SELECTION OF Beauveria bassiana AND Metarhizium anisopliae ISOLATES FOR THE CONTROL OF Blissus antillus ( Hemiptera : Lygaeidae )

Chinch bugs of the genus Blissus (Hemiptera: Lygaeidae) are serious pests of maize, wheat, sorghum and various grasses. This study is the first to screen isolates of Beauveria bassiana and Metarhizium anisopliae against 4 instar and adult Blissus antillus aiming to develop a biological control program for this important pasture pest. Ten fungal isolates were initially screened and three isolates were chosen for further investigation. To determine virulence, insects were inoculated by immersion in concentrations of 5 × 10 conidia mL. Mortality was evaluated for 10 days. B. bassiana ARSEF 792 was the most virulent isolate to both nymphs and adults, causing 53 and 78% infection, respectively, and values for LT 50 of 7.8 and 5.0 days, respectively. Germination studies were carried out to confirm viability and determine speed of germination as a pathogenicity factor. The production of conidia on the cadavers of insects infected with the three selected isolates was determined. The production of conidia on rice media was also evaluated. B. bassiana CG 24 produced the highest number of conidia on insects cadavers (14.9 × 10 conidia per insect) and also on rice media (10.6 × 10 conidia per g).


INTRODUCTION
Chinch bugs of the genus Blissus (Hemiptera: Lygaeidae) are serious pests of maize, wheat, sorghum and various grasses in the USA (Spike et al., 1994).In Brazil, a chinch bug recently identified as Blissus antillus (Leonard), causes serious damage to pastures planted with Brachiaria arrecta (T.Dur and Schinz) (tanner grass) and tangola grass (hybrid of B. arrecta and Brachiaria angola), in the states of Minas Gerais, Mato Grosso do Sul and Rio de Janeiro (Valério et al., 2000).To date, the only control method available for these insects is the application of insecticides.However, the use of insecticides on pasturelands presents restrictions, such as the need for removal of grazing cattle during chemical applications and observing a latency period, when the pasture cannot be used.
Biological control of chinch bugs using entomopathogenic fungi may provide an alternative to chemical agents.The possibility of using entomopathogenic fungi against chinch bugs is supported by reports of naturally-occurring epizootics caused by B. bassiana in populations of B. leucopterus (Ramoska, 1984), and by the fact that these insects were found to be highly susceptible to fungal infections under laboratory conditions (Krueger et al., 1991).Fungi are capable of infecting all phases of the insect's life cycle and a previous study has shown that eggs of B. antillus are highly susceptible to isolates of Metarhizium anisopliae (Metsch.)Sorok (Samuels et al., 2002).This work investigates for the first time the pathogenicity and virulence of isolates of B. bassiana and M. anisopliae to nymphs and adults of B. antillus, with the aim of developing a biological control strategy for this pest.

METERIAL AND METHODS
Blissus antillus (adults and nymphs) were collected in pastures of tanner grass (B.arrecta) in the State of Rio de Janeiro, Brazil and maintained in the laboratory under controlled conditions (25o C, 70 ± 10 % RH and 14L:10D).Insects were fed on tanner grass plantlets grown from cuttings of plants collected in the field onto which the adult females laid their eggs.Grass stems were changed every two days and at that time, the eggs were carefully collected using a fine brush.Eggs were maintained in Petri dishes with damp cotton wool to avoid desiccation.On hatching (approx.15 days), 1 st instar insects were transferred to plastic pots with tanner grass stems.When insects reached the 3 rd instar they were transferred to glass jars with plantlets of B. arrecta, where they were maintained until adult.Two-day-old 4 th instar and 3-5day-old adult female B. antillus were used in the experiments described below.
Isolates of M. anisopliae and B. bassiana were obtained from the collection at CENARGEN-EMBRAPA in Brasília, ESALQ in Piracicaba (São Paulo), USDA-ARS in Ithaca NY, or the Laboratory of Plant Protection (LPP-UENF) in Campos, Rio de Janeiro (Table 1).Fungi were grown on Sabouraud dextrose agar (dextrose 10 g; peptone 2.5 g; yeast extract 2.5 g; agar 20 g; 1L H 2 0) at 27°C for 15 days.Fungal suspensions were prepared by scrapping conidia from the surface of 15 day-old cultures using a sterile loop and added to Tween 80 (0.5% in ster-ile distilled water).Conidial concentration was determined using a Neubauer chamber.The viability of all isolates was confirmed before use in experiments using standard techniques (Goettel & Inglis, 1997) and was at all times greater than 90%.
Insects (4 th instar insects and adult females) were inoculated with fungal suspensions at a concentration of 5 × 10 8 conidia mL -1 .Ten fungal isolates were screened for pathogenicity but only three isolates where selected for more detailed study: B. bassiana ARSEF 792; B. bassiana CG 24; M. anisopliae ESALQ 818.Thirty nymphs or 30 adults were chosen and anaesthetised using a stream of CO 2 for 5 sec, and groups of five individuals were then placed in a glass funnel lined with Whatman #1 filter paper and inoculated with 1 mL of the fungal suspension.Insects were then transferred to small plastic pots lined with damp filter paper and containing tanner grass stems.Pots were maintained at >90% plus relative humidity for 48h in an incubator at 25°C and 14L:10D photoperiod.Control treatments received 0.5% Tween 80 .After 48h, treated insects were transferred to pots maintained at 70 ± 10 % RH; plant stems were changed every 2 days.Insect mortality was monitored daily and dead insects were surface sterilized (1 min in 92% ethanol followed by 2 rinses in sterile distilled water) and then transferred to Petri dishes lined with damp filter paper.Petri dishes containing the dead insects were maintained in a humidity chamber at room temperature (>90% RH; 25°C).Insects were confirmed to have been infected by the entomopathogenic fungi only following verification of conidiogenesis.All experiments were repeated three times and Abbot´s formula (Abbot, 1925) was applied to the results for mortality (control mortality was always less than 10%).Results for percentage mortality were transformed using arcsine-squareroot to normalise distributions.Results were analysed using Analysis of Variance followed by Tukey's test.LT 50 was calculated using Probit Analysis (SPSS 10).The rate of germination of the three most virulent isolates selected from the preliminary study was determined.Small, plastic Petri dishes (6 cm × 1.5 cm) with SDA were inoculated with fungal suspensions (5 × 10 6 conidia mL -1 in 0.5% Tween 80).The cultures were maintained at 27°C in the dark and germination was evaluated every 2h for a total period of 24h.The Petri dishes were removed from the incubator to observe germination by placing the Petri dishes directly under an optical microscope.Following evaluation the dishes were returned to the incubator.One hundred conidia were counted in each field of view and three fields of view were chosen from each Petri dish (three Petri dishes per treatment).The criteria for conidial germination was the presence of a germ tube which was equal or greater in size than the conidia itself.The experiment was repeated five times to enable calculating of mean and standard deviation.The results were analysed using Analysis of Variance followed by Tukey's test (P = 0.05).
Differences in the production of conidia on the insect cuticle were evaluated 20 days following initial exposure of adult females to three selected fungal isolates.Five insects infected by each isolate were evaluated individually by immersion in 1 mL of Tween 80 (0.5%) in a 1.5 mL Eppendorf tube.The tube was then vortex mixed and the resulting suspension evaluated for conidial concentration using a Neubauer chamber.The experiment was repeated three times.
The production of conidia on rice media was determined 15 days after fungal inoculation of flasks (250 mL) containing 25 g of sterilized rice, as described by Alves (1998); three flasks were used per isolate.Fungal suspensions, 1.25 mL of 5 × 10 6 conidia mL -1 in 0.5% Tween 80, were added to each flask, and flasks were incubated at 27°C in the dark.One gram of rice was sampled from each flask.Three flasks were sampled three times for each isolate.Tween 80 (5 mL of 0.5%) was added to each 1-g rice sample in a test tube, which was subsequently vortex mixed for 1 min.Concentration of conidia in the suspensions was determined as previously described.

RESULTS AND DISCUSSION
Only three of ten isolates tested were considered to be virulent (>40% adult mortality) (Table 2).These three isolates were selected for further screening of the pathogenicity to 4 th instar and adult insects, and studies on production of conidia.
B. bassiana ARSEF 792 was the most virulent of the three isolates, causing 53.3 % mortality of nymphs (Table 2).B. bassiana ARSEF 792, was originally obtained from naturally infected B. leucopterus, and has been previously used for laboratory bioassays (Ramoska, 1984;Ramoska & Todd, 1985).The second most virulent isolate was M. anisopliae ESALQ 818, which caused 35.5 % mortality.Differences were observed in the rate of mortality caused by all three isolates (P < 0.05).However, another naturally pathogenic isolate of B. bassiana from B. leucopterus, ARSEF 2430, was considered to be of low virulence to B. antillus (Table 2).B. bassiana ARSEF 792 was again the most virulent isolate against adult female chinch bugs, causing higher mortality than the two other isolates (P < 0.05).In these experiments, there was no difference between CG 24 and ESALQ 818, when mortality caused by fungal infection was compared (P < 0.05).A large range of con-  Values followed by the same small letter within the same column were not different using Tukey's test (P < 0.05%).Values followed by the same capital letter were not different when comparing mortality between nymphs and adults infected with the same isolate.The results were the mean of three repetitions for each isolate.Thirty 4 th instar nymphs and adults were used for each treatment/ repetition.Results for mortality were calculated using Abbots formula.TL 50 was calculated using Probit analysis.CL= upper and lower confidence limits.-data not obtained firmed mortality was also seen when screening the red mite Oligonychus yothersi (Acari: Tetranychidae) against different isolates of B. bassiana (Oliveira et al., 2002) Results herein shown for mortality are difficult to compare with those of previous studies on Blissus because of differences in infection techniques.In previous studies, insects were infected by direct contact with dry spores for long periods of time (Ramoska, 1984;Kruger et al., 1991).Using direct contact, Ramoska (1984) achieved 100% infection of B. leucopterus at a concentration of 4.4 × 10 6 viable conidia per insect.

Germination studies
Germination was first observed between 6 and 8h following inoculation of media.The highest percentage of germinating conidia at this time was seen in cultures of ESALQ 818 (Table 3).Conidia of B. bassiana ARSEF 792 and CG24 were slower to initiate germination.However, at 14h, 90% of the conidia had germinated in cultures of all three isolates.
The rate of germination is an important factor in the infection process, with faster germinating conidia generally considered to have a greater chance of causing infection.However, germination rates on the cuticle may not reflect those seen on artificial media because of the presence of anti-fungal compounds associated with the cuticle (Charnley, 1989).In this work, the most virulent isolate, B. bassiana ARSEF 792, was not in fact the fastest to germinate.
The production of secondary inoculum on infected insects is considered to be important in the spread of infection throughout a population (Carruthers & Soper, 1987).Conidial production on cadavers of infected, adult female insects was quantified for the three isolates (Table 4), and there were differences among the three isolates (F = 10.80;df = 2,12; P < 0.05).B. bassiana CG 24 produced the highest number of conidia, 14.88 × 10 7 conidia per insect.The same isolate (CG 24) produced 7.71 × 10 7 conidia per insect when infecting Triatoma infestans at 20ºC (Luz et al., 1999).
Because the production of microbial insecticides based on entomopathogenic fungi in Brazil is routinely carried out using solid substrates such as rice, the production of conidia of the three isolates on this substrate was evaluated.Conidial production on rice grains differed among isolates (F = 59.68; df = 2,6; P < 0.0001).B. bassiana CG 24 produced the highest number of conidia, 106.5 × 10 8 conidia g -1 (Table 4).This result was different (P < 0.05) from the production of conidia by B. bassiana ARSEF 792 and M. anisopliae ESALQ 818.
The lower production of conidia on rice by B. bassiana ARSEF 792 indicates the need for further studies to optimise conditions for sporulation of this isolate if it is to be further developed as a microbial control agent.Influence of environmental factors on the virulence and viability of the three candidate isolates is under evaluation.

Table 1 -
Details of the isolates of entomopathogenic fungi used in tests for pathogenicity and virulence against Blissus antillus.

Table 2 -
Mean percentage confirmed mortality and values for TL 50 of 4 th instar and adult female Blissus antillus inoculated by immersion in conidial suspensions (5 × 10 8 conidia mL -1 ), assessed ten days following inoculation.

Table 4 -
Conidial production on the body surface of Blissus antillus adults and on rice media by isolates of Beauveria bassiana and Metarhizium anisopliae.Values followed by different letters in the same column were different (P < 0.05) using Tukey's Test.The data represent the means of five repetitions. *