key: cord-020714-h1fevqcw
authors: Compans, Richard W.; Kemp, Maurice C.
title: Membrane Glycoproteins of Enveloped Viruses
date: 2008-05-30
journal: Curr Top Membr Transp
DOI: 10.1016/s0070-2161(08)60750-9
sha: 
doc_id: 20714
cord_uid: h1fevqcw

This chapter focuses on the recent information of the glycoprotein components of enveloped viruses and points out specific findings on viral envelopes. Although enveloped viruses of different major groups vary in size and shape, as well as in the molecular weight of their structural polypeptides, there are general similarities in the types of polypeptide components present in virions. The types of structural components found in viral membranes are summarized briefly in the chapter. All the enveloped viruses studied to date possess one or more glycoprotein species and lipid as a major structural component. The presence of carbohydrate covalently linked to proteins is demonstrated by the incorporation of a radioactive precursor, such as glucosamine or fucose, into viral polypeptides, which is resolved by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis. Enveloped viruses share many common features in the organization of their structural components, as indicated by several approaches, including electron microscopy, surface-labeling, and proteolytic digestion experiments, and the isolation of subviral components. The chapter summarizes the detailed structure of the glycoproteins of four virus groups: (1) influenza virus glycoproteins, (2) rhabdovirus G protein, (3) togavirus glycoprotein, and (4) paramyxovirus glycoproteins The information obtained includes the size and shape of viral glycoproteins, the number of polypeptide chains in the complete glycoprotein structure, and compositional data on the polypeptide and oligosaccharide portions of the molecules.

search. The most important advantage is structural simplicity, in that lipid-containing viruses possess a small number of virus-coded polypeptides and glycoproteins as membrane components. Sufficient quantities of many enveloped viruses can be purified for detailed biochemical analysis, including primary structure studies of viral membrane proteins like those now underway. Purification procedures are rapid and simple, and a fairly homogeneous population of particles can be obtained free of contaminating cellular membranes. For the simpler lipid-containing viruses, the possibility therefore exists for complete structural determination of a biological membrane, and such information should contribute greatly to an understanding of the molecular details of membrane structure and interactions among membrane components in general.

Other advantages of enveloped viruses in studies of membrane structure and biogenesis include the ease of biosynthetic labeling of viruses grown in cell culture with specific radioactive precursors and the availability of mutants in defined gene products, some of which are proving to be useful in the analysis of viral membrane assembly.

Since envelope components are integral parts of cellular membranes during the assembly of virus particles (Compans et al., 1966 ; Compans and Choppin, 1971) , studies of the synthesis and mechanism of insertion of these components into membranes should provide useful information pertaining to the biogenesis of cellular membranes. Many viral systems offer the additional advantage that host cell biosynthesis is inhibited during infection, so that it is possible to analyze the synthesis of membrane components in a cell in which only a small number of virus-coded membrane components are being produced and to follow the intracellular migration of specific membrane glycoproteins by radiolabeling procedures.

Finally, the ability to prepare viruses with specific modifications in either lipid or protein composition arises from the fact that viral envelope lipids are derived from the plasma membrane of the host cell, whereas viral proteins are entirely virus-coded. Thus by growing different virus types in the same cell it is possible to prepare a membrane in which the protein composition can be varied while the lipid composition remains constant. Similarly, by growing the same virus in a series of different cells, viral membranes are obtained which possess the same set of proteins but vary in lipid composition. These approaches have been useful in studies of lipid-protein interactions in viral membranes (Landsberger et al., 1973) .

Several recent reviews are available concerning both the experimental and theoretical aspects of viral envelope structure and assem-bly (Lenard and Compans, 1974 ; Blough and Tiffany, 1975 ; Wagner, 1975 ; Choppin and Compans, 1975 ; Compans and Choppin, 1975 ; Klenk, 1974) . In this chapter, we emphasize recent information on the glycoprotein components of enveloped viruses and endeavor to point out specific findings on viral envelopes which we believe to be of broad significance.

In Table I are listed the 10 established major groups, or families, of lipid-containing viruses of vertebrates, and examples of the best studied members of certain groups. For several families, only limited information is available on viral envelope structure, whereas other groups have been the object of intensive study. We concentrate our discussion on the glycoproteins of the most intensively studied virus groups, which are the first four groups of RNA viruses listed in Table   I : myxoviruses, paramyxoviruses, rhabdoviruses, and togaviruses. We also discuss selected findings for other virus groups where these have revealed unusual or important aspects of viral membrane structure or assembly.

Although enveloped viruses of different major groups vary in size and shape, as well as in the MWs of their structural polypeptides, there are general similarities in the types of polypeptide components present in virions. The types of structural components found in viral membranes are summarized briefly in the following discussion, and further detailed information is available in the recent reviews cited above.

All the enveloped viruses studied to date possess one or more glycoprotein species. The presence of carbohydrate covalently linked to proteins has usually been demonstrated by the incorporation of a radioactive precursor such as glucosamine or fucose into viral polypeptides, which is resolved by sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis. The specificity of labeling by these precursors is shown by the lack of their incorporation into carbohydrate-free polypeptides of the virion (Straws et al., 1970) . The number of distinct glycoprotein species present in the virion varies for various virus groups. As shown in Table 11 , there is also marked variation in the glycoprotein content of virions when expressed as a percent of (Gard et al., 1977) , whereas the data given are for Pichinde virions which contain two glycoproteins. The values represent our calculations from the autoradiograph scan data presented in the reference.

The values given represent our calculations from the data presented in the reference. total viral protein. As discussed in Section V, these data suggest that glycoproteins of the various virus groups may play a greater or lesser role in assembly and maintenance of the viral envelope structure. However, glycoproteins are essential components for viral infectivity even where they constitute only a minor fraction of the mass of the virion, since they are necessary for the initial events in the viral replication cycle. The fine structure, biological functions, and orgapization of glycoproteins in the viral envelope are discussed in detail in Section IV.

The carbohydrates of viral glycoproteins are specified in large part by host cell transferases, whereas the amino acid sequences are coded b y the viral genome. Glycoproteins of the same virus may exhibit host cell-dependent differences in electrophoretic mobility, which are due to differences in the carbohydrate components (Haslam et al., 1970; Compans et al., 1970; Schulze, 1970) . Thus viral glycoproteins may be useful probes to detect differences in glycosylation among various cell types. An important exception to the general finding of host cell-specified carbohydrates is found in myxoviruses and paramyxoviruses. Glycoproteins of these viruses lack sialic acid (Klenk et al., 1970a,b) , which is believed to be a result of the virus-specified neuraminidase incorporated as a structural component of these virions.

All enveloped viruses contain lipid as a major structural component, and available evidence indicates that all the lipid is contained in a single bilayer structure which forms the matrix of the limiting membrane of the virion. Investigations using biophysical methods including xray diffraction (Harrison et al., 1971) and electron spin resonance (Landsberger et al., 1971 (Landsberger et al., , 1973 Landsberger and Compans, 1976) indicate that viral lipids are arranged in a bilayer. With electron microscopy, a well-defined unit membrane is observed at the location of the bilayer.

The virus derives its lipids from the host cell membrane where virus maturation occurs, which is the plasma membrane in most instances. Therefore a given virus can exhibit marked variation in lipid composition when grown in different cell types, which reflects the composition of the host cell plasma membrane in each instance (Klenk and Choppin, 1969, 1970; Quigley et al., 1971) . Apart from minor differences in carbohydrates of glycoproteins, virion proteins are indistinguishable when the virus is propagated in a variety of cells; therefore there appears to be little or no determining influence of viral proteins on the composition of the lipid bilayer. Since the viral envelope is continuous with the host cell membrane during morphogenesis, and membrane lipids are characterized by a high rate of lateral diffusion in the plane of the membrane, the similarity in lipid composition between viral envelopes and plasma membranes is not surprising.

The distribution of various lipid classes on the internal and external sides of the bilayer has recently been investigated in influenza virions using phospholipase digestion and phospholipid exchange proteins (Tsai and Lenard, 1975; Rothman et al., 1976) . The results indicate that lipids are distributed asymmetrically, with the majority of the phosphatidylinositol and about half of the phosphatidylcholine in the external half of the bilayer and most of the phosphatidylethanolamine and phosphatidylserine in the inner half. These distributions are thought to reflect a similar distribution in the host cell plasma membrane. Since less than half of the total phospholipids were accessible b y either approach, it was suggested that glycolipids constitute a significant part of the outer monolayer of the viral membrane. The presence of glycolipids in various enveloped viruses has been demonstrated by chemical analyses , and the types of glycolipids present also reflect those of the host cell. Myxoand paramyxoviruses lack neuraminic acid residues in their glycolipids, which again is a likely result of their neuraminidase activity Klenk et al., 1970b) . The reactivity of viral glycolipids toward specific lectins demonstrates that their carbohydrates are exposed on the external surface of the bilayer (Klenk et al., 1972a) .

All polypeptides located internal to the viral lipid bilayer are devoid of carbohydrate, and they may be subdivided into several types according to their structural location and functions. All enveloped viruses possess at least one polypeptide species closely associated with the nucleic acid to form a nucleocapsid or core structure, which is termed the nucleocapsid protein. In addition, viruses with helical nucleocapsids possess another major internal protein, termed the membrane protein, that appears to be associated with the internal surface of the lipid bilayer. Most nucleocapsid proteins are not directly involved in interaction with the lipid bilayer, although the nucleocapsids of togaviruses are roughly spherical structures that appear to be closely apposed to the inner surface of the bilayer. The internal membrane (M) proteins of myxo-, paramyxo-, and rhabdoviruses occupy a similar location in the viral envelope. It appears that these proteins are primarily responsible for conferring considerable rigidity on the viral envelope, as compared to that of host cell membranes of similar lipid composition (Stoeffel and Bister, 1975; Landsberger and Compans, 1976; Lenard et al., 1976) .

More complex viruses possess additional internal polypeptides as major components, but their precise locations in the virion have not been established. In addition, minor internal protein components are frequently observed in virions which possess enzymic functions, such as transcriptase activity. Such minor polypeptides are not thought to play a direct role in the structure or assembly of the viral envelope.

It is well established that carbohydrate chains are covalently linked to glycoproteins and glycolipids of the viral envelope, and both of these components were described above. However, it is uncertain whether or not other complex carbohydrates serve a function in viral envelope structure. The presence of sulfated mucopolysaccharides in highly purified virus preparations of several major groups was recently observed by radiolabeling virions with sulfate Pinter and Compans, 1975) . These sulfated components are derived from the host cell, since they can be labeled selectively by the growth of cells in 35S0,-containing medium prior to virus infection. In the case of influenza virus, it is possible that these components represent a host cell antigen described to be associated with purified virus particles (Knight, 1944) . Labeled sulfated mucopolysaccharides can be removed from virus preparations by digestion with hyaluronidase or trypsin without any effect on viral infectivity, suggesting that these components are not essential . It is likely that they associate with the external surface of the virion during the process of maturation.

Enveloped viruses share many common features in the organization of their structural components, as indicated by several approaches which include electron microscopy, surface-labeling and proteolytic digestion experiments, and isolation of subviral components. Much of the evidence pertaining to the arrangement of membrane components has been discussed in other reviews cited above and is only briefly summarized in this section.

The organization of the viral membrane corresponds in many respects to the fluid mosaic membrane model . Viral glycoproteins appear to be integral membrane proteins which are exposed on the external surface of a lipid bilayer. The external location of the glycoproteins is indicated by their sensitivity to proteolytic digestion and reactivity toward surface-labeling reagents (see review in Lenard and Compans, 1974) . Partial penetration of the glycoproteins into the bilayer is suggested by the fact that segments of the glycoproteins of several virus groups are found to be resistant to protease (Gahmberg et al., 1972; Mudd, 1974, Schloemer and Wagner, 1975b; Lenard et al., 1976) .

In contrast to the glycoproteins, the carbohydrate-free polypeptides of enveloped viruses are located internally to the lipid bilayer, as indicated by their resistance to protease treatment and lack of reactivity toward surface-labeling reagents. These internal proteins are also not reactive toward specific antibodies without disruption of the viral envelope. The best examples of nonglycosylated proteins which appear to be intimately associated with the internal surface of the envelope are the M proteins of myxo-, paramyxo-, and rhabdoviruses. However, whether these polypeptides associate with the bilayer as integral membrane proteins, or as peripheral membrane proteins such as spectrin in erythrocyte membranes, has not been clearly established.

In Fig. 1 the schematic cross section of an influenza virion shows the general arrangement of its structural components. The extent of penetration of the glycoproteins into the bilayer is not indicated but, FIG. 1. Schematic cross section of an influenza virion, depicting the arrangement of the major structural polypeptides. The rodlike HA spikes and the more complex NA spikes are composed of HA and NA polypeptides, respectively. These glycoproteins are exposed at the external surface of the lipid bilayer, and segments may penetrate the bilayer (not shown). The internal membrane-associated M protein appears to form a closely packed layer beneath the bilayer. The helical nucleocapsids are found as multiple discrete segments, each containing an RNA molecule coated with a major nucleocapsid protein, (NP). The nucleocapsids probably interact directly with the M protein during virion assembly. (From Compans and Choppin, 1975.) as discussed in Section IV, it is likely that a hydrophobic segment penetrates into the bilayer and may even traverse it. A closely packed layer of M protein is depicted on the internal surface of the bilayer in accordance with available information, and the helical nucleocapsids are depicted as twisted hairpinlike structures . It is likely that the nucleocapsids interact with the M proteins during assembly and in so doing participate in determining virion size and shape. Little is known about these interactions, and they are not indicated in Fig. 1 .

In this section we summarize the available information on the detailed structure of the glycoproteins of four virus groups. The infonnation obtained includes the size and shape of viral glycoproteins, the number of polypeptide chains in the complete glycoprotein structure, and compositional data on the polypeptide and oligosaccharide portions of the molecules. Detailed structural information on glycoproteins has been obtained for only a few virus types and is obviously incomplete even for the best studied viral systems.

Influenza viruses possess two glycoproteins with distinct biochemical and morphological properties: hemaggluti-nin (HA) and neuraminidase (NA). The HA of influenza virus iii one of the best studied viral glycoproteins. Electron microscope studies of the isolated spike structure have shown that it is a rod-shaped triangular prism approximately 14 nm long and 5 nm wide (Laver and Valentine, 1969) . The HA spike projects radially from the virion envelope, and it is probably a trimer having a MW of about 225,000 daltons (Schulze, 1975; Wiley et al., 1977) . HA glycoproteins can exist in one of two alternative forms: a single polypeptide with a MW of 75,000 daltons designated HA, or two polypeptides with MWs of approximately 50,000 and 25,000 daltons designated HA1 and HA2, respectively. HA, and HA, are proteolytic cleavage products of HA, which remain cross-linked by disulfide bonds (Lazarowitz et al., 1971; Laver, 1971) . The extent of cleavage of HA to HA, and HA2 may vary from none to complete cleavage of all HA polypeptides (Lazarowitz et al., 1973) . The extent of cleavage de-pends upon several factors including the host cell, the virus strain, and the presence or absence of proteases in the culture medium. The cleavage of HA may not be a requirement for infectivity of all virus strains, and spike morphology is not altered detectably by cleavage of HA to HA, and HA,, but recent studies have shown that cleavage enhances the infectivity of some influenza virus strains (Klenk et al., 1975 ; Lazarowitz and Choppin, 1975) .

The tryptic peptide maps of HA, and HA, are distinct, and the only notable difference in the amino acid composition of the two polypeptides is the proline content (Laver, 1971) . The proline content of HA, is at least six times higher than that of HA,. HA, is tightly folded on itself, probably reflecting the high proline content. A portion of the HA spike is released by protease treatment as a soluble protein which can be crystallized (Brand and Skehel, 1972) .

Functional HA spikes have been isolated from several influenza strains by treating virions with ionic or nonionic detergents. After removal of the detergent, the spikes aggregate, forming rosettelike structures, suggesting that the base of the HA spike is hydrophobic in nature (Laver and Valentine, 1969) . Available evidence indicates that the HA spike is amphipathic and that a portion of it is buried within the lipid bilayer of the viral envelope. This is indicated b y the finding that protease-treated virions retained HA, or portions thereof while having no identifiable spikes (Compans et nl., 1970; Lenard et al., 1976) .

The HA, glycoprotein therefore contains the hydrophobic end of the spike. This glycoprotein aggregates even in the presence of guanidine hydrochloride (Laver, 1971) . The composition of a 50-residue peptide associated with the lipid bilayer, which is removed from the carboxyl terminus of HA2 by bromelain, has been estimated, and 21 of the amino acids are hydrophobic in nature, whereas 14 may be charged (Skehel and Waterfield, 1975) . The carboxyl terminus of this peptide is thought to be buried within the lipid bilayer. The aminoterminal end of HA, seems to extend radially from the virion surface, and a sequence of 10 residues from the amino-terminal end of HA, is highly conserved in both type-A and type-B influenza virions (Skehel and Waterfield, 1975) . The hydrophobicity of HA, suggests that it is similar to other membrane-associated glycoprotein segments (Segrest et al., 1972; Nakashima et al., 1975) , but freeze-fracture studies of the influenza virion indicate that extensive portions of HA and NA do not penetrate through the lipid bilayer (Bachi et al., 1969) . Circular dichroism studies of HA spikes isolated intact and those cleaved from the virion surface may be useful in further defining the secondary structure of the hydrophobic regions.

Until recently, little information was available concerning the carbohydrate portion of the HA glycoprotein. About 80% of the carbohydrate attached to viral proteins was shown to be linked to HA; the gly-cosy1 moieties were shown to be comprised of glucosamine, mannose, galactose and fucose at a molar ratio of 6 :4 : 1 : 1, respectively, and the total carbohydrate of HA, determined by chemical means, was estimated to be approximately 12,000 daltons (Laver, 1971) . It was also found that HA, of influenza A virions contains fucose, whereas HA, of influenza B (GL1760) does not . Recently the glycopeptides obtained by pronase digestion of influenza A viruses have been characterized (Schwarz et al., 1977; Nakamura and Compans, 1978b) . Both groups showed that HA possesses type I (oligosaccharide side-chain comprised of glucosamine, mannose, galactose, fucose and sialic acid) and type I1 (oligosaccharides comprised of glucosamine and mannose) glycopeptides, but the distribution of each type on HA, and HA, was shown to depend upon the virus strain. Schwarz et al. (1977) showed that HA, of two avian influenza viruses possessed only type I glycopeptides, whereas both type I and I1 glycopeptides were shown to be present on HA,. However, Nakamura and Compans (197813) using the WSN strain of influenza showed that HA, possessed both type I and I1 glycopeptides whereas HA, contained only type I glycopeptides. Since it is likely that HA, occupies a similar structural location in WSN and avian hemagglutinin proteins, these results suggest that specific amino acid sequences determine whether a type I or type I1 oligosaccharide is added at a particular site on the glycoprotein.

The host cell type may also determine the type of oligosaccharide chains added to a given viral glycoprotein. Thus, although HA, of WSN strain influenza virions grown in MDBK cells contained only type I glycopeptides, both type I and I1 glycopeptides were found in HA, when grown in CEF cells (Nakamura and Compans, 1978~) . Schwarz et al. (1977) showed that the type I HA glycopeptides of fowl plague and N influenza strains grown in chicken embryo fibroblasts (CEF) had a molecular weight of approximately 2,600 daltons, whereas type I1 glycopeptides from the same source has a molecular weight of approximately 2,000 daltons. The molecular weights of type I and I1 glycopeptides of the WSN strain of influenza virus were also found to have approximately this size (Nakamura and Compans, 1978b) . HA glycopeptides of virus grown in MDBK cells were slightly larger in size than those of virus grown in CEF cells. Based upon the estimated carbohydrate content (12,000 daltons) of the HA glycoprotein obtained by Laver (1971) and Schwarz and Klenk (1974) and the size estimates of the type I and I1 glycopeptides of influenza virus grown in MDBK cells, it was estimated that HA, contains a single type I glycopeptide whereas HA, possesses two type I and one or two type I1 oligosaccharide side-chains for the WSN strain (Nakamura and Com pans, 1978b) .

The NA spike, as shown in Fig. 1 , also projects radially from the lipid bilayer of the influenza virion but is morphologically distinguishable from the HA spike (Laver and Valentine, 1969) . The NA spike is elongated and has a square knoblike structure at one end. The oblong head is approximately 5 x 8.5 nm, and it is attached to a fiber approximately 10 nm long. The NA spike has a MW of approximately 240,000 daltons and is comprised of four NA polypeptides each having a MW of approximately 58,000 daltons; two NA polypeptides appear to be linked by disulfide bonds to form dimers, which are thought in turn to aggregate by noncovalent bonds to form the tetrameric spike (Bucher and Kilbourne, 1972; Lazdins et al., 1972) . The amino acid composition of the NA has been determined, and it has a cysteine content significantly higher than that of the other viral polypeptides (Laver and Baker, 1972) . Influenza NA is in some ways morphologically similar to enzymes involved in sugar metabolism in the gut, as discussed by Forstner and Riordan in this volume.

After trypsin treatment a tetramer of four coplaner subunits of 4 X 4 x 4 nm was isolated (Wrigley et al., 1973) . This structure appears to correspond to the knob, and it is enzymically active. The knob structure can no longer aggregate with itself or with HA molecules, indicating that the hydrophobic regions of the molecule have been removed by proteolysis. The MW of the portion of the NA monomer which remains associated with the envelope after trypsin treatment was estimated to be 7000 daltons by Lazdins et al. (1972) and 12,000 daltons by Wrigley et al. (1973) . Presumably, this portion functions to attach the NA molecule to the virion and plays no role in the enzymic properties. Furthermore, the portion of the spike which remains associated with the viral envelope is more highly glycosylated than the knob-shaped portion of the NA spike with enzymic activity (Lazdins et al., 1972) .

NA is present in influenza virions in smaller amounts than HA, in a ratio of about three to four HA polypeptides for each NA polypeptide. Little information is available on the carbohydrate components of the NA.

Recently, both the HA and NA of influenza virions were shown to be sulfated glycoproteins . The sulfate appears to be covalently linked to the oligosaccharide chains of viral gly-coproteins Compans, 1977, 1978a) . Glycoproteins of enveloped viruses of all the other major groups studied also are sulfated, whereas carbohydrate-free polypeptides are not Kaplan and Ben-Porat, 1976 ).

Rhabdoviruses are covered with closely spaced glycoprotein spikes approximately 10 nm in length as shown by electron microscope studies. Rhabdoviruses possess only a single glycoprotein species termed the G protein, and for vesicular stomatitis virus (VSV) it has a MW of approximately 69,000 daltons (Wagner, 1975) . Carhvright et d . (1972) have postulated that only a single glycoprotein molecule constitutes the spike structure.

The G proteins of rhabdoviruses are amphipathic, like the glycoproteins of influenza virions. After protease treatment of VSV a fragment of the G protein was demonstrated to be associated with the intact virion (Mudd, 1974) . More recently, Schloemer and Wagner (1975a) isolated a small nonglycosylated portion of the G protein from the envelope of protease-treated VSV virions. The fragment was found to have a MW of 5200 daltons, approximately equivalent to 50 amino acids. This is similar to the estimated size of the portion of the HA2 protein of the influenza virion thought to be buried within the lipid bilayer. Amino acid analysis of the G-protein fragment showed that it contained a preponderance of hydrophobic amino acids. As is the case for the membrane-associated portions of influenza HA, the hydrophobic fragment of the G protein is long enough to penedate the lipid bilayer. Conclusive evidence for such penetration has not been obtained, but cross-linking experiments with glutaraldehyde suggest that interactions may occur between G proteins and internal M proteins (Brown et d . , 1974) .

The carbohydrates linked to the G protein of VSV have been studied in detail. The glycoprotein is 9-10% carbohydrate by weight and contains mannose, galactose, N-acetylglucosamine, and neuraminic acid as the major sugar components, with lesser amounts of N-acetylgalactosamine and fucose (McSharry and Wagner, 1971 ; Huang, 1970. Etchison and Holland, 1974a) . The size and composition of the carbohydrate moieties per VSV glycoprotein are variable, depending upon the cell type in which the virions are grown (Burge and Huang, 1970; Etchison and Holland, 1974a) . Likewise, the sequence of the carbohydrates within the glycosyl side-chains may exhibit cell depen-dence (Moyer and Summers, 1974) . The monosaccharide composition of the oligosaccharide side-chains is similar to that of influenza virus (Etchison and Holland, 1974b) , with the exception that sialic acid is present on the termini. Klenk et aZ. (1970b) demonstrated the presence of sialic acid on the envelope of VS virions by treating the virus with colloidal iron hydroxide which stains sialic acid residues. The stain was shown to bind to the envelope of VSV, whereas it did not bind to influenza virions or paramyxoviruses as shown by electron microscopy. More recent studies have shown that the glycosyl moieties are not terminated by sialic acid when VSV is grown in mosquito cells, because these cells lack sialyl transferase (Schloemer and Wagner, 1975b) .

Preliminary studies indicated that from 3-5 cyanogen bromide peptides of the VSV G-protein may be glycosylated (Wagner, 1975) . However, Etchison and Holland (1974a,b) have calculated that there are only 2 glycopeptides of 3000-3400 daltons in each G protein molecule. More recent studies by Etchison et al. (1977) have indicated that the VSV G-protein possesses two identical glycopeptides. The number average molecular weight of the glycopeptides was estimated to be 3150 by gel filtration analysis and 3450 based on composition of the amino acid and sugar residues. Additional information concerning the structure of the glycopeptide was obtained by sequential chemical and enzymatic degradation. These results indicate that the glycopeptides are acidic type I glycopeptides with two or three mannose branches terminating in sialic acid. Moyer and co-workers (1976) have obtained evidence that the glycosyl residues are covalently linked to an asparagine residue of the G protein, and the sequence of the monosaccharides that constitute the oligosaccharide side-chains has been determined (Hunt and Summers, 1976b) .

Togavinises are small, spherical enveloped viruses 50-60 nm in diameter with a core structure that appears to be icosahedral. Although there are many members of the togavirus group, the best studied are Sindbis virus and Semliki Forest virus (SFV). Sindbis virus possesses two glycoproteins designated El and E,; both have a MW of approximately 50,000 daltons, and they are not linked by disulfide bonds because they can be separated under nonreducing conditions and without alkylation (Schlesingeret al., 1972) . SFV is similar to Sindbis virus in that it also has 2 glycoproteins of similar molecular weight designated El and E2, but a third glycoprotein designated E, has also been detected. The molecular weights of El, E,, and E, were estimated to be 49,000,52,000 and 10,000 daltons, respectively (Garoffet al., 1974) . E, and E, are synthesized as a common precursor protein (NVP68) that is cleaved to yield E, and E,. El, EP, and E, of SFV are present in equimolar ratios as are E, and E2 of Sindbis virus.

The arrangement and relationship of the glycoproteins within the spike structure of the Sindbis and SFV virion are yet to be resolved. and Garoff (1974) have used the cross-linking agent dimethyl suberimidate to study the interrelationships of the glycoproteins of SFV. El and E, were most readily cross-linked, but steric hindrance may have prevented the cross-linking of El, E2, and E,. Recent studies by Jones et al. (1977) have suggested that El and E, of Sindbis virus, like El, EZ, and E, of SFV, probably constitute the glycoprotein spike, since the cleavage of PEP, a precursor of E,, does not occur in temperature sensitive mutants of complementation groups including that thought to represent El. Furthermore, cleavage is inhibited by antibodies directed against either El or E,. These data suggest that PE, and El may exist as a complex in the membrane of the infected cell, and presumably also that El and E, remain as a complex in the viral envelope.

Protease treatment of SFV cleaves the glycoproteins, El and E,, and residual segments can be isolated from the envelope of spikeless particles (Utermann and Simons, 1974) . The amino acid composition of each of these peptides demonstrates that they are enriched in hydrophobic amino acids. The residual peptides have a MW of approximately 5000 daltons; hence they are comprised of about 50 amino acids. Thus, like the glycoproteins of influenza virus and VSV, togavirus glycoproteins are amphipathic in nature and appear to possess a peptide of similar size embedded within the viral envelope. When SFV was treated with high concentrations of dimethyl suberimidate, both the tail fragments of El and Ez were cross-linked with the nucleocapsid protein, which supports the conclusion that the hydrophobic segments, of El and E, may penetrate through the lipid bilayer and interact with the nucleocapsid .

Carbohydrate analysis of SFV showed that El contains about 18 moles of monosaccharide, E2 about 28 moles and E, about 22 moles . E2 seemed to be particularly rich in mannose. It has recently been reported that the SFV glycoproteins El, E,, and E, are differentially glycosylated (Mattila et al., 1976) . El and E, were shown to contain, on the average, one type A glycosyl side chain. E2 was shown to contain one type A glycosyl side chain and possibly one or two B-type glycosyl side chains. A-type oligosaccharides are complex structures containing fucose, galactose, mannose, and N-acetyl-glucosamine, whereas B-type oligosaccharides contain only mannose and N-acetylglucosamine (Johnson and Clamp, 1971 (1975) showed that one type-A and one type-B residue were attached to each of the two different glycoproteins. The extent of completion of the glycosyl side chains was shown to be dependent in part upon the cell in which the virus was grown (Burge and Huang, 1970; Keegstra et ul., 1975) . BHK-21 cells were able to complete type-A side chains (i.e., to add galactose and terminal sialic acid), whereas chicken embryo fibroblasts were quite inefficient in adding sialic acid (Keegstra et al., 1975) .

Both SFV and Sindbis virus possess hemagglutinating activity. Recent studies by Dalrymple et al. (1976) localized this activity to the El glycoprotein of Sindbis virus, whereas the E2 glycoprotein possesses the antigenic determinants which react with neutralizing antibody.

The virions of paramyxoviruses are covered with glycoprotein surface projections approximately 10 nm in length. Two functionally distinct types of glycoproteins have been isolated from several members of the paramyxovirus group (Scheid et al., 1972; Scheid and Choppin, 1973; . In contrast to the situation in influenza viruses, which possess hemagglutinating and neuraminidase activities on distinct glycoprotein molecules, both of these activities are associated with the larger of the two glycoprotein species in paramyxoviruses, which is designated the HN glycoprotein. The smaller glycoprotein component in paramyxoviruses is associated with cell fusion and hemolysis activities and is designated the F glycoprotein.

Purified glycoproteins of SV5 form rosettelike clusters in the absence of detergents, suggesting that the spikes have hydrophobic bases (Scheid et al., 1972) . The aggregates formed are morphologically distinguishable, the HN protein forming berrylike aggregates while the F protein forms distinct rosettelike clusters consisting of radiating spikes 10-13 nm in length with distinct terminal knobs.

In the three best studied paramyxoviruses, SV5, Newcastle disease virus (NDV), and Sendai virus, the H N glycoprotein has a MW range of 65,000-74,000 daltons, and the HN spike solubilized by detergent treatment sediments at 8.9s (Scheid et al., 1972) . Although the exact size and fine structure of the HN spike remain to be determined, it has been suggested that each morphological spike contains at least two glycoprotein monomers . In some strains of NDV, an 82,000-dalton precursor of the HN glycoprotein designated HNo is incorporated into virions . These particles have reduced hemagglutinating and neuraminidase activities which are activated upon proteolytic cleavage, which produces the 74,000 dalton HN molecule.

In SVS, NDV, and Sendai virions, the F glycoprotein has a MW of -56,000 daltons, and the F spike that is solubilized by treatment of SV5 virions with Triton X-100 has a sedimentation coefficient of approximately 6.7s (Scheid et al., 1972) . In Sendai virions (Homma and Ohuchi, 1973; Scheid and Choppin, 1974) and certain strains of NDV (Nagai et al., 1976) grown in some cell types, the F glycoprotein is not found, and a larger precursor molecule designated Fo is observed.

Recent studies have indicated that proteolytic cleavage of the Fo glycoprotein yields two cleavage products, which have bekn designated F1 and F, (Shimizu et al., 1974; Nagai et al., 1976; Scheid and Choppin, 1977) . The F2 cleavage product is more highly glycosylated than F1, and appears to be located on the distal end of the spike; it contains a blocked N-terminal as is also found in the uncleaved F, glycoprotein (Scheid and Choppin, 1977) . Thus the cleavage of F, generates a free N-terminal on the F, segment of the glycoprotein, and this appears to expose a new hydrophobic region of the molecule which may be important for virus-induced cell fusion.

The F, glycoprotein is inactive in cell fusion and hemolysis and can be converted into the active F glycoprotein by proteolytic cleavage, with concomitant activation of cell fusion and hemolysis. It is of particular interest that virions containing the Fo precursor are not infective and gain infectivity upon such proteolytic cleavage.

In the case of enveloped viruses, glycoproteins located on the surface of the virion are the components involved in adsorption to cellular receptors, which may or may not be host cell glycoproteins. Hemagglutination by influenza virus has been studied as a model system for the adsorption of a virus to receptors, and considerable information has been obtained. Adsorption to the erythrocyte occurs by the HA spike binding to sialic acid-containing components on the cell surface. The receptor molecule has been isolated from chick red blood cells, and it is a major glycoprotein containing M and N blood group anti-gens. A detailed summary of the properties of this receptor is given by Schulze (1975) , as well as by Tanner in this volume. Removal of sialic acid from the receptor molecule by NA prevents the agglutination of erythrocytes by influenza virus. Hemagglutination can also be inhibited by pretreating the virus with specific antibodies.

While the model for adsorption of influenza virus to erythrocytes seems straightforward, its applicability to the adsorption of viruses to other cell types is uncertain, and little information is available on the nature of receptors for viruses. Liposomes containing gangliosides may be capable of acting as receptors for Sendai virus, a paramyxovirus (Haywood, 1974 (Haywood, , 1975 . Sialyoglycoproteins inserted in liposomes can also act as receptors, but the fact that gangliosides can act as receptors raises the possibility that glycolipids may be involved in viral attachment.

Further evidence that viral glycoproteins specify virus-host cell interactions has been obtained from the studies of Bishop et al. (1975) . They produced spikeless particles of the Indiana serotype of VSV b y treating the virus with bromelain or pronase. These particles were shown to be noninfectious but, when they were reconstituted with purified G protein isolated from the same strain or from the New Jersey serotype, infectivity was restored. Antibody directed against the homologous G protein used for reconstitution effectively neutralized the virus, but antibody directed against the serotype of the spikeless particle was ineffective in neutralization when glycoproteins from a different serotype were used for reconstitution.

The mechanism of adsorption of other enveloped viruses has not been studied to the same degree as that of ortho-and paramyxoviruses. Herpes viruses do not exhibit hemagglutinating activity, but they possess at least 13 glycoproteins which are asymmetrically located on the external surface of the envelope (Roizman and Furlong, 1974; O'Callaghan and Randall, 1976) . Removal of the envelope by nonionic detergents irreversibly alters the infectivity of these viruses (Abodeely et al., 1970) .

Oncornavinis glycoproteins react specifically with host cell receptors and in the case of avian leukosis viruses they define the host range as well as the classification into subgroups based on interference and neutralization properties Ishizaki and Vogt, 1966; Duff and Vogt, 1969) . In addition, Tozawa et al. (1970) showed that the viral glycoproteins absorbed homologous neutralizing antibody but not antisera prepared against heterologous virions. The purified glycoproteins were shown to interfere with the early steps in infection by the homologous virus. The determination of host range by the glycoproteins of avian leukosis viruses was further demonstrated by phenotypic mixing experiments. Defective Rous sarcoma virions that lack the envelope gene cannot infect cells. However, when grown in the presence of an avian leukosis virus, they acquire the glycoproteins of that virus, and the host range of the sarcoma virus reflects that of the leukosis virus (Hanafusa, 1965; Vogt, 1965 ; Kawai and Hanafusa, 1973) . The major glycoprotein of Rauscher murine leukemia virus (gp71) binds specifically to receptor molecules found on murine cells but not on other mammalian cells (DeLarco and Todaro, 1976).

The phenomenon of cell fusion may be caused by members of several groups of enveloped viruses of which paramyxoviruses are the best studied. The properties of hemolysis and cell fusion are associated with the F glycoprotein as described above. Virus-induced cell fusion may occur in the absence of virus replication, in the presence of high concentrations of virus particles. Both infectious and noninfectious viruses are equally adept at causing such cell fusion which sometimes has been termed fusion from without (Bratt and Gallaher, 1972) . In contrast, fusion with low multiplicities of virus has been called fusion from within and is dependent upon replication of the virus. It is likely that the F glycoprotein is involved in both types of fusion phenomena.

The process of fusion through the direct action of concentrated virus has been studied intensively since it was first observed by Okada (1958) . Sendai virus-induced fusion involves several separate identifiable steps, as indicated by Maeda et al. (1977) : (1) adsorption of the virus to the cell, which seems to occur at the tips of the spikes located on the virion surface; (2) aggregation of cells; (3) fusion of the viral envelope with the cell membrane and finally fusion of the cells. These studies suggest that the envelope bilayer of the virion and the membrane of the target cell are brought into contact by the action of the viral HN glycoprotein. The F protein then causes destabilization of the lipids, and an intermixing of lipids occurs. The mechanism of action of the F protein remains to be determined.

A neuraminidase activity has been shown to be associated with orthomyxoviruses and paramyxoviruses. The functional role of this enzyme was uncertain for many years, and there was even doubt at some point that it was a viral gene product, since similar enzyme ac-tivity is present in normal cells (White, 1974); however, these doubts have been resolved by studies of the biochemical, genetic, and antigenic properties of viral neuraminidase (see review in Bucher and Palese, 1975) . Several distinct functions have been proposed for this enzyme. It is postulated that neuraminidase-containing virus lodges in the upper respiratory tract and binds to mucin via HA. The glycosyl residues of muchin are terminated by sialic acid (N-acetylneuraminic acid) to which HA binds, and one role of the neuraminidase may be to cleave the sialoglycoprotein bond, freeing the bound virion. It is presumed that in this way the virus is released, and that underlying cell receptors are exposed to which HA can attach (Davenport, 1976 ).

It has also been postulated that neuraminidase is involved in an early event such as penetration, but this is unlikely because virions remain infectious after inhibition of neuraminidase activity by specific antibody (Bucher and Palese, 1975) . However, the possibility that neuraminidase participates in release of the budding virion from the infected cell surface has gained support from several types of experiments. By using influenza strains with different levels of enzyme activity, it was shown that virus strains having low activity were released from cells more slowly than strains with higher enzyme activity (Palese and Schulman, 1974) . Further, in the presence of antibody to viral neuraminidase, which inhibited enzyme activity, virions were formed but release of virus into culture media was inhibited (Seto and Rott, 1966; Compans et al., 1969; Webster, 1970 More conclusive information on the function of the enzyme has been obtained with temperature-sensitive mutants of influenza virus which are defective in neuraminidase activity . At the nonpermissive temperature, no neuraminidase activity is detected; virus particles are produced by cells, despite the fact that infectivity titers are markedly reduced. However, the virus particles form large aggregates, and in contrast to wild-type virions these particles contain sialic acid as shown by colloidal iron hydroxide staining.

Since influenza virions bind to sialic acid residues, these results indicate that the mutant virus particles aggregate to each other, because sialic acid is added to viral carbohydrates, and that the essential function of viral neuraminidase is to remove or prevent the addition of such sialic acid. In support of this conclusion, the addition of bacterial neur-aminidase to cells infected with these mutants cause a marked enhancement of virus release.

Further evidence supporting this role for the neuraminidase was obtained with a neuraminidase inhibitor, 2-deoxy-2,3-dehydro-Ntrifluoracetylneuraminic acid (FANA). Influenza virus grown in the presence of FANA contains neuraminic acid on its envelope, and the particles undergo extensive aggregation (Palese and Compans, 1976) . A marked reduction in virus yield is observed because of this aggregation, and treatment with purified neuraminidase results in a marked enhancement in progeny virus yields, apparently through disaggregation of virus. Thus viral neuraminidase is not required for assembly of progeny virions but appears to be essential for the removal of sialic acid from the surface of the virion itself.

From the studies described above it is evident that the membranes of enveloped viruses are asymmetrically constructed, with the glycoproteins that comprise the spikes or surface projections physically located on the exterior of the viral envelope. Thus the glycoproteins are exposed to the immune surveillance system of the host and, being good immunogens, may elicit humoral and cellular immune responses (Evans, 1976) . The classification of virus isolates into specific strains depends largely upon serological procedures, and for enveloped viruses, surface glycoproteins are usually the relevant antigens in such tests. Neutralizing antibodies are usually directed against the viral proteins involved in attachment to receptors, e.g., HA glycoprotein of influenza virus (Webster and Laver, 1975) , G protein of VSV (Wagner, 1975) , gp69/71 of murine leukemia viruses (Fischinger et aZ., 1976; Strand and August, 1976) , gp85 of avian leukosis viruses (Bolognesi, 1974) , and E, of Sindbis virus (Dalrymple et d., 1976) . High concentrations of antibody can prevent attachment of the virus to receptors, but low concentrations can also neutralize infectivity by a mechanism that is not understood.

The specific determinants of viral glycoproteins recognized by virus-neutralizing antibodies have not been chemically characterized. It has been postulated that antibody molecules may recognize only determinants on the tip of the HA spike of the influenza virion (White, 1974) . In reaching these conclusions, it has been assumed that the size of the antibody molecule precludes the possibility that it could make contact with any other part of the spike, since the spaces between the spikes are too small for the immu'noglobulin to interact with other regions. Electron microscopic observations (Lderty and Oertilis, 1963) indicate that antibody molecules interact with the tips of surface spikes, and analysis of the tryptic peptides of HA molecules isolated from *closely related influenza strains has shown that they rarely differ b y more than one or two peptides Webster and Laver, 1972) , suggesting that the strain differences are indeed restricted to small regions or determinants of the HA spike.

The antigenic character of many viral glycoproteins is stable, but the determinants exhibited by the glycoproteins of influenza virions are characteristically variable, as evidenced by the many different strains of type-A and -B influenza. The HA and NA glycoproteins of influenza viruses are antigenically distinct, and they undergo antigenic changes independently of each other. The antigenic changes that occur may be gradual, in which case the different virus strains are clearly related to each other with respect to both surface antigens. Antigenic changes of this nature are termed antigenic drift, and they result from the interplay of viral mutability and immunological selection (Webster and Laver, 1975) . The presence of antibody of low avidity may select for single-step mutants, which have an altered amino acid in the key area of the antigenic determinant, giving rise to a new viral strain.

At intervals of 10-15 years sudden and complete changes in the determinants of type-A influenza glycoproteins occur; the changes are such that the viruses that arise possess glycoproteins that appear completely distinct on peptide mapping . Dramatic changes in the antigenic determinants of viral glycoproteins are termed antigenic shifts, and it is these new viruses that cause worldwide influenza pandemics. It has been postulated that antigenic shift may occur because type-A strains of human origin may undergo recombination with type-A strains of avian and animal origin. The antigenic determinants of the new HA or NA are sufficiently different that the human host does not possess immunity, and the virus gains a selective advantage. Interestingly type-B influenza viruses do not undergo antigenic shift. The reason for this may lie in the fact that type-B influenza viruses have not been isolated from other animal species; thus it is probable that the type-B viruses can not undergo similar recombination events (Webster and Laver, 1975) .

Viral glycoproteins provide excellent systems for analysis of the function of carbohydrates in membrane glycoproteins. Several approaches have been used to modify the carbohydrates, including treatment of virions with specific glycosidases, growth of virus in cells with specific sugar transferase defects, and treatment of virus-infected cells with inhibitors of glycosylation.

Removal of sialic acid from the G protein of VSV has been reported to reduce the infectivity of VSV virions (Schloemer and Wagner, 1974) , whereas the infectivity of SFV (Kennedy, 1974) and Friend leukemia virus (Schafer et al., 1977) were reportedly unaltered by such treatment. Moreover, enzymic addition of sialic acid to the glycoproteins of Sindbis virus did not alter the infectivity of this virus (Stollar et al., 1976) , while enhancing the infectivity of influenza virus (Schulze, 1975) and restoring infectivity to neuraminidase-treated VSV (Schloemer and Wagner, 1974) .

Treatment of influenza virions with glycosidases alter hemagglutinating activity, but neuraminidase activity was unaffected; however, similar treatment of NDV, a paramyxovirus, resulted in no alteration in hemagglutinating or neuraminidase activities (Bike1 and Knight, 1972) . Schiifer et al. (1977) showed that the indirect hemagglutinating activity of Friend leukemia virus was inhibited by glycosidase treatment but that viral infectivity and determinants involved in viral interference and absorption of neutralizing antibody were unaltered. Sindbis virus were grown in these cells, the apparent MWs of their glycoproteins were lower. The infectivity of VSV and Sindbis virus grown in these cells was not altered from that of fully glycosylated virions. However, glycosidase treatment of SFV was shown to decrease the infectivity of this virus (Kennedy, 1974) . The apparent difference between these results remains to be resolved. It is possible that the core of the glycosyl side chain is added to the Sindbis virus glycoproteins in the enzyme-deficient cells, whereas the glycosidase treatment used by Kennedy may have removed more of the glycosyl moieties from the SFV glycoproteins.

Inhibitors of glycosylation have also been used to assess the role of glycosyl side chains of glycoproteins. Primarily, three different inhibitors have been used for this purpose: 2-deoxy-~-glucose (2-dG), an analog of glucose which substitutes for mannose, preventing the further addition of monosaccharides to the glycosyl side chain; Dglucosamine, which is thought to inhibit glycosylation at high concentrations by decreasing UTP pools in the cell and subsequent activation of other sugars (Scholtissek, 1971) ; and tunicamycin (TM), a glucosamine-containing antibiotic that inhibits the formation of N-acetyl-glucosamine-lipid intermediates which serve as donors for the synthesis of the oligosaccharide side chains of glycoproteins (Tkacz and Lampen, 1975) . Kilbourne (1959) and Kaluza et al. (1972) were among the first investigators to utilize inhibitors of glycosylation to study the role of carbohydrates in influenza virions. They showed that 2-dG and D-glucosamine inhibited the biosynthesis of active HA, NA, and mature infectious influenza virions. Subsequent biochemical studies revealed that high concentrations of 2-dG or Dglucosamine prevented the synthesis of influenza virus glycoproteins (Klenk et al., 1972b Nakamura and Compans, 1978a) . Instead, an unglycosylated or incompletely glycosylated hemagglutinin precursor HA,, was detected. Once synthesized, HAo was found associated with cytoplasmic membranes, as is the case with the normal glycoprotein. HAo glycoprotein in cells infected with the fowl plague strain (FPV) was cleaved by cellular proteases to yield a heterogeneous product . However, the comparable protein synthesized in cells infected with the WSN strain was processed and incorporated into virions (Nakamura and Compans, 1978a) .

Because 2-dG and D-ghcosamine interfere with metabolic reactions other than the glycosylation of glycoproteins, they may cause side effects that can affect virus replication. Therefore TM, a compound that seems to affect only the glycosylation of glycoproteins, has been employed to extend these studies. T M inhibited virion formation in FPVinfected cells and the unglycosylated glycoprotein HA,, appeared to be degraded by cellular proteases (R. T. ; HA,, of the WSN strain synthesized in the presence of TM also appeared to be completely unglycosylated, whereas even at high concentrations of 2-dG and Dglucosamine some glycosylation occurred (Nakamura and Compans, 1978a) . Nonetheless, TM did not inhibit virion formation to the same extent as 2-dG (Nakamura and Compans, 1978) . The surface spike layer of WSN virions produced in the presence of 2-dG, D-glucosamine, and TM was altered morphologically, and the hemagglutinating activity of the virions was significantly reduced. These results suggest that glycosylation of virion glycoproteins is not required for influenza virion formation but is needed for biological activity of viral glycoproteins. These glycosylation inhibitors have also been used with several other enveloped viruses. When herpes viruses are grown in the presence of 2-dG, the infectious virus yield is decreased by greater than 95%, but the yield of viral particles is not reduced (Courtney et al., 1973) . The reduction in infectivity was attributed to an inability of the virions to attach to the host cell and penetrate, implying that oligosac-charide residues of herpes virus glycoproteins play a role in the attachment and recognition of host cell receptors.

The glycoproteins of VSV, Sindbis virus, and SFV are not glycosylated in the presence of TM (R. T. Leavittet al., 1977) . Mature VSV and Sindbis virions are not released from TMtreated cells, but nonglycosylated precursors are synthesized and seem to be stable within the cell (Leavitt et al., 1977) . Similarly, R. T.

Schwarz et al. (1976) showed that unglycosylated SFV glycoprotein precursors were synthesized in TM-treated cells. Moreover, they were not degraded by host cell proteases and virion assembly was completely inhibited. It is interesting that, as noted above, HA, of FPV is degraded when it is grown in TM-treated chicken embryo fibroblasts, but the glycoprotein precursors of SFV synthesized in the same cells are stable. These results may be due to greater release of protease by FPV infection or, alternatively, the polypeptide backbone of SFV glycoproteins may be less susceptible to proteolytic degradation.

High concentrations of glucosamine rapidly shut off the production of infectious avian sarcoma virus particles (Hunter et al., 1974) . However, avian sarcoma virus particles were assembled and released at about 60% of the control level in TM-treated cells, and such particles appeared to lack glycoproteins (R. T. . The infectivity titer of the virus produced in TM-treated cells decreased by only 15%.

These results indicate that the effects of glycosylation inhibitors may vary with the virus and host cell, as well as with the specific inhibitor used. Side effects of some inhibitors may have marked effects on virus replication. However, the fact that in some systems virus particles are produced with unglycosylated or incompletely glycosylated glycoproteins clearly demonstrates that the complete glycosylation process is not essential for intracellular migration of glycoproteins or their incorporation into the plasma membrane and subsequently into virus particles. The biological activities of some glycoproteins such as those of influenza virus and SFV appear to require glycosyl moieties, whereas for other viruses, such as murine leukemia virus, this may not be the case.

The membranes of enveloped viruses, particularly those of singlestranded RNA viruses, are unique in their simplicity of construction and are useful systems for studying the interactions of specific proteins with the lipid bilayer. Hence comparative studies of the effects of viral proteins of the lipid bilayer structure have been made using the techniques of electron spin resonance (ESR), nuclear magnetic resonance (NMR), and fluorescence polarization (FP). ESR studies have provided evidence that the lipids of influenza virus, SV5, Rauscher leukemia virus, VSV, and SFV (Landsberger et uZ., 1971 (Landsberger et uZ., , 1972 (Landsberger et uZ., , 1973 ; Sefton and Gaffney, 1974) are bilayer structures with fluid lipid phases similar to those observed for other biological membranes, but all viral membranes were shown to be substantially more rigid than the corresponding host cell plasma membrane. Proteolytic removal of the glycoprotein spikes from the surface of SV5 and influenza virions did not appreciably alter the fluidity of the lipid bilayer of these viruses. Lenard et al. (1976) have provided further evidence that the glycoproteins of the influenza virion contribute little to the rigidity of the lipid bilayer. The phospholipid composition of standard influenza particles, and "incomplete" virus produced upon serial undiluted passage, were compared and found to be indistinguishable, as were the ESR spectra of the two types of particles. The incomplete particles were shown to contain approximately twice the amount of glycoproteins relative to the complete particles. Thus it was concluded that the rigidity of viral membranes may be determined b y the M protein and not by the viral glycoproteins. This may not be entirely the case for VSV and SFV, since Sefton and Gaffney (1974) and Landsberger and Compans (1976) observed that, when the glycoproteins of these viruses were removed by proteases, the envelope became more fluid, indicating that the viral glycoproteins may contribute to the rigidity of the envelope. However, Landsberger and Compans (1976) postulated that the major effect on VSV bilayer fluidity was exerted by the M protein, since the fluidity of the lipid bilayer was altered only slightly when the G protein was removed by protease, whereas vesicles prepared from extracted viral lipids were much more fluid than lipids in virions.

Using the technique of fluorescence depolarization, Moore et aZ. (1976) and Barenholz et al. (1976) showed that the envelope of SFV, Sindbis virus, and VSV has a higher microviscosity than that of the plasma membranes from which the virions budded. The increased microviscosity was attributed in part to insertion of the hydrophobic regions of the glycoproteins into the envelope bilayer. Stoffel and Bister (1975) , using NMR spectra of '3C-labeled lipids, also demonstrated that the envelope lipids of VSV are highly rigid, as a result of either lipid-lipid or lipid-protein interactions. In general it may be concluded therefore that the lipid bilayer of enveloped virions is more rigid than the host cell membrane from which the virion buds. While virion glycoproteins do interact with the lipid bilayer and may affect the rigidity of the membrane to some extent, the internal membrane protein(s) may also be of equal or greater importance in determining membrane rigidity.

Most enveloped viruses form by a process of budding at the plasma membrane, with little or no participation of other membrane structures in the final steps of maturation. However, there are important exceptions in the case of certain virus groups. The capsids of herpes viruses are assembled in the nucleoplasm and are observed to bud through the inner nuclear membrane, acquiring their envelopes in the process (Roizman and Furlong, 1974; O'Callaghan and Randall, 1976) . They appear to be transported as enveloped particles through cytoplasmic channels to the cell surface. Bunyaviruses (Murphy et aZ., 1973) and coronaviruses (Oshiro, 1973) appear to form primarily by budding into cytoplasmic cisternae, and extracellular virions are observed associated with the plasma membrane. For the rhabdovirus group, several modes of maturation have been reported. VSV, the most widely studied member, usually forms by budding at the cell surface, but maturation at intracellular membranes has been observed. The New Jersey strain of VSV was observed to bud primarily from plasma membranes of L or Vero cells, and almost entirely at intracytoplasmic membranes of pig kidney cells (Zee et al., 1970) .

These reports indicate that the site of maturation of a specific virus type may vary depending on the host cell. Rabies virus, which resembles VSV morphologically and biochemically, is exceptional in that assembly of the virion appears to occur through a process involving de n o w formation of membranes in the cytoplasmic matrix (Hummeler et al., 1967) . De novo formation of membranes is the usual process of assembly for members of the pox virus group (Dales and Mosbach, 1968) .

The appearance of viral proteins on the cell surface, as well as the cellular site of virus assembly, can be modified by external agents including antibodies and lectins, The phenomenon of antigenic modulation involves altering the expression of cell surface antigens by specific antibody, and such antigens may be of viral origin (Lampert et d . , 1975) . This undoubtedly has an effect on viral maturation, although the precise effects have not been determined. Exposure of influenza virus-infected cells to convanavalin A appears to alter the site of virus maturation (Stitz et al., 1977) . In the presence of this lectin, normal maturation at the plasma membrane is not observed, but instead large numbers of virions are observed in intracellular vacuoles.

It is evident therefore that the maturation site for enveloped viruses can vary with the virus type as well as the host cell, and can be altered in response to specific stimuli. The process is likely to be determined as a result of interactions between virus-specific proteins and host cell membranes. I t is remarkable that in most instances only a single type of cellular membrane is selected as the site of virus assembly in a particular virus-infected cell. An understanding of the mechanisms which govern the selection of the assembly site may provide new insights into the assembly of cellular membrane components and organelles, since similar interactions are likely to be involved in determining the location of subsets of cellular proteins in specific cellular organelles.

The synthesis and assembly of viral membrane components have been analyzed in cells in which host cell synthesis is inhibited as a result of virus infection. Similar conclusions have been made for several virus types. Viral glycoproteins appear to be synthesized on membrane-bound polyribosomes and remain associated with various cellular membranes (Spear and Roizman, 1970; Compans, 1973a,b; Stanley et al., 1973; Klenket al., 1974; David, 1973; Hay, 1974; Atkinsonet al., 1976; Nagai et al., 1976 ; Knipe et al., 1977) . Glycosylation occurs in association with cytoplasmic membranes. Glycoproteins appear to migrate from rough endoplasmic reticulum to smooth or Golgi complex membranes to the cell surface and are then incorporated into virions; at no time are they found as ''soluble'' cytoplasmic components. This general scheme for the synthesis and migration of proteins through intracellular membranes to the cell surface has been termed membrane flow. Although the precise intracellular location of viral glycoproteins remains to be established, a scheme may be envisaged in which these components remain associated with the cisternal side of membranes of the endoplasmic reticulum and Golgi complex after synthesis. Incorporation into the plasma membrane by a process of vesicle fusion would then result in the correct orientation on the cell surface for viral as-sembly. A similar scheme has been suggested for the incorporation of glycoproteins into the plasma membrane, based on the distribution of oligosaccharides in different membrane fractions (Hirano et al., 1972) . Viral systems allowed investigators to follow the intracellular migration and glycosylation of specific glycoprotein species for the first time, whereas previous studies with cellular glycoproteins Kad dealt with cell fractions essentially uncharacterized as to the specific proteins present.

Recently, Katz et al. (1977) and Rothman and Lodish (1977) have described a new model system for studies of the incorporation of viral glycoproteins into membranes. Katz et al. (1977) have shown that the mRNA for VSV G-protein translated in vitro by wheat germ extracts in the presence of dog pancreas rough endoplasmic reticulum is associated with the membrane. Furthermore, the newly synthesized G protein was shown to span the membrane with the amino-terminal asymmetrically located. In addition, the polypeptide portion which penetrated the membrane was shown to be glycosylated. Glycosylation did not occur in the absence of membranes. These studies were extended by Rothman and Lodish (1977) , who showed that the insertion of G-protein into the membrane begins when 80 or fewer amino acid residues are polymerized. Evidence was also obtained that the nascent chain is glycosylated while still attached to the ribosomes on the cytoplasmic side of the endoplasmic reticulum vesicle. Since the mechanism by which viral glycoproteins are inserted into membranes is in all probability similar to that of the insertion of host cell glycoproteins, further studies using viral systems may provide more insight into the mechanism of insertion of glycoproteins into membranes.

The process of glycosylation does not appear to play an important role in determining the intracellular migration of glycoproteins. Although this has been suggested as a possible function for carbohydrate components of glycoproteins, the available data using inhibitors of glycosylation suggest that it is possible to inhibit or extensively modify the gl ycosylation process without preventing the migration of viral glycoproteins to the cell surface (Courtney et al., 1973; Nakamura and Compans, 1978a) .

The incorporation of carbohydrate-free M proteins into membranes appears to involve a distinct pathway in which cytoplasmic synthesis is followed by rapid association with the plasma membrane (Lazarowitz et d., 1971; Meier-Ewert and Compans, 1974; Hay, 1974; Nagai et al., 1976; Knipe et al., 1977) . No evidence for migration through cytoplasmic membrane structures has been obtained' for these components, aad it has been suggested that they are inserted directly into membranes after synthesis. Cytoplasmic synthesis followed by direct insertion into membranes has also been proposed for some classes of cellular membrane proteins (Lodish and Small, 1975) .

Although there are several unanswered questions concerning the precise steps in assembly even for the best studied enveloped viruses, the available information from electron microscope studies as well as the biochemical approaches described above suggest a scheme like that depicted in Fig. 2 for influenza virus. Glycoproteins appear to be inserted into the plasma membrane as the first step in assembly, after migration through cytoplasmic membranes. After they are inserted into the plasma membrane, initial random distribution may occur in which the proteins are free to undergo lateral diffusion in the plane of the membrane. Such random distribution is illustrated for the glycoproteins of parainfluenza virus in Fig. 3 , and similar observations have been reported for other viruses (Birdwell and Strauss, 1974; H. Schwarz et al., 1976) . Random distribution of viral antigens (Fig. 3 ) is observed only when ferritin-antibody conjugates are applied to cells after glutaraldehyde fixation. In previous studies of the distribution of antigen, in which unfixed cells were used ), antigens were observed in discrete patches. It is likely that under these conditions they undergo lateral redistribution and aggregation into patches as a result of bivalent antibody.

Schematic diagram of the assembly process of an influenza virion. Clycoproteins are thought to be inserted into the plasma membrane by a process called memhrane flow and are initially found randomly dispersed in the membrane. Following in--sertion of the M protein, glycoproteins are thought to accumulate in discrete regions from which host cell membrane proteins are excluded. The association of the ribonucleoprotein (RNP) with such regions of modified membrane is followed by budding and release of the completed virion. In the case of influenza virus, the M protein may form a domain on the internal surface of the plasma membrane, stabilized by proteinprotein interactions. Specific recognition of the M protein by the glycoproteins could then produce an accumulation of glycoproteins in a circumscribed region of the cell surface. Alternatively, it is possible that M protein monomers associate with glycoprotein monomers at the plasma membrane, and that these complexes undergo lateral diffusion and aggregation into domains. In either case the lack of host cell membrane proteins in the viral envelope indicates that cellular proteins are efficiently excluded from the region of the plasma membrane which becomes the viral envelope.

Association of the nucleocapsid with regions of the cell surface containing viral envelope protein may stimulate the process of budding. Fig. 4 . The presence of virus-specific glycoproteins on the external surface is indicated by tagging with ferritin-conjugated antibody. In Fig. 5 , the emerging virus particles tagged with ferritin-antibody are shown at higher magnifica- For viruses with icosahedral nucleocapsids and no M protein, a similar mechanism for virus assembly has been proposed in which the nucleocapsid itself binds to glycoproteincontaining membranes, with subsequent lateral diffusion and clustering of the glycoproteins in this region.

Although information is accumulating about the pathways by which viral membrane proteins are incorporated into plasma membranes, there is little direct evidence concerning the precise interactions which lead to the formation of domains on the cell surface which contain virus-specific proteins and lack host cell proteins. The lack of significant amounts of host cell protein in the virion indicates that such domains must he intermediates in assembly. Further, the available data on viral protein composition (Table 11 ) and morphology suggest that assembly interactions may differ for various virus groups. In all cases it is likely that protein-protein interaction serves to create a patch of virus-specific proteins in a cellular membrane, but this interaction may occur on the external or internal surface of the lipid bilayer. As discussed above, for viruses that contain M proteins or welldefined icosahedral nucleocapsids, some evidence has been obtained for the initial random distribution of glycoproteins on the cell surface, which may be followed by lateral diffusion and accumulation of glycoproteins in juxtaposition to the M protein or icosahedral nucleocapsid, with transmembrane interactions between the external and internal proteins serving to anchor the viral glycoproteins in place. With other virus types that contain a large amount of glycoprotein and no obvious M protein or well-defined icosahedral capsid, it is possible that lateral interactions between the glycoproteins are important in assembly.

The glycoproteins of togaviruses and bunyaviruses have been observed in a regular surface arrangement (von Bonsdorff and Harrison, 1975; von Bonsdorff and Pettersson, 1975) . The latter viruses lack an M protein or an icosahedral internal component, and it has been suggested that direct interactions between glycoproteins may be involved in assembly and maintenance of the viral structure (von Bonsdorff and Pettersson, 1975) . Similar interactions may be important in other virus groups in which glycoproteins are major protein constituents of the virion and there is no known internal membrane protein, such as B-type oncomaviruses and coronaviruses.

Pox viruses are unique in that formation of their lipid-containing membrane occurs de novo in the cytoplasm, rather than on a preexisting membrane structure. These viruses thus provide an unusual system for investigation of the molecular interactions involved in the formation of a highly organized structure within the cytoplasmic matrix (Dales and Mosbach, 1968 ). However, these viruses are structurally very complex, and limited information has been obtained on their molecular organization. Shape and size determination may also be controlled by viral proteins at various levels; nucleocapsids, M proteins, or glycoproteins could be the determining factor for different virus groups. The phenomenon of phenotypic mixing of envelope glycoproteins in cells doubly infected with VSV and the parainfluenza virus SV5 clearly demonstrates that the internal proteins and not the glycoproteins determine the particle size and shape of rhabdoviruses McSharry et al., 1971 ). This conclusion is supported by the observation that particles with the internal proteins of VSV possess the characteristic bullet shape of VSV while containing mixtures of envelope glycoproteins derived from VSV and SV5. A similar analysis of phenotypically mixed particles produced by dual infections with viruses of other major groups may provide further insights into the macromolecular interactions involved in virion assembly.

The fixed shape and size of many lipid-containing viruses stands in contrast to the situation in membranous cellular organelles, which generally exhibit pleomorphism. Similar pleomorphism is observed in some enveloped viruses. The assembly of such membranes of organelles and pleomorphic viruses may be regulated more by the production of materials than by precise constraints imposed by macromolecular interactions.

Morphology and entry of enveloped and deenveloped equine abortion (herpes) virus

The amino acid and carbohydrate composition of the neuraminidase of B/Lee/40 influenza virus

Assembly of vesicular stomatitis virus glycoprotein and matrix protein into HeLa cell plasma membranes

Morphogenesis of influenza A virus in Ehrlich ascites tumor cells as revealed by thin-sectioning and freeze-etching

Enveloped viruses as model membrane systems: Uicroviscosity of vesicular stomatitis virus and host cell membranes

Undisturbed release of influenza virus in the presence of univalent antineuraminidase antibodies

Differential action ofAspergillus glycosidase activities of influenza and Newcastle disease viruses

Replication of Sindbis virus. IV. Electron microscopic study of the insertion of viral glycoproteins into the surface of infected cells

Dissociation of vesicular stomatitis virus and relation of the virion proteins to the viral transcriptase

Restitution of infectivity to spikeless vesicular stomatitis virus by solubilized viral components

Theoretical aspects of structure and assembly of viral envelopes

Structural components of RNA tumor viruses

Crystalline antigen from the influenza virus envelope

Biological parameters of fusion from without

Lipid and protein organization in vesicular stomatitis and Sindbis viruses

A,(N,) neuraminidase of the X-7 influenza virus recombinant: Determination of molecular size and subunit composition of the active unit

The biologically active proteins of influenza virus: Neurarninidase. I n "The Influenza Viruses and Influenza

Comparison of membrane protein glycopeptides of Sindbis virus and vesicular stomatitis virus

Model for vesicular stomatitis virus

Phenotypic mixing of envelope proteins of the parainfluenza virus SV5 and vesicular stomatitis virus

Reproduction of paramyxoviruses. I n "Comprehensive Virology

Studies on proteolytic cleavage and glycosylation of the hemagglutinin of influenza A and B viruses. I n "Negative Strand Viruses

Influenza virus proteins. 11. Association with components of the cytoplasm

Distinct carbohydrate components of influenza virus glycoproteins in smooth and rough cytoplasmic membranes

The structure and assembly of influenza and parainfluenza viruses

Reproduction of myxoviruses. I n "Comprehensive Virology

Incorporation of sulfate into influenza virus glycoproteins

An electron microscopic study of moderate and virulent virus-cell interactions of the parainfluenza virus SV5. Virolog!j 30

Effect of antibody to neuraminidase on the maturation and hemagglutinating activity of an influenza A2 virus

Influenza virus proteins. I. Analysis of polypeptides of the virion and identification of spike glycoproteins

Structure of the ribonucleoprotein of influenza virus

Assembly of lipid-containing viruses

Effects of 2-deoxy-Dglucose on herpes simplex virus replication

Vaccinia as a model for membrane biogenesis

Antigenic characterization of two Sindbis envelope glycoproteins separated by isoelectric focusing

Influenza virus

Assembly of the vesicular stomatitis virus envelope: Incorporation of viral polypeptides into the host plasma membrane

Membrane receptors for murine leukemia viruses: Characterization using the purified viral envelope glycoprotein, gp71

Characteristics of two new avian tumor virus subgroups

Carbohydrate composition of the membrane glycoprotein of vesicular stomatitis virus grown in mammalian cell lines

Carbohydrate composition of the membrane glycoprotein of vesicular stomatitis virus

Partial structural analysis of the oligosaccharide moieties of the vesicular stomatitis virus glycoprotein by sequential chemical and enzymatic degradation

Epidemiological concepts and methods

Presence of murine leukemia virus envelope proteins gp70 and p15(E) in a common polyprotein of infected cells

Neutralization of homologous and heterologous oncornaviruses by antisera against the p 15(E) and gp7l polypeptides of Friend murine leukemia virus

The membrane proteins of Semliki Forest virus have a hydrophobic part attached to the viral membrane

Structural proteins of Tacaribe and Tamiami virions

Cross-linking of the spike glycoproteins in Semliki Forest virus with dimethyl suberimidate

Location of the spike glycoproteins in the Semliki Forest virus membrane

Isolation and characterization of the membrane proteins of Semliki Forest virus

Determining influence of helper virus on the host range and susceptibility to interference of RSV

Lipid and protein organization in Sindbis virus

The polypeptides of influenza virus. 11. Interpretation ofpolyacrylamide gel electrophoresis patterns

Studies on the formation of the influenza virus envelope

Characteristics of Sendai virus receptors in a model membrane

Model membranes and Sendai virus: Surface-surface interactions. I n "Negative Strand Viruses

Proteins specified by herpes simplex virus. XII. The virion polypeptides of type 1 strains

Protein composition of coronavirus OC-43

Distribution of saccharide residues on membrane fragments from a myeloma-cell homogenate: Its implications for membrane biogenesis

Trypsin action on the growth of Sendai virus in tissue culture cells. 111. Structural difference of Sendai viruses grown in eggs and tissue culture cells

Regulation of herpes virus macromolecular synthesis. I. Cascade regulation of the synthesis of three groups of viral proteins

Structure and development of rabies virus in tissue culture

Glycosylation of vesicular stomatitis virus glycoprotein in virus-infected HeLa cells

Inhibition of avian sarcoma virus replication by glucosamine

Immunological relationships among envelope antigens of avian tumor viruses

The oligosaccharide units of human type L immunoglobulin M (macroglobulin)

Interaction of Sindbis virus glycoproteins during morphogenesis

Inhibition of the multiplication of enveloped RNA-viruses by glucosamine and 2-deoxy-D-glucose

Synthesis of'proteins in cells infected with herpes virus. XI. Sulfated structural proteins

Membrane assembly In Vitro: Synthesis, glycosylation, and asymmetric insertion of a transmembrane protein

Isolation of defective mutant of avian sarcoma virus

Sindbis virus glycoproteins: Effect of the host cell on the oligosaccharides

The effects of enzymes on structural and biological properties of Semliki Forest virus

Inhibition of influenza virus multiplication with a glucose antimetabolite (2-deoxy-D-glucose)

Viral envelopes and their relationship to cellular membranes

Lipids of plasma membranes of monkey and hamster kidney cells and of parainfluenza virions grown in these cells

Glycosphingolipids of plasma membranes of cultured cells and an enveloped virus (SV5) grown in these cells

The proteins of the parainfluenza virus SV5 11. The carbohydrate content and glycoproteins of the virion

An electron microscope study of the presence or absence of neuraminic acid in enveloped viruses

On the structure of the influenza virus envelope

Inhibition of glycoprotein biosynthesis of influenza virus by D-glucosamine and 2-deoxy-~-glucose

Association of influenza virus proteins with cytoplasmic fractions

Activation of influenza A viruses by trypsin treatment

A sedimentable component of allantoic fluid and.its relationship to influenza viruses

Separate pathways of maturation ofthe major structural proteins of vesicular stomatitis virus

The interaction between virus and antibody. HI. Examination of virus-antibody complexes with the electron microscope

The polypeptides and RNA of Sendai virus

Antibody-induced capping of measles virus antigens on plasma membrane studied by electron microscopy

Effect of membrane protein on the lipid bilayer structure: A spin label ESR study of vesicular stomatitis virus

Spin label ESR study of the lipid-containing membrane of influenza virus

Structure of the lipid phase of Rauscher murine leukemia virus

Organization of the lipid phase in viral membranes: Effects of independent variation of the lipid and the protein composition

Separation of two polypeptide chains from the hemagglutinin subunit of influenza virus

Amino acid composition of polypeptides from influenza virus particles

Morphology of the isolated hemagglutinin and neuraminidase subunits of influenza virus

Studies on the origin of pandemic influenza. 11. Peptide maps of the light and heavy polypeptide chains from the hemagglutinin subunits of A, influenza viruses isolated before and after the appearance of Hong Kong influenza

Enhancement of the infectivity of influenza A and B viruses by proteolytic cleavage of the hemagglutinin polypeptide

Influenza virus structural and nonstructural proteins in infected cells and their plasma membranes

Proteolytic cleavage of the hemagglutinin polypeptide of influenza virus. Function of the uncleaved polypeptide HA

The polypeptide of influenza virus

Composition of the neuraminidase

Tunicamycin inhibits glycosylation and multiplication of Sindbis and vesicular stomatitis viruses

The membrane structure of lipid-containing viruses

Organization of the membrane of standard and incomplete influenza virus

Membrane proteins synthesized by rabbit reticulocytes

Carbohydrate composition of vesicular stomatitis virus

Proteins ofvesicular stomatitis virus and of phenotypically mixed vesicular stomatitis virus-simian virus 5 virions

Transmembrane phospholipid motions induced by F glycoprotein in hemagglutinating virus of Japan

Protein-bound oligosaccharides of Semliki Forest virus

Time course of synthesis and assembly of influenza virus proteins

Microviscosity of togavirus membranes studied by fluorescence depolarization: Influence of envelope proteins and the host cell

Vesicular stomatitis virus envelope glycoprotein a]-terations induced b y host cell transformation

Oligosaccharide moieties of the glycoprotein of Vesicular Stomatitis virus

Glycoprotein fragment associated with vesicular stomatitis virus after proteolytic digestion

Bunyaviridae: Morphologic and morphogenetic similarities of Bunyamwera serologic supergroup viruses and several other arthropod-borne viruses

Activation of precursor to both glycoproteins of Newcastle disease virus by proteolytic cleavage

Studies on the assembly of the envelope of Newcastle disease virus

Proteolytic cleavage of the viral glycoproteins and its significance for the virulence of Newcastle disease virus

Effects of inhibitors on glycosylation sulfation, and assembly of influenza virus glycoproteins

Glycopeptide components of influenza viral glycoproteins

Host cell dependent glycosylation of influenza virus glycoproteins

Primary structure and side chain interactions of PF, filamentous bacterial virus coat protein

The structural proteins of La Crosse virus

Molecular anatomy of herpes viruses: Recent studies

The fusion of Ehrlich's tumor cells caused by HVJ virus in uitro

Ultrastructure of Animal Viruses and Bacteriophages

Inhibition of influenza virus replication in tissue culture by 2-deoxy-2,3-dehydro-N-trifluoracetylneuraminic acid. (FANA): Mechanism of action

Isolation and characterization of influenza virus recombinants with high and low neuraminidase activity: Use of 2-(3'-methoxy-pheny1)-N-acetylneuraminic acid to identify cloned populations

Characterization of temperature sensitive influenza virus mutants defective in neuraminidase

Sulfated glycoproteins and polysaccharides of enveloped viruses

Phospholipid composition of Rous sarcoma virus, host cell membranes, and other enveloped viruses

The replication of herpes viruses

Transbilayer phospholipid asymmetry and its maintenance in the membrane of influenza virus

Role of carbohydrate in biological functions of Friend murine leukemia virus gp7lJ

Isolation and purification of the envelope proteins of Newcastle disease virus

Identification of biological activity of paramyxovirus glycoproteins: Activation of cell fusion, hemolysis and infectivity by proteolytic cleavage of an inactive precursor protein of Sendai virus

Isolation of paramyxovirus glycoproteins: Association of both hemagglutinating and neuraminidase activities with the larger SV5 glycoprotein

Identification of a second glycoprotein in Sindbis virus

Grbwth of enveloped RNA viruses in a line of Chinese hamster ovary cells with deficientN-acetylglucosaminyl transferase activity

Sialoglycoprotein of vesicular stomatitis virus: Role of the neuraminic acid in infecti0n

Mosquito cells infected with VSV yield unsialylated virions of low infectivity

Association of vesicular stomatitis virus glycoproteins with virion membrane: Characterization of the lipophilic tail fragment

Detection of an unstable RNA in chick fibroblasts after reduction of the UTP pool by glucosamine

The struchire of influenza virus. I. The polypeptides of the virion

The biologically active proteins of influenza virus: The hemagglutinin

Properties of mouse leukemia virus. XI. Immunoelectron microscopic studies on viral antigens on the cell surface

Inhibition of glycosylation of influenza virus hemagg1utinin

Carbohydrates of influenza virus I. Glycopeptides derived from viral glycoproteins after labeling with radioactive sugars

Suppression of glycoprotein formation of Semliki Forest, influenza, and avian sarcoma virus by tunicamycin

Effect of the viral proteins on the fluidity of the membrane lipids in Sindbis virus

Red cell membrane glycoprotein: Amino acid sequence of an intramembranous region

Functional significance of sialidase during influenza virus multiplication

Isolation and characterization of two distinct types of HVJ (Sendai virus) spikes

The fluid mosaic model of the structure of cell membranes

Studies on the primary structure of the influenza virus hemagglutinin. Proc. N a t l

The proteins specified by herpes simplex virus. IV. The site of glycosylation and accumulation of viral membrane proteins

The polypeptides of influenza virus. VII. Synthesis of the hemagglutinin

Studies on the inhibitory effect of lectins on myxovirus re1ease

SC nuclear magnetic resonance studies on the lipid organization of enveloped virions (vesicular stomatitis virus)

Sialic acid content of Sindbis virus from vertebrate and mosquito cells

Structural proteins of ribonucleic acid tumor viruses

Carbohydrate content of membrane protein of Sindbis virus

Tunicamycin inhibition of polyisoprenyl N-acetylglucosaminyl pyrophosphate formation in calf liver microsomes

Strain-specific antigen of the avian leukosis sarcoma virus group

Asymmetry of influenza virus membrane bilayer demonstrated with phospholipase C

Studies on the amphipathic nature of the membrane proteins in Semliki Forest virus

Structural components of the arenavirus Pichinde

A heterogeneity of Rous sarcoma virus revealed by selectively resistant chick embryo cells

Patterns of viral interference in the avian leukosis and sarcoma complex

Sindbis virus glycoproteins form a regular surface lattice

Surface structure of Uukuniemi virus

Reproduction of rhabdoviruses

Estimation of the molecular weights of the polypeptide chains from the isolated hemagglutinin and neuraminidase subunits of influenza viruses

Studies on the origin of pandemic influenza. I. Antigenic analysis of A2 influenza viruses isolated before and after the appearance of Hong Kong influenza using antisera to the isolated hemagglutinin subunits

Antigenic variation of influenza viruses

Influenza viral proteins: Identification and synthesis

Evidence from studies with a cross-linking reagent that the haemagglutinin of influenza virus is a trimer

Structural components of mouse mammary

Vesicular stomatitis virus maturation og!/ 79,446-448. shape of influenza virus neuraminidase

Research by the authors was supported by Grants No. A1 12680 and CA 18611 from the USPHS, PCM76-09711 from the National Science Foundation, and VC 149B from the American Cancer Society. M.C.K. was supported b y a fellowship from the Anna Fuller Fund.