Pathogenicity of the Entomopathogenic FungiPaecilomycesspp. andBeauveria bassianaagainst the Silverleaf Whitefly,Bemisia argentifolii

JOURNAL OF INVERTEBRATE PATHOLOGY ARTICLE NO.

71, 217–226 (1998)

IN974734

Pathogenicity of the Entomopathogenic Fungi Paecilomyces spp. and Beauveria bassiana against the Silverleaf Whitefly, Bemisia argentifolii S. P. Wraight,*,1 R. I. Carruthers,† C. A. Bradley,‡ S. T. Jaronski,‡ L. A. Lacey,§ P. Wood,‡ and S. Galaini-Wraight‡ *USDA-ARS-Subtropical Agricultural Research Laboratory, Weslaco, Texas 78596; †USDA, National Program Staff, BARC-West, Beltsville, Maryland 20705; ‡Mycotech Corporation, Butte, Montana 59702; and §USDA-ARS, Yakima Agricultural Research Laboratory, Yakima, Washington 98902 Received February 28, 1997; accepted November 5, 1997

INTRODUCTION

Pathogenicities of three species of entomopathogenic fungi against preimaginal Bemisia argentifolii were measured and compared. Third-instar nymphs on excised leaves of Hibiscus rosa-sinensis were exposed to spray applications of 14 isolates of Beauveria bassiana, 22 isolates of Paecilomyces fumosoroseus, and five isolates of Paecilomyces farinosus. B. bassiana and P. fumosoroseus isolates of diverse origins were highly pathogenic to the whitefly nymphs; median lethal doses of 14 of the 22 P. fumosoroseus and four of the 13 B. bassiana isolates ranged between 50 and 150 conidia/mm2. Five isolates of P. farinosus were also pathogenic; however, LC50s were relatively high, ranging between 350 and 4000 conidia/mm2. Nymphs infected with all but one isolate of B. bassiana displayed a pronounced red pigmentation. Postmortem hyphal growth and sporulation of B. bassiana was relatively slow and usually confined to the region immediately surrounding the dead host. Whitefly nymphs patently infected with P. fumosoroseus and P. farinosus were lightly pigmented yellow or orange. Postmortem hyphal growth and sporulation of P. fumosoroseus rapidly covered the dead host and extended several millimeters onto the surrounding leaf surface. The results indicate that highly virulent strains of P. fumosoroseus and B. bassiana with considerable whitefly control potential are widespread and numerous. r 1998 Academic Press

Key Words: Beauveria bassiana; Paecilomyces fumosoroseus; Paecilomyces farinosus; conidia; pathogenicity of; whitefly; Bemisia argentifolii; nymphs; biosassay; biological control.

1 To whom correspondence and reprint requests should be addressed at present address: USDA-ARS, U.S. Plant, Soil, and Nutrition Laboratory, Tower Road, Ithaca, NY 14853.

Entomopathogenic deuteromycete fungi of the genera Beauveria and Paecilomyces have been recognized as important biocontrol agents of aleyrodid pests of field and greenhouse crops for more than 20 years (Nene, 1973; Borisov and Vinokurova, 1983; Treifi, 1984; Fang et al., 1986; Osborne and Landa, 1992; Wright, 1992; Carruthers et al., 1993; Garza and Arredondo, 1993; Eyal et al., 1994). However, few studies investigating the whitefly-control potential of these pathogens were initiated in the U.S. until the last decade when a highly aggressive strain of Bemisia tabaci of unknown origin identified as strain B appeared in the country. Researchers in Florida, where the new pest was initially identified, isolated a highly virulent strain of Paecilomyces fumosoroseus from a naturally infected mealybug (Osborne and Landa 1992). The biological control potential of this pathogen was considered sufficiently great that a patent on the use of P. fumosoroseus strains exhibiting the virulence, agrochemical resistance, and identifying characteristics of strain ATCC 20874 (Pfr 97) was applied for and issued in 1990 (Osborne, 1990). Morphological, biochemical, genetic, and behavioral studies conducted on whiteflies collected in the southern U.S. prior to 1986 and after subsequent outbreaks suggested that the unprecedented pest problem was the result of an invasion by a new species ultimately described as the silverleaf whitefly, Bemisia argentifolii (Bellows et al., 1994). Experience with other species of whiteflies and evaluation of their population dynamics strongly suggested that biological control might provide the greatest capacity to reduce damage caused by B. argentifolii in both greenhouse and field crops (van Lenteren and Woets, 1988; Baumgartner and Yano, 1990; Gerling, 1990). Consequently, extensive surveys for natural enemies of whiteflies were pursued. These

217

0022-2011/98 $25.00 Copyright r 1998 by Academic Press All rights of reproduction in any form reserved.

218

WRAIGHT ET AL.

explorations revealed that P. fumosoroseus, the pathogen found in Florida, is also one of the most ubiquitous and important pathogens of Bemisia spp. worldwide (Lacey et al., 1993; Kirk et al., 1993; Garza and Arredondo, 1993; T. Poprawski, personal communication). Nene (1973), Humber (1992), and Carruthers et al. (1993) identify Paecilomyces farinosus as an important naturally occurring pathogen of B. tabaci in India and the U.S. However, the identity of the Indian isolates is questionable in light of the recent discoveries in India and throughout Asia of extensive epizootics of Paecilomyces fumosoroseus (Lacey et al., 1993). Furthermore, the U.S. isolates from Texas and California (Humber, 1992) have been reidentified (by S.P.W.) as P. fumosoroseus based on the size and color of the conidia from in vitro cultures (measuring 1–2 3 3–5 µm and colored varying shades of pink to mauve, turning gray with age). Beauveria bassiana is not a natural regulator of whitefly populations and is notably absent from the recent listings of fungal isolates from Bemisia spp. (see Fransen, 1990; Humber, 1992). Nevertheless, surveys have discovered Bemisia individuals incidentally infected with this fungus (T. Poprawski, personal communication), and laboratory and field studies have revealed it to be an excellent pathogen of B. argentifolii when applied directly as a concentrated spore suspension (Wright, 1992; Carruthers et al., 1993; Garza and Arredondo, 1993; Eyal et al., 1994). In 1991, USDA scientists studying the microbial control potential of B. bassiana for boll weevil control discovered that the strain under investigation was also pathogenic to Bemisia (Wright, 1992) and began development of a product comprising a mixture of B. Bassiana (ATCC 70490) conidia, vegetable oil, and other plant-derived ingredients claimed to possess arrestant and feeding stimulant properties. This biopesticide formulation, was patented in 1995 and has been shown to reduce B. argentifolii populations in various crops (Wright and Chandler, 1995). Because the aforementioned P. fumosoroseus and B. bassiana strains being developed for whitefly control in the U.S. were discovered more or less fortuitously, the collaborative work reported here was undertaken with the goal of identifying new isolates of these pathogens possessing maximum virulence against B. argentifolii. Several isolates of P. farinosus of diverse origin were also screened to verify pathogenicity of this species against B. argentifolii. MATERIALS AND METHODS

Test Insects Leaves of Hibiscus rosa-sinensis infested with 15-dayold nymphs of B. argentifolii were obtained regularly

from the USDA-APHIS-PPQ Biological Control Laboratory in Mission, Texas, between the dates of November 1992 and May 1994. The colony originated from whiteflies collected in Arizona on cotton in 1991, and was maintained under simulated natural conditions including gradually increasing and decreasing light intensities within a 17-h photoperiod and daily temperature fluctuations between 21 and 27°C. Fungal Isolates The isolates of P. fumosoroseus selected for screening originated from diverse sites in North America, Asia, and Europe (Table 1); all except one were isolated from Bemisia spp. The B. bassiana isolates originated from nonwhitefly hosts collected in the United States; however, most were reisolated from B. argentifolii nymphs infected in the laboratory (Table 1). Also screened were one European and four North American isolates of P. farinosus from thysanopteran and lepidopteran hosts. Fungal Preparations All assays utilized dry conidia from solid substrate fermentations conducted by Mycotech Corp., using proprietary methods and ingredients. Spore preparations were received in the form of unformulated (technical) powders containing between 5 3 1010 and 1.5 3 1011 conidia/g with .85% viability. These were stored in sealed containers at approx. 4°C. Test insects were sprayed with conidia suspended in 0.01% aqueous Tween 80. Suspensions were prepared by introducing 5–25 mg of technical powder into 10 ml of Tween solution, adding 1 g glass beads (2 mm diam.) and shaking vigorously by wrist action for 2 min. All suspensions were used within a few hours of preparation. Bioassay Protocol The basic measure of virulence used in this study was the estimated dose required to kill 50% of the test insects (LC50 expressed as conidia/mm2 ). Each individual assay of a selected fungal strain comprised three doses of conidia, each applied to two replicate leaves infested with whitefly nymphs. Spore deposits targeted for the high, medium, and low doses were 500–1000, 100–200, and 20–40 spores/mm2, respectively. Five candidate isolates were usually assayed in a single trial conducted on a given day. Each trial also included six control leaves sprayed with a solution of 0.01% Tween. A total of 40 trials were conducted over the course of the study. Most isolates were assayed on two or more occasions using different generations of whiteflies. Because new isolates and spore preparations were made available for screening at various times, it was not possible to employ an experimental design incorporating complete randomization and blocking of the indi-

219

PATHOGENICITY OF Paecilomyces spp. AND B. bassiana

TABLE 1 Identification of Fungal Isolates Assayed against Bemisia argentifolii Species/isolate (Holding) a Paecilomyces fumosoroseus 2429 (ARSEF) 3083 (ARSEF) 3313 (ARSEF) 3322 (ARSEF) 3572 (ARSEF) 3581 (ARSEF) 3594 (ARSEF) 3658 (ARSEF) 3660 (ARSEF) 3663 (ARSEF) 3699 (ARSEF) 3700 (ARSEF) 20874 (ATCC) 612 (Mycotech) 613 (Mycotech) 614 (Mycotech) 10 (SARC) 36 (SARC) 39 (SARC) 40 (SARC) 42 (SARC) 43 (SARC) Paecilomyces farinosus 3 (ARSEF) 1225 (ARSEF) 3522 (ARSEF) 3526 (ARSEF) 3564 (ARSEF) Beauveria bassiana 252 (ARSEF) 3543 (ARSEF) 74040 (ATCC) 648 (Mycotech) 654 (Mycotech) 655 (Mycotech) 656 (Mycotech) 657 (Mycotech) 658 (Mycotech) 659 (Mycotech) 660 (Mycotech) GHA (Mycotech) NC-3 (Mycotech) MLC8 (UVM)

Insect host of origin

Site and date of origin

Niliparvata lugens Bemisia tabacib Bemisia sp. Bemisia tabacib Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia tabaci Bemisia tabaci Pseudococcus sp. Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia argentifolii Bemisia tabaci (ARSEF 4500) c Bemisia tabaci (ARSEF 3843) c Bemisia tabaci (ARSEF 4495) c Bemisia tabaci (ARSEF 3871) c Bemisia tabaci (ARSEF 4502) c

Cikampek, Java, Indonesia, 1987 Alva, Florida, 1990, Tecoma´n, Colima, Mexico, 1990 Fort Meyer, Florida, 1990 McAllen, Texas, 1992 McAllen, Texas, 1992 McAllen, Texas, 1992 El Centro, California, 1992 Calexico, California, 1992 Calexico, California, 1992 Padappai, India, 1992 Padappai, India, 1992 Apopka, Florida, 1989 Weslaco, Texas, 1993 Weslaco, Texas, 1993 Weslaco, Texas, 1993 Santa Maria, Texas, 1992 Multan, Pakistan, 1992 Padappai, India, 1992 Padappai, India, 1992 Kathmandu, Nepal, 1993 Multan, Pakistan, 1992

Malacosoma americanum Lespeyresia medicagicus Lymantria dispar Lymantria dispar Taeniothrips inconsequens

Dennis, Massachusetts, 1972 Lusignan, Vienne, France, 1981 Quarry Run, West Virginia, 1991 Piclic Run, Maryland, 1991 Warren, Vermont, 1991

Leptinotarsa decemlineata Soil, Galleria mellonella (trap host) —d Leptinotarsa decemlineata (ARSEF 252) c Diuraphis noxia (ARSEF 2864) c Diuraphis noxia (ARSEF 2879) c Schizaphis graminum (ARSEF 2881) c Diuraphis noxia (ARSEF 2882) c Schizaphis graminum (ARSEF 2336) c Thrips calcaratus (ARSEF 3216) c Myzus persicae (ARSEF 3385) c —d Alphitobius diaperinus Paraclemensia acerifoliella

Orono, Maine, 1978 Warren, Vermont, 1991 —d Orono, Maine, 1978 Parma, Idaho, 1988 Parma, Idaho, 1988 Parma, Idaho, 1988 Parma, Idaho, 1988 Parma, Idaho, 1986 Rusk County, Wisconsin, 1991 Yakima, Washington, 1991 —d North Carolina, 1993 Randolf, Vermont, 1993

aARSEF, U.S. Department of Agriculture, Agricultural Research Service, Collection of Entomopathogenic Fungus Cultures, Ithaca, NY; ATCC, American Type Culture Collection, Rockville, MD; Mycotech, Mycotech Corp., Butte, MT; UVM, University of Vermont, South Burlington, VT; SARC, U.S. Department of Agriculture, Agricultural Research Service, Subtropical Agricultural Research Center, Weslaco, TX. bProbably Bemisia argentifolii (5Bemisia tabaci strain B) cStrain passed through and reisolated from Bemisia argentifolii in the laboratory. dStrain under commercial development, information restricted.

vidual assays over time. Between-assay comparisons were therefore provided for by inclusion of a standard fungus in each trial. B. bassiana was selected as the standard in light of Mycotech’s considerable experience in producing highly stable, dry-conidia preparations of this pathogen. Isolate ARSEF 252 was selected as the specific standard because it has an extensive history of experimental biocontrol applications (Campbell et al.,

1985; Feng et al., 1985; Hajek et al., 1987; Wright and Chandler, 1992). A standard technical powder of ARSEF 252 was produced in October 1992. The small size and sedentary habit of immature whiteflies makes them difficult subjects for laboratory assay, and a considerable effort was initially invested in methods development. There follows a detailed description of our standard bioassay protocol.

220

WRAIGHT ET AL.

Immediately upon excision, the petioles of the selected hibiscus leaves infested with whitefly nymphs were embedded in water-saturated cotton contained within a covered 35-mm-diameter plastic petri dish (petiole inserted through a small hole in the side of the dish) affixed to the bottom of a large 150 3 15-mm plastic petri dish (hereafter referred to as the incubation dish). The leaves were maintained ventral side up to facilitate microscopic observations and to reduce leaf surface humidity (see below). Fungal applications to individual leaves were made using a Potter Precision Laboratory Spray Tower (Berkard Manufacturing Co. Ltd, Hertfordshire, England). One-milliliter aliquots of spore suspension were sprayed at 10 psi (0.7 kg/cm2 ) using the fine nozzle (orifice diameter 0.25 mm) provided with the tower. This deposited approximately 0.01-µl spore suspension per square mm of target surface. The three doses of conidia of each candidate fungus were sprayed in ascending order. From the initial high dose suspension, a series of two fivefold dilutions was prepared, representing medium and low doses. In making each application, a leaf was removed from its incubation dish and placed on a holding platform made from the lid of a standard 10-cm-diameter plastic petri dish. Two small blocks of 1.5% water agar (approximately 5 mm deep 3 5 mm wide 3 20 mm long) were placed along opposite sides of the leaf to collect spore deposit samples. The holding platform with the leaf was then centered on the target surface of the spray tower. Finally the test spore suspension was agitated and the 1-ml aliquot was drawn and sprayed. Immediately after spraying, the holding platform was removed from the tower and the leaf was returned to its incubation dish (ventral surface up as before). A piece of absorbent paper was placed beneath the leaf and saturated with water. The dish was then covered, sealed in a plastic bag, and incubated at 25°C under a 16-h daily photoperiod. The agar blocks were collected and stored in sealed dishes at 4°C until processing. The spray tower was cleaned with 80% ethanol after the applications of each fungus strain. Postapplication Monitoring The treated leaf was incubated in the water-saturated environment of the plastic bag for 24 h. After this time, the dish was removed from the bag and placed open on the bench to allow the leaf to dry. Test insects were then carefully selected to form a group as uniformly aged as possible. Although we use only lightly infested leaves for treatment, there were usually many more insects than required for the assay, and these were of mixed instars. Therefore, each leaf was scanned under a stereoscope, and 40–50 mid-third-instars were selected for monitoring by marking a small spot on the leaf near each nymph using a fine, black technical pen.

After selection was completed, the dish was covered with a lid having an 8-cm-diameter hole covered with fine screen (0.02 mm mesh openings). If the ventilated lid was not used (or if the leaf was incubated in a natural position with the dorsal surface up), high humidity supported rapid saprophytic growth of the fungi on honeydew excreted by the whiteflies and on leaf exudates; this growth produced large numbers of conidia representing an unquantifiable secondary inoculum. Humidity sensors (Phys-Chem Scientific Corp., New York, NY) situated directly above the leaves indicated that relative humidities of 25–30% prevailed inside the ventilated dishes after they were returned to the 25°C incubator. Infected whitefly nymphs began to die 3 days after application, usually as fourth instars. Under the dry conditions of the incubation dishes, the killed insects desiccated quickly, and the fungus did not grow out of the host. When desired, dead insects were removed from the leaf using a fine needle dipped in 50% glycerin solution (sticker) and placed on water agar (1.5%) for evaluation of fungal outgrowth and sporulation capacity. The ventilated incubation dishes were monitored for 8–9 days. Some of the hibiscus leaves wilted and turned chlorotic within 4–5 days; however, nymphs that reached the fourth instar usually survived on the senescing leaves. Emerged adults (indicated by the remaining exuviae) were counted as having survived the treatment. No attempt was made to monitor survivorship of emerging adults. Dose Quantification After spraying the leaves with the medium dose, a plate of Sabouraud dextrose agar with 1% yeast extract (SDAY) was sprayed with an additional aliquot for viability assessment. The SDAY plate was incubated at 25°C for approximately 16 h and then 100 conidia from each of three areas of the plate were observed at 4003 magnification and scored for germination. The dose applied to each leaf was estimated by counting (at 4003 magnification) conidia deposited on the surfaces of the agar sample blocks described above. The following is the only practical method we have found to rapidly quantify the broad range of spore densities encountered on the large number of sample blocks. The surface area scanned is varied with spore density; areas of approximately 3.25, 0.50, and 0.125 mm2 are counted at the low, medium, and high doses, respectively. Each area is represented by the path (swath) traced by a measured line on the microscope eyepiece reticle viewed against the spore-contaminated substrate as it is slowly moved a distance of 10 mm on the microscope stage. The area counted for the high dose is thus a narrow rectangle 0.0125 mm wide by 10 mm long. One count is made on each of the two agar

PATHOGENICITY OF Paecilomyces spp. AND B. bassiana

blocks sprayed with each leaf. If the coefficient of variation of the two counts exceeds 50%, an additional count is made on each block to improve the estimate. All spore counts are ultimately adjusted for percentage of viability and expressed on a per mm2 basis. Using this technique in ancillary studies, we have been able to measure doses from a few conidia to several thousand conidia/mm2. Statistical Analysis

221

conditions (Mycotech Corp., unpublished data). Dry technical spore powders of both species were highly stable under 4°C storage. For example, viabilities of P. fumosoroseus isolate 3663 Lot No. 921005 and B. bassiana isolate 252 Lot No. 921020 spores were estimated at 87 and 94%, respectively, in November 1992 and were not significantly different when last measured in September 1994 (89 and 95%, respectively). Pathogenicity

Median lethal doses (LC50s) were estimated using the personal computer version of the POLO program developed by Russell et al. (1977). Dose responses were corrected for control mortality by the POLO program. Within-assay potency ratios were calculated by dividing the LC50 of the standard isolate by the LC50 of each test isolate. Other regressions, ANOVAs, and nonparametric tests were conducted using statistical analysis software by Wilkinson et al. (1992). RESULTS

Fungus Production Few difficulties were encountered in producing spore numbers adequate for the bioassays. However, spore yields of B. bassiana were approximately fourfold greater per unit of solid-substrate fermentation than those of Paecilomyces spp. under pilot-scale production

B. bassiana and P. fumosoroseus were highly pathogenic to B. argentifolii nymphs; median lethal doses of many isolates (14 of the 22 P. fumosoroseus and four of the 13 B. bassiana isolates tested) ranged between 50 and 150 conidia/mm2 (Fig. 1). Among the four most pathogenic B. bassiana isolates were two originally isolated from nonwhitefly hosts and not reisolated from B. argentifolii (252 and MLC8). All five of the P. farinosus isolates were pathogenic, but the LC50s were relatively high, ranging between approximately 350 and 4000 conidia/mm2 (Fig. 1). Control mortality over all assays averaged 10.2% and ranged from 2.0 to 26.9% (control mortality exceeded 20% in only two of the 40 trials). Regression coefficients from all assays were low, with mean values for the individual isolates varying from 0.5 to 2.0. The unweighted grand mean slope (6SE) from regressions of the B. bassiana isolates was equal

FIG. 1. Mean LC50s (6SE) of 14 isolates of Beauveria bassiana, 22 isolates of Paecilomyces fumosoroseus, and five isolates of Paecilomyces farinosus applied against third-instar nymphs of Bemisia argentifolii.

222

WRAIGHT ET AL.

to that from regressions of the P. fumosoroseus isolates, viz., 1.09 6 0.119 vs 1.16 6 0.048, respectively. The mean slope from the P. farinosus isolates was lower (0.77 6 0.083); however, in the single trial of the five P. farinosus isolates, the slopes did not differ significantly from that of the B. bassiana standard (chi-square 5 3.0; 5 df ). All of the B. bassiana isolates except NC-3 produced an unidentified pigment within the host hemocoel that colored most infected nymphs various intensities of red, making infected individuals readily identifiable by the unaided eye against the dark green background of the hibiscus leaf. The color was most evident in the larger third and fourth-instar nymphs (Fig. 2). The red pigment persisted after death and desiccation of the host; but when moisture levels increased, the color rapidly faded to white as the fungus resumed hyphal growth inside the killed host and then emerged and underwent

sporulation. When dried cadavers were removed from the leaves and placed on water agar, fungal outgrowth and initial sporulation was usually observed within 24 h at 25°C. The initial external growth of B. bassiana arose primarily from the pleural regions of the dead host and formed a characteristic white circle of dense sporulation (Fig. 3). Colonization of the substrate surrounding the nymph (water agar or leaf surface) was generally slow and not extensive. If maintained under continuous high humidity conditions, however, the sporulating growth would gradually cover the killed host and spread outward. Nymphs infected with Paecilomyces spp. exhibited no overt signs of infection until near death when they typically displayed a yellow-orange color. The intensity of coloration imparted by the fungus varied greatly among isolates, being most pronounced in nymphs infected with orange-pigmented isolates of P. farinosus.

FIGS. 2–5. (2) Healthy fourth-instar nymph of B. argentifolii compared to one infected with B. bassiana strain GHA; note nymph on right side of photograph with dark (red) pigmentation produced by infection. (3) Fourth-instar nymph of B. argentifolii encircled by sporulating growth of B. bassiana strain GHA (photograph taken after 48 h of incubation at 20°C, 100% RH); note limited growth of fungus onto surrounding leaf surface. (4) Healthy fourth-instar nymph of B. argentifolii compared to one infected with P. fumosoroseus isolate 612; note nymph on right side of photograph with internal white masses of fungus. (5) Fourth-instar nymph of B. argentifolii overgrown with sporulating hyphae of P. fumosoroseus isolate 612 (photograph taken after 48 h of incubation at 20°C, 100% RH); note extensive growth of fungus onto surrounding leaf surface. All figures are 325.

PATHOGENICITY OF Paecilomyces spp. AND B. bassiana

In most cases, the color was not intense enough to distinguish the mycosis from other causes of mortality. However, at the time of death of third- and fourthinstar nymphs, the large amount of fungus present in the hemocoel usually become visible as white, irregularly shaped patches when the hemolymph dried and the body wall collapsed around it (Fig. 4). Like B. bassiana, the Paecilomyces spp. rapidly underwent external hyphal development and sporulation under moist conditions. Paecilomyces fumosoroseus initially emerged from the anal region of the killed host and then quickly covered the cadaver with diffuse hyphal growth and sporulation. In most cases, the fungus also rapidly colonized several millimeters of the surrounding substrate (Fig. 5). Relative Potencies Potencies of the 40 selected fungal isolates relative to the B. bassiana (ARSEF 252) standard are presented in Fig. 6. As in the case of the LC50 values, the potency ratios indicated similar pathogenicity of B. bassiana and P. fumosoroseus against B. argentifolii. Of the B. bassiana and P. fumosoroseus isolates examined, 23.1% (3/13) and 22.7% (5/22), respectively, exhibited potency ratios greater than one. Potency ratios varied from near zero to more than two; however, considerable heterogeneity of potency ratios was encountered in

223

many cases (as indicated by the large between-assay standard errors), and a standard ANOVA comparing the 34 isolates that were tested more than once indicated significant differences only at a 90% confidence level (P 5 0.07). Removal of the isolate with the largest standard error (ARSEF 3581) reduced the P value to 0.049. A logarithmic transformation to stabilize variance did not reduce the P value. Because ratio variables tend to be nonnormal (Anderson and Lydic, 1977), a nonparametric analysis was also conducted. A Kruskal– Wallis comparison of the 34 mean potency ratios indicated significant differences at the P 5 0.02 level. These generally high P values indicate that true differences likely exist only between isolates ranked high vs low in Fig. 6. In fact, application of the Kruskal–Wallis analysis to the top 18 isolates (those with ratios greater than approximately 0.75) shows no significant differences (P 5 0.36). The potencies of the five P. farinosus isolates were among the lowest of all the 40 isolates tested; four ranked below 35 and the most pathogenic isolate (ARSEF 3526 originally from a gypsy moth) ranked only number 31 (Fig. 6). No correlation between regression coefficient and potency was observed. For example the mean potency ratios of the 25 isolates assayed an approximately equal number of times (2–3 replicates) regressed on

FIG. 6. Whitefly nymphicidal potencies (6SE) of 13 isolates of Beauveria bassiana, 22 isolates of Paecilomyces fumosoroseus, and five isolates of Paecilomyces farinosus relative to the standard, Beauveria bassiana isolate ARSEF 252 (relative potency 5 LC50 of standard isolate divided by LC50 of test isolate).

224

WRAIGHT ET AL.

mean slope produced a nonsignificant correlation coefficient of 0.339. DISCUSSION

The strategy of using a standard isolate to generate a parameter (potency ratio) with low between-assay variance was not successful in this study, and the factors causing the high variability have not been identified. The generally random fluctuations in overall responses observed during the long course of the study might have resulted from changing susceptibility of the test insects; however, it is equally possible that an unknown element (possibly associated with the hibiscus plants) may have influenced fungal infectivity. We anticipated that difficulties might arise in measuring Paecilomyces potency using a Beauveria standard, and in fact, the weighted mean variance of the Paecilomyces potency ratios was 0.640 compared to 0.285 for Beauveria. This indicates that an unidentified factor may have affected the virulence of P. fumosoroseus more than B. bassiana (or affected the insects’ susceptibility to P. fumosoroseus more than their susceptibility to B. bassiana). While this suggests a need to include a standard isolate of each fungus species tested, limitations to the use of standards in the screening of infectious agents must be recognized. The method of using a standard in bioassay was developed for determination of relative potencies of test preparations containing the same active ingredient as the standard. If the active ingredient in a preparation of unknown potency is the same as that in the standard, the preparation will exhibit the same mode of action against the treated insects as the standard (producing equal regression coefficients), and any changing factors associated with the insects or test environment will affect the activity of the ingredient in all preparations equally (producing stable potency ratios). However, different isolates of even the same species of a pathogen may differ markedly in numerous traits; they may not produce identical pathologies in a particular host or react similarly to changes in the host or environment. Important but undetected changes occurring in a test system over the course of a long-term experiment could translate into highly variable potency ratios such as those recorded in this study. A difficult problem inherent to the bioassay of nymphal stages of whiteflies is the impracticality of randomly assigning individuals to the treatments. The fact that we were able to randomize only the relatively large groups of insects associated with each leaf certainly contributed to within-assay heterogeneity and diminished the benefit of employing an internal standard with each series of assays. An improved protocol would involve treating a greater number of replicate leaves with fewer insects per leaf. We were unable to adopt this approach because the amount of infested plant material available to us was limited.

Despite the variability of assay results, this screening study has clearly demonstrated that isolates of P. fumosoroseus and B. bassiana of diverse origin are equally pathogenic to B. argentifolii nymphs. This conclusion is supported by recent field trials in southern Texas in which isolates of B. bassiana and P. fumosoroseus from various regions of the U.S. have repeatedly produced nearly identical levels of control of whitefly nymphs infesting melons, cucumbers, and tomatoes (Wraight et al., 1996). Also, Garza and Arredondo (1993) reported statistically equal virulence of several Mexican isolates of B. bassiana and P. fumosoroseus tested in the laboratory against adult B. tabaci. Vidal et al. (1997) recently reported substantially lower LC50s of P. fumosoroseus (five isolates) against B. argentifolii than reported here (viz., 6–22 versus 50– 150 conidia/mm2 ). These lower values likely derive from differences in instar of the test insects and environmental conditions in the assay dishes; spray-inoculated, second-instar nymphs were incubated on leaves held in a natural position (dorsal surface up) in petri dishes with smaller ventilation holes (1-cm diameter) in an incubator environment maintained at higher relative humidity (60%). Virulences of B. bassiana and P. fumosoroseus against Bemisia spp. reported by other investigators are not readily compared to those presented here, because of even greater differences in assay protocol. Landa et al. (1994) report extensive bioassay of P. fumosoroseus strain ATCC 20874 (Pfr 97) against B. argentifolii nymphs. However, virulence is expressed in terms of a fungus growth development index (FGDI) rather than LC50s or rates of infection. Eyal et al. (1994) inoculated B. argentifolii nymphs by dipping for 10 s in B. bassiana and P. fumosoroseus spore suspensions and reported 52–98% mortality following treatments with concentrations of approximately 1–4 3 106 conidia/ml. Herna´ndez and Garza (1994) reported an LC50 of 4.3 3 108 conidia/ml for P. fumosoroseus isolate PF2 sprayed against B. tabaci nymphs. However, the amount of suspension sprayed was not indicated. In our assays doses of 100 spores/ mm2 are produced by spraying suspensions containing approximately 1 3 107 spores/ml, but only 0.01 µl is deposited per mm2 of leaf surface (equivalent to 100 L/ha of a planar surface). We initially planned to screen a greater number of B. bassiana isolates. However, when many isolates in the first test group proved as pathogenic as the best whitefly isolates of P. fumosoroseus, the effort was discontinued and resources were redirected into field studies. One of the most virulent fungi identified by the screening assays was B. bassiana strain GHA (Fig. 6). This strain was originally developed by Mycotech Corp. for control of grasshoppers and locusts. Its strong biological control potential derives from combined attributes of high virulence and a stable capacity for

PATHOGENICITY OF Paecilomyces spp. AND B. bassiana

efficient sporulation under industrial-scale production conditions. Following several seasons of successful field trials (to be reported elsewhere), a wettable powder formulation was developed for application against whiteflies and other homopteran pests. This product was recently registered in the U.S. under the trade name Mycotrol. The red pigment produced in the infected whitefly nymphs by the B. bassiana isolates is presumed to be oosporein. This dibenzoquinone has been reported from a variety of ascomycete and deuteromycete species and is produced by many isolates of B. bassiana and Beauveria brongniartii (5B. tenella) (El-Basyouni et al., 1968). It possesses antibiotic properties which probably aid the host infection process. Eyal et al. (1994) identified this pigment from whiteflies infected with B. bassiana isolate ATCC 74037 and discussed various aspects and implications of its production by microbes being developed for biological control. Although synthesized at highly visible levels in the whitefly hemolymph, significant amounts of this compound were not detectable in the dry spore preparations produced for this study nor in liquid cultures of this strain (Mycotech Corp., unpublished data). Although conidia of P. fumosoroseus have thus far proved more difficult to mass produce than those of B. bassiana, interest in commercial development remains strong because of the importance of this pathogen as a natural whitefly control agent with considerable epizootic potential (Lacey et al., 1993; Carruthers et al., 1993). The ability of this fungus to grow extensively over the leaf surface under humid conditions is a characteristic that certainly enhances its ability to spread rapidly through whitefly populations. Our findings confirm that P. farinosus is pathogenic to B. argentifolii, but suggest that highly virulent strains may be uncommon. More extensive testing of this fungus will be required to assess its whiteflycontrol potential. Passage through Bemisia might enhance virulence against this host; although, a single passage of the Colorado potato beetle isolate of B. bassiana, ARSEF 252, through B. argentifolii did not alter virulence of this fungus against the latter insect (reisolate identified as 648 in Fig. 6). This study reveals that virulent strains of P. fumosoroseus and B. bassiana with considerable potential for whitefly control are widespread and numerous. This suggests that national biocontrol efforts need not rely upon exotic isolates; although, continued exploration will almost certainly discover strains with unique combinations of desirable traits such as resistance to agrochemicals (especially fungicides), restricted or expanded host range, or elevated mass culture capacity. In addition, extensive international collection could provide the diverse germplasm necessary to support future genetic engineering efforts.

225

ACKNOWLEDGMENTS We gratefully acknowledge Andrew Sattler, Joel Garza, and Frank De La Garza for technical assistance with the bioassays, Dr. Leeda Wood for providing whiteflies from the Mission Biological Control Laboratory, Mary-Lynn Cummings for assisting with fungus culture maintenance, Richard Humber for providing fungal isolates from the ARSEF collection, and Michael Brownbridge for providing isolates from the University of Vermont culture collection. We thank Drs. Michael Brownbridge, Tad Poprawski, and John Vandenberg for their critical reviews of the manuscript. Trade names are necessary to report factually on available data; however, the USDA neither guarantees nor warrants the standard of the product, and the use of the name by USDA implies no approval of the product to the exclusion of others that may also be suitable. REFERENCES Anderson, D. E., and Lydic, R. 1977. On the effect of using ratios in the analysis of variance. Biobehav. Rev. 1, 225–229. Baumgartner, J., and Yano, E. 1990. Whitefly population dynamics and modelling. In ‘‘Whiteflies, Their Bionomics, Pest Status and Management’’ (D. Gerling, Ed.), pp. 123–146. Intercept Ltd., Andover, UK. Bellows, T. S., Perring, T. M., Gill, R. J., and Headrick, D. H. 1994. Description of a species of Bemisia (Homoptera, Aleyrodidae). Ann. Entomol. Soc. Am. 87, 195–206. Borisov, B. A., and Vinokurova, T. P. 1983. Increasing the effectiveness of entomopathogenic fungi. Zashch. Rast. 9, 20–22. Campbell, R. K., Anderson, T. E., Semel, M., and Roberts, D. W. 1985. Management of the Colorado potato beetle using the entomogenous fungus Beauveria bassiana. Am. Pot. J. 62, 29–37. Carruthers, R. I., Wraight, S. P., and Jones, W. A. 1993. An overview of biological control of the sweetpotato whitefly, Bemisia tabaci. In ‘‘Proceedings Beltwide Cotton Conferences’’ (D. J. Herber and D. A. Richter, Eds.), Vol. 2, pp. 680–685. National Cotton Council of America, Memphis, TN. El-Basyouni, S. H., Brewer, D., and Vining, L. C. 1968. Pigments of the genus Beauveria. Can. J. Bot. 46, 441–448. Eyal, J., Mabud, M. D. A., Fischbein, K. L., Walter, J. F., Osborne, L. S., and Landa, Z. 1994. Assessment of Beauveria bassiana Nov. EO-1 strain, which produces a red pigment for microbial control. Appl. Biochem. Biotechnol. 44, 65–80. Fang, Q. X., Yang, S. F., Hu, Y. M., Zhou, Y. Y., Gong, Y. X., and Kang, Z. J. 1986. Control of whitefly, Trialeurodes vaporariorum by Paecilomyces fumosoroseus var. beijingensis Fang et Q. T. Chen, in greenhouses. Chin. J. Biol. Control 2, 129–131. Feng, Z., Carruthers, R. I., Roberts, D. W., and Robson, D. S. 1985. Age-specific dose-mortality effects of Beauveria bassiana (Deuteromycotina, Hyphomycetes) on the European corn borer, Ostrinia nubilalis (Lepidoptera, Pyralidae). J. Invertebr. Pathol. 46, 259– 264. Fransen, J. J. 1990. Natural enemies of whiteflies, Fungi. In ‘‘Whiteflies, Their Bionomics, Pest Status and Management’’ (D. Gerling, Ed.), pp. 187–210. Intercept Ltd., Andover, UK. Garza, E., and Arredondo, H. C. 1993. Sensibilidad in vitro de adultos de mosquita blanca de las hortalizas Bemisia tabaci (Genn.) (Homoptera, Aleyrodidae) a diferentes aislamientos de Beauveria bassiana (Bals.), Paecilomyces spp. y Verticillium lecanii (Vuill.). In ‘‘XVI Congreso Nacional de Control Biolo´gico,’’ pp. 31–32. Monterrey, N. L., Mexico, October 1993. Sociedad Mexicana de Control Biolo´gico. Gerling, D. 1990. Natural enemies of whiteflies, predators and parasitoids. In ‘‘Whiteflies, Their Bionomics, Pest Status and Management’’ (D. Gerling, Ed.), pp. 147–185. Intercept Ltd., Andover, UK.

226

WRAIGHT ET AL.

Hajek, A. E., Soper, R. S., Roberts, D. W., Anderson, T. E., Biever, K. D., Ferro, D. N., LeBrun, R. A., and Storch, R. H. 1987. Foliar applications of Beauveria bassiana (Balsamo) Vuillemin for control of the Colorado potato beetle, Leptinotarsa decemlineata (Coleoptera, Chrysomelidae), an overview of pilot tests results from the northern United States. Can. Entomol. 119, 959–974. Hernandez, V. M., and Garza, E. 1994. Patogenicidad de Paecilomyces fumosoroseus a mosquita blanca Bemisia tabaci (Genn.) (Homoptera, Aleyrodidae). In ‘‘XVII Congreso Nacional de Control Biolo´gico,’’ pp. 42–43. Oaxaca, Oax., Mexico, October, 1994. Sociedad Mexicana de Control Biolo´gico. Humber, R. A. 1992. Collection of entomopathogenic fungal cultures, Catalog of strains. U.S. Department of Agriculture, Agricultural Research Service, ARS-110, 177 pp. Kirk, A. A., Lacey, L. A., Roditakis, N., and Brown, J. K. 1993. The status of Bemisia tabaci (Hom., Aleyrodidae), Trialeurodes vaporariorum (Hom., Aleyrodidae) and their natural enemies in Crete. Entomophaga 38, 405–410. Landa, Z., Osborne, L., Lopez, F., and Eyal, J. 1994. A bioassay for determining pathogenicity of entomogenous fungi on whiteflies. Biol. Control 4, 341–350. Lacey, L. A., Kirk, A. A., and Hennessey, R. D. 1993. Foreign exploration for natural enemies of Bemisia tabaci and implementation in integrated control programs in the United States. In ‘‘Third International Conference on Pests in Agriculture,’’ pp. 351–360. Association Nationale de Protection des Plantes, Paris, France. Nene, Y. L. 1973. Note on a fungus parasite of Bemisia tabaci Genn., a vector of several plant viruses. Indian J. Agric. Sci. 43, 514–516. Osborne, L. S. 1990. Biological control of whiteflies and other pests with a fungal pathogen. United States Patent Number 4,942,030. Osborne, L. S., and Landa, Z. 1992. Biological control of whiteflies with entomopathogenic fungi. Fla. Entomol. 75, 456–471.

Russell, R. M., Robertson, J. L., and Savin, N. E. 1977. POLO, a new computer program for probit analysis. Bull. Entomol. Soc. Am. 23, 209–213. Treifi, A. M. 1984. Use of Beauveria bassiana (Bals.) to control the immature stages of the whitefly Trialeurodes vaporariorum (Westw.) (Homoptera, Aleyrodidae) in the greenhouse. Arab. J. Plant Prot. 2, 83–86. van Lenteren, J. C., and Woets, J. 1988. Biological and integrated pest control in greenhouses. Annu. Rev. Entomol. 33, 239–69. Vidal, C., Lacey, L. A., and Fargue, J. 1997. Pathogenicity of Paecilomyces fumosoroseus (Deuteromycotina: Hyphomycetes) against Bemisia argentifolii (HomopteraL Aleyrodidae) with a description of a bioassay method. J. Econ. Entomol. 90, 765–772. Wilkinson, L., Hill, M., and Vang, E. 1992. Systat, Statistics, Version 5.2 Edition. Systat, Inc., Evanston, IL. Wraight, S. P., Carruthers, R. I., and Bradley, C. A. 1996. Development of Entomopathogenic fungi for microbial control of whiteflies of the Bemisia tabaci complex. In ‘‘V SICONBIOL, Simpo´sio de Controle Biolo´gico, Anais: Confereˆncias e Palestras,’’ pp. 28–34. Foz do Iguac¸u, Brazil. Wright, J. E. 1992. Whiteflies: Development of Naturalis, a biorational mycoinsecticide for control. In ‘‘Proceedings Beltwide Cotton Conferences’’ (D. J. Herber and D. A. Richter, Eds.), Vol. 2, pp. 887–888. National Cotton Council of America, Memphis, TN. Wright, J. E., and Chandler, L. D. 1992. Development of a biorational mycoinsecticide, Beauveria bassiana conidial formulation and its application against boll weevil populations (Coleoptera, Curculionidae). J. Econ. Entomol. 85, 1130–1135. Wright, J. E., and Chandler, L. D. 1995. Biopesticide composition and process for controlling insect pests. United States Patent Number 5,413,784.