of liposomes (1) with the immobilized enzyme or antibody and (2) covalent coupling via a "spacer" on their surface. In the case of incorporated protein (a) high affinity binding with macromolecular antigen or substrate (3) is impossible. Covalently bound through the "spacer' antibody or enzyme (b) preserves binding ability completely. 406 enzyme activity (relative units) 5 107 30 pancreatic trypsin inhibitor, uM FIGURE 2. Inhibition of native a-chymotrypsin and o-chymotrypsin immobilized on the surface of liposomes by pancreatic trypsin inhibitor. For experimental study of the binding of antibodies with liposomes, lipo- somes were produced by the method of Torchilin et al. (20). The same method was used to activate their surface with glutaraldehyde. Excess dialdehyde was re- moved by dialysis, and 1251_1abeled antimyosin antibodies, unlabeled antibodies, or unlabeled nonimmune rabbit IgG were added to the activated liposomes. The mixture was incubated overnight to allow time for the reaction to occur between the amino groups on the lysine residues of the protein and the aldehyde groups of the activated liposomes. The unbound protein was removed by gel chromatog- raphy on a Sepharose-4B column. The quantity of bound protein was then deter- mined either by the radioactivity associated with the liposome fraction, or by the optical density of the protein at 280 nm (in this case, after preliminary breakdown of the liposomes by the addition of Triton X-100). Data presented in figure 3 show that, given the concentrations of protein used in the reaction mixture (on the order of 10-4 M), from 40 to 60 percent of the antibodies or nonspecific immunoglobulin added could be bound to the liposomes. For every 1,000 lipid molecules, 0.1 to 0.2 molecules of protein could be bound, depending on whether the liposome fraction is monolamellar or multilamellar (on the aver- age, we were dealing with liposomes on the order of 800 & in diameter, the composition of which included up to 10,000 lipid molecules) (see reference 20). 407 0.5¢ 1°50 = E 04r 4140 > o => 8 o3f 130 S = 9 02} 120 5 oI} {10 N°N°® fractions FIGURE 3. Gel chromatography of antimyosin antibodies bound to liposomes on Sepharose-4B column. 1 = 125I-labeled antibodies (mr/hr). 2 = Dyggp for non- labeled antibodies. 3 = Dpgp for nonlabeled nonspecific IgG. Table 2 presents the results of the study comparing the specific activity of antibodies bound to the liposomes, and of nonimmune immunoglobulin bound to liposomes, for 125I-labeled canine cardiac myosin (24). Radiolabeling of the antigen was performed by the method of Marchalonis (25). The results show that, under standard conditions of the radioimmunoprecipitation reaction, antibody covalently bound to liposomes through a "spacer" demonstrated a high degree of precipitation of radioiodine-labeled antigen, approximating the degree of pre- cipitation with native antibodies, which was in excess of the radioactivity precipitated by the control, nonimmune immunoglobulin. These results cannot be related to the splitting of immobilized antibodies from the surface of lipo- somes, since rechromatography of the antibody-liposomes after 3 days of storage showed that over 90 percent of the protein remains bound to the liposomes. Furthermore, antibodies immobilized on the liposomes by this method have the capability to interact not only with dissolved antigen, but also with anti- gen in the form of an immunosorbent. Thus, chromatography of liposome prepara- tions containing 3H-labeled cholesterol and a nonlabeled antibody immobilized on the surface, on a column containing myosin coupled to Sepharose 4B, showed that a significant portion of the 3H radioactivity is retained in the column (see figure 4). This radioactivity can be eluated only with 6 M guanidine chloride, which dissociates the antigen-antibody complex. These results suggest that this system may be useful for directed transport of drugs to the necrotic zone in myocardial infarction, since the presence of liposomes does not influence recognition of the antigen by the bound antibody 408 TABLE 2. Properties of Antimyosin Antibodies and Nonspecific Immunoglobulin Bound to Liposomes by Immunoprecipitation With 125I-labeled Myosin % Precipitation, Related Mean No. of Protein to Same Quantity of Pro- Molecules per 1,000 tein (precipitation of Molecules of Phos- native antibodies taken Preparation pholipid as 100 percent) Native antibodies —— 100 Pure rabbit IgG - 8.5 Proteins bound to the surface of lipo- somes through glutar- aldehyde: Antibodies on multi- lamellar liposomes 0.08 35 Antibodies on mono- lamellar liposomes 0.2 72 Nonspecific immuno- globulin on mono- lamellar liposomes 0.2 8 For experimental conditions, see reference 24. or the process of specific binding. Preliminary experiments performed on dogs with infarctions caused by ligation of a coronary artery showed that, if a system consisting of liposomes with a radioactive label plus unlabeled antimyosin anti- body, covalently bound to the surface of the liposomes, is administered by the intracoronary route, radioactivity is concentrated in the necrotic zone. This concentration cannot be caused by nonspecific binding, since the replacement of the antibody with nonspecific immunoglobulin does not produce this concentration effect. In other words, in a real system, the antibodies are capable of trans- porting heavy liposomes to the target organ, which indicates the possibility of creating a complex system also suitable for diagnostic visualization of the region of damage and for transporting therapeutic agents to that region. The system studied is not the only one of interest in the treatment of cardiovascular diseases. For example, it is well known that the initial stage of many pathological processes in vessels, leading to the formation of thrombi or atheromas, is the disruption of the integrity of the endothelial layer, re- sulting in an uncovering of the collagen substrate to which thrombocytes begin to adhere (26). This process suggests a possible means of blocking the pathology in the initial stage. Thus, if agents can be delivered by some method to the point of disruption of the endothelium which could either prevent adherence of 409 1000 T 800T imp min dD 8 6M Gu HCI 400+ 1 4 8 12 1&6 20 36 40 44 48 = fraction number FIGURE 4. Gel chromatography of antibodies to myosin and nonspecific rabbit IgG bound to (3H cholesterol)-labeled liposomes on a Sepharose-4B column with myosin. the thrombocytes or accelerate healing of the damaged section with new endo- thelium or both, atherogenesis might be prevented. Such a system might consist of liposomes containing the appropriate drug, with anticollagen antibodies bound to their surfaces. In the initial stage of developing such a system, rabbits were immunized with type 1 collagen, after which the serum taken from the immunized animals was subjected to preliminary purification on a polyacrylamide sorbent with type 1 bovine collagen using the method of Carrel and Barandun (27). The immunoglobu- lin fraction separated, which had an increased degree of precipitation in the immunoprecipitation reaction with 1251-1abeled collagen, was immobilized by the method described above on the surface of the liposomes, which contained lac- cholesteryloleate. To check the ability of the conjugates produced to bind specifically with collagen, the method of Voller et al. (28) was used to produce polystyrene microcapsules, the inner surface of which was coated with a layer of collagen. The experiment was then conducted using three series of five microcapsules each. In the first series, 100 pl of a suspension of control liposomes treated with glutaraldehyde was added. In the second series, 100 pl each of activated liposomes with immobilized nonspecific immunoglobulin was added, and in the third series, 100 ul each of liposomes with immobilized immunoglobulins having increased affinity for collagen was added. After 3 hours of incubation, all 410 capsules were repeatedly washed until no radioactivity was found in the wash water, indicating that the main mass of nonspecifically adsorbed, labeled lipo- somes had been washed from the collagen substrate. The containers were then washed with 6 M guanidine chloride, which disrupted the antigen-antibody inter- action, and a 1 percent solution of Triton X-100, which carried the phospho- lipids and other components of the liposomal membranes into solution. The radioactivity of the wash fluids was then measured in a scintillation counter. The data presented in table 3 show that significantly increased quantities of radioactivity can be washed from the collagen of the capsules containing lipo- somes with specific immunoglobulins. In other words, as in the case of anti- myosin antibodies, when immunoglobulins are immobilized on the liposome surface through a "spacer," the activity of the immunoglobulins, with increased affinity for collagen, is retained. The above experiments are but a first step toward creating preparations intended for specific binding with the bare collagen of vessels. In summary, we can assume that liposomes which carry specific antibodies on their surface are a promising means for directed transport of drugs in the treatment of cardiovascular and many other diseases. Further studies are needed in this direction. TABLE 3. Binding of Various Preparations of Labeled Liposomes With Collagen Substrate Preparation Relative Effectiveness of Binding Liposomes without protein 1 Liposomes with nonspecific IgG immobilized on their surface 1.05 Liposomes with immunoglobulin having high affinity for collagen immobilized on their surface 2.80 411 10. 11. 12. 13. REFERENCES Bangham AD, Standish MM, Weissmann G: The action of steroids and strepto- lysin S on the permeability of phospholipid structures to cations. J Mol Biol 13:253-259, 1965 Gregoriadis G: Medical applications of liposome-entrapped enzymes. In Immobilized Enzymes, edited by K Mosbach. Methods in Enzymology, vol 44, edited by SP Colowick and NO Kaplan. New York, Academic Press, 1976, pp 698-709 Gregoriadis G: Liposomes in the therapy of lysosomal storage diseases. Nature 275:695-696, 1978 Fendler JH, Romero A: Liposomes as drug carriers. Life Sci 20:1109-1120, 1977 Gregoriadis G, Ryman BE: Fate of protein-containing liposomes injected into rats. An approach to the treatment of storage diseases. Eur J Biochem 24:485-491, 1972 McDougall IR, Dunnick JK, McNamee MG, Kriss JP: Distribution and fate of synthetic lipid vesicles in the mouse: A combined radionuclide and spin label study. Proc Natl Acad Sci USA 71:3487-3491, 1974 Juliano RL, Stamp D: The effect of particle size and charge on the clear- ance rates of liposomes and liposome encapsulated drugs. Biochem Biophys Res Commun 63:651-658, 1975 Deshmukh DS, Bear WD, Wisniewsky HM, Brockerhoff H: Long-living liposomes as potential drug carriers. Biochem Biophys Res Commun 82:328-334, 1978 Torchilin VP, Berdichevsky VR, Goldmacher VS, Smirnov VN, Chazov EI: [Dynamics of clearance from serum of liposomes from nonhydrolyzed diester analog of phosphatidylcholine upon i/v administration to mice] (Rus). Biull Eksp Biol Med No. 8:161-163, 1979 Gregoriadis G: Targeting of drugs. Nature 265:407-411, 1977 Goldberg EP: Polymeric affinity drugs for cardiovascular, cancer and urolithiasis therapy. In Polymeric Drugs: Proceedings of the Interna- tional Symposium on Polymeric Drugs, 173rd National Meeting of the American Chemical Society, March 20-25, 1977, New Orleans, Louisiana, edited by LG Donaruma and O Vogl. New York, Academic Press, 1978, pp 239-262 Gregoriadis G, Neerunjun ED: Homing of liposomes to target cells. Biochem Biophys Res Commun 65:537-544, 1975 Weissmann G, Bloomgarten D, Kaplan R, Cohen C, Hoffstein S, Collins T, Gotlieb A, Nagle D: A general method for the introduction of enzymes, by means of immunoglobulin-coated liposomes, into lysosomes of deficient cells. Proc Natl Acad Sci USA 72:88-92, 1975 412 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Finkelstein M, Weissmann G: The introduction of enzymes into cells by means of liposomes. J Lipid Res 19:289-303, 1978 Tyrrell DA, Heath TD, Colley CM, Ryman BE: New aspects of liposomes. Biochim Biophys Acta 457:259-302, 1976 Solomon B, Miller IR: Interaction of glucose oxidase with phospholipid vesicles. Biochim Biophys Acta 455:332-342, 1976 Berezin IV, Antonov VK, Martinek K (ed.): [Immobilized Enzymes], vol 2 (Rus). Moscow, Moscow State University, 1976, pp 76-79 Mosbach K (ed.): Immobilized Enzymes. Methods in Enzymology, vol 44, edited by SP Colowick and NO Kaplan. New York, Academic Press, 1976 Torchilin VP, Goldmacher VS, Smirnov VN: [Protein Binding to Liposomes] (Rus). Bioorg Khim 4:1560-1562, 1978 Torchilin VP, Goldmacher VS, Smirnov VN: Comparative studies on covalent and noncovalent immobilization of protein molecules on the surface of lipo- somes. Biochem Biophys Res Commun 85:983-990, 1978 Ban-An Khaw, Beller GA, Haber E: Experimental myocardial infarct imaging following intravenous administration of iodine-131 labeled antibody (Fab')j fragments specific for cardiac myosin. Circulation 57:743-750, 1978 Beller GA, Ban-An Khaw, Haber E, Smith TW: Localization of radiolabeled cardiac myosin-specific antibody in myocardial infarcts. Circulation 55: 74-78, 1977 Haber E, Ban—-An Khaw, Beller G, Gold HK: Cardiac myosin-specific antigen in the localization of myocardial infarction. In Proceedings of the Third Joint Symposium on Myocardial Metabolism, Williamsburg, Virginia, May 9-11, 1977. Washington, D.C., U.S. Department of Health, Education, and Welfare, Public Health Service, National Institutes of Health, DHEW Publication No. (NIH) 78-1457, pp 453-470 Torchilin VP, Ban—-An Khaw, Berdichevsky VR, Locke ER, Smirnov VN, Haber E, Chazov EI: [Retention of specific binding capacity by antibodies covalently bound to the surface of liposomes] (Rus). Dokl Akad Nauk SSSR 246 (No. 3): 746-749, 1979 Marchalonis JJ: An enzymic method for the trace iodination of immuno- globulins and other proteins. Biochem J 113:299-305, 1969 Ross R, Glomset JA: The pathogenesis of atherosclerosis (first of two parts). N Engl J Med 295:369-377, 1976 413 27. 28. Carrel S, Barandun S: Protein-containing polyacrylamide gels: their use as immuno-adsorbents of high capacity. Immunochemistry 8:39-48, 1971 Voller A, Bidwell DE, Bartlett A: The Enzyme Linked Immunosorbent Assay (ELISA): A review with a bibliography of microplate applications. Cuernsey, Europe, Flowline Publishers, 1977 414 MORPHOFUNCTIONAL DESCRIPTION OF THE EFFECT OF CYTOCHROME C ON VARIOUS ZONES OF THE MYOCARDIUM IN EXPERIMENTAL INFARCTION K. A. Zufarov, R. A. Katsenovich, Sh. B. Irgashev, M. F. Khudayberdyeva, and K. N. Azizov INTRODUCTION In accordance with current understanding, treatment of myocardial infarc- tion is complex and directed toward various aspects of the pathology. A tre- mendous number of preparations are presently used, and because of the complex approach followed in treatment of this disease, one must understand clearly the mechanism of action of each preparation. We know that when there are dis- orders in coronary circulation the pathology includes not only the ischemic area, but also the entire myocardium; thus, the specific features in the course of the disease in various sections of the myocardium become important (1-5). In order to maintain metabolism at a level sufficient to assure survi- val of myocardial cells, it is important to increase the delivery of oxygen and to stimulate the capability of the cell for maximum utilization of oxygen by activation of the oxidation-reduction enzyme systems. The natural compo- nents of oxidative phosphorylation, including cytochrome C, are therefore, of interest. Many studies have shown that the content of cytochrome C and the activity of the cytochrome system both decrease significantly with myocardial infarction (6-10). Depression of the cytochrome system occurs primarily because of the loss of cytochrome C, while normalization is probably facilitated by restora- tion of the system's activity (8), which can include formation of rich high- energy compounds. Therefore, study of the singular influence of cytochrome C in various sec- tions of the myocardium after damage is an interesting subject for further re- search. This is particularly true since there is a lack of comparative studies of metabolic processes in necrotic and undamaged areas of the myocardium under the influence of this preparation. From the Clinical-Experimental Biophysical Laboratory, Institute of Cardiology, Tashkent, USSR. 415 MATERIALS AND METHODS The experiments were performed on random-bred male white rats weighing 180-200 g. Myocardial infarction was evoked by ligature of the descending branch of the left coronary artery. Cytochrome C produced by Reanal Co. (Hungary) was administered intramuscularly twice daily at 5 mg/kg of body weight for 30 days. The animals were sacrificed by decapitation after 1, 3, 7, 15, and 30 days of treatment. Untreated animals with experimental myocar- dial infarction, which were administered saline solution, were used as controls. A group of animals with undamaged myocardium was also studied. A total of 140 animals was studied. The myocardium was examined in the zone of necrosis which was usually located on the anterior wall of the left ventricle, in the infarct zone, and on the posterior wall of the left ventricle. Materials were taken for general morphological (hematoxylin-eosin) and electron-microscopic studies. These were fixed in glutaraldehyde with subse- quent postfixation in osmic acid, and then eluated into epon. 3H-methionine was administered 4 hours before decapitation for radiologic purposes, and the number of counts per minute per mg of protein in the myocar- dium in the zone of necrosis and the posterior ventricle were counted on a Beckman B-mate II scintillation counter. A biochemical method was used to study the cytochrome oxidase activity in the zone of necrosis and in the pos- terior wall of the left ventricle. This activity was expressed in units of optical density per mg of moist tissue; the quantity of cytochrome C was ex- pressed in mg% moist tissue. The polarographic method was used to study the respiratory coefficient in the left ventricle, including the necrotic zone. Furthermore, electrocardiograms were taken in three standard and three ampli- fied leads from the extremities. In some cases, at 15-30 days, the area of the cross-section of cardiomyo- cytes was determined in the zone of the posterior wall of the left ventricle by planimetric methods, by drawing contours with a total magnification x 2,900. In each case, the average of 100 measurements was expressed in arbitrary units. RESULTS AND DISCUSSION It is known that the course of myocardial infarction, particularly in the ischemic zone, passes through several stages. The results of our observations and the data of other researchers (2) indicate significant changes in the early stages both in the ligature zone of the vessel and in the remaining areas of the myocardium. Within 1-3 days, clear dystrophic reactions with edema of the stroma, loss of transverse and longitudinal exhaustion of myofibrils, and sub- sequent granular and lump breakdown were observed. The ultrastructure indicated the presence of residues of the cellular structures--myofilaments, mitochondria, expansion of the sarcotubular apparatus, and the appearance of elements of mesenchymal origin. Later, during the period of organization of the myocardial infarction, resorption of necrotic masses, fibroblastic reaction with increased collagenization, and the formation of scar tissue occurred. 416 Biochemical study of cytochrome C and cytochrome oxidase showed a signifi- cant reduction in the zone of necrosis at all times (figures 1 and 2). The low- est level occurred on day 15 at 29.03 percent of the initial level. Cytochrome oxidase activity was also lower at all times, with the minimum activity on day 15 at 26.37 percent of the initial level. Some authors (8) believe that the depression of cytochrome oxidase occurs primarily due to the loss of cytochrome C, which is the main limiting factor. All of these changes lead to a decrease in the intensity of oxidative processes, since we know that in cardiac muscle under normal conditions 97 percent of the oxidative reactions involve partici- pation of the cytochrome system (9). The peripheral segments of the zone of necrosis, i.e., the areas adjacent to the infarct area, are of great significance in terms of its condition which also undergoes significant changes manifested as partial atrophy of the muscles, intracellular structural disorders, and as shown by our results as well as by the data of other researchers (3), structural changes typical of most muscle cells in the earliest periods of myocardial infarction in the necrotic zone. SONOS ANONNNONNNNNNANNBSH— ASONUONNONONN NNN oO RRR Q 7d oO a 30d 3 FIGURE 1. Content of cytochrome C in the zone of necrosis in rats with experi- mental myocardial infarction. The ordinate shows the quantity of cytochrome C in mg% of moist tissue. The abscissa shows the experimental time in days. @ = Quantity of cytochrome C in the myocardium of the left ventricle of intact rats. (J0= Quantity of cytochrome C in the zone of necrosis in rats not receiving cyto- chrome C. @ = Quantity of cytochrome C in the zone of necrosis in rats receiv- ing cytochrome C. 417 Sl NNNNNANNNNNNNaN 7d 15d 30d ow Q FIGURE 2. Cytochrome oxidase activity in the zone of necrosis in rats with experimental myocardial infarction. The ordinate shows activity of cytochrome oxidase in units of optical density. The abscissa shows the experimental time in days. BB = Activity of cytochrome oxidase in the myocardium of the left ven- tricle of intact rats. [J = Activity of cytochrome oxidase in the zone of necro- sis in rats not receiving cytochrome C. [J = Activity of cytochrome oxidase in the zone of necrosis in rats receiving cytochrome C. The zone farthest from the infarct also participates in this process. During various stages of the disease, morphological disorders occur, extending to dystrophic and necrobiotic changes of individual muscle fibers and groups of fibers, with subsequent development of focal cardiosclerosis. Biochemical study of respiratory enzymes also demonstrated a decrease in the content and activity of these enzymes throughout the entire period of the study (figures 3 and 4). By analyzing the dynamics of experimental myocardial infarction, we note once more the nonuniformity of development and organization of this disease in various segments of the myocardium, which is natural with such serious metabolic disorders. Pharmacological correction of the course of this disease is vital. As the data following myocardial infarction show, severe changes in the respi- ratory chain occur, leading to disruption of the formation of energy-rich com- pounds, particularly adenosine triphosphate (ATP). Decrease in ATP synthesis 418 1d 3d 7d 15d 30d FIGURE 3. Content of cytochrome C in the cardiac muscle of the left ventricle in rats with experimental myocardial infarction. See figure 1 for designation of the columns. : 1d 3 7d sd 30d FIGURE 4. Cytochrome oxidase activity in the myocardium of the left ventricle in rats with experimental myocardial infarction. See figure 2 for designation of the columns. 419 is reflected in the contractile capacity of the myocardium in plasticity, and therefore inhibition or loss of components of the respiratory chain is an im- portant mechanism in disruption of heart function following infarction. Administration of exogenous cytochrome C within 1-3 days indicates a more favorable morphofunctional picture. Due to the increase in the number of hy- pertrophic muscles located around the periphery of the necrotic area, this zone is somewhat smaller than in the control animals, and the transverse and longitudinal exhaustion of myofibrils is reduced. Increased fibroblastic re- action and new vessel formation are noted. In addition to the greater preser- vation of cardiomyocytes (figure 5), there are cells which manifest active intracellular regeneration in the form of myoblastic elements. These are characterized by a large number of ribonucleoprotein granules and elements of the rough endoplasmic reticulum, with the appearance of myofilaments act- ing as a prerequisite for the formation of myofibrils (figure 6). This status of the intracellular structures is prerequisite for hypertrophy of the cells. FIGURE 5. Myocytes in preserved sections in the zone of necrosis 1 day after treatment with cytochrome C. Some mitochondria (M) are altered. The sarco- lemma (S) is clearly seen. Magnification x 2,300. 420 FIGURE 6. Myoblast from the zone adjacent to the focus of necrosis 3 days after treatment with cytochrome C. The cytoplasm shows many ribosomes (R) and mito- chondria (M) of various sizes. Elements of rough endoplasmic reticulum (RER) are seen, as well as bundles of protofibrils (Pr). Magnification x 15,000. This positive morphological picture is also reflected in the status of respiratory enzymes. As shown in figure 1, the quantitative content of cyto- chrome C increases in the early stages of myocardial infarction in treated animals both in comparison to the intact animals and particularly in compari- son to the untreated rats. Cytochrome oxidase activity, as shown in figure 2, does not reach the level of the intact animals or the treated animals, but is higher than in the control untreated rats. Polarographic study of the respiratory coefficient, indicating the cou- pling of oxidation and phosphorylation, also showed some increase in the treated rats, reaching 2.5 * 0.12 (and 1.92 * 0.2 in the untreated rats) versus 4.57 + 0.3 in the intact animals. Activation of intracellular structures and rapid development of connective tissue elements during the early stages of myocardial infarction lead to a sig- nificant increase in the inclusion of labeled methionine in the zone of necrosis, to 157.2 percent in treated animals (and 157.6 percent in untreated animals), compared to the activity of inclusion in the myocardium of intact rats (figure 7). And, as several researchers have noted (11), the increase in protein syn- thesis in the infarct zone apparently begins initially in the nuclei, and then follows in the mitochondria, microsomes, connective tissue elements, and intact muscle cells. It does not extend to the myofibrillar apparatus. In the pos- terior wall of the left ventricle during the early stages of infarction, certain 421 T 30 000 T 20 000 10 OOOF Counts per min SUOMI NNN NN NNN NAH 2 Mmmm imtmmiimimamtaR Id 3d 7d 15d 3 FIGURE 7. Inclusion of 3H-methionine in the total protein of the zone of necro- sis in rats with experimental myocardial infarction. The ordinate shows inclu- sion of 3H-methionine in counts per minute per mg of protein. The abscissa shows the experimental time in days. [ = Inclusion of 3H-methionine in the intact left ventricle. (J = Inclusion of 3H-methionine in the zone of necrosis in rats not receiving cytochrome C. {4 = 3H-methionine in the zone of necrosis in rats re- ceiving cytochrome C. dystrophic changes also occur: Both the optical picture and the ultrastructure indicate significant preservation and activation of intracellular structures in the treated animals. Incorporation of labeled methionine is slightly increased in comparison with intact animals--105.8 percent (89.0 percent in untreated animals) (see figure 8). The few observations of the intensity of protein synthesis outside the focus of necrosis characterize the course of this process differently. Some researchers (9) believe that the intensity of protein synthesis in areas not directly in contact with the area of necrosis does not change, whereas in the boundary zone the inclusion of labeled amino acids into the proteins of the myocardium is significantly increased. Biochemical study of respiratory enzymes in treated animals indicated an in- crease in their activity in comparison with untreated animals (figures 3 and 4). In treated animals during the stage of organization of the myocardial in- farction, which lasts in rats from 7 to 30 days, accelerated scar development 422 1S 000T + er min FR ly I ANONONNNNNNNNNNN NG Q 10 000 5 000 Counts SONONNNSN— SOUONMOONONNNNNOO— 0 / Id 3 7d 15d 30d Q FIGURE 8. Inclusion of 3H-methionine in the total protein of the cardiac muscle of the left ventricle. See figure 7 for designation of the columns. was noted and regenerative processes were also significant in the cardiomyocytes which were preserved in the necrotic zone and particularly in the periinfarction zones, where a large number of hypertrophic mitochondria and myofibrillar struc- tures appear (figure 9). The number of glycogen and ribosome granules increased, as did the number of endothelia of vessels with a large number of pinocytous vacuoles (figure 10). Changes in sectors of the left ventricle far from the zone of necrosis in the treated rats were slight in nature and were mainfested as greater conserva- tion and activation of the intracellular structures in the early stages, whereas in the later stages (15-30 days), they indicated hypertrophy of the cardiomyo- cytes. This confirms our planimetric data. Thus, by day 15 in the treated ani- mals the mean indexes of dimensions of the cross-section of the cardiomyocytes of the posterior wall of the left ventricle reached 13.6 * 0.29 versus 8.9 = 0.3 in the controls. The histogram was significantly different from normal in these animals, with an obvious shift to the right. This occurs due to a decrease in the number of small cells and an increase in the number of large cells. By day 30, the planimetric picture was approximately analogous to the controls. Radiologic studies in the zone of necrosis of the treated animals first showed a decrease, and then by 30 days an increase in the incorporation of labeled amino acids, indicating an apparently more adaptive (in the line of subsequent compensation) protein synthesis in the treated animals (figure 7). 423 Periinfarction zone 15 days after treatment with cytochrome C, show- FIGURE 9. Magnification x 33,300. ing hypertrophy of mitochondria (M) and myofibrils (MF). FIGURE 10. Periinfarction zone 7 days after treatment with cytochrome C. The capillary (C) is in a state of elevated functional activity with pinocytous vacuoles (PV) in the cytoplasm. The myofibrils (MF) contain many cytogranules (Gr). Magnification x 26,300. 424 The content of cytochrome C and cytochrome oxidase activity also increased at all stages of the organization of the myocardial infarction (figures 1 and 2). An analogous picture was also observed in areas remote from the zone of necrosis (figures 3, 4, and 8). Positive dynamics on the electrocardiogram were accelerated in the treated animals. Approach of the S-T conjunction to the isoline was clearly expressed in the treated animals and we noted an acceleration in the formation of the nega- tive T wave, which is characteristic of a subacute electrocardiogram in the myo- cardial infarction stage. In treated animals, the negative T wave forms by 3 days, whereas in the control it only forms after 7 days. At later periods in the treated animals, this wave smoothed in the first lead and AVL lead. The Q wave, particularly in the first lead, was seen less clearly at 15-30 days in the treated animals than in the untreated animals, all indicating positive electro- cardiogram dynamics. Thus, exogenous cytochrome C, the terminal link in the chain of oxidative phosphorylation, has a rather favorable effect on the course of experimental myocardial infarction in rats. The zone of necrosis which is usually considered to be lost, and healed only by a scar, is characterized not only by accelerated development of connective tissue elements, but also by significant preservation of cardiomyocytes located primarily on the periphery of the necrosis, which hypertrophy, and have a favorable effect on the functional capability of the heart as a whole. Naturally, this facilitates a favorable morphofunctional picture of the remaining areas of the left ventricle. 425 10. 11. REFERENCES Vail SS: [Functional Morphology of Disturbances of Cardiac Activity] (Rus). Leningrad, Medgiz, 1960 Strukov AI, Lushnikov EF, Gornak KA: [Histochemistry of Myocardial Infarc- tion] (Rus). Moscow, Meditsina, 1967 Mitin KS: [Electron Microscope Analysis of Changes in the Heart Upon Infarction] (Rus). Moscow, Meditsina, 1974 Chuchulin IuS: [The Damaged Heart] (Rus). Moscow, Meditsina, 1975 Hecht A: [Introduction to the Experimental Bases of Modern Pathology of Heart Muscle] (Russian translation). Moscow, Meditsina, 1975 Manoylov SE, Firsova VI: [The content of cytochrome C in the cardiac mus- cle in acute hypoxia] (Rus). [Biophysical Principles of Pathologic Status of the Heart and Energetic Support of the Contractile Apparatus]. Tbilisi, Metznieraba, 1973, pp 146-148 Frolkis RA: [Some components of the respiratory chain of the cardiac muscle in experimental myocardial infarction in dogs] (Rus). Vopr Med Khim 14 (No. 3):302-306, 1968 Fetisova TV, Frolkis RA: [Biochemistry of Myocardial Infarction] (Rus). Kiev, Zdorovia, 1976 Challoner DR: Respiration in the myocardium. Nature 217:78-79, 1968 Gudbjarnason S, Fenton JC, Wolf PL, et al.: Stimulation of reparative processes following experimental myocardial infarction. Arch Intern Med 118:33-40, 1966 Bing RJ: Reparative processes in heart muscle following myocardial in- farction. Cardiology 56:314-324, 1972 426 COMPARISON OF THE EFFECTS OF B-BLOCKERS AND B-STIMULATORS ON MYOCARDIAL FUNCTION B. I. Tkachenko, R. A. Katsenovich, S. Z. Kostko, Kh. A. Khashimov, A. Sh. Kasymkhodzhaev, and Z. Z. TIunusov SUMMARY The effect of a B-adrenoreceptor blocker (propranolol) and a stimulator (nonaclazine) on the arterial and venous vessels of the heart and its func- tional activity was studied. The experiments were performed on 25 isolated animal hearts. The resistance of the coronary arteries was studied by changes in perfusion pressure during constant flow. Change in myocardial function was studied in 127 experimental hearts in situ using a strain gauge and cal- culating dP/dt. Propranolol and nonaclazine upon peroral administration caused initial dilation and then constriction of the coronary arteries. It was established that the dilation occurs as propranolol enters the vessel lumen. In contrast to propranolol, nonaclazine constricts the veins of the heart. Propranolol had negative inotropic and chronotropic effects, decreas- ing the consumption of oxygen by the myocardium. Nonaclazine first reduced and then increased the contraction of the myocardium. The negative effect of nonaclazine on myocardium was more pronounced in experiments on desympa- thized animals and when the preparation was administered following previous blockage of the B-adrenoreceptors. Possible mechanisms of these changes are discussed. INTRODUCTION When pharmacological agents appeared which had blocking or stimulating effects on the B-adrenoreceptors, a fundamentally new direction developed in the treatment of a number of cardiovascular diseases. Use of these prepara- tions in ischemic heart disease was found to be particularly effective (1-8). However, the nature and mechanism of the effect of these substances on the circulation in connection with changes in the functional activity and metabolic status of the heart remain unclear. Solution of this problem may allow us to uncover the mechanism for the favorable therapeutic effect of these prepara- tions, given the variation of effect in terms of the state of circulation and From the Institute of Experimental Medicine, Academy of Medical Sciences, Leningrad, and from the Institute of Cardiology, Tashkent, USSR. 427 the intensity of metabolism in the myocardium. There is a pressing need for such studies because these preparations can have opposite effects for the same disease. The current literature contains no information on the nature of changes in the venous portion of the coronary vascular bed upon blockage or stimulation of B-adrenoreceptors. The present study was focused on these questions. MATERIALS AND METHODS Study of the effect of preparations on the interaction of circulation with changes in the contractile capacity of the heart and absorption of oxygen was performed on 25 isolated animal hearts. The coronary vessels of the iso- lated hearts were perfused with the blood of a cat donor through an aortic can- nula by means of a constant delivery pump. Changes in the resistance of the coronary vessels were judged from changes in the perfusion pressure. Changes in drainage from the coronary vessels perfused with a constant volume of blood reflected changes in the capacity of the vascular bed of the heart. Since the capacity of a limited vascular bed is determined primarily by its venous seg- ment, an increase or decrease reflects corresponding changes in the lumina of the veins of the heart. The method of heart extraction, perfusion, and record- ing of the outflow from the coronary vessels has been described previously in detail (9). The blood flowing from the coronary vessels was carried off through wide catheters inserted into the right and left ventricles. Two glass cuvettes car- rying transducers from a type 036 M oxyhemograph recorded the oxygen content of the arterial donor blood and the venous coronary blood flowing from the right half of the heart. A type MN-7 analog computer was used to calculate the arteriovenous difference in the content of hemoglobin oxygen; changes in this value during perfusion of the coronary vessels with a constant volume of blood reflected the change in oxygen absorption by the myocardium. The blood circulating in the system was heated by an ultrathermostat to 37-38° C. The effect of propranolol (67 experiments) and nonaclazine (60 experiments) on myocardial function was studied in experiments on the heart in situ. Myocardial stress was recorded by means of a strain gauge attached to the ventricle; pulse pressure in the aorta and the pressure in the left ven- tricle (LVP) were also recorded. To evaluate the fluctuations in end-diastolic pressure in the left ventricle, the LVP signal was amplified and cut off elec- tronically at 20 mm Hg on an MN-7 analog computer. The positive dP/dt was determined by differentiation of LVP using the MN-7 computer. The Veraguta index was calculated. Pressure shifts in the vessels were recorded with a type ID-2 electromanometer with a mechanotron sensor. Propranolol, a B-adrenoreceptor blocker (Obsidan, GDR), was administered all at once in doses of 0.001, 0.01, 0.1, and 1 mg; the first two doses and epi- nephrine (1 pg) were administered in 0.1 ml of saline. Intracoronary infusion of propranolol at 0.5 mg/min and epinephrine at 5 ug/min was performed by an automatic injector (this experiment was performed on 18 hearts). Nonaclazine (USSR), which has a stimulating effect on the B-adrenoreceptors, was administered 428 at 1.2 mg/kg (this experiment was performed on eight hearts). The chemical structure of nonaclazine is 10'[(1,4-diazabicyclo-(4,3,0)-nonanyl-4)propionyl]- 2-chlorophenothiazine hydrochloride (10). Preparations were injected into the main vessel by a perfusion pump. Control intracoronary administration of 0.1 ml saline or its infusion at 0.5 ml/min had no significant effect on the ves- sels or on heart activity. The a-adrenoreceptors were blocked by administra- tion of dehydroergotoxin (DH-ergotoxin, SPOFA, Poland) or phentolamine (Regitin, CIBA) in a dose of 1 mg. The experiments were performed on mature cats of both sexes weighing 3.5-4.5 kg, anesthetized with a solution of urethane (500 mg/kg) and chloralose (50 mg/kg). The data obtained were statistically analyzed for significant differences using Student's t-test. RESULTS Immediate intracoronary administration of propranolol caused an initial brief drop in coronary artery tone, with a subsequent long-term elevation (figure 1). As the dose increased, the phases of dilation and constriction became more clearly expressed. Propranolol caused an increase in the capacity of the vascular bed of the heart, i.e., expansion of the veins. It should be noted that expansion of the coronary arteries occurred before the change in force and frequency of cardiac contractions, i.e., consumption of oxygen by the myocardium. In contrast, the increase in resistance of the coronary arteries was accompanied by a decrease in functional activity of the heart and in the intensity of metabolism in the myocardium. The change in these last two factors depended on dose. *30 +3030 '30 '30 +30 ‘29 +2928 2828282828 2828 +28:28 28-28 28" 28: ——— 10 SEC FIGURE 1. Effect of intracoronary administration on propanolol. +4 = Moment of administration of 0.1 mg propanolol. Recordings top to bottom: perfusion pres- sure in the coronary arteries (mm Hg), venous return (ml), myocardial stress (% of initial level), and arteriovenous difference of oxygen content in coronary blood (% HbO2). The numbers indicate the number of contractions per 10 seconds. 429 Infusion of propranolol caused qualitatively the same effect as immediate administration (figure 2). However, the resistance of the coronary arteries, which decreased at the beginning of infusion, remained below the initial level throughout the entire period of infusion. In all experiments, when infusion of propranolol was stopped, the resistance increased greatly, exceeding the initial level before infusion. With blockage of o- and B-adrenoreceptors, infusion of propranolol caused greater dilation than infusion before adrenoreceptor blockage (36.0 * 5.6 per- cent and 18.4 *+ 2.6 percent, respectively), whereas constriction was less pro- nounced (26.5 * 4.3 percent and 41.8 * 6.7 percent, respectively). Consequently, the a-adrenoreceptors play a significant role in increasing the tone of the coronary arteries with B-adrenoreceptor blockage. Nonaclazine also caused a two-phased change in the resistance of the coro- nary artery upon intracoronary administration: An initial drop in resistance was followed by an increase. The resistance of the coronary arteries was de- creased by an average of 44.8 * 2.18 percent, and then increased by 33.9 % 0.3 percent. The capacity of the coronary bed decreased by an average of 0.59 # 0.15 ml. Nonaclazine decreased the frequency (by 13.8 * 6.9 percent) and force (by 36.4 * 9.3 percent) of coronary contractions. The consumption of oxygen by the myocardium decreased (by 11.4 * 0.9 percent). After 3-5 minutes, there was a moderate positive inotropic effect (by an average of 22.8 * 15.8 percent). Thus, both propranolol and nonaclazine upon intracoronary administration cause a two-phased change in coronary artery tonus: Initial coronary dilation was replaced by coronary constriction. In contrast to propranolol, nonaclazine decreased the capacity of the coronary bed, i.e., it constricted the venous vessels of the heart. A difference was also noted in the effect on the func- tional capacity of the myocardium. Nonaclazine first decreased and then in- creased the contractile capacity of the myocardium, whereas propranolol only decreased the contractile capacity of the cardiac muscle. [- J I 120 ! | tT _— 30 _ +05 E of rn —-05 —— 10 sec — 10 sec 0 * ol - 27 27 27 27°27 27 27 27'27°27:26°25'25'24'23'23' 17 17 17 17:17'16'16'17'18:'20'20' 21 23 So +10 5 f—— R® _g SN ———— . ee —— FIGURE 2. Intracoronary infusion of propanolol. The symbols are the same as in figure 1. +4 = Beginning of infusion. + = End of infusion. 430 The decrease in coronary resistance caused by propranolol cannot be ex- plained by a decrease in extravascular compression of the vessels of the heart, since the change in resistance occurred before a change occurred in the force and frequency of cardiac contractions. Nor was dilation related to the specific blocking effect of propranolol on the B-adrenoreceptors, since disruption of the B-adrenoreceptors by propranolol reinforces the effect of catecholamines circu- lating in the donor blood and in the perfusion system on the a-adrenoreceptors, which agrees with the available data (11-14), and may lead to constriction but not dilation of the coronary arteries. Furthermore, dilation of the coronary arteries under the effect of propranolol was also observed under the conditions of blockage of o- and B-adrenoreceptors. This allows us to conclude that dila- tion of the coronary arteries upon administration of propranolol results from the direct effect of propranolol on the smooth musculature of the vessels. This con- clusion confirms the data of Whitsitt et al. (15) who observed coronary dilation upon administration not only of dl and l-forms of propranolol, which have a B- blocking effect, but also of d-propranolol, which has practically no such effect. In our experiments, attention was drawn to the fact that, although the resistance of the coronary vessels could increase gradually during infusion of propranolol, abrupt increases were observed only after administration of the sub- stance was stopped. This increase in resistance could not be caused by a change in extravascular compression of coronary vessels or metabolic processes in the heart, since it occurred with unaltered force and frequency of cardiac contrac- tions and oxygen absorption by the myocardium. The decrease in contractile activity of the heart during administration of propranolol into the coronary vessels could lead to a decrease in its metabolism. We can assume that, dur- ing infusion of propranolol, the constriction effect of the decrease in meta- bolic dilation influences was to some extent compensated by the direct dilator effect of propranolol on the coronary vessels. When infusion was stopped, this effect was demasked, causing a sudden increase in the tonus of the coro- nary vessels as a result of the level of metabolism in the myocardium. In contrast to propranolol, coronary dilation upon administration of non- aclazine occurred with a decrease in the functional activity of the myocardium and consumption of oxygen, indicating a decrease in metabolic coronary dilation effects. This corresponds to the appearance of coronary constriction, not coro- nary dilation. Consequently, dilation of the external arteries caused by non- aclazine results from the direct effect of the preparation on the vascular wall. At the same time, we cannot exclude the possible participation in coronary dila- tion of the decrease in extravascular compression of the vessels, since dilation was accompanied by a decrease in myocardial contractile capacity. Based on these experiments performed on isolated animal hearts, we can pre- sent data on the differences found in inotropic and chronotropic reactions of the heart to blockage and stimulation of B-receptors. Propranolol at a dose of 0.001 mg actually had no influence on the positive inotropic effect of epineph- rine (69.9 * 20.1 percent before and 67.9 * 13.2 percent after propranolol); the chronotropic effect was decreased by 11.5 + 3.7 percent (p < 0.01). At a dose of 0.01 mg, propranolol limited the positive effect of epinephrine, though this was not significant (p > 0.05). In contrast, the chronotropic effect was decreased by 21.3 + 3.8 percent (p < 0.01). The capability of epinephrine to 431 increase consumption of oxygen by the myocardium was significantly decreased in both cases (by 17.9 * 5.3 percent and 21.3 * 3.8 percent, respectively). This agrees with the data of Blinks (16) who performed experiments on muscle strips from the atrium and capillary muscle, determining that propranolol limited the positive inotropic effect of catecholamines to a somewhat greater extent than the chronotropic effect. Administration of propranolol (0.1 mg) with an in- crease in epinephrine content in the blood was accompanied by a more clearly expressed inotropic effect, but the difference was not significant (p > 0.1). In contrast, the negative chronotropic effect of propranolol in these experi- ments was more clearly expressed (by 14.3 + 3.9 percent, p < 0.01), which is also associated with a greater decrease in oxygen consumption by the myocardium. Epinephrine, as a rule, increases the force and frequency of heart contractions under ordinary conditions, but when administered with infusion of propranolol, caused a decrease in the frequency of heart contractions (by 12.2 * 2.9 percent, p < 0.001) and a significant increase in their force (p < 0.001). It is char- acteristic that before propranolol, epinephrine increased oxygen consumption by the myocardium by an average of 38.8 + 1.8 percent, and with propranolol, by an average of 0.77 * 0.45 percent. Thus, according to our data, there is a definite difference in the inotropic and chronotropic reaction of the heart to adrenergic effects. It is quite impor- tant that, with a certain degree of blockage of the B-adrenoreceptors, the posi- tive chronotropic effect of adrenergic stimulation can be significantly limited, whereas the degree of expression of the positive inotropic effect is not signif- icantly changed. The change in oxygen absorption by the cardiac muscle is most closely related to a change in the frequency of heart contractions. This is of great significance in interpreting the antianginal mechanism of B-adrenoblockers. Experiments in isolated hearts can clarify the interrelationship of blood flow in the myocardium with its function. However, the functional state of car- diac muscle is closely related to central hemodynamics. Therefore, experiments were performed on hearts in situ. Intravenous administration of propranolol at 0.03, 0.05, 0.07, and 0.7 mg/kg caused a decrease in the force and frequency of cardiac contractions and systemic arterial pressure, and an increase in central venous pressure; a tendency for these effects to vary with dose was noted. Nonaclazine, when administered at 5-6 mg/kg, caused a brief decrease in the contractile capacity of the myocardium, followed by an extended increase in the functional activity of the myocardium and hypertension. A significant change was noted 15 minutes after administration, corresponding to the moment of accu- mulation of norepinephrine in the myocardium under the effect of nonaclazine (10,17). Based on the direction of the effect of these preparations, we can expect that the nature and manifestation of their influence on the cardiovascular sys- tem will depend on the adrenosympathetic background. Administration of propranolol (0.07 mg/kg) with an increase in epinephrine content in the blood (infusion of 5 pg/min) was accompanied by a less clearly expressed drop in the contractile capacity indexes than in the control (45.8 + 3.8 percent and 57.5 * 5.3 percent, respectively). When nonaclazine was 432 administered (6 mg/kg) with infusion of norepinephrine, a clear increase in contractile capacity of the myocardium was observed, with no decrease at all. The data indicate that the effect of blockers and stimulators of the B- adrenoreceptors on the function of the myocardium and the circulatory system as a whole depends on the initial adrenosympathetic activity of the organism and the possible functional state of the adrenoreceptors. This assumption is confirmed by experiments on desympathized animals. Preliminary administration of ornid is accompanied by a decrease in the content of norepinephrine in the synaptic cleft, i.e., a decrease in the adrenosympa- thetic activity of the organism. Nonaclazine under these conditions has a clearer and longer negative inotropic effect. The negative effect of nonacla- zine on myocardial function was significantly increased, both in strength and in length, upon administration of the preparation following blocking of the B- adrenoreceptors, making the desirability of their combined application doubtful. Comparative evaluation of the strength of the negative inotropic effect of nonaclazine under various conditions of administration is shown in table 1, from which we see that the negative chronotropic and inotropic effect of nona- clazine was increased in desympathized animals and still further increased fol- lowing administration of B-adrenoreceptor blockers. TABLE 1. Comparative Evaluation of the Degree of Negative Inotropic Effect of Nonaclazine Under Various Conditions of Administration Ornid- Propranolol- Parameter Control nonaclazine nonaclazine Pulse rate -2.12 + 1.03 -6.29 + 0.49 -14.31 + 1.7 Veragut contraction index 4.72 + 1.84 -20.18 + 1.18 -41.8 ££ 1.91 Myocardial stress -5.67 += 0.12 -57.35 + 15.21 -34.28 + 7.32 End-diastolic pressure in left ventricle 16.15 + 3.47 25.2 + 5.48 47.42 + 1.95 Aortic diastolic pressure -16.46 + 4.07 -30.63 + 2.91 -16.23 + 1.9 Length of negative inotropic effect (based More than on Veragut index) 3-4 minutes 12-15 minutes 30 minutes Note: Maximum deviations of the values of the parameters recorded are given in percent of the initial values. 433 CONCLUSIONS 1. Propranolol and nonaclazine upon intravenous administration cause an initial decrease in the tone of the coronary arteries with a subsequent in- crease. Coronary dilation is a manifestation of the direct nonspecific influ- ence of these preparations on the arterial walls. In contrast to propranolol, nonaclazine constricts the venous coronary vessels. 2. Coronary dilation caused by propranolol is retained as the substance enters the coronary bed, which is a new fact in the spectrum of the pharmaco- logical effect of this preparation. Coronary dilation is caused both by acti- vation of the a-adrenoreceptors with disruption of the B-adrenoreceptors, and by a decrease in metabolic coronary dilator effects. 3. Propranolol has a negative inotropic and chronotropic effect. Non- aclazine causes an initial brief decrease in the contraction of the myocardium which is replaced by a long-term increase. The effect of preparations on myo- cardial function and the circulatory system depends on the initial activity of the adrenosympathetic system of the organism. 4. The change in oxygen absorption by the myocardium is most closely re- lated to the change in frequency of cardiac contractions, which is probably the leading mechanism of the antianginal effect of B-adrenoblockers. 434 10. 11. 12. REFERENCES Myasnikov LA, Krasnikov IuA, Nikolaeva E, Grigorinats RA: [Clinical effect and hemodynamic changes associated with the use of inderal] (Rus). Kardiologiia No. 7:67-72, 1977 Myasnikov LA, Metelitsa VI: [Possibilities of differentiated treatment of the chronic form of ischemic heart disease] (Rus). Ter Arkh 45(No. 6): 8-21, 1973 Billimoria AR, Haveliwala HK, Vohra JK, Shah SJ: Propranolol in angina pectoris. Indian Heart J 24(1):20-24, 1972 Sivkov II, Kukes VG, Ziselman SB, Chebyshev NS, Shedov VV, Semenov VN: [Clinical use of some agents that block the adrenergic B-receptors] (Rus). Klin Med 52(No. 6):51-56, 1974 Komarov FI, Olbinskaya LI, Abinder AA, Iankin VV, Kitaeva IT, Kun IS, Ovchinnikova LV: [Stimulators of B-adrenergic structures in the treat- ment of ischemic heart disease patients] (Rus). Kardiologiia No. 7:37-43, 1975 Khamidova MKh, Akhmedova FA, Gitin IIa, and Kholmatov BKh: [The Soviet beta-blocking agent anaprilin in the treatment of ischemic heart disease patients] (Rus). Med Zh Uzb No. 11:8-10, 1975 Shkhvatsabaya IK: [Treatment of chronic ischemic heart disease in light of data on the pathophysiology of coronary circulation] (Rus). Kardio- logiia No. 7:5-15, 1975 Zamotaev IP, Lozinsky LG, Sandomirsky BL, Maximova LN: [Comparative evaluation of the clinical action of a number of B-adrenergic blocking agents] (Rus). Kardiologiia 16(No. 7):75-81, 1976 Tkachenko BI, Dvoretsky DI, Ovsiannikov VK, Samoilenko AV, Krasilnikov VG: [Regional and Systemic Vasomotor Reactions] (Rus). Leningrad, Meditsina, 1971 Kaverina NV, Markova GA, Chichkanov GG, Chumburidze VB, Basaeva AI: [Non- aclazine--a new drug for the treatment of ischemic heart disease] (Rus). Kardiologiia 15(No.7):43-48, 1975 Gaal PG, Kattus AA, Kolin A, Ross G: Effects of adrenaline and noradrena- line on coronary blood flow before and after beta-adrenergic blockade. Br J Pharmacol 26:713-722, 1966 Bohr DS: Adrenergic receptors in coronary arteries. Ann NY Acad Sci 139:799-807, 1967 435 13. 14. 15. 16. 17. Ovsiannikov VI: [Effect of catecholamines on the resistance and capacity of the coronary vessels before and after blockade of adrenergic receptors] (Rus). Fiz Zh SSSR 58 (No. 9):1415-1424, 1972 Orlova NN: [Coronary circulation under a-adrenergic block] (Rus). Kar- diologiia 18(No. 3):93-97, 1978 Whitsitt LS, Lucchesi BR: Effects of propranolol and its stereoisomers upon coronary vascular resistance. Circ Res 21:305-317, 1967 Blinks JR: Evaluation of the cardiac effects of several beta-adrenergic blocking agents. Ann NY Acad Sci 139:673-685, 1967 Kaverina NV, Griglewski R, Basaeva AI, Markova GA, Chumburidze VG: [Mechanism of action of nonaclazine on the blood supply and activity of the heart] (Rus). Biull Eksp Biol Med 80(No. 11) :48-50, 1975 436 # U. S. GOVERNMENT PRINTING OFFICE : 1980 629-160/2823 Ey U.C. BERKELEY LIBRARIES C095461113