Genetic variation in the cuticle-degrading protease activity of the entomopathogen Metarhizium flavoviride

Extracellular proteases have been shown to be factors of virulence in fungal entomopathogenicity. We examined the production of the cuticle-degrading extracellular proteases chymoelastase (Pr1) and trypsin (Pr2) in isolates of the fungus Metarhizium flavoviride. Fungal growth was in a mineral medium (MM) containing nitrate, and in MM supplemented with either cuticle from Rhammatocerus schistocercoides or with the non-cuticular substrate casein. The substrates used for growth influenced the expression of both analyzed proteases, the highest protease activities of nearly all isolates having been observed in the medium containing insect cuticle, with more Pr1 than Pr2 being produced. There was a natural variability in the production of cuticle-degrading proteases among isolates, although this was less evident for Pr2. Our data support the hypothesis that the production of Pr1 on insect cuticle is a useful characteristic for the analysis of intraspecific variability of M. flavoviride isolates.


Introduction
The South-American grasshopper, Rhammatocerus schistocercoides, is a major problem in Central Brazil, causing severe agricultural losses (Lecoq and Pierozzi, 1994).The entomopathogenic fungus Metarhizium flavoviride has been suggested as a biological control agent for locusts and grasshoppers (Prior et al., 1992), and it is known that some M. flavoviride isolates are highly virulent against R. schistocercoides (Magalhães et al., 1997).Fungal proteases are the enzymes most frequently considered critical in facilitating invasive mycosis of insects (Charnley and St. Leger, 1991), the studies on cuticle-degrading proteases in entomopathogenic fungi having mainly focused on Metarhizium anisopliae, which produces a chymoelastase (Pr1) (St. Leger et al., 1987a), a metalloprotease (St. Leger et al., 1994), a trypsin-like enzyme (Pr2) belonging to the serine protease group and a cysteine protease (Pr4) (Cole et al., 1993).
The most extensively studied protease of M. anisopliae is Pr1, whose role in host invasion was clearly demonstrated by St. Leger et al. (1988).According to Charnley and St. Leger (1991), this enzyme provides the best understood model of the fungal entomopathogenicity determinant .Gillespie et al. (1998) found that the Pr1 protease from different isolates of Metarhizium hydrolyzes different types of insect cuticle to different degrees, although there was no correlation between Pr1 activity and the mean lethal time of the isolates in the desert locust (Schistocerca gregaria).
The role of the Pr2 protease in insect parasitism was not yet elucidated, although St. Leger et al. (1994) reported that Pr2 as well as the other cuticle-degrading proteases may complement each other in the splitting of peptide bonds in the insect cuticle.Results obtained by Gillespie et al. (1998) are consistent with a role of Pr2 in the induction or activation of Pr1, since Pr2 was detected before Pr1 in a culture containing ground locust cuticle.Both Pr1 and Pr2 are also produced by other entomogenous fungi, including Beauveria bassiana (St. Leger et al., 1987b;Gupta et al., 1992), Nomuraea rileyi, and Verticillim lecanii (St. Leger et al., 1987b).
The only study on the production of cuticle-degrading proteases by M. flavoviride used cuticle of the desert locust, S. gregaria (Gillespie et al., 1998).We compared the production of Pr1 and Pr2 in M. flavoviride isolates, after growth on R. schistocercoides cuticle, with to protease production on non-cuticular substrate (casein and nitrate).
Conidia were obtained by harvesting sporulated spread plates of minimal medium (MM) (Pontecorvo et al., 1953), containing nitrate as nitrogen source, and MM supplemented with either 1.0% casein (MM+cas) or 0.5% (w/v) insect cuticle (MM+cut) prepared from adult Rhammatocerus schistocercoides using an aqueous solution of 1% potassium tetraborate (Andersen, 1980).The prepared cuticle was added to previously sterilized MM and autoclaved for 5 min at 115 °C.For enzyme production, conidia were inoculated into 100 mL of liquid MM+cas and MM+cut at a concentration of 1 x 10 7 mL -1 and incubated as shake cultures at 150 rpm, 28 °C, for 72 h.The mycelia were harvested by filtration and the supernatants assayed for enzyme activity.

Statistical analyses
Statistical analyses were performed using the Statistical Analysis System software (SAS Institute Inc., 1989).The experiment was set up as a randomized complete block design, with two replicates.The effect of treatment combinations (isolates and media) was assessed by factorial analysis.The variance analyses were followed by a comparison of the proteolytic activity means, using the Tukey test.

Results and Discussion
Chymoelastase (Pr1) activity was present in all the supernatants analyzed, and was highly expressed in the presence of insect cuticle (Figure 1).Isolate CG288 had the lowest activity, and isolate CG441 had the highest activity, about 27 times that of CG288 on the media tested.
All the isolates, with the exception of CG288, showed a significantly different Pr1 activity (p < 0.01) as a function of the growth substrate (Table I), the enzyme activity being about 2 to 8 times higher in MM+cas and about 4 to 20 times higher in MM+cut than in MM.The high levels of Pr1 found in MM+cut cultures may be due to the induction of Pr1 by insect cuticle components, since Paterson et al. (1994) determined that the Pr1 activity of starved M. anisopliae mycelium is specifically induced by cuticle from S. gregaria.
Among the Brazilian isolates, there was a significant enhancement in the Pr1 activity of CG423 after growth on R. schistocercoides cuticle, which may be related to its high virulence against this grasshopper species ( Magalhães et al., 1997), since quantitative differences in in vivo protease production between isolates can influence virulence (Gillespie et al., 1998).
Trypsin (Pr2) was also produced by all isolates (Figure 2), but at lower levels than Pr1and with less variability between isolates than Pr1 (Table I).All isolates except CG288 and CG291 produced significantly different levels of trypsin (p < 0.01) as a function of the substrate used for growth (Table I), the highest production having been in MM+cut.Paterson et al. (1993) showed that M. anisopliae Pr2 is nonspecifically induced by any protein, if the fungus is starved of nitrogen, but not if starved of carbon in the presence of nitrogen.In our experiments with MM+cut, the Pr2 activity was only 29% (isolate CG442) and 24% (isolate CG423) of the Pr1 activity, similar results having been observed for cuticle-degrading proteases produced by M. anisopliae against S. gregaria (Cole et al., 1993).
In our study, the substrates used for growth significantly influenced the expression of Pr1 and Pr2, except in isolate CG288 (Table I).We found a natural variability in the production of insect cuticle-degrading proteases among M. flavoviride isolates .Strain variability has also been observed in the production of proteases in M. anisopliae (St. Leger et al., 1986;Gillespie et al., 1998) and in B. bassiana (Gupta et al., 1992), while Leal et al. (1997) demonstrated significant variations in Pr1 sequences among strains of entomopathogenic fungi.It has also been described that M. anisopliae proteases comprise multiple isoforms (St. Leger et al., 1994).
A consistent pattern of Pr1 expression was apparent among isolates, with all isolates except CG423 and CG430 producing statistically different levels of Pr1 on cuticle, which gave a 100% homology on RAPD analysis (Martins, 1998).The production of Pr1 in a culture containing R. schistocercoides cuticle may be useful to analyze strain variability of M. flavoviride and related entomopathogenic fungi.Leal et al. (1997) used the PCR product derived from the M. anisopliae Pr1 gene to differentiate between many Metarhizium strains.Fungal strain variability in Pr1 production may be directly related to variability in virulence, because the regulation of the Pr1 gene expression may determine the capacity of the fungus to cause insect disease (St. Leger et al., 1988).

Table I -
Mean values of chymoelastase (Pr1) and trypsin (Pr2) activities from Metarhizium flavoviride isolates grown in minimal medium plus cuticle (MM+ cut), minimal medium plus casein (MM+ cas), and minimal medium (MM), respectively.The results are from two experiments with three replicates for each sample.
Different letters represent a significant difference (p < 0.05) according to Tukey's test.Lower case letters indicate significant differences within a column, and capital letters within a row.