*For offprint requests: fax: (504)388-1643, email: jfuxa@unix1.sncc.lsu.edu

SUMMARY

The purpose of the research was to test the capability of recombinant insect viruses to compete with a wild-type virus for a host-insect resource in a greenhouse microcosm. A wild-type Autographa californica nucleopolyhedrovirus (AcNPV.WT) was tested with two recombinant baculoviruses, AcNPV expressing a scorpion toxin (AcNPV.AaIT) and AcNPV expressing the mutant form of an insect-derived juvenile hormone esterase (AcNPV.JHEKK). The greenhouse microcosm for the viral competition experiments consisted of 16 mini-plots, each plot with 16 collards plants, arranged in 4 treatments with 4 replications. Larvae treated in four different ways were released into the plots in week 1. The four treatments included: 1) uninfected Trichoplusia ni (cabbage looper) larvae; 2) 100% AcNPV.WT-infected larvae; 3) 100% recombinant-infected larvae; and 4) 1:1 ratio of WT-infected and recombinant-infected. On a weekly basis, larvae were sampled and new, uninfected larvae were added to all 16 plots. Sampled larvae were reared until death and then subjected individually to DNA-DNA dot-blot hybridization assay for the WT and recombinant viruses to determine the proportion of insects infected with each virus in each plot. After 5 generations in the experiment with AcNPV.AaIT, disease prevalence rates were: control plots, 1% AcNPV.WT and 0% AcNPV.AaIT; WT plots, 57% AcNPV.WT and 0% AcNPV.AaIT; AaIT plots, 1% AcNPV.WT and 0% AcNPV.AaIT; and WT + AaIT plots, 58% AcNPV.WT and 0% AcNPV.AaIT. After 3 generations in a preliminary experiment with AcNPV.JHEKK, disease prevalence rates were: control plots, no virus detected; WT plots, 62% AcNPV.WT and 0% AcNPV.JHEKK; JHEKK plots, 4% AcNPV.WT and 16% AcNPV.JHEKK; and WT + JHEKK plots, 13% AcNPV.WT and 2% AcNPV.JHEKK. Thus, AcNPV.WT out-competed the recombinant viruses for a niche in the greenhouse microcosm, which reduces the probability that the recombinant viruses will persist in an agroecosystem.

Key words: Nucleopolyhedrovirus, recombinant baculovirus, viral persistence, viral competition, risk assessment

INTRODUCTION

Nucleopolyhedroviruses (NPVs), a subgroup of the Baculoviridae, have major advantages for insect pest management. The primary reason for interest in the NPVs for pest management is their environmental safety, which, in turn, is due to their extreme host specificity. They are safe to humans, wildlife, livestock, plants, and beneficial insects (Fuxa, 1989; Laird et al., 1990, Heinz et al., 1995; McCutchen et al., 1996). In addition, the NPVs are virulent and cause massive disease epizootics in their insect hosts, primarily the larvae of sawflies and moths (Fuxa, 1991).

The NPVs, though virulent in terms of their case-fatality rates, generally require approximately one week to kill the host insect. This slow speed is very deleterious when little damage from the pest can be tolerated. It also is disadvantageous to a "knockdown" or "firefighting" approach in which a user expects quick action (Fuxa, 1991).

There have been several approaches to increasing the speed of kill of recombinant baculoviruses. One of the two major subjects of the current research is the virus expressing a highly specific insect toxins from the Algerian scorpion Androctonous australis Hector (AaIT). This toxin has been expressed in AcMNPV termed AcMNPV.AaIT ( Maeda et al., 1991; McCutchen et al., 1991; Stewart et al., 1991). This toxin is highly specific for the insect sodium channel (the same site of action of pyrethroids). It fails to bind to the mammalian sodium channel and shows no toxicity even when injected at very high levels intracerebrally. AcMNPV.AaIT kills noctuid larvae that are over 2,000x resistant to pyrethroid insecticides even faster than susceptible larvae, making this virus very attractive in resistance managment programs (McCutchen et al., 1996).

Another approach involves the insertion of a gene for an enzyme into the baculovirus. When the gene for juvenile hormone esterase (JHE) was cloned into a baculovirus vector there was a reduction in food consumed by first instar larvae of T. ni (Hammock et al., 1990), but the virus offered no advantage with later instars. By making just two mutations in the JHE in regions thought needed for lysozomal recognition, a JHE protein was obtained which is as active as AaIT in reducing feeding damage of virus infected insects (Bonning and Hammock 1994). This virus AcMNPV.JHEKK reduces feeding damage over 50% when compared to wildtype virus on lettuce plants. Thus the current research with AcMNPV.AaIT and AcMNPV. JHEKK involves two of the best recombinant viruses from two distinct lines of research.

In spite of the safety record of natural strains of NPVs, there has been reluctance on the part of regulatory agencies to allow releases (field tests) of recombinant NPVs. Much of the concern about the release of recombinant-DNA microorganisms is due to uncertainty about their potential for environmental persistence and subsequent transport and dispersal (Jones 1988). Even an agent harmful to some environmental component might have low risk if it cannot persist or spread. Because there is reluctance to release recombinant NPVs to obtain data on environmental safety, other types of data must be collected. Microcosms, or contained systems ranging from test tubes to greenhouses in which a portion of an ecosystem is duplicated, can provide data concerning how a recombinant might perform or affect a portion of the environment without actually doing an environmental release (Omenn, 1986).

The purpose of the current research is to determine whether the recombinant viruses AcNPV.AaIT and AcNPV.JHEKK can compete with wild Autographa californica NPV (AcNPV) for a niche in an plant/insect/virus microcosm.

MATERIALS AND METHODS

Collard plants were grown in a greenhouse at a spacing of 30 cm between plants. There were 16 plots on four greenhouse benches (one replication of four plots per bench). Plots were 1.35 x 1.35 m with buffers of 45 cm between plots. Each plot was surrounded by a wooden, plastic-covered wall 20 cm high. Each plot contained 16 evenly-spaced plants. The plants were watered with sprinklers at a rate of 5 mm every 2 d to simulate rainfall on the foliage.

The three viruses (wild-type AcNPV [AcNPV.WT], AcNPV.AaIT, and AcNPV.JHEKK) were grown in laboratory-reared Trichoplusia ni. AcNPV.WT was provided by Biosys, Inc. (Hanover, Maryland). Viruses were purified by standard techniques of homogenization of dead insects followed by filtration and centrifugation. NPV preparations were quantified on an improved Neubauer hemacytometer. The T. ni were reared on artificial diet with standard techniques.

For the experiment, one group of third instars of T. ni was infected by diet surface treatment with AcNPV.WT and another group with AcNPV.AaIT. High doses of NPV were used to ensure 100% infection. The T. ni were infected after the collards plants grew to a height of approximately 15 cm. One day after infection, the T. ni were released into the greenhouse plots in a randomized block design experiment (each bench = one replication with four treatments or plots). Six larvae were placed on each plant with a camel's hair brush, for a total of 96 insects/plot. Only 4 larvae were placed on plants in control plots, because the plants could not survive the damage from 6 healthy insects through the course of the experiment. The four treatments of generation-1 larvae released into the plots were: 1) uninfected T. ni larvae; 2) 100% AcNPV.WT-infected larvae; 3) 100% AcNPV.AaIT-infected larvae; and 4) 1:1 ratio of WT-infected and AaIT-infected.

One week after the release of the insects in Treatments 1-4, and weekly thereafter, healthy second instars were released in all 16 plots, again at a density of 6 larvae/plant or 96/plot (controls, 4/plant or 64/plot). Two days after each release of insects (including the original "treatment" insects), two larvae per plant were randomly sampled. At each time of sampling, any insects that reached the pupal stage were removed from the experiment. The sampled larvae were reared individually in 30 ml cups with artificial diet until they died or pupated.

NPV from every larva killed by virus was analyzed to determine whether that insect was infected with AcNPV.WT or AcNPV.AaIT. The dead insects were frozen and held below -20C until sent from Louisiana to California on dry ice. A DNA-DNA dot-blot hybridisation assay was then used to determine the identity of the virus present in each dead insect (Method 1 in Ward et al. 1987). The sensitivity of this technique, in which the whole homogenized larvae is probed with a high specific activity ³²P DNA probe, was shown to be more sensitive than differential staining, phase-contrast microscopy, or indirect solid-phase radioimmunoassay (Ward et al., 1987). To detect wild-type virus, a fragment of cloned DNA encoding the essential gene ORF 1629 (open reading frame) was used. This probe hybridises with all three of the viruses in the current research. To detect the two recombinant viruses, the coding sequences for the introduced genes were used as probes. These probes are specific for the engineered viruses.

The releases and sampling were continued for up to 5 weeks. After the last larval sampling, one random soil sample was collected per plot, with a sampler 10 cm diameter and 5 cm deep. Each soil sample was bioassayed with a leaf-disk method with second-instar T. ni (Fuxa et al., 1985). Briefly, each sample was homogenized in distilled water, centrifuged, and the pellet resuspended in 0.5% Triton X-100. Forty larvae were each fed 5-mm collards leaf disks which had been dipped in the soil suspension and dried. Insects that ate the entire disk within 2 d were transferred to artificial diet. Controls for the soil bioassay were treated identically except that the leaf disks were dipped only in 0.5% Triton X-100. Percent mortality was recorded at pupation. Soil-bioassay larvae killed by NPV will be analyzed, as above, to determine which virus had infected individual insects (work still in progress).

The entire experiment was repeated with AcNPV.JHEKK in the place of AcNPV.AaIT. The soil and greenhouse benches were decontaminated prior to the second experiment.

RESULTS

Wild-type AcNPV out-competed both recombinant NPVs on the insect's host plant in the greenhouse plots. In the experiment with AcNPV.AaIT (Table 1), AcNPV.WT produced epizootics in T. ni for five generations, whereas in the AcNPV.AaIT treatment the prevalence of the latter virus decreased to zero by the fourth generation of insects. In the treatment with both viruses released, prevalence of AcNPV.WT was virtually identical to that in the AcNPV.WT treatment after the second generation, whereas AcNPV.AaIT prevalence again decreased to zero, this time in the third generation. In the experiment with AcNPV.JHEKK (Table 2), the AcNPV.WT treatment again produced epizootics of wild-type virus over several generations, whereas prevalence of recombinant virus steadily decreased in the AcNPV.JHEKK treatment. In the treatment with both viruses released, AcNPV.WT was out-competing AcNPV.JHEKK, but the prevalence of the wild-type virus was not as high as in the AcNPV.WT treatment. The AcNPV.JHEKK experiment, which ended prematurely due to insect rearing problems, will be repeated in future research.

Insect cadavers were observed in all the viral treatment plots in both experiments. In both experiments, cadavers in the AcNPV.WT treatments as well as those in treatments with both viruses exhibited typical signs of nucleopolyhedrosis. Cadavers virtually became liquified on the leaves, presumably releasing viral polyhedral occlusion bodies. In the AcNPV.AaIT treatment plots (Table 1), insects died as small larvae and fell from leaves onto the soil surface, where they remained intact. Numerous insects in the AcNPV.JHEKK treatment plots (Table 2) died on leaves; these cadavers disintegrated to some degree, but not as extensively or as quickly as AcNPV.WT cadavers.

DNA probes have not yet been completed for the soil bioassay at the conclusion of the experiment with AcNPV.AaIT. However, viral mortality rates of soil bioassay insects from the various treatments (Table 2) indicate that greater amounts of AcNPV.WT than AcNPV.AaIT accumulated in the soil.

It is virtually certain that all of the insects in generation 1 of the AcNPV.AaIT treatment (Table 1) were infected with AcNPV.AaIT due to confirmation of the inoculum as well as symptomatology of cadavers. However, the shipment of generation-1 insects was lost in the mail and remained at room temperature ca. 1 week. A corollary experiment indicated that the WT probe hybridized with 100% of samples of insects treated with AcNPV.AaIT regardless of whether the insects were frozen immediately after death or kept at room temperature for 6 days. On the other hand, the AaIT probe hybridized with 100% of samples of insects that had been frozen but only 88% of those at room temperature.

Statistical analyses of the data in Tables 1-3 have not yet been completed.

DISCUSSION

It is not surprising that AcNPV.WT out-competed AcNPV.AaIT and AcNPV.JHEKK for the insect resource on host plants. AcNPV.AaIT and AcNPV.JHEKK produce only 20 and 60% as many viral polyhedra, respectively, as AcNPV.WT (Kunimi et al., 1996). Additionally, only 5% or 23% of T. ni killed by AcNPV.AaIT or AcNPV.JHEKK, respectively, disintegrate by day 4 after death, compared to 100% by AcNPV.WT after 1 day (JA Fuxa, JR Fuxa, AR Richter, unpublished data). Thus, it is likely that very little recombinant virus, particularly AcNPV.AaIT, is released onto the leaves of host plants (Hoover et al., 1995) where new generations of insects can ingest it and become infected.

Dynamics of the recombinant NPVs in soil are not easy to predict. The wild-type NPV could increase to greater levels than the recombinants in soil, because the wild-type virus continues to infect and replicate in host insects (Tables 1, 2). On the other hand, the recombinants could increase to greater levels than wild-type virus in soil, because the wild-type virus in cadavers disintegrating on leaves is more exposed to sunlight, which inactivates NPVs within hours, than the recombinant viruses in intact cadavers beneath plants. The current, incomplete results (Table 3) indicate that the wild-type virus increased to greater levels in soil than AcNPV.AaIT.

Our results reduce the probability of environmental harm due to release of AcNPV.AaIT or AcNPV.JHEKK. The possible harmfulness of recombinant organisms to non-target organisms in the environment is difficult to predict (Fuxa, 1990). Therefore, the probability that a released recombinant will contact non-target organisms becomes a critical issue. In the case of living organisms, this probability of contact is largely a function of the recombinant's capability to persist, increase in numbers, and move or be transported away from the release site (Fuxa, 1990). The inability of the recombinant viruses to compete with wild-type virus in the current research indicates that the populations of the recombinant viruses are likely to decrease after release, which in turn reduces their environmental risk.

Further research in the current project will focus on the possibility that AcNPV.AaIT and AcNPV.JHEKK will be transported outside of release sites by scavenging and predatory arthropods.

ACKNOWLEDGMENTS

This research was supported by a grant (no. 95-33120-1977) from the USDA Biotechnology Risk Assessment program. Additional support was provided by USDA 94-37302-0567, US/Israel BARD (N. Chejanovsky) #IS-2139-92, and the University of California Systemwide Biotechnology Research and Education Program. The University of California, Davis, is an EPA Center for Environmental Health Research (CR 819658) and a NIEHS Center for Environmental Health (P30ES05707). This paper was approved for publication by the Director of the Louisiana Agricultural Experiment Station. We thank the following for technical support: Yasuhisa Kunimi, Kelli Hoover, Christine M. Schultz, Haim Moskowitz, and Kiyoko Taniai.

REFERENCES

Bonning BC, Hammock BD. 1994. Insect control by use of recombinant baculoviruses expressing juvenile hormone esterase. In: Hedin PA, Menn JJ, Hollingworth RM (eds) Natural and engineered pest management agents. ACS Symposium Series #551, American Chemical Society, Washington, D.C., pp 368-383.

Fuxa JR. 1989. Fate of released entomopathogens with reference to risk assessment of genetically engineered microorganisms. Bull Entomol Soc Am 35: 12-24.

Fuxa JR. 1990. Environmental risks of genetically engineered entomopathogens. In: Laird M, Lacey LA, Davidson EW (eds) Safety of microbial insecticides. CRC Press, Boca Raton, FL.

Fuxa JR. 1991. Insect control with baculoviruses. Biotech Adv 9: 425-442.

Fuxa JR., Warren GW, Kawanishi C. 1985. Comparison of bioassay and enzyme-linked immunosorbent assay for quantification of Spodoptera frugiperda nuclear polyhedrosis virus in soil. J Invertebr Pathol 46: 133-138.

Hammock BD, Bonning BC, Possee RD, Hanzlik TN, Maeda S. 1990. Expression and effects of the juvenile hormone esterase in a baculovirus vector. Nature 344: 458-461.

Heinz KM, McCutchen BF, Herrmann R, Parrella MP, Hammock BD. 1995. Direct effects of recombinant nuclear polyhedrosis viruses on selected non-target organisms. J Econ Entomol 88: 259-264.

Hoover K, Schultz CM, Lane SS, Bonning BC, Duffey SS, McCutchen BF, and Hammock BD. 1995. Reduction in damage to cotton plants by a recombinant baculovirus that knocks moribund larvae of Heliothis virescens off the plant. Biological Control 5: 419-426.

Jones DD. 1988. USDA safety review of biotechnology research in agriculture. American Biotechnology Laboratory, January.

Kunimi Y, Fuxa JR, Hammock BD. 1996. Comparison of wild type and genetically engineered nuclear polyhedrosis viruses of Autographa californica for mortality, virus replication and polyhedra production in Trichoplusia ni larvae. Entomol Experimentalis Applicata (In Press).

Laird M, Lacey LA, Davidson EW (eds). 1990. Safety of microbial insecticides. CRC Press, Boca Raton, FL.

Maeda S, Volrath SL, Hanzlik TN, Harper SA, Maddox DW, Majima K, Hammock BD, Fowler E. 1991. Insecticidal effects of an insect-specific neurotoxin expressed by a recombinant baculovirus. Virology 184: 777-780.

McCutchen BF, Choudary PV, Crenshaw R, Maddox D, Kamita SG, Palekar N, Volrath S, Fowler E, Hammock BD, Maeda S. 1991. Development of a recombinant baculovirus expressing an insect-selective neurotoxin: potential for pest control. Bio/Technology 9: 848-852.

McCutchen BF, Herrmann R, Heinz KM, Parrella MP, Hammock BD. 1996. Effects of recombinant baculoviruses on a nontarget endoparasitoid of Heliothis virescens. Biol. Control 6: 45-50.

McCutchen BF, Hoover K, Preisler HK, Betana MD, Herrmann R, Robertson JL, Hammock BD. 1996. Interactions of recombinant and wild-type baculoviruses with classical insecticides and pyrethroid-resistant tobacco budworm (Lepidoptera: Noctuidae). J Econ Entomol (In Press).

Omenn GS. 1986. Controlled testing and monitoring methods for microorganisms. In: Fiksel J, Covello VT (eds) Biotechnology risk assessment. Pergamon, New York, pp 144-163.

Stewart LMD, Hirst M, Ferber ML, Merryweather AT, Cayley PJ, Possee RD. 1991. Construction of an improved baculovirus insecticide containing an insect-specific toxin gene. Nature 352: 85-88.

Ward VK, Fleming SB, Kalmakoff J. 1987. Comparison of a DNA-DNA dot-blot hybridisation assay with light microscopy and radioimmunoassay for the detection of a nuclear polyhedrosis virus. J Virological Methods 15: 65-73.

Table 1. Mean mortality ± SE (n = 4 replications) of Trichoplusia ni due to wild-type (AcNPV.WT) or recombinant (AcNPV.AaIT) virus on collards plants in a greenhouse microcosm.

Insect generation

Control %WT %AaIT

Wta %WT %AaIT

AaITa %WT %AaIT

WT + AaITb %WT %AaIT

1	0	0	90.0± 5.7	0	15.0 ±3.7	71.0± 7.2	47.8 ±12.5	44.0 ±12.8
2	0	0	75.8 ±13.3	0	0	0.8 ±1.3	51.8 ±22.8	3.0 ±2.1
3	0	0	92.5 ±3.4	0	0.8 ±1.3	3.8 ±2.5	91.5 ±4.8	0
4	0.8 ± 1.3	0	97.8 ±1.3	0	0	0	97.0 ±2.1	0
5	0.8 ± 1.3	0	57.3 ± 9.1	0	0.8 ±1.3	0	57.8 ±11.3	0

^aAll released generation-1 insects infected with AcNPV.WT or AcNPV.AaIT, respectively. Subsequent generations were released as healthy insects.
^bHalf of released generation-1 insects infected with AcNPV.WT and half with AcNPV.AaIT. Subsequent generations were released as healthy insects.

Table 2. Mean mortality ± SE (n = 4 replications) of Trichoplusia ni due to wild-type (AcNPV.WT) or recombinant (AcNPV.JHEKK) virus on collards plants in a greenhouse microcosm.

Insect generation

Control %WT %JHEKK

Wta %WT %JHEKK

JHEKKa %WT %JHEKK

WT + JHEKKb %WT %JHEKK

1	0	0	97.0 ± 1.9	0	1.8 ± 1.8	96.8 ± 2.5	51.0 ± 7.0	42.0 ± 10.3
2	0	0	76.5 ± 14.8	0	2.3 ± 2.5	76.8 ± 15.0	7.0 ± 4.1	2.3 ± 2.5
3	0	0	61.5 ± 14.4	0	3.8 ± 4.9	15.8 ± 10.0	12.5 ± 9.0	1.5 ± 1.5

^aAll released generation-1 insects infected with AcNPV.WT or AcNPV.JHEKK, respectively. Subsequent generations were released as healthy insects.
^bHalf of released generation-1 insects infected with AcNPV.WT and half with AcNPV.JHEKK. Subsequent generations were released as healthy insects.

Table 3. Mean mortality ± SE (n = 4 replications) of Trichoplusia ni due to wild-type (AcNPV.WT) or recombinant (AcNPV.JHEKK) virus in bioassay of soil from greenhouse microcosm plots.^a

Contol plots

Plots with insects infected by: AcNPV.WT AcNPV.AaIT AcNPV.WT+AcNPV.JHEKK

0

61.0 ± 3.0

25.5 ± 4.1

60.0 ± 2.8

^aDNA probes are still in progress to determine proportions of insects in each treatment killed with each type of NPV.