Introduction

Maize (Zea mays L.) is one of the main crops in Argentina. Over the sixty percent of maize produced in this country is exported (INAI 2009). To maintain or increase this percentage, maize producers and stockpilers must assure the importer countries good quality grains and free of contaminants. The saprophytic activity of different fungus, like species of the genus Aspergillus, is the main cause of degradation of agricultural products before and after harvest. Fungi of this genus can grow in a wide range of environmental conditions; therefore, preharvest, harvest, and storage conditions are ideal (Payne 1998). Potentially toxigenic Aspergillus section Flavi strains were extensively distributed in non-rhizospheric soil, debris, and insects of maize agroecosystem (Nesci and Etcheverry 2002). Moreover, Aspergillus section Flavi was constantly recovering from soil under different tillage practices (Nesci et al. 2006). The agroecosystem of stored grain has its own dynamic which depends on environmental, biologic, and substrate-specific factors. The interaction between the substrate, the biologic, and the abiotic factors may favor the invasion of Aspergillus species and lead to the concomitant production of mycotoxins called aflatoxins. Aflatoxins, especially aflatoxin B1 (AFB1), are considered the most carcinogenic, mutagenic, and teratogenic substance found naturally in foods and feeds (IARC 1993).

Insects are involved in the colonization of grains. They can provide sites for fungal infection through the damage that can result in grain. It is therefore essential to understand that in this biologic system, insects are a common problem. Sitophilus zeamais (Motschulsky), Rhyzopertha dominica (Fabricius), and Tribolium confusum (Jacquelin du Val) cause significant damage to stored maize (Mejía 2007). Damage may be direct, such as weight loss, reduced germination, and reduced nutritional value of grain. Indirect damage can be heat and moisture migration, reservoir of disease, and distribution of microorganisms (White 1995). Previous studies show that certain insects that attack stored grains have the ability to disperse toxigenic Aspergillus flavus in those grains (Nesci et al. 2011a, b).

Knowing that many insects, besides damaging the grain, are vectors of Aspergillus section Flavi which facilitates increased levels of aflatoxins, integrated control strategies are necessary. Currently, most of the post-harvest management of insect pests is made with synthetic chemical insecticides. However, the use of synthetic insecticides allowed in argentine grains for export is very limited (Casini and Santajuliana 2012). In recent times, the research is aimed at finding ways to prevent the entry of xenobiotics in the food chain (Jayashree and Subramanyam 1999). Therefore, the search for natural methods of crop protection is still valid despite the fact that the market offers a wide variety of products. There are biologic agents with natural potential to protect crops. This derives from the intrinsic richness of species and their struggle for survival (Stoll 1989). Entomopathogenic fungi are natural enemies of a wide range of insects and some species are used as microbial bio pesticides (Tanada and Kaya 1993; James and Elzen 2001). Entomopathogenic fungi are distributed in a wide range of habitats including aquatic forest, agricultural, pasture, desert, and urban habitats (Sánchez-Pena 1990; Lacey et al. 1996; Chandler et al. 1997). The selection of virulent isolates adapted to local components of the agroecosystem is one of the most important aspects in the development of mycoinsecticides (Cortez-Madrigal et al. 2003). Soil is considered an excellent environmental shelter for entomopathogenic fungi since it is protected from UV radiation and other adverse abiotic and biotic influences (Keller and Zimmermann 1989). Isolation of entomopathogenic fungi involves soil sampling since that is their natural habitat (Asensio et al. 2003). It is also interesting to compare the diversity of entomopathogenic fungi in different components of the same ecosystem to obtain greater variety of antagonistic microorganisms (Cross et al. 1999; Bidochka et al. 1998; Asensio et al. 2003).

Numerous studies show that isolates of Beauveria bassiana (Balsamo) Vuillemin and Metarhizium anisopliae (Metschnikoff) Sorokin are potential microbial control agents against some stored product pests (Tanya and Doberski 1984; Adane et al. 1996; Hidalgo et al. 1998; Dal Bello et al. 2001; Ekesi et al. 2001; Padin et al. 2002; Khashaveh et al. 2008). Nevertheless, there is a constant search for other fungal biologic control agents, including nematophagous fungi like Paecilomyces lilacinus (Thom) Samson (Fiedler and Sosnowska 2007). P. lilacinus is a soil-inhabiting fungus that has shown promising results for use as an effective biocontrol agent (Morgan-Jones et al. 1984; Jatala 1986; Dube and Smart 1987; Khan et al. 2006).

The objective of this study was to isolate and identify native entomopathogenic fungi from different components of maize agroecosystem, and evaluate their virulence against Tribolium confusum, Rhyzopertha dominica, and Sitophilus zeamais, three insect pest vectors of aflatoxigenic fungi.

Materials and methods

Sampling site

Soil samples were collected from the University of Río Cuarto Experimental Field Station in Río Cuarto, Córdoba, Argentina (30°57′S latitude, 64°50′W longitude, 562 m altitude) during the maize growing season (2009–2010). Soil consisted of a sandy loam texture (pH 6.1 in water 1:1 v/v, 1.4 % organic matter, 86 ppm of nitrates).

Collection procedures

Sampling was performed during the first 20 days of development of the maize crop. Fifty samples of non-rhizospheric soil were collected. Each of the samples was taken from the top 3 cm of soil at different places within the field. The samples were taken in double diagonal section at 4 m intervals. Samples were individually placed in plastic bags and stored at −20 °C until analysis.

From 50 soil samples, 25 samples were selected randomly and dried at 50 °C during 24 h. Dried samples were passed through a testing sieve (2-mm mesh size) and the debris separated from the soil. The debris were placed in plastic bags and stored at −20 °C until use.

Isolation and identification of entomopathogenic fungi

Insect baiting method

Cultures of one strain of the confused flour beetle Tribolium confusum (Jacquelin du Val) were obtained from the Department of Agricultural Zoology, Faculty of Agronomy, University of Buenos Aires, Argentina. Mixed-sex adults 1–3 weeks old were used in the test. Insects were reared on a diet of wheat flour, corn starch, and yeast (10:10:1.5) in plastic containers containing 200 g of the mixture. Insects were reared at 27 ± 1 °C and 70 ± 5 % relative humidity (R.H.) and photoperiod of 12:12 h light:dark cycle. These insects were used to entomopathogenic fungi isolation. Three types of substrates were used; twenty five dried soil samples (250 g each) without debris, 10 not-dried soil samples (250 g each) with debris, and 5 maize grain samples (500 g each). Maize grains were harvested from plants grown in the same experimental maize field in the growing season 2009–2010. All samples were weighed in plastic jars of 500-ml capacity. Twenty adults of insects were placed per jar. All jars were placed in a chamber with controlled conditions 27 ± 1 °C and 70 ± 5 % relative humidity (R.H.) and incubated during 15 days. After 15 days, all samples were frozen at −20 °C to kill insects. Then, the soil and maize samples were sieved to recover the insects. Recovery insects were surface-disinfected using 1 % sodium hypochlorite for 1 min. The disinfected agent was eliminated and insects were washed two times in sterile distilled water (5 min each time). Insects were finally plated directly on semi-selective isolation media (SM) (maize meal 17 g l−1, ClNa 17.5 g l−1, rose bengal 75 mg l−1, triton X-100 0.3 ml l−1, chloramphenicol 50 mg l−1, streptomycin 0.1 ml l−1, cycloheximide 2.5 ml l−1, and agar-agar 15 g l−1, 1,000 ml) and incubated at 25 °C during 10 days. After the incubation period, colonized insects were analyzed. The colonies isolated from insects, suspected to be entomopathogenic fungi, were subcultured on Sabouraud Dextrose Agar (SDA) (trypticase 5 g l−1, peptone beef 5 g l−1, glucose 20 g l−1, and agar–agar 15 g l−1, 1,000 ml) and incubated at 25 °C during 10 days. The control treatment consisted of taking 80 insects from containers in which they were kept, and frozen at −20 °C to kill them. Then, the insects were surface disinfected and finally plated directly and incubated as described above. After microscopic observation, genus was assigned according to Samson (1974), Samson et al. (1988), Humber (2005) and Luangsa-ard et al. (2011).

Soil samples

The soil plating method was adapted from Garrido-Jurado et al. (2011). Enumeration of fungal propagules was carried out on solid medium, by the surface spread method, by blending 10 g of soil from each of the 25 samples randomly selected from a total of 50 samples with 90 ml 0.1 % peptone water solution. Serial dilutions from each sample and 0.1 ml aliquots were inoculated in SM medium and incubated at 25 °C for 13 days. Fungal count was expressed as log10 per g of soil. The fungi were subcultured to SDA medium to obtain pure cultures and identified by microscopic characteristics as described above.

Debris samples

The experience was conducted according to the methodology proposed by Nesci and Etcheverry (2002) with some modifications. Twenty five pieces of debris from each of the 25 samples were surface disinfected using 1 % sodium hypochlorite for 1 min. The disinfected agent was eliminated and debris pieces were washed two times in sterile distilled water (5 min each time). Debris pieces were finally plated directly on SM medium and incubated at 25 °C for 10 days. After the incubation period, colonized debris was analyzed. Count of possible entomopathogenic fungi was expressed as percentage of contaminated pieces. The fungi were subcultured and identified as described above.

Extraction of DNA from fungi grown in culture

The fungal cultures isolated in the previous assays were maintained in sterile soil (Abreu et al. 2003). All fungal cultures were grown on PDA medium for 7 days at 25 °C. Mycelial biomass was extracted for DNA analyses according to the procedure of Passone et al. (2010) with same modifications. An aliquot of 100 mg mycelium of each fungal isolate was transferred into microtubes. The mycelium was vortexed for 1 min in the presence of 300 μl of sterile water plus 700 μl of extraction buffer (100 mM Tris–HCl, 2 % CTAB, 1.4 mM NaCl) and glass beads (425–600 μm diameters) to favor the disruption of fungal material. After incubation at 65 °C for 60 min, 500 μl of chloroform was added to the sample, homogenized and centrifuged for 10 min at 13,000 rpm. The aqueous phase was recovered and 500 μl of chloroform was added. The sample was homogenized and centrifuged again for 10 min at 13,000 rpm. The aqueous phase was recovered and precipitated with 2 volumes of precipitation buffer (14 mM CTAB, 40 mM NaCl, pH 8). After incubation at room temperature for 1 h, the sample was homogenized and centrifuged for 5 min at 13,000 rpm. The sample was homogenized by inversion in presence of NaCl (1 M) 350 μl and chloroform 350 μl and centrifuged for 5 min at 13,000 rpm. Chloroform phase was recovered and precipitated with 0.6 volumes of isopropanol at −20 °C. After incubation at room temperature for 20 min, it was centrifuged for 10 min at 13,000 rpm and the aqueous phase was discarded. Finally, the DNA pellet was washed with 70 % ethanol and suspended in 25 μl of nuclease-free H2O. Polymerase chain reaction (PCR; TECHNE TC-512, Barloworld, Scientific Ltd., UK) was conducted in 50 μl reactions with the following concentrations: 5 U μl−1 of Invitrogen (Brazil) Taq DNA polymerase, 5× Invitrogen Buffer, 2 mM of dNTPs Invitrogen, 1.5 mM Mg2+, and 3 pmol μl−1 of each primer. The primers used were the fungus-specific forward primer EF1T (ATGGGTAAGGACAAGAC) and the reverse primer EF2T (GGAAGTACCAGTGATCATGTT) (O’Donnell et al. 1998) manufactured by Invitrogen Custom Primers (Carlsbad, CA). PCR was carried out by the following protocol: initial denaturation at 94 °C for 1 min, followed by 31 cycles of denaturation at 94 °C for 30 s, annealing at 56 °C for 45 s, extension at 72 °C for 1 min, and a final extension at 72 °C for 5 min. The reaction was held at 4 °C. The presence of PCR products was confirmed by gel electrophoresis on agarose gel.

Purification, sequencing, and identification of fungal DNA

Purification of PCR products was conducted using DNA Wizard Clean-Up Kit (product A9282, Promega, Madison, WI). To estimate the concentration of DNA needed for sequencing, purified PCR products were compared with Invitrogen’s Low DNA Mass Ladder. All samples were sequenced using an Applied Biosystem ABI 3730 sequencer (Applied Biosystem). Each PCR product was sequenced in the forward and reverse directions, and consensus sequences were created by means of the BioEdit program version 7.0.9.0. (Thompson et al. 1994) (http://www.mbio.ncsu.edu/RNasaP/info/programs/BIOEDIT/bioedit.html).

Lethal time and percentage of mortality

The lethal time (LT50), the number of days until 50 % of insects were dead, was determined for each of the 20 fungal suspensions. For this purpose, subsamples of maize grains of 500 g were put into plastic jars and then 20 adults of each insect pest Tribolium confusum, Sitophilus zeamais, and Rhyzopertha dominica treated by immersion for 30 s with 600 μl of 107 spores ml−1 suspensions of each entomopathogenic fungi (Goettel and Inglis 1997) were introduced in each jar (Padín et al. 1997). All jars were placed in a chamber with controlled conditions (27 ± 1 °C, 70 ± 5 % R.H., with photoperiod of 12:12 h light:dark cycle) (Wicklow et al. 1998). Mortality was analyzed during 20 days and compared with the untreated control samples. All dead insects were placed directly on plates containing water agar medium (agar–agar 15 g l−1, distilled water 1,000 ml), which were incubated at 25 °C for 7 days to confirm that the inoculated fungus was the causal agent of the death of insects. The lethal time (LT50) for each of the fungal isolates was analyzed by probit analysis.

Statistical analysis

Analysis of variance was performed on LT50 and mortality percentage using SigmaStat for windows version 3.0 (SPSS, Chicago, IL, USA). To establish differences, Holm-Sidak Method was performed.

Results

Isolating entomopathogenic fungi by the insect bait method

The total number of insects bait plated on semi-selective isolation medium were 800: 200 prior contact with 10 not-dried soil samples with debris, 500 prior contact with 25 dried soil samples without debris, and 100 prior contact with 5 maize grain samples. Table 1 shows the percentage of samples with fungal contamination, in which insects were in contact. In 90 % of soil samples and in 100 % of grain samples, insects showed fungal contamination. However, all insects placed in maize grains were colonized by Fusarium.

Table 1 Fungal contaminations with potential entomopathogenic fungi from insects bait prior contact with not-dried soil samples, dried soil samples and maize grain, and percentage of debris contaminated
Full size table

Samples of not-dried and dried soil used to place insects bait showed between 5–25 and 5–45 %, respectively, of fungal contamination. The highest percentages of infection were detected in dried soil samples. Only insects of 4 samples were completely free of fungal contamination.

Figure 1 shows the frequency of different fungal genera isolated from insect bait. They were Penicillium, Fusarium, and Paecilomyces. Genus Penicillium was isolated with more frequency reaching 56.3 % of infection. The 40.5 % of fungal isolates were identified as species from genera Fusarium; only 3.2 % were identified as species from genera known to include potential entomopathogenic species, such as Paecilomyces.

Fig. 1
figure 1

Frequency of different fungal genera isolated from insects bait

Full size image

Entomopathogenic fungi isolated from soil and debris samples

Our results showed that micobiota population was different in soils and debris samples. Densities of filamentous fungi in soil samples varied between 2 × 102 and 1.1 × 104 ufc g−1 (Fig. 2a). Mycoflora analysis showed that the 64 % of total fungal population isolated was confirmed as fungus from possible entomopathogenic genera. Figure 2b shows that the main entomopathogenic genus isolated was Paecilomyces, comprising 75 % of the total isolates, while the remaining 25 % corresponded to isolates belonging to the genus Metarhizium. The counts were 4 and 3 log for Paecilomyces and Metarhizium, respectively.

Fig. 2
figure 2

a Incidence of potential entomopathogenic fungi in soil samples. b Incidence of different fungal genera isolates from soil samples

Full size image

Micobiota analysis of 25 debris samples showed that 40 % of samples (10 samples) were contaminated with fungi with percentage of debris contaminated between 4 and 36 % (Table 1). All fungi isolated were identified as Fusarium and Penicillium, none of interest in this study.

Identification of entomopathogenic fungus

The 35 Paecilomyces isolates were identified as Paecilomyces lilacinus. These isolates showed colonies on MEA of fast growing, attaining a diameter of 20–37 mm after 7 days at 25 °C. Colonies consisting of a basal felt with or without floccose aerial overgrowth, the color is white at first and becoming vinaceous, reverse mostly in shades of purple or sometimes uncolored. Conidiophores are erect with roughened thick walls around 3-μm wide consisting of verticillate branches with two to four phialides. The phialides measure 8 × 2.1 μm and have a swollen basal portion tapering into a short neck. Conidia in divergent chains are ellipsoidal to fusiform, hyaline, measuring 2.8 × 2.4 μm. No or restricted growth was observed at 37 °C with colonies of 2–3 mm of diameter.

All these isolates were selected for DNA analysis. Seventy-seven cultures isolates provided clean sequence results. All culture isolates, when sequenced and compared with those in GenBank, were identified as Purpureocillium lilacinum (Table 2). Fifteen isolates showed a 99 % homology to the BLAST identities for sequences from P. lilacinum. Nine isolates are identical to the BLAST identities and 3 isolates showed homologies of 98, 97, and 95 %.

Table 2 Homology of translation elongation factor 1 alpha gene sequences from Purpureocillium lilacinum isolates with sequences in GenBank. Sequence informations and BLAST results are shown for closest accessions found in GenBank
Full size table

Determination of lethal time (LT50)

The mortality in control treatment was low (<10 %) and no mycosis was detected on any insects. LT50 values for 20 isolates of P. lilacinum against T. confusum varied from 4.66 to 17.41 days (Table 3). However, LT50 values against S. zeamais and R. dominica were higher with values ranging between 11.13 to >62.5 days and 9.38 to 48.38 days, respectively. The isolate JQ926223 showed the lowest LT50 for T. confusum and R. dominica, but the highest LT50 for S. zeamais. The isolate JQ926212 demonstrated similar LT50 for the three insects evaluated with a range of 11.7–14.95 days. Maximum mortality rates, 20 days after fungus application for T. confusum and R. dominica, were observed for the isolate JQ926223 (90 and 65 %, respectively). The isolate JQ926212 showed a mortality rate of 45, 50, and 45 % for T. confusum, S. zeamais, and R. dominica, respectively. Fungal infection with P. lilacinum was confirmed for all dead insects in each treatment.

Table 3 LT50 and mortality percentage for Tribolium confusum, Sitophilus zeamais and Rhyzopertha dominica adults exposed during 20 days to 20 isolates of P. lilacinum
Full size table

Discussion

The methods used to isolate entomopathogenic fungi using bait insects (T. confusum) on three types of substrates, plating soil, and plating debris showed differences in growing potentially entomopathogenic fungi on semi-selective isolation media plates. Zimmermann (1986) suggests that the bait method as a standard isolation of entomopathogenic fungi is using the larvae highly susceptible Galleria mellonella (Linnaeus) (Lepidopera: Pyralidae). However, other studies showed that entomopathogenic strains more virulent were isolated more frequently from larval and adult stages of the insect to be controlled (Prior 1991; Klingen and Haukeland 2006) than with G. mellonella. All bait insects placed in maize kernels were contaminated with Fusarium. While bait insects placed in not-dried and dried soil showed contamination with Penicillium, Fusarium, and Paecilomyces. The lowest percentage corresponded to potentially entomopathogenic fungal colonies. No entomopathogenic fungus of interest was isolated from debris samples. In contrast, the higher number of possible entomopathogenic fungi like Paecilomyces and Metarhizium were obtained from soil samples. Thus, data suggest that these types of fungus are not distributed equally in different components of maize agroecosystem. Our results are similar to Demirci and Denizhan (2010) who while evaluating a potential biocontrol agent on apple rust mite showed that much of the fungi isolated from mite cadavers were saprophytic or common contaminants such as Rhizopus spp. and Penicillium spp., and Paecilomyces was the only one genera isolated known to be entomopathogenic.

Soil is the natural habitat of entomopathogenic fungi (Hajek 1997) and studies have shown differences in the relative abundance present in different soils like the agricultural use and soils adjacent to crops and forest (Bidochka et al. 1998; Keller et al. 2003). Many of the species belonging to the Phylum Ascomycota, such as Beauveria spp., Metarhizium anisopliae, and Paecilomyces spp., develop in the soil part of their life cycle in which are outside the insect hosts (Keller and Zimmermann 1989). Jenkins and Grzywacz (2000) proposed that there are competing microorganisms that produce active metabolites that may affect the viability of fungal propagules of interest. In a previous study in which evaluated soil fungal populations of fields for the cultivation of maize in Río Cuarto, Córdoba, Argentina, it was found that the predominant genera were Aspergillus, Fusarium, and Penicillium (Nesci et al. 2006). Pereira et al. (2010, 2011) found that the count of Fusarium spp., obtained from the same field, where we take the different samples for this study, was in the order of 4 and 2 log per g of maize and soil, respectively. It is possible that the high incidence of the genus Fusarium and Penicillium in different components of maize agroecosystem have influenced the colonization of the insects bait. Most Fusarium species are saprophytic and relatively abundant members of soil microbiota (Leslie and Summerell 2006). Many species are known as plant, insects, and humans pathogens, and there are species of Fusarium pathogenic for insects and non-pathogenic for plants (Majumbar et al. 2008). Some species isolated from insect larvae and adults were reported as pathogens of these insects (Claydon and Grove 1984; Sur et al. 1999) and in other studies as an opportunistic pathogen for insects in soil (Ali-Shtayeh et al. 2002; Sun Bing et al. 2008; Abdullah and Mohamed Amin 2009).

Tuininga et al. (2009) analyzing different field samples showed important deficiencies in environmental conditions needed to maintain entomopathogenic fungi in dry leaf litter. Probably, something similar occurs with our debris samples in which no fungi of interest for this study were detected. In agreement with other authors (Asensio et al. 2003) in our study, we confirm that soil is an important reservoir of entomopathogenic fungi, potential antagonist for controlling insect pests.

Paecilomyces belongs to the family of filamentous fungi, like Aspergillus, Penicillium, and Fusarium (Humber 1997), is present in soil and decaying organic matter, and is usually recognized as an infectious agent in animals and insects (Luangsa-ard et al. 2011). In this genus, there are nematophagous species like Paecilomyces fumosoroseus (Wize) (Hoodle 2011; Rodriguez Dos Santos and del Pozo Núñez 2003) and Paecilomyces lilacinus (Borisov and Ushchekov 1997; Gökce and Er 2005; Fiedler and Sosnowska 2007). P. lilacinus was evaluated as biologic control agents against maize weevil (Sitophilus zeamais) (Ahmed 2010). P. lilacinus can be a human pathogen (Takayasu et al. 1977), but there is a large differentiation in host infection between strains of this fungus (Fiedler and Sosnowska 2007). This fungus has been found as the causal agent of infections in patients with compromised immune systems (Luangsa-ard et al. 2011). In our study, according to macro- and micromorphological characteristics, Paecilomyces isolates were identified as P. lilacinus. These characteristics are consistent with those described by Samson (1974), Samson et al. (1988), and Humber (2005). Identification of fungal cultures by DNA extraction, amplification, and sequencing showed that all isolates macro- and micromorphologically identified as P. lilacinus were Purpureocillium lilacinum. Luangsa-ard et al. (2011) show that species previously assigned to Paecilomyces like P. lilacinus was accommodated in the genus Purpureocillium.

In vitro assessment of potential entomopathogenic fungi against insect pests is one essential step in the selection of virulent strains before large scale application. Results of this study indicated that there was variability in the virulence of the 21 isolates of P. lilacinum against the three insect pest evaluated. Only one isolate, JQ926212, showed a similar virulence activity against the three insects assayed. Similar lethality levels were observed with different isolates of M. anisopliae against thrips (Sánchez-Pena et al. 2011). Rodriguez-Kabana et al. (1984) established significant differences in pathogenicity among different strains of Paecilomyces lilacinus against the nematode Meloidogyne arenaria (Chitwood). Adane et al. (1996) demonstrated that several isolates of B. bassiana, obtained from different coleopetan insects, against S. zeamais showed significant differences with respect to virulence. The isolate JQ926223 used in our study showed the highest mortality against T. confusum and R. dominica. Similar results with high mortality rates were observed with M. anisopliae against Sitophilus granaries (Linnaeus) (Khashaveh et al. 2008). Golnaz et al. (2011) showed that one isolate of B. bassiana had a LT50 of 10.45 days against S. granarius, similar to the average LT50 (10.23 days) of the 21 isolates of P. lilacinum against T. confusum evaluated in this study.

In conclusion, this study shows that the soil destined for cultivation of maize is an important reservoir of entomopathogenic fungi. The combination of methods, macro- and microscopic identification and DNA extraction and sequencing, are necessary to characterizing these fungi. The native soil P. lilacinum JQ926223 and JQ926212 will be evaluated as biologic control agents in the ecosystem of stored maize. Enhancement of the efficacy of different entomopathogens may still be required for successful use (Hallsworth and Magan 1994). These studies should be coupled with host-pathogen assays in the presence of biotic and abiotic stress factors. Major parameters for fungal growth, such as water requirements and humidity, should be examined in depth for their effect. Assays with induced changes in water stress are currently taking place for the above-mentioned isolated entomopathogenic fungi. In addition, sensitivity and tolerance assays should test the compatibility of theses entomopathogenic fungi with natural fungicides. This last point could contribute to an integrated management to reduce aflatoxins in stored grains.