key: cord-0038493-5m7s2a62 authors: Mollier, P.; Lagnel, J.; Quiot, J.M.; Aloun, A.; Riba, G. title: Cytotoxic activity in culture filtrates from the entomopathogenic fungus Beauveria sulfurescens date: 2005-01-13 journal: J Invertebr Pathol DOI: 10.1016/s0022-2011(94)90200-3 sha: d89693593042ec4bfe906dd08eee4c630705f3fe doc_id: 38493 cord_uid: 5m7s2a62 The hyphomycete fungus Beauveria sulfurescens secretes into the culture medium several components toxic for Galleria mellonella larvae. Concentrated and dialyzed (cutoff 12,000) B. sulfurescens culture filtrate also contains a highmolecular-weight compound cytotoxic for the Mamestra brassicae cell line. A fractionation in two chromatographic steps separated a cytotoxic fraction (IC(50)= 1.15 ± 0.05 μg/ ml), which was poorly toxic to larvae (LC(50) = 150 ± 10 μg/ ml) and an insecticidal fraction (LC(50) = 2.1 ± 0.1 μg/ml) devoid of detectable cytotoxicity. The two kinds of activity, i.e., cytotoxicity in vitroand toxicity in vivo are thus independent. The cytotoxic activity, initially detected on a M. brassicae cell line, also affected two other insect cell lines (G. mellonella and Spodoptera frugiperda). The cytotoxic fraction provoked not only the inhibition of cell proliferation, but also cell death. Adhesive mammalian cell lines were not sensitive. Nonadhesive mammalian cell lines of murine myeloma and hybridomas were poorly sensitive. tein of oligomannosidic type which is highly toxic for Galleria mellonella (Mollier et al., submitted for publication) . This work shows that concentrated and dialyzed (cutoff 12,000) culture filtrates of B. sulfurescens possess a cytotoxic activity against a cell line from the insect Mamestra brassicae. Dialysis eliminates components of molecular weight theoritically below 12,000, possibly including toxic cyclic peptides analogous to those described for the other Beauveria genus strains (mean M~ = 600) (Hamil et al., 1969; Suzuki et al., 1977) . The observed cytotoxicity may thus be due to another type(s) of molecule(s) of higher molecular weight. We tested this cytotoxic activity on cell lines from both insects and mammals. We also investigated whether the cytotoxicity was correlated to the in vivo insecticidal activity. Chromatographic fractionation indicated that the in vivo and in vitro activities were unrelated. INTRODUCTION Entomopathogenic fungi produce several insecticidal toxins and are thus of interest as potential biological control agents. These toxins can be classified in two groups: (i) low-molecular-weight compounds including cyclodepsipeptides (cyclic peptides of 6 or 7 amino acids) and (ii) higher-molecular-weight protein molecules . A number of cyclodepsipeptides have been purified and extensively studied (Lysenko and Kucera, 1971; Roberts, 1981; Vey et al., 1986; Kleinhauf and von DShren, 1986) . Some of them were found to exhibit cytotoxic effects on insect cell lines (Quiot et al., 1985) . The high-molecular-weight protein toxins were initially described as proteases in Metarhizium anisopliae and Beauveria bassiana (Kucera and Samsinakova, 1968; Kucera, 1980) but have not been purified or further characterized. Recently, we described a high-molecular-weight glycopro- The strain used in these experiments (from the La Mini~re culture collection) was originally isolated from soil in France and designated as B. sulfurescens (ATCC 7159). The strain was maintained in 90-mm petri plates at 25°C on complete agar medium (Ravallec et al., 1989) . The fungus was grown as a submerged culture in a M2M medium containing 5 g/liter casein hydrolysate N3 PEM, 30 g/liter glucose, 4 drops/liter Berthelot solution, L-amino acids (in g/liter: alanine 0.13, arginine 0.18, asparagine 0.2, aspartic acid 0.1, glutamine 2.0, glycine 0.16, histidine 0.8, lysine 0.8, proline 2.0), vitamins (in mg/liter: p-aminobenzoic acid 2, biotin 0.02, calcium panthotenate 20, nicotinic acid 4, pyridoxin 4, riboflavin 2, thiamin 4), salts (in g/liter: KC1 0.8, NaC1 0.1, CaCI 2 • 2H20 0.1, MgSO 4 • 7H20 0.5, KH2PO4 1). An inoculum was prepared by washing the conidia on the surface of a sporulating aerial myce-lial mat with M2M and collecting the liquid. This conidia suspension (5 x 106 total conidia per milliliter of M2M) was cultured with shaking (80 rpm) on a reciprocal shaker for 4 days at 25°C. Fifty-milliliter aliquots of this culture were used to inoculate 1 liter of M2M in 5-liter Erlenmeyer flasks. Subcultures were agitated for 3 days at 25°C. The protocol for fractionation of toxic activity is shown in Fig. 1 . Step 1. Extraction and precipitation with ammonium sulfate. Fungal cultures were centrifuged at 1500g for 15 min and the supernatant (containing the toxic activity) was collected. The supernatant was filtered through No. 2 Whatman paper and then through 1.2-, 0.6-, and 0.22-~m pore size filters (Millipore, Bedford, MA). Crystalline ammonium sulfate was added to the culture filtrate to 90% saturation (600 g/liter) and the sample was incubated at 4°C for 1 hr, then centrifuged at 7200g for 30 min. The pellet.was resuspended in a minimal volume of distilled water, dialyzed (membrane cutoff 12,000) against three changes of water at 4°C for 16 hr, and sterilized by filtration. Step 2. Precipitation with ethanol. Ethanol (precooled to -80°C) was added to the ammonium sulfateconcentrated filtrate (ethanol: 4.5, filtrate: 1, v/v) under continuous magnetic stirring. The mixture was allowed to stand for 30 min in ice to equilibrate and then centrifuged at 7200g for 15 min in a precooled centrifuge. The protein pellet was then vacuum-dried in a Speed-Vac apparatus (Savant, Farmingdale, NY), resuspended in a minimal volume of distilled water, and dialyzed (membrane cutoff 12,000) against three changes of water at 4°C for 16 hr. Step 3. Affinity chromatography on immobilized lectin: concanavalin A. Following ethanol precipitation, the concentrated sample was made up to a final concentration of 0.02 M phosphate buffer-0.2 M NaCl-1 mM CaC12, pH 7, sterilized by filtration, and loaded on a concanavalin A (ConA)-Ultrogel (IBF, Villeneuve-la-Garenne, France) column (2.6 x 7 cm) equilibrated with the above buffer. The ConA-unreactive form was removed by washing the column with the equilibration buffer. The ConA-reactive form was eluted stepwise with the above buffer containing 10 mM methyl a-Dglucopyranoside and then 250 mM methyl a-D-mannopyranoside. The flow rate was 0.24 mYmin during sample loading and 0.8 ml/min during the washing and elution steps. Fractions of 10 ml were collected. Step 4. Hydroxyapatite chromatography. The fraction eluted from Con A with methyl ~-D-mannopyranoside was dialyzed at 4°C against three changes of water and then made up to 1 mM NaC1 and applied to a hydroxyapatite column (BDH Chemicals, Ltd., Poole, UK) (1.2 x 13 cm), previously equilibrated with 1 mM NaC1. The column was washed with 1 mM NaC1 and eluted stepwise with 1 mM MgC12 and 1 M MgC12. Then the column was washed with 1 mM sodium phosphate, pH 7, and an elution gradient (1 to 150 mM sodium phosphate, pH 7, in 25 ml) was applied. The flow rate was 1 ml/min and the fraction size was 5 ml. The median lethal concentration (LCso) of the B. sulfurescens crude supernatant and chromatographic fractions were calculated by the graphic method of probit analysis (Finney, 1952) . Each fraction tested was first dialyzed against three changes of water, vacuumdried in a Speed-Vac apparatus, and resuspended in sterile water at the convenient concentration. Several dilutions were made. The dilutions were sterilized by filtration and injected (20 ~l) into the hemocoel of G. melloneUa sixth-instar larvae (20 larvae per dose, mean wt. = 205 -+ 10 mg). The LCso values were scored 7 days after the injection. Protein concentrations in samples were determined by the Bradford method (Bradford, 1976) . The M. brassicae (SPCMb92) and G. melloneUa' (SPCGm17) cell lines were obtained from Dr. J.M. Quiot. SPCMb92, SPCGm17, and the Spodoptera frugiperda line (Sf9, ATCC CRL 1711) were grown in Hink's medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal bovine serum (Gibco) at 26°C. The porcine fetal testis cell line ST (ATCC CRL 1746) and the dog kidney epithelial cell line MDCK (Misfeldt et al., 1976) were grown in MEM (Gibco) supplemented with 10% fetal bovine serum, at 37°C. The human carcinoma cell line CaCO2 (Fogh et al., 1977) was grown at 37°C in DMEM supplemented with 2 mM glutamine, 1% nonessential amino acid (Gibco) and 20% fetal bovine serum. The myeloma cell line SP20 (Shulman et al., 1978) , and the hybridoma G43 (Delmas et al., 1992) and 70K5Cll (produced using Barley yellow mosaic virus, obtained from T. Delaunay, Jouy-en-Josas) were grown at 37°C in RPMI 1640 supplemented with 2 mM glutamine, 2 mM sodium pyruvate, 3 ~M 2-mercaptoethanol (for hybridomas), and 10% fetal bovine serum. Each fraction to be tested was dialyzed at 4°C againgt three changes of water for 16 hr, vacuumdried, resuspended at various concentrations in the appropriate cell media, and sterilized by filtration. Cells were plated in appropriate media in 96-well plates at a density of 10S/ml (100 ixl per well). After a resting time of 1 hr (SPCMb92, SPCGm17, sfg, SP20, G43, and 70K5Cll), 24 hr (ST, MDCK), or 72 hr (CaC02), the medium wa~ removed and replaced with the medium containing the fractions to be tested for toxicity (3 wells per concentration, with 12 untreated cell control wells). The cells were incubated for a further 72 hr and total cellular protein content was determined as described by Knox et al. (1986) . SPCMb92 cells were treated as described for the KBP assay. Then, the MTT method was applied as described by Hansen et al. (1989) . Cytotoxicity tests were carried out in triplicate for each concentration. The results for tested samples are expressed as percentages of untreated control cells value -+ SEM. The ICso values (i.e., the concentration of sample that reduced total cellular protein to 50% of that of untreated cell control wells) are the mean -+ SEM of the values from at least two independent experiments. Insecticidal Activity The insecticidal activity in B. sulfurescens crude filtrate was fractionated by a four-step fractionation procedure: precipitation of most proteins with 90% ammonium sulfate, ethanol precipitation (4.5 v/v), affinity chromatography on immobilized ConA, and hydroxyapatite chromatography (Table 1) . Chromatography on ConA gave three fractions: (i) a ConA-unreactive fraction (Fraction A), (ii) a ConAreactive fraction, eluted with a low concentration (10 mM) of methyl a-D-glucopyranoside (Fraction B), and (iii) a ConA-reactive fraction, eluted with a high concentration (250 mM) of methyl a-D-mannopyranoside (Fraction C). These fractions contained 38.5, 5.4, and 4.2% of the applied protein, respectively. Fraction C was the most toxic for insects (cf. Table 2 ). This fraction contained 3% of the ammonium sulfate-precipitated protein. Hydroxyapatite chromatography also gave three fractions: (i) a non-retained fraction (Fraction D), (ii) a fraction retained and eluted by 1 M MgC12 (Fraction E), and (iii) a fraction retained and eluted by a phosphate gradient (Fraction F). The insecticidal activity was recovered in fraction F, which contained 7.4% of the protein applied to the column and 0.2% of the ammonium sulfate-precipitated protein. a LCso is the median lethal concentration. b The ICso value is the concentration of the tested fraction that reduces the final cellular protein content to 50% of that of control well. Each partially purified fraction was tested for toxicity on larvae and for cytotoxicity on the M. brassicae (SPCMb92) cell line. The toxicity for insect larvae is expressed as LCso, scored 7 d~ys after injection. The cytotoxicity was quantified by the KBP method (Knox et al., 1986) . This method evaluates cell proliferation by measuring total cellular protein after 72 hr, and the results are expressed as the ICso. For each tested fraction, the ratio of the LCso to the ICso was then calculated. The results are shown in Table 2 . The LCso of the starting fraction (B. sulfurescens culture filtrate concentrated with ammonium sulfate and ethanol) was 40 -+ 10 ~g/ml and the ICso was 4.2 +-0.1 ~tg/ml. The ratio of LCso to ICso was thus 9.5. The LCso and ICso of each of the fractions from ConA chromatography were determined. Fraction A had an LCso of 150 -+ 10 ~g/ml and an ICso of 1.15 -+ 0.05 ~Lg/ml. For this fraction, the insecticidal activity was 3.75-fold lower, whereas the cytotoxicity was 3.6-fold higher, than that of the starting fraction. The ratio of LCso to ICso for fraction A was 116.6. Fraction B was not toxic either to larvae or to cells, even when concentrated up to 225 ~g/ml. Fraction C had an LCso 2.1-fold lower and an ICso 4.75-fold higher than that of the starting fraction. Thus, Fraction C was enriched in insecticidal activity but was less cytotoxic, compared to the starting fraction. The ratio of LCso to ICso for Fraction C was i. Fraction C was further fractionated by hydroxyapatite chromatography. The fraction toxic for larvae obtained from this step (Fraction F) had an LCso of 2.1 +-0.1 ~g/ml. Fraction F displayed no detectable cytotoxic activity at concentrations up to 4.2 ~Lg/ml. The ratio of LCso to ICso was thus below 0.5. The cytotoxicity of Fraction A for various cell lines, including insect lines and adhesive and nonadhesive mammalian lines, was assayed by the KBP method. Fraction A was cytotoxic to the insect cell lines (Fig. 2) with ICso values of 0.70 +-0.05, 1.15 -+ 0.05, and 36 -+-2 ~g/ml for cell lines from G. meUonella (SPCGml7), M. brassicae (SPCMb92), and S. frugiperda (Sf9), respectively. In contrast, 108 ~g/ml of Fraction A was not cytotoxic for adhesive mammals cell lines from porcine fetal testis (ST), dog kidney (MDCK), or human carcinoma (CaCO2) (not shown). It weakly inhibited cell proliferation of the nonadhesive mammal cell lines SP20 (murine myeloma) and two murine hybridomas (Fig. 3) . Even at concentrations of Fraction A that produced the maximal cytotoxic effect, the cellular protein was not reduced to zero (Fig. 2) . This could be explained either by the presence of cellular debris, which was not eliminated by washing, or by the presence of remaining living cells. The MTT method is based on the reduction of the soluble yellow tetrazolium salt to a blue insoluble formazan product by mitochondrial succinic dehydrogenase and thus indicates both cell proliferation and cell survival. In the presence of 144 ~g/ml of Fraction A, SPCMb92 cells did not produce any formazan (Fig. 4) . The samples thus contained no living cells. In this work, B. sulfurescens was cultivated on M2M medium. The composition of M2M was empirically defined to mimic the composition of G. mellonella hemocoel (Grenier et al., 1974) . It has been shown that the secretion of toxins is highest in the presence of complex organic nitrogen sources such as ground maize, skim milk, yeast extract, or G. meUonella homogenate (Kucera, 1971 (Kucera, , 1981 . In a previous study, we cultivated the fungal strain on a medium containing yeast extract (MY20) (Mollier et al., submitted for publication). Because of the high viscosity of MY20, sterile filtration was difficult. Moreover, MY20 contained pigments that affected the chromatography columns used for fractionation. We therefore used MeM in this study. Affinity chromatography on ConA and hydroxyapatite chromatography were used to fractionate the insecticidal activity in culture filtrate concentrated with ammonium, sulfate and ethanol. As previously observed (Mollier et al., submitted for publication) the ConA-reactive, methyl ~-D-mannopyranoside-eluted fraction (Fraction C) displayed the highest insecticidal activity. This activity was further purified by hydroxyapatite chromatography. The procedure included two elution steps: with 1 M MgC12, to elute proteins with isoelectric points theoretically ranging from 5 to 8, and then with a phosphate gradient to elute acidic proteins (Gorbunoff, 1990) . The insecticidal activity was eluted from the hydroxyapatite column by the phosphate gradient, indicating that the toxin has an acidic isoelectric point. The insecticidal fraction finally obtained from successive ConA and hydroxyapatite chromatographies (Fraction F) was 0.2% of the starting protein and possessed an LCso value of 2.1 -+ 0.1 ~Lg/ml ( Table 2) . The concentrated and dialyzed culture filtrate of B. sulfurescens was shown to be cytotoxic for a cell line from M. brassicae ovary (SPCMb92). In order to determine if this cytotoxicity was due to the same molecule(s) as the in vivo insecticidal activity, each partially purified fraction was tested for cytotoxicity against the SPCMb92 cell line. This line was highly sensitive to B. sulfurescens concentrated filtrate and easy to cultivate. The KBP assay used (Knox et al., 1986) measures cell proliferation. The rationale for this choice is that cells in continuous culture proliferate at a known rate, and this rate would be reduced by products that affect one or more essential cytological functions (DNA or protein synthesis, or mitochondrial or membrane properties). The ratio of LCso to ICso of the insecticidal fraction decreased with each step of the fractionation. Thus, the fractions enriched in larval toxicity were less cytotoxic. At the end of the fractionation procedure, the insecticidal fraction (Fraction F) displayed no detectable cytotoxicity at the tested concentrations (Table 2 ). In the previous study (Mollier et al., submitted for publication), the insecticidal activity was fractionated by a different procedure: culture filtrate concentrated by tangential ultrafiltration was submitted to anionexchange chromatography and ConA chromatography. The insecticidal fraction obtained possessed no detectable cytotoxicity at a concentration equal to 17-fold the LC5o (not shown). These results indicate that the in vivo toxicity and the in vitro cytotoxicity are unrelated activities and presumably due to different agents. In consequence, the in vitro assay could not be used to replace the assay using insect larvae to follow the purification of the insecticidal product. The cytotoxic activity of the ConA-unreactive fraction (Fraction A) has not previously been described. It is unlikely to be due to putative cyclodepsipeptides which were eliminated by dialysis of the culture filtrate: the active compound is presumably a highermolecular-weight species. This high-molecular-weight cytotoxic species was not equally active against all cell types. In the same concentration range, Fraction A was cytotoxic for in-sect cell lines (Fig. 2) , but not adhesive mammalian cell lines (not shown). A slight inhibition of nonadhesive mammalian cell proliferation was observed with murine myeloma and hybridomas (Fig. 3) . The KBP assay measures cell proliferation. The tetrazolium assay (MTT) measures both cell proliferation and survival (Mosmann, 1983) . MTT assays showed that Fraction A not only inhibited cell proliferation but also killed cells (Fig. 4) . Cell lysis was confirmed by optical microscopy (results not shown). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding Aminopeptidase N is a major receptor for the enteropathogenic coronavirus TGEV Probit analysis One hundred and twenty seven cultured human tumor cell lines producing tumors in nude mice Persistance des activit~s enzymatiques du corps adipeux de Galleria mellonella L. (Lepidoptera) en milieux d~finis Protein chromatography on hydroxyapatite columns The structure of beauvericin: A new cyclodepsipeptide antibiotic toxic to Artemia salina Re-examination and further development of a precise and rapid dye method for CYTOTOXIC ACTIVITY 213 measuring cell growth/cell kill Peptide Antibiotics The frame multicentre project on in vitro cytotoxicity Toxins of the entomogenous fungus Beauveria bass/ana. II. Effect of nitrogen sources on formation of the toxic protease in submerged culture Proteases from the fungus Metarhizium anisopliae toxic for Galleria mellonella larvae The production of toxic protease by the entomopathogenous fungus Metarhizium anisopliae in submerged culture Toxins of the entomophagous fungus Beauveria bassiana Microorganisms as sources of new insecticidal chemicals: Toxins Transepithelial transport in cell culture Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays Effects of mycotoxins on invertebrate cells in vitro Infection ofAedes albopictus by Tolypocladium cylindrosporum Toxins of entomopathogenic fungi• In "Microbial Control of Pests and Plant Diseases A better cell line for making hybridomas secreting specific antibodies Bassianolide: A new insecticidal cyclodepsipeptide from Beauveria bassiana and VerticiUium lecanii Tox~mie d'origine fongique chez les invert~br~s et ses consequences cytotoxiques: Etude sur l'infection ~ Metarhizium anisopliae (Hyphomyc~te: Moniliales) chez les L~pidopt~res et les Col~opt~res Toxines insecticides issues de champignons entomopathog~nes This research was partially supported by ZENECA Agrochemicals (Jealott's Hill Research Station, Bracknell, Berkshire RG 12 6EY, UK). We are grateful to Drs. J. Grosclaude and T. Delaunay for having suggested cytotoxic assays and offering the appropriate laboratory facilities in VIM (Virologie et Immunologie Mol~culaire) INRA, Jouy-en-Josas, France. We thank Mrs. S. Labiau, L. Abinne-Molza, S. de Vaureix, and A. Charpilienne for their expert technical assistance. We are grateful to Dr. S. Chwetzoff for his critical review of the manuscript and to Dr. A. Vey for his constructive suggestions. We thank Mrs. C. Cormenier for typing the manuscript.