Effects of double-stranded RNA in Metarhiziumanisopliae var. acridum and Paecilomycesfumosoroseus on protease activities, conidiaproduction, and virulence

Patricia Vieira Tiago, Maria Helena Pelegrinelli Fungaro, Marcos Rodrigues de Faria,and Marcia Cristina Furlaneto

Abstract: Isogenic strains (with and without dsRNA) of the entomogenous fungi Metarhizium anisopliae var. acridumand Paecilomyces fumosoroseus were investigated for correlation between the presence of dsRNA and the production ofcuticle-degrading proteases that play an important role in host parasitism, total secreted protein, and conidia production.Similar levels of cuticle-degrading subtilisin-like (Pr1) protease were observed for isogenic strains of M. anisopliae var.acridum after growth in medium supplemented with the cuticle of the grasshopper Rhammatocerus schistocercoides.Similarly, no statistical differences were observed for protease production, detected using the chromogenic substrateazocasein. For P. fumosoroseus isogenic strains, no significant differences in protease activity were observed aftergrowth in the presence of either Euschistus heros or Nezara viridula (Hemiptera: Pentatomidae) cuticle. Similarly, nostatistical differences were observed in virulence against E. heros. A comparison of mean conidia production showed asignificantly higher production in the dsRNA-free isogenic strains of M. anisopliae var. acridum. Although, for most ofthe fungal phenotypes analysed, no overt effects were associated with the presence of these dsRNA infections, the reductionin conidia production by the isogenic strains of M. anisopliae var. acridum with dsRNA suggested that it may not beentirely accurate to describe these infections as latent.

Key words: Metarhizium anisopliae var. acridum, Paecilomyces fumosoroseus, double-stranded RNA (dsRNA), pro-teases, conidia production, virulence.

Résumé : La corrélation entre la présence d’ARNdb et la production de protéases dégradant la cuticule a été examinéechez des souches syngéniques (avec ou sans ARNdb) des champignons entomogènes Metarhizium anisopliae var. acri-dum et Paecilomyces fumosoroseus. Ces protéases jouent un rôle important dans le parasitisme de l’hôte, la productiontotale de protéines sécrétées et la production de conidies. Des taux analogues de la protéase semblable à la subtilisine(Pr1) dégradant la cuticule ont été observés chez les souches syngéniques de M. anisopliae var. acridum suite à uneculture dans un milieu supplémenté avec de la cuticule de la sauterelle Rhammatocerus schistocercoides. De même, au-cune différence statistique significative dans la production de la protéase n’a été observée, en utilisant pour sa détectionl’azocaséine comme substrat chromogène. Dans le cas des souches syngéniques de P. fumosoroseus, aucune différencedans l’activité de la protéase n’a été observée suite à une culture en présence de la cuticule de Euschistus heros ou deNezara viridula (Hemiptera: Pentatomidae). De même, aucune différence statistique significative n’a été observée dansla virulence envers E. heros. Une comparaison de la production moyenne de conidies a démontré une production signi-ficativement plus élevée chez les souches sans ARNdb de M. anisopliae var. acridum. Bien que, pour la plupart desphénotypes analysés, aucun effet néfaste n’a été associé avec la présence de ces infections d’ARNdb, la diminution dela production de conidies par les souches syngéniques de M. anisopliae var. acridum renfermant de l’ARNdb indiquequ’il ne serait pas tout à fait exact de définir ces infections comme étant latentes.

Mots clés : Metarhizium anisopliae var. acridum, Paecilomyces fumosoroseus, ARN à double brin (ARNdb), protéases,production de conidies, virulence.

[Traduit par la Rédaction] Tiago et al. 339

Can. J. Microbiol. 50: 335–339 (2004) doi: 10.1139/W04-023 © 2004 NRC Canada

335

Received 31 July 2003. Revision received 20 February 2004. Accepted 1 March 2004. Published on the NRC Research Press Website at http://cjm.nrc.ca on 1 June 2004.

P.V. Tiago and M.C. Furlaneto.1 Departamento de Microbiologia, Universidade Estadual de Londrina, P.O. Box 6001, 86051-990Londrina-PR, Brazil.M.H.P. Fungaro. Departamento de Biologia Geral, Universidade Estadual de Londrina, C.P. 6001, 86051-990 Londrina-PR, Brazil.M.R. de Faria. EMBRAPA Genetic Resources and Biotechnology, Parque Estação Biológica s/n, CEP 70770-900 Brasília-DF, Brazil.

1Corresponding author (e-mail: mcfurlan@sercomtel.com.br).

Introduction

Double-stranded RNA (dsRNA) viruses are common infungi (Buck 1998). In the case of entomopathogenic fungi,the presence of dsRNA genetic elements has been describedin Metarhizium anisopliae var. anisopliae, Metarhiziumanisopliae var. acridum (formerly Metarhizium flavoviride),Beauveria bassiana, and Paecilomyces fumosoroseus (Lealet al. 1994; Bogo et al. 1996; Melzer and Bidochka 1998;Martins et al. 1999; Azevedo et al. 2000).

Benign phenotypes appear to be the most common pheno-types associated with fungal viruses (Buck 1998). However,for plant pathogenic species, deleterious phenotypes associ-ated with virus infection have been described (McCabe et al.1999). A dsRNA virus associated with hypovirulence inCryphonectria parasitica (hypovirus 1 CHV1) affects fungaldevelopment, including asexual sporulation and virulence(Nuss 1992). Recent studies have analysed the association ofvirus replication with the impairment of protein secretionthat could cause the hypovirulence phenotype (McCabe etal. 1999).

It is often difficult to associate a particular phenotype withthe presence of a particular dsRNA genome because fungican contain multiple and complex patterns of dsRNA ele-ments, even within a single species (McCabe et al. 1999),and also because of the genotypic variability of different iso-lates. The employment of isogenic strains, one with and onewithout dsRNA, seemed to be more appropriate for suchstudies.

A comparison between two isogenic strains of M.anisopliae revealed that the presence of dsRNA resulted inreduced insect (Grillus domesticus) virulence and variationin fungal growth rate and sporulation (Melzer and Bidochka1998), which is similar to findings in phytopathogenic fungi(Elliston 1982; Boland 1992). A mutant of M. anisopliae,which had lost part of its dsRNA components also presentedalterations in colony morphology and conidia production(Giménez-Pecci et al. 2002). Recently, our group obtainedisogenic strains of M. anisopliae var. acridum (Martins et al.1999) and P. fumosoroseus (Azevedo et al. 2000). Virulencetests against the grasshopper Rhammatocerus schistocercoidesand the silverleaf whitefly Bemisia tabaci, respectively,revealed that these dsRNA elements did not cause hypo-virulence of their hosts.

Although the effects of dsRNA elements on fungal viru-lence have been analysed, no attempts have been made tocorrelate the presence of a particular dsRNA with a changeon fungal determinants of entomopathogenicity. Entomo-pathogenic fungi exhibit many attributes, which determinevirulence towards their hosts. Proteases are involved in deg-radation of insect cuticular proteins during pathogenesis, as-sisting penetration and providing nutrients for furthergrowth. The subtilisin-like (Pr1) protease is the best under-stood model of fungal determinants of entomopathogenicity(St. Leger 1995). Its production was recently examined inisolates of M. anisopliae var. acridum after growth on cutic-ular substrate (Pinto et al. 2002; Tiago et al. 2002).

In the present study, isogenic strains of M. anisopliae var.acridum and P. fumosoroseus were investigated for associationof phenotypic parameters, such as production of extracellularproteases, total secreted protein, and conidia production,

with the presence or absence of dsRNA. We also sought todetermine whether the dsRNA present in P. fumosoroseuscould affect its virulence to the stinkbug Euschistus heros.

Materials and methods

Fungal isolatesThe M. anisopliae var. acridum isolate CG442 (ARSEF

3391; IMI 324673) was provided by the entomopathogenicfungal collection of EMBRAPA. Isogenic strains, withoutdsRNA (CG442 met) and with dsRNA (CG442 met7 andCG442 met8) were obtained via hyphal anastomosis betweenthe auxotrophic mutants CG291 arg and CG442 met byMartins et al. (1999). A cured strain of P. fumosoroseus(Pf92 L-19), isolated from a dsRNA-containing isolateCNPSo-Pf92 (ARSEF 3638), was obtained by singleconidium subculture (Azevedo et al. 2000).

Culture conditionsConidia from M. anisopliae var. acridum CG442 and the

isogenic strains were obtained by harvesting sporulated cul-tures grown on agar plates composed of minimal medium(MM; Pontecorvo et al. 1953) supplemented with 0.5%(m/v) insect cuticle (MM+cut), prepared from the adultgrasshopper Rhammatocerus schistocercoides using an aque-ous solution of 1% (m/v) potassium tetraborate (Andersen1980). The cuticle extract was added to previously sterilized(121 °C for 15 min) MM and autoclaved for 15 min at115 °C. For enzyme production conidia were inoculated into30 mL of liquid MM+cut at a concentration of 5 × 106

conidia·mL–1 and grown in submerged culture (180 r·mim–1)at 28 °C for 96 h. The same procedure was performed on theisogenic strains of P. fumosoroseus except that the cuticlesource was from the stinkbugs Euschistus heros or Nezaraviridula. Following growth, mycelium was harvested bycentrifugation at 8000g for 15 min. The supernatants(extracellular fraction) obtained were stored at –20 °C andassayed for protease activity and total protein. All experi-ments were repeated four times, and the results representmean values ± SD.

Enzyme assaysMetarhizium anisopliae var. acridum subtilisin-like activ-

ity (referred to as Pr1) was assayed by modification of themethod of St. Leger et al. (1987), using succinyl-(alanine)2-proline-phenylalanine-p-nitroanilide as the substrate. Eachassay consisted of 0.05 mL substrate (1 mmol·L–1), 0.85 mL30 mmol·L–1 Tris-HCl buffer (pH 7.0), and 0.1 mL of crudeenzyme. The mixture was incubated for 1 h at 28 °C, and thereaction was terminated by adding 0.25 mL of 30% aceticacid, left to stand for 15 min in ice, after which sampleswere centrifuged at 1250g for 5 min at 4 °C. Thesupernatants were read at 410 nm. Activity was expressed asmicromoles of p-nitroanilide per millilitre per hour.

The proteolytic activity of the crude proteases from P.fumosoroseus culture was quantified using the chromogenicsubstrate azocasein, as described by Charney and Tomarelli(1947). Each assay consisted of 0.025 mL of substrate(10 mg·mL–1), 0.1 mL 0.2 mol·L–1 Tris–HCl buffer (pH 7.0),and 0.75 mL of crude enzyme. The mixture was incubatedfor 1 h at 28 °C. Undigested substrate was precipitated by

© 2004 NRC Canada

336 Can. J. Microbiol. Vol. 50, 2004

adding 0.25 mL of 10% percloric acid and left to stand for15 min in ice, after which samples were centrifuged at1250g for 5 min at 4 °C. Supernatants were neutralized byadding 0.25 mL 2 mol·L–1 NaOH, left for 10 min at roomtemperature, and absorbance was measured at 440 nm. Ac-tivity was expressed as micromoles of sulfanilamide permillilitre per hour. Assays were performed in duplicate foreach sample. Protein was measured using Coomassie BrilliantBlue G-250, according to Bradford (1976), with bovine se-rum albumin as the standard. Assays were performed in du-plicate for each sample.

Determination of conidia productionA volume of 100 µL of conidial suspensions in sterile wa-

ter (1 × 107 conidia·mL–1) were spread evenly onto completemedium agar plates with a sterile glass spreader, and theplates were incubated at 28 °C for 7 days. Three agar plugs(7 mm diam.) were taken from a random location on eachplate and placed in 10 mL of distilled water. Conidia werethoroughly washed from the surface of the agar plug, and theresulting conidial suspension was counted in a Neubauerhaemocytometer. The number of conidia per square centi-metre of agar was calculated based on the area of the agardisk. The experiment was repeated four times, and the re-sults represent mean values ± SD.

Virulence testsOne- to seven-day-old Euschistus heros adults were topically

inoculated with 3 µL of conidial suspensions of the isogenicstrains of P. fumosoroseus (1 × 105 spores·µL–1) prepared inpure soybean oil. Control insects were inoculated with puresoybean oil as a control. Insects, five per Petri dish, were in-cubated at 28 °C for 9 days. Food was replaced every otherday. There were 12 replicates per treatment. The experimentwas repeated twice. Mortality was assessed daily and thedata were subjected to statistical analysis (t test).

Results and discussion

To analyse whether a particular dsRNA could be corre-lated with extracellular protease activity in M. anisopliaevar. acridum, we analysed the production of subtilisin-like(Pr1) protease from isogenic strains CG442 met (-dsRNA),CG442 met7, and CG442 met8 (+dsRNA) after growth inthe presence of grasshopper (R. schistocercoides) cuticle.Pr1 activities from the wild strain CG442 and isogenic

strains grown on cuticle-containing medium are shown inTable 1. No significant differences (P < 0.05) in Pr1 activi-ties were observed among the different strains. This cuticle-degrading protease activity represents the most extensivelystudied fungal determinant of entomopathogenicity. Its rolein host invasion has been clearly demonstrated in M.anisopliae var. anisopliae (St. Leger et al. 1988). The Pr1protease is adapted to extensively degrade cuticular protein(St. Leger et al. 1987) and has been ultrastructurally locatedduring host penetration (Goettel et al. 1989). Our group re-cently described the production of Pr1-like activity by M.anisopliae var. acridum CG442 grown on medium supple-mented with R. schistocercoides cuticle (Pinto et al. 2002).Until recently, this grasshopper species was a major agricul-tural problem in Central Brazil, and M. anisopliae var.acridum had been described as one of the promising speciesfor development as a biological control agent for grasshop-pers. Protease activities from isogenic strains were also as-sessed using the chromogenic substrate azocasein (Table 1).Similar to the results found for Pr1, a comparison of meanprotease activity showed no statistical differences (P < 0.05)among the isogenic strains examined, indicating that therewas no correlation of protease activity with the presence orabsence of dsRNA.

In the case of P. fumosoroseus, proteolytic activities froma dsRNA-containing isolate and a cured strain (PfP92 L19)grown on medium supplemented with the cuticle of thestinkbugs E. heros, a naturally occurring host for P.fumosoroseus, and N. viridula were assessed using the sub-strate azocasein (Table 2). For this fungal species, its Pr1 ac-tivity has not yet been described. Similar levels of enzymeactivities were observed for both cuticle sources. No signifi-cant differences (P < 0.05) in protease activities were ob-served among the isogenic strains.

Isogenic strains from both fungal species were further ex-amined for association between the presence of dsRNA andtotal secreted protein from cuticle-containing cultures (Ta-bles 1 and 2). Similar to the observed proteolytic activities,differences in total protein were not statistically significant(P < 0.05) among the isogenic strains. The results of totalprotein could be more conclusive if they were expressed interms of fungal dry weight. However, this was not possiblebecause the insoluble nature of the cuticle made dry weightmeasurements impractical.

Additionally, no differences on SDS–PAGE patterns ofculture supernatants from M. anisopliae var. acridum

© 2004 NRC Canada

Tiago et al. 337

Strain Pr1a Proteaseb Protein (mg·mL–1)

CG442 (–dsRNA) 0.24±0.013 0.09±0.022 0.24±0.10CG442 met (–dsRNA) 0.20±0.039 0.10±0.024 0.27±0.10CG442 met7 (+dsRNA) 0.20±0.051 0.10±0.028 0.18±0.058CG442 met8 (+dsRNA) 0.19±0.065 0.10±0.021 0.22±0.079

Note: Each result is the means of four experiments ± standard error of the mean. Pairwise t test (P <0.05) revealed no significant differences among strains.

aPr1 activity expressed as micromoles of p-nitroanilide per millilitre per hour.bProtease activity expressed as micromoles of sulfanilamide per millilitre per hour.

Table 1. Subtilisin-like (Pr1) activity and protease activity detected by azocasein assay andtotal protein in culture supernatants from Metarhizium anisopliae var. acridum CG442 andisogenic strains (with and without dsRNA) after 96 h of growth on minimal mediumsupplemented with cuticle from the grasshopper Rhammatocerus schistocercoides.

isogenic strains were observed, nor were there differences inprotease activity as revealed by clear zones in a R.schistocercoides cuticular-soluble protein-containing gel(substrate-SDS–PAGE) (Garcia-Carreno et al. 1993) (datanot shown).

Isogenic strains were also investigated for association ofconidia production with the presence of dsRNA, as shown inTable 3. A comparison of mean conidia production showedstatistical differences (P < 0.05) among the strains of M.anisopliae var. acridum with and without dsRNA. Conidiaproduction was greater for CG442 met (–dsRNA) than forCG442 met7 and CG442 met8 (+dsRNA). Similar resultswere found for the isogenic strains of M. anisopliae (Melzerand Bidochka 1998). In contrast, for strains of P.fumosoroseus with and without dsRNA, no statistical differ-ences were observed in conidia production. For biocontrolpurposes, conidia production by entomogenous fungi is animportant characteristic, since conidia are commonly used incommercial myco-insecticide-formulated products.

Except for the observed reduction in conidia productionby the isogenic strains of M. anisopliae var. acridum withdsRNA, there was no association of the remaining phenotypicparameters analysed with the presence or absence of dsRNA.

Results from bioassays using adults of the stinkbugEuschistus heros and strains of P. fumosoroseus (Pf92 andPf92-L19) revealed no significant differences in virulenceamong the isogenic strains (t test; P = 0.157), which is in ac-cordance with findings for R. schistocercoides and the

isogenic strains of M. anisopliae var. acridum (Martins et al.1999).

Few fungal viruses, with a dsRNA genome, have beendemonstrated conclusively to cause a specific phenotype.Melzer and Bidochka (1998) found that the phenotypic vari-ability observed among isolates of M. anisopliae was notcorrelated with the absence or presence of dsRNA infection.However, phenotypic effects associated with a particulardsRNA infection were observed using isogenic strains (onewith and one without dsRNA).

In the present study, the employment of isogenic strains,of both species, allowed us to analyse whether a particulardsRNA could cause changes in their host’s phenotype. Al-though, for most of the fungal phenotypes analysed, no overteffects were associated with the presence of these dsRNA in-fections, the reduction in conidia production by the isogenicstrains of M. anisopliae var. acridum with dsRNA suggestedthat it may not be entirely accurate to describe these infec-tions as latent. The biological significance of dsRNA infec-tion in these fungi, as well as in many other fungal species,remains unknown.

Acknowledgements

This study was supported in part by Conselho Nacional deDesenvolvimento Científico e Tecnológico (CNPq) andCoordenadoria de Pesquisa e Pós-Graduação/ UniversidadeEstadual de Londrina–Brazil. The authors thank ProfessorRobert F.H. Dekker for critically reading this manuscript andDr. José Carlos Dalmas for the statistical analyses.

References

Andersen, S.O. 1980. Cuticular sclerotization. In Cuticle tech-niques in arthropods. Edited by T.A. Miller. Springer-Verlag,New York. pp. 185–217.

Azevedo, A.C.S., Sosa-Gómez, D.R., Faria, M.R., and Fungaro,M.H.P. 2000. Effects of double-stranded RNA on virulence ofPaecilomyces fumosoroseus (Deuteromycotina: Hyphomycetes)against the silverleaf whitefly, Bemisia tabaci strain B(Homoptera: Aleyrodidae) Genet. Mol. Biol. 23: 61–63.

Bogo, M.R., Queiroz, M.V., Silva, D.M., Giménez, M.P., Azevedo,J.L., and Schrank, A. 1996. Double-stranded RNA and isometricvirus-like particles in the entomopathogenic fungus Metarhiziumanisopliae. Mycol. Res. 100: 1468–1472.

Boland, G.J. 1992. Hypovirulence and double-stranded RNA inSclerotinia sclerotiorum. Can. J. Plant Pathol. 14: 10–17.

© 2004 NRC Canada

338 Can. J. Microbiol. Vol. 50, 2004

MM + insect cuticle(Euschistus heros)

MM + insect cuticle (Nezaraviridula)

Strain Proteasea Proteinb Proteasea Proteinb

Pf92 (+dsRNA) 0.06±0.013 0.18±0.045 0.07±0.007 0.18±0.031Pf92 L19 (–dsRNA) 0.06±0.005 0.21±0.076 0.06±0.005 0.19±0.057

Note: Each result is the means of four experiments ± standard error of the mean. Pairwise t test (P <0.05) revealed no significant differences between strains.

aProtease activity expressed as micromoles of sulfanilamide per millilitre per hour.bProtein expressed in milligrams per millilitre.

Table 2. Protease activity detected by assay against azocasein and total protein in culturesupernatants from isogenic strains (with and without dsRNA) of Paecilomycesfumosoroseus after 96 h of growth in minimal medium (MM) supplemented with insectcuticle from either Euschistus heros or Nezara viridula.

Strain Conidia (×106)·cm–2

M. anisopliae var. acridumCG442 123.5±29.62aCG442 met (–dsRNA) 68.2±2.81bCG442 met 7 (+dsRNA) 39.9±5.12cCG442 met 8 (+dsRNA) 39.8±7.21cP. fumosoroseusPf92(+dsRNA) 78.4±13.35aPf92L19(–dsRNA) 71.1±7.87a

Note: Means followed by a different letter differ are significantlydifferent according to the pairwise t test (P < 0.05). Each result is themeans of four experiments ± standard error of the mean.

Table 3. Conidia production by isogenic strains of Metarhziumanisopliae var. acridum and Paecilomyces fumosoroseus after7 days of growth on complete medium.

© 2004 NRC Canada

Tiago et al. 339

Bradford, M.M. 1976. A rapid and sensitive method for thequantitation of microgram quantities of protein utilizing theprinciple of protein-dye binding. Anal. Biochem. 72: 248–254.

Buck, K.W. 1998. Molecular variability of viruses of fungi. In Mo-lecular variability of fungal pathogens. Edited by P.D. Bridge, Y.Couteaudier, and J.M. Clarkson. CAB International, Wallingford,UK. pp. 53–72.

Charney, A.K., and Tomarelli, R.M. 1947. A colorimetric methodfor the determination of the proteolytic activity of duodenaljuice. J. Biol. Chem. 171: 501–505.

Elliston, J.E. 1982. Hypovirulence. In Advances in plant pathology.Edited by D.S. Ingram and P.H. Williams. Academic Press,Toronto, Ont. pp. 1–33.

Garcia-Carreno, F.L., Dimes, L.E., and Haard, N.F. 1993. Sub-strate-gel electrophoresis for composition and molecular weightof proteinases or proteinaceous proteinase inhibitors. Anal.Biochem. 214: 65–69.

Giménez-Pecci, M.P., Bogo, M.R., Santi, L., Moraes, C.K., Corrêa,C.T., Vainstein, M.H., and Schrank, A. 2002. Characterisation ofmycoviruses and analyses of chitinase secretion in thebiocontrol fungus Metarhizium anisopliae. Curr. Microbiol. 45:334–339.

Goettel, M.S., St. Leger, R.J., Rizzo, N.W., Staples, R.C., andRoberts, D.W. 1989. Ultrastructural localization of a cuticle de-grading protease produced by the entomopathogenic fungus,Metarhizium anisopliae during penetration of host cuticle. J.Gen. Microbiol. 135: 2223–2239.

Leal, S.C.M., Bertioli, D.J., Ball, B.V., and Butt, T.M. 1994. Pres-ence of double-stranded RNAs and virus-like particles in theentomopathogenic fungus Metarhizium anisopliae. BiocontrolSci. Technol. 4: 89–94.

Martins, M.K., Furlaneto, M.C., Sosa-Gómez, D.R., Faria, M.R.,and Fungaro, M.H.P. 1999. Double-strand RNA in the entomo-

pathogenic fungus Metarhizium flavoviride. Curr. Genet. 36:94–97.

McCabe, P.M., Pfeiffer, P., and Van Alfen, N.K. 1999. The influ-ence of dsRNA viruses on the biology of plant pathogenic fungi.Trends Microbiol. 7: 377–381.

Melzer, M.J., and Bidochka, M.J. 1998. Diversity of double-stranded RNA viruses within populations of entomopathogenicfungi and potential implications for fungal growth and viru-lence. Mycologia, 90: 586–594.

Nuss, D.L. 1992. Biological control of chestnut blight: an exampleof virus-mediated attenuation of fungal pathogenesis. Microbiol.Rev. 56: 561–576.

Pinto, F.G.S., Fungaro, M.H.P., Ferreira, J.M., Valadares-Inglis,M.C., and Furlaneto, M.C. 2002. Genetic variation in the cuticle-degrading protease activity of the entomopathogen Metarhiziumflavoviride. Genet. Mol. Biol. 25: 231–234.

Pontecorvo, G., Roper, J.A., Hemons, L.M., MacDonald, K.D., andBufton, A.W.J. 1953. The genetics of Aspergillus nidulans. Adv.Genet. 5: 141–238.

St. Leger, R.J. 1995. The role of cuticle-degrading proteases infungal pathogenesis of insects. Can. J. Bot. 73: 1119–1125.

St. Leger, R.J., Charnley, A.K., and Cooper, R.M. 1987. Character-ization of cuticle degrading proteases produced by the entomo-pathogen Metarhizium anisopliae. Arch. Biochem. Biophys.253: 221–232.

St. Leger, R.J., Durrands, P.K., Charnley, A.K., and Cooper, R.M.1988. Role of extracellular chymoelastase in the virulence ofMetarhizium anisopliae for Manduca sexta. J. Invertebr. Pathol.52: 285–293.

Tiago, P.V., Fungaro, M.H.P., and Furlaneto, M.C. 2002. Cuticle-degrading proteases from the entomopathogen Metarhiziumflavoviride and their distribution in secreted and intracellularfractions. Lett. Appl. Microbiol. 34: 91–94.