Bot. Bull. Acad. Sin. (2005) 46: 11-20

TUAN et al. — Engineered baculovirus for better plant protection

Improved plant protective efficacy of a baculovirus using an early promoter to drive insect-specific neurotoxin expression

Shu-Jen TUAN^{1, 2}, Roger F. HOU³, Suey-Sheng KAO¹, Chi-Fen LEE¹, and Yu-Chan CHAO^2,*

¹Biopesticide Department, Taiwan Agricultural Chemicals and Toxic Substances Research Institute, Wufeng, Taichung 413, Taiwan, ROC

²Institute of Molecular Biology, Academia Sinica, Nankang, Taipei 115, Taiwan, ROC

³Department of Entomology, National Chung Hsing University, Taichung 402, Taiwan, ROC

(Received April 5, 2004; Accepted September 30, 2004)

Abstract. Genetically improved Autographa californica (Speyer) multiple nucleopolyhedrovirus (AcMNPV), the type species of the genus Nucleopolyhedrovirus, was constructed with an insect-selective toxin gene (LqhIT₂) derived from the venom of the Israeli yellow scorpion, Leiurus quinquestriatus (Ehrenberg). LqhIT₂ was expressed by two distinct and temporally regulated promoters. These two promoters were the early promoter p-PCm (a pu-enhanced minimal CMV promoter) and the very late p10 promoter, resulting in recombinant viruses named vAPcmIT₂ and vAP10IT₂, respectively. Western blot and bioassay analysis showed that the toxin gene under the control of the early p-PCm promoter was expressed 24 h earlier than under the very late p10 promoter control, and vAPcmIT₂ killed faster than vAP10IT₂. The EC₅₀ (effective concentration, EC) values of wild type AcMNPV, vAP10IT₂, and vAPcmIT₂ to the 3^rd-instar larvae of Trichoplusia ni (Hübner) were 1.00, 0.19, and 0.17 polyhedral inclusion bodies (PIBs)/mm³, respectively. Compared to the very late promoter, the early promoter could shorten the effective time (ET₅₀) by 14~17 h. Furthermore, for the 4^th-instar larvae, compared to the wild-type virus and vAP10G, a 19~23% and 30~33% reduction in ET₅₀ were found using vAP10IT₂ and vAPcmIT₂, respectively. Field trials of these two viruses showed an effective paralysis of the infected larvae which then stopped feeding, resulting in a decrease in the leaf area eaten compared to those larvae infected with wild-type virus and another control virus vAP10G, the GFP-expression recombinant AcMNPV. Scorpion toxin driven by the early p-PCm promoter shortened the time-to-paralysis of insect larvae, and thus provided an economical pest control advantage over that driven by the very late p10 promoter.

Keywords: AcMNPV; Autographa californica multicapsid nucleopolyhedrovirus; Baculovirus; Cabbage; Depressant neurotoxin; Leiurus quinquestriatus hebraeus; LqhIT₂; Plant protection; Recombinant virus; Scorpion toxin; Trichoplusia ni.

Introduction

Since chemical pesticides frequently pose environmental and health risks, carefully selected biological control agents, like predators, parasitoids, and pathogens, have become attractive alternatives for the suppression of insect pests in the field. Baculoviruses (Baculoviridae) are arthropod-specific pathogens that have served as microbial biopesticides for control of many lepidopteran pests for several decades (Entwistle and Evans, 1985). These pathogens are characterized by large, circular double-stranded DNA genomes and rod-shaped virions that are enveloped within polyhedrin to form polyhedra (Tinsley and Kelly, 1985; Whitt and Manning, 1988). While baculoviruses have the potential for insect control, infected lepidopteran larvae still cause considerable crop damage due to continued feeding for several days after the initial infection, resulting in reduced efficacy in the field (Bonning and Hammock, 1996; Tuan et al., 1997; 1998).

A number of studies have established that the efficacy of baculovirus for insect pest control is greatly improved by the insertion of foreign genes into the viral genome. These include genes encoding the Manduca sexta (Linnaeus) diuretic hormone (Maeda, 1989), the Heliothis virescens (Fabricius) juvenile hormone esterase (Hammock et al., 1990), the Bacillus thuringiensis (Berliner) d-endotoxin (Martens et al., 1990; 1995), mite or spider toxins (Tomalski and Miller, 1991; Hughes et al., 1997; Priknod'ko et al., 1998), proteases (Harrison and Bonning, 2001), and the scorpion insect-specific neurotoxins (Stewart et al., 1991; Maeda et al., 1991; McCutchen et al., 1991; Fuxa et al., 1998; Gershburg et al., 1998; van Beek et al., 2003). All these studies reveal the significant contribution of exogenous gene products to the efficacy of baculovirus in pest control.

The LqhIT₂ depressant insect toxin, derived from the Israeli yellow scorpion Leiurus quinquestriatus hebraeus (Ehrenberg), is a polypeptide of 61 residues (Zilberberg et al., 1991). Injection of LqhIT₂ toxin into Sarcophaga falculata (Pandelle) blowfly larvae induced symptoms typical of the transient excitatory effect that preceeds the onset of prolonged flaccidity. The larvae are immobilized and

*Corresponding author. Phone: 886-02-2788-2697; E-mail: mbycchao@imb.sinica.edu.tw

Botanical Bulletin of Academia Sinica, Vol. 46, 2005

paralyzed and become completely flaccid (Zlotkin et al., 1993; Gershburg et al., 1998; Harrison and Bonning, 2000). The depressant toxin`s mode of action involves depolarization of the axonal membrane, blockage of the evoked action potential, and changes in the amplitude and the kinetics of the sodium current. Flaccid properties of the host insect are attributed to the opening of sodium channels and inability to inactivate them (Benkhalifa et al., 1997). The higher efficacy of the virus expressing the depressant toxin LqhIT₂ over the excitatory toxin LqhIT₁ suggests that pharmacokinetic factors and/or promoter efficiency may play a role during infection of insect pest larvae by recombinant baculoviruses (Gershburg et al., 1998).

When the scorpion toxins were tested using recombinant baculoviruses, the insecticidal activity of the very late p10 promoter was higher than that of the early p35 promoter (Gershburg et al., 1998). These data suggest that promoter efficiency may play a role in the infection of insect pest larvae by recombinant baculoviruses since an early promoter should exert an effect earlier than the very late promoter. If the early promoter is strong enough, it should show more lethality in the recombinant virus. Previously, the human cytomegalovirus minimal (CMVm) promoter, ligated in cis with a newly discovered activator, the polyhedrin upstream (pu) sequence, and the early p-PCm promoter (pu plus CMVm) was shown to result in high level expression of foreign genes. In these experiments, although less luciferase was expressed by the early p-PCm promoter rather than by the very late p10 promoter, the former expressed earlier with better total luciferase activity (Wu et al., 2000; Lo et al., 2002).

The Autographa californica nucleopolyhedrovirus (AcMNPV) is the type species of the Nucleopolyhedrovirus (Volkman, 1995). Its broad insect infectivity spectrum of more than 30 species lepidopteran also made this virus a more acceptable microbial insecticide (Granados and Williams, 1986). The pathogenicity of AcMNPV against nine important agricultural pests was studied previously (Tuan et al., 1997). Recombinant AcMNPV expressing insect-specific toxins under different temporally regulated promoters showed various efficacies. The immediate early ie1 promoter expressed AaIT earlier than the very late p10 promoter and significantly reduced the feeding activity of infected larvae (Jarvis et al., 1996). The time required for paralysis or death was promoter-dependent, and the late 6.9K DNA binding protein gene promoter was generally more effective than P_synXIV, HSP70, and DA26 promoters for expression of Tox34 (Lu et al., 1996). The hybrid late and very late promoter were able to drive expression of a tox34 gene and showed superior reductions in kill times compared to the very late promoter only (Tomalski and Miller, 1992). The hr5/ie1 promoter lead to a faster response than the very late p10 promoter for expression of either LqhIT₂ or AaIT (van Beek et al., 2003). In order to further improve the efficacy of recombinant AcMNPV for earlier insect pest control, a newly developed early promoter p-PCm (Lo et al., 2002) was used to drive a scorpion depressant toxin gene, LqhIT₂. Our results demonstrate that the scorpion toxin is expressed earlier by the

early p-PCm promoter than the very late p10 promoter and that insecticidal efficacy of recombinant AcMNPV expressing LqhIT₂ was significantly improved compared to wild-type AcMNPV.

Materials and Methods

Cell Cultures and Viruses

The cell line Spodoptera frugiperda IPLB-Sf21AE (Sf21, Vaughn et al., 1977) was maintained at 26°C in a modified TNM-FH medium that contained 8% fetal bovine serum (Lee et al., 1998; Lin et al., 1999). The GFP-expression recombinant virus, vAP10G (Chao et al., 1996) and the C6 strain of wild-type AcMNPV were propagated in the Sf21 cell line. The vAP10G was here used as control because it was constructed by the cotransfection with commercial linearized genomic DNA of AcMNPV, and we used the same material and methodology to construct both neurotoxin-expressing recombinant viruses vAP10IT₂ and vAPcmIT₂. At 5~7 days post infection (d p.i.), the infected cells were pelleted and resuspended in 0.5% SDS solution for 10 min, and the number of polyhedra were counted by a hemocytometer and stored at 4°C in sterile H₂O.

Insects

The larvae of cabbage loopers, Trichoplusia ni (Hübner), were collected from Shi-hu or Wu-feng County in Taiwan and reared on an artificial diet (Tuan et al., 1997). These insects were reared at 25±1°C with 70±5% relative humidity (RH) and 14:10 h (L:D) photoperiod. All tested insects were the Shi-hu strain 4^thor Wu-feng strain 3^rd instar larvae after three generations of rearing. The mass-rearing artificial diet contained 0.1% sorbic acid and p-methyl-benzoate for anti-bacteria and anti-fungi purposes. Pupae and eggs were washed for 10-15 min in 0.5% NaClO or 4% formalin solution, after which they were rinsed for 15 min with running water followed by air-drying. During the course of experiments, all precautions to ensure against contamination with microorganisms (equipment autoclaving, surface wiping with 70% ethanol, instrument flaming, formalin fumigating for all rearing apparatus, and the use of disposable containers and instruments) were applied. The insects in preliminary work and in all experiments were observed daily and by smears under light or phase-contrast microscopy to determine if they were pathogens in origin.

Construction of Plasmids and Recombinant Virus

We excised a complete LqhIT₂-coding sequence (included Bombyxin signal sequence and LqhIT₂ mature peptide sequence, 250 bp in length) by digestion with BglII and EcoRI (PharMingen) from pBmLqhIT₂ (provided by Maeda's Lab., Maeda et al., 1991; Gershburg et al., 1998). The sequence then was inserted into a polyhedrin gene containing transfer vector pAcUW21 between a blunt-ended BglII site and the EcoRI site. The polyhedrin gene was useful for later experiments in forming polyhedral inclusion bodies for insect infection by feeding. The scor

TUAN et al. — Engineered baculovirus for better plant protection

pion depressant toxin gene was expressed under the control of the very late p10 promoter of AcMNPV, and named pAP10IT₂. For the early expression of scorpion depressant toxin gene, the very late p10 promoter of plasmid pAP10IT₂ was replaced with an early p-PCm promoter, which contains a CMVm promoter that is activated in cis by a upstream pu element already situated in the plasmid (Lo et al., 2002) to generate pAPcmIT₂ (Figure 1). The CMVm promoter was excised from the vector pTRE (Clontech) between XbaI and BamHI sites. The 3' end of CMVm promoter fragment was ligated into the BglII cloning site through the BamHI site. The 5' end of the CMVm promoter fragment was ligated into XbaI site with the aid of an XbaI-XhoI (5'-3') adaptor, resulting in the XbaI-XhoI-CMVm-BamHI (BglII) linker. Recombinant viruses, that expressed scorpion depressant toxin, were constructed by the cotransfection of transfer vectors pAP10IT₂ or pAPcmIT₂ separately with linearized genomic DNA of AcMNPV (vAcRP23.LacZ, PharMingen), using Lipofectin (Life Technologies). Polyhedra-positive recombinant

viruses, vAP10IT₂ and vAPcmIT₂ were isolated by four rounds of end-point dilution. They were verified by restriction fragment length analysis and then confirmed by PCR analyses. The sequences of these two primers were primers polh: 5' CCG ATG TAA ACG ATG GGC TT 3', and EcoIT 5' GAA TTC TTA TCC ACA GGT ATT CGT 3'.

Western Analysis

Isolated recombinant viruses were further confirmed and the timing of toxin expression was analyzed. AcMNPV-mediated expression of LqhIT₂ under the control of the very late p10 and early p-PCm promoters in insect cells was analyzed by Western blotting. The polyclonal antibody was from rabbits immunized by a sub-dermal multi-injection of synthetic peptide mixed with adjuvant (amino acid sequence: H-DGYIKRRDGCDDKTWK-NH₂, conjugated with Keyhole Limpet Hemocyanin as a carrier protein). SF 21 cells in 6-well microplates were inoculated with vAP10IT_2, vAPcmIT₂, and vAP10G recombinant viruses or with wild-type AcMNPV at a multiplicity of infection (moi) = 1. Infected cells were collected at 4, 8, 12, 18, 24, 36, 48, 72 and 96 h post infection (h p.i.), and were centrifuged at 1000 g for 5 min, after which the resulting supernatants were collected. The cell pellets were resuspended in double-distilled H₂O treated with 0.5% SDS, then vortexed for 5 min and washed twice. Cell extracts and medium samples were mixed with equal volumes of sample buffer and analyzed by SDS-PAGE in 4% stacking gel and 15% resolving gel. Gels were electroblotted onto polyvinylidene fluoride membranes (PVDF, Millipore) and blocked with Tris-buffer saline with 0.1% Tween-20 containing 3% skim milk at 4°C overnight. Subsequently, the membrane was incubated with polyclonal rabbit anti-LqhIT₂ antibody for 1 h at room temperature followed by horseradish peroxidase (HRP)-conjugated secondary antibody for 1 h at room temperature. Finally, the HRP spots on the membrane were detected by an enhanced chemiluminscence kit (ECL, Amersham Life Science) following the protocol provided by the manufacturer.

Lethal Dose and Time Measurements

Each candidate recombinant virus was screened by a bioassay of activity against the host larvae. The pathogenicity of scorpion toxin containing recombinant viruses (vAP10IT₂ and vAPcmIT₂) and wild-type AcMNPV were compared. The vAP10G, a recombinant virus containing a GFP coding region driven by the very late p10 promoter, was also used as another control. The median lethal concentration (LC₅₀) and median lethal time (LT₅₀) were determined on the 3^rd- and 4^th-instar larvae of T. ni. Molting larvae at the same age were selected and enclosed individually in a 30-well plate overnight, and the next morning the newly molted larvae were starved for 8 h to synchronize larval growth. These larvae were bioassayed with a diet-contaminated method (Tuan et al., 1997) and maintained at 25±1°C, 70±5% RH, and under a 14:10 h photoperiod. Tested larvae were fed with a set of five serial 10-fold dilutions of each virus stock (from 780 to 0.078 polyhedral

Figure 1. Construction of the transfer vectors pAp10IT₂ and pAPcmIT₂. The LqhIT₂ gene was excised from the transfer vector pBmLqhIT₂ by Bgl II and EcoR I digestion, and inserted into pAcUW21. The scorpion toxin gene was cloned juxtaposed to the p10 promoter yielding the vector pAp10IT₂. The vector pAp10IT₂ was modified by replacing the p10 promoter with the CMVm promoter using the restriction sites of Xho I, BamH I, Xba I and Bgl II, and an Xba I-Xho I linker, resulting in a new plasmid pAPcmIT₂.

Botanical Bulletin of Academia Sinica, Vol. 46, 2005

inclusion bodies, PIBs/mm³) on a small plug of artificial diet. Sixty larvae in three replicates were inoculated in each treatment. For lethal time or effective time determining assay, 4^th-instar larvae were treated with 78 and 7.8 PIBs/mm³, and 3^rd-instar larvae were treated with 78 and 0.78 PIBs/mm³ (to obtain ~90-100%, and <75% mortality within seven days). Larvae that consumed the diet were then fed another plug diet without virus and examined two or three times daily for a response. Tested larvae were scored as responders if they were paralyzed or dead. In doubtful cases, the larva was touched slightly with a paint brush five times and scored as responder only if it failed to move within 30 s. All data were recorded until the 10^th d p.i., corrected with Abbott's formula, and analyzed using probit analysis (Finney, 1971) to determine the LC₅₀ (concentration of PIBs/mm³ to kill 50% of the test larvae), EC₅₀ (concentration of PIBs/mm³ necessary to resulting in 50% of the test larvae to paralyze or cease feeding), ET₅₀ (mean time at which 50% of test larvae cease to respond to a stimulus), ET₉₀ (time at which 90% of test larvae cease to respond to a stimulus), and LT₅₀ (mean time at which 50% of test larvae are killed) values. All of the bioassays were performed at least three times with different virus batches. The resulting values for each virus were averaged and analyzed by one-way ANOVA. The means were separated by the Fisher's Least Significant Difference Test (Steel and Torrie, 1980)

Field Trial

Newly molted 4^th-instar T. ni larvae were starved for 16 h, and then transferred to potted cabbage with 12~14 leaves. Thirty larvae on each pot were confined in a fine-mesh nylon cage. The cabbages were treated individually with sterile water (as control), recombinant viruses with (vAP10IT₂ and vAPcmIT₂) or without (vAP10G) scorpion toxin gene, and wild-type AcMNPV, that were applied at 10⁷ PIBs/ml using a handheld sprayer. Tween 20 (0.05%) was added to all viral suspensions and control. Three potted cabbages were tested for each treatment. The insects were checked twice daily to record signs and symptoms of disease and calculate mortality. The dead larvae were

collected and stored at 4°C, then checked by microscopy at a magnification of 400X. Six days after application, all larvae, living and dead, were collected from potted cabbages. The surviving larvae were weighed, reared individually on artificial diet at 24±1°C, and checked twice daily for death or pupation. Larvae failing to respond to gentle stimulation by the paint brush within 30 s were scored as dead. Mean leaf area eaten per cabbage (sum of 12 leaves) was determined in infested control and virus-treated cabbage pots at the 6^th day after virus application. All leaves were removed from the plant and Xeroxed. Each intact leaf-image was cut out and scanned with a leaf area meter (Model Li-3100 Area Meter, Li-Cor, USA). The holes in each leaf were then trimmed off carefully and scanned again. The differences between two scans were recorded as the area consumed for that leaf, and these differences in all leaves for each plant were summed up. All field trials were performed at least three times with different virus batches. Leaf areas eaten by tested larvae for each treatment group were compared by one-way ANOVA followed by Fisher's Least Significant Difference Test (Steel and Torrie, 1980).

Results

Construction and Expression of LqhIT₂ by Recombinant AcMNPV

Transfer vectors that contained the AcMNPV polyhedrin gene and the scorpion toxin gene LqhIT₂ under the control of the very late p10 or the early p-PCm promoters (Lo et al., 2002) were constructed (Figure 1). The production of 6 kD LqhIT₂ toxin proteins by individual isolates was subsequently examined. Western analysis showed that LqhIT₂ was first detectable in vAPcmIT₂-infected cells at 12 h p.i. Expression of LqhIT₂ peaked at 36 h p.i. in vAPcmIT₂-infected cells and subsequently declined by 48 h p.i. The scorpion toxin expressed by vAP10IT₂ was first clearly detectable at 36 h p.i., reached a plateau around 48 ~ 72 h p.i., and declined to 96 h p.i. (Figure 2). Based on size (6 kDa) and its specific expres

Figure 2. Time course expression of LqhIT₂ in insect cells. Sf21 cells were infected with vAPcmIT₂, vAP10IT₂ and AcMNPV at moi = 1. Immunoblot analysis of scorpion toxin expression was assayed at various times p.i. Cell extracts were prepared at the indicated times (h), separated by 15% SDS-PAGE, and immunoblotted with anti-LqhIT₂ polyclonal antibody. Lanes 1, 2, 4, 6, 8, 10, 12, and 14 were cell extracts from vAPcmIT₂-infected Sf21 cells. Lanes 3, 5, 7, 9, 11, 13, and 15 were cell extracts from vAP10IT₂-infected Sf21 cells, and lane 16 was cell extracts from AcMNPV-infected Sf21 cells at 48 h post infection. Kaleidoscope pre-stained standards (Bio-Rad, catalog 161-0324) were bovine serum albumin 85 kDa, carbonic anhydrase 41.8 kDa, soybean trypsin inhibitor 31.8 kDa, lysozyme 18 kDa, and aprotinin 6.4 kDa.

TUAN et al. — Engineered baculovirus for better plant protection

sion in infected cells (vs wild-type AcMNPV-infected cells), we feel confident that LqhIT₂ is recognized by the antibody used in this experiment (Figure 2, lane 16). Even though the detectable level of LqhIT₂ toxin in the vAPcmIT₂-infected cells was not as abundant as in vAP10IT₂-infected cells after 48 h p.i., the efficacy of insect killing was faster and higher for the former than for the latter viruses. This might be due to the better secretion of the toxin protein by the p-PCm promoter early in the viral infection. Since the vast volume of medium is 50 × greater than that of the cells, we tried and failed to probe the LqhIT₂ toxin from the medium of the cells infected with recombinant viruses by Western blotting.

Dose- and Time-Mortality Responses of Infected Insect

Bioassays were carried out to measure the insecticidal efficacy of the recombinant viruses against T. ni. The larvae infected with the scorpion toxin-containing recombinant viruses showed a prolonged flaccid paralysis, and they were effectively immobilized, rather than swollen and liquefied as with the wild-type AcMNPV and vAP10G. The LC₅₀ values of wild-type AcMNPV, vAP10G, vAP10IT₂, and vAPcmIT₂ to the early 4^th-instar T. ni larvae were 1.43,

2.76, 1.40, and 1.36 PIBs/mm³, respectively, and differences were not significant (Table 1). Although toxin-expressing recombinant viruses killed insect larvae much more quickly than non-toxin-expressing viruses, the host LC₅₀ values of their 4^th-instar larvae were not significantly different. However for the 3^rd-instar larvae, the EC₅₀ values of wild-type AcMNPV, vAP10IT₂, and vAPcmIT₂ were 1.00, 0.19, and 0.17 PIBs/mm³, respectively (Table 2). This is a more than fivefold difference in virulence between the recombinants and wild-type AcMNPV. In addition, the ET₅₀ values against the 3^rd-instar larvae of T. ni at a concentration of 78 PIBs/mm³ were 80, and 66 h, respectively for vAP10IT₂, and vAPcmIT₂. For either lower or higher concentrations as 0.78 or 78 PIBs/mm³, vAPcmIT₂ resulted 18% (17 h) or 12% (14 h) reductions, respectively, in effective time compared to vAP10IT₂ (Table 2). The time of paralysis of the 4^th instar T. ni fed with the recombinant baculoviruses (measured by immobility of the larvae) was determined. In general, the effective time (ET₅₀) was 9~12 h earlier than the lethal time (LT₅₀). Insects infected with vAPcmIT₂ and vAP10IT₂ provided significantly lower ET₅₀ values, 83.2 and 87.1 h, respectively, and they were 19~41 h earlier than those of larvae infected with wild type AcMNPV and vAP10G. The ET₉₀ values also demonstrated that toxin-expressing recombinants resulted in 36~61 h ear

Botanical Bulletin of Academia Sinica, Vol. 46, 2005

lier response times than the non-recombinant viruses. The LT₅₀ values against the 4^th-instar T. ni at a high concentration of 78 PIBs/mm³ were 112.2, 130.4, 98.6, and 92.9 h, for infection by AcMNPV, vAp10G, vAP10IT₂, and vAPcmIT₂, respectively (Table 3). The LT₅₀ of these viruses was also significantly different at 7.8 as well as at 78 PIBs/mm³. A 20-73% reduction in LT₅₀ or effective time (ET₅₀) was found using vAPcmIT₂ compared to wild-type AcMNPV (Table 3). The baculoviruses expressing the depressant toxin showed improved insecticidal efficacy (faster paralysis) compared to the toxin-negative viruses (Figure 3). At 60 to 84 h p.i. of the 4^th instar larvae, 10~20% higher mortality resulted from vAPcmIT₂ compared to vAP10IT₂. At 96 h p.i., more than 85% of tested insects were killed by either vAPcmIT₂ or vAP10IT₂ while only 10% of the larvae were killed by either vAP10G or wild-type AcMNPV at the same incubation time. It took 156 h for vAP10G and wild type AcMNPV to kill 85-98% of the insect larvae (Figure 3).

Field Experiments

In the field, both toxin-expressing recombinant viruses killed the 4^th-instar larvae of T. ni faster than the parental or the GFP-expression recombinant viruses. Toxin expression by the recombinant virus induced larval paralysis in

the field, resulting in many larvae ceasing to feed and falling off the plants before death. Over 20% of larvae treated with vAPcmIT₂ were found paralyzed/dead on the soil, compared to none in wild-type treatments at the 3^rd day post application (p.a.) treated at 10⁷ PIBs/ml. On the 6^th day p.a., more than 80% of larvae infected with vAPcmIT₂ and vAP10IT₂ fell off the plants before death while most larvae infected with vAp10G and wild-type AcMNPV kept eating and grew into 5^th-instar larvae. The mortality and average body weight of larvae on the 6^th day p.a. were significantly different between toxin-expressing recombinant viruses and non-toxin viruses. By 6 days p.a., the mortalities of larvae infected with wild-type AcMNPV, vAp10G, vAP10IT₂, and vAPcmIT₂ were 96.7, 67.4, 96.7, and 95.1%, respectively (Table 4).

Virus treatment significantly reduced insect damage compared to the untreated controls, but vAPcmIT₂ and vAP10IT₂ showed no significant difference. Larvae fed toxin-expressing recombinants ate less than larvae treated with wild-type AcMNPV, resulting in a 22% reduction of body weight (Table 4). Untreated larvae caused 842 cm² loss in leaf area while the vAPcmIT₂ and vAP10IT₂-treated plots exhibited a dramatic reduction in cabbage plant damage, with up to 64% less leaf area consumed compared

TUAN et al. — Engineered baculovirus for better plant protection

Figure 4. Comparison of leaf area eaten by T. ni 4^th-instar larvae infected with recombinant and wild-type AcMNPV at the 6^th day post application. All viral suspensions were adjusted to a concentration of 10⁷ polyhedral inclusion bodies (PIBs)/ml with 2,000 fold-diluted Triton^®. Means labeled with different letters are significantly different at the 5% level analyzed at P<0.05, Least Significant Difference. F value = 102.6>F (4, 10, 0.05).

Figure 3. Cumulative percentage of T. ni 4^th-instar larvae that had ceased to respond to stimuli after inoculation with recombinant or wild type AcMNPV at a concentration of 78 polyhedral inclusion bodies (PIBs)/mm³.

to control (Figure 4). Toxin-expressing recombinant-treated plots had significantly less crop damage, about 16~18% less than those treated with the wild-type AcMNPV, which resulted in about 50% less loss compared to vAp10G. Most insects infected with vAPcmIT₂ and vAP10IT₂ fell onto the ground while other insects infected with wild-type AcMNPV and vAp10G usually remained on the plant after death, where they liquefied and released of large quantities of viruses (Figure 5).

Discussion

The scorpion toxin-expressing recombinant viruses always killed host larvae more quickly than the wild-type AcMNPV regardless of larval size, and the maximal efficacy would occur in the early stages of pest development (Smits and Vlak, 1988; McCutchen et al., 1991; Hoover et al., 1995; Ignoffo and Garcia, 1996; Hughes et al., 1997; Ignoffo and Garcia, 1997; Gershburg et al., 1998; Milks et al., 1998; Harrison and Bonning, 2001). To speed up the time required to kill or paralyze pest insects, however depends upon the action of the toxin and on the species and instars of the tested insects. The timing and efficacy of the promoter used to drive the toxin gene is also very important (Lu et al., 1996; Prikhod'ko et al., 1998; Harrison and Bonning, 2000). The appropriate choice and combination of different signal sequences or promoters could maximize the efficacy of insect-selective toxin-expressing recombinant viruses (van Beek et al., 2003). Furthermore, the speed of kill of toxin-expressing recombinant could be further reduced by coproducing synergistic toxins (Prikhod'ko et al., 1998; Regev et al., 2003).

The very late p10 promoter is a popular choice for driving a high-level of foreign gene expression in baculoviruses. Although it is slightly weaker than the polyhedrin promoter, it activates foreign gene expression

Figure 5. Treatment of T. ni larvae with wild-type or recombinant AcMNPV. Figures showed the infection of larvae with different viruses. A: wild-type AcMNPV; B: vAP10G; C: vAP10IT2; D: vAPcmIT2; E: uninfected larvae as a control.

Botanical Bulletin of Academia Sinica, Vol. 46, 2005

gested that viruses, whether recombinants or wild type, should be applied at early stages of pest development to maximize the bioinsecticidal activity and plant protection efficacy.

All the recombinant viruses used in our studies were derived from linear commercial viral DNA rather than the genome of wild-type baculovirus. As the bioassay data showed, the highest LC₅₀ for the larvae (Table 1), the highest rate of leaf area eaten by the treated larvae (Figure 4), and the lowest level of larval mortality with the highest body weight (Table 4) were all performed by the vAP10G. The efficacy of vAP10G was not only worse than our toxin-gene containing viruses, it also significantly underperformed the wild type virus. This suggests that if a field-collected wild type virus were used for the insertion of LqhIT₂ in later experiments, significant better control of pest insects would be achieved using the same p-PCm promoter.

Acknowledgment. We thank Mr. D. Chamberlin for careful reading and editing of the manuscript. This work was supported by grants from Academia Sinica, AS93-AB-IMB-01, and from National Science Council, Taiwan, ROC, NSC 90-2313-B-001-014.

Literature Cited

Benkhalifa, R., M. Stankiewicz, B. Lapied, M. Turkov, N. Zilberberg, M. Gurevitz, and M. Pelhate. 1997. Refined electrophysiological analysis suggests that a depressant toxin is a sodium channel opener rather than a blocker. Life Sci. 61: 819-830.

Bonning, B.C. and B.D. Hammock. 1996. Development of recombinant baculoviruses for insect control. Annu. Rev. Entomol. 41: 191-210.

Chao, Y.C., S.L. Chen, and C.F. Li. 1996. Fluorescent insects by infection. Nature 380: 396-397.

Entwistle, P.F. and H.F. Evans. 1985. Viral control. In G.A. Kerkut and L.I. Gilbert (eds.), Comprehensive Insect Physiology Biochemistry and Pharmacology. Pergamon Press, New York, Vol. 7, pp. 347-412.

Finney, D.J. 1971. Probit analysis. 3^rd ed. Cambridge Univ. Press. London, 333 pp.

Fuxa, J.A., J.R. Fuxa, and A.R. Richter. 1998. Host-insect survival time and disintegration in relation to population density and dispersion of recombinant and wild-type nucleopolyhedroviruses. Biol. Control 12: 143-150.

Gershburg, E., D. Stockholm, O. Froy, S. Rashi, M. Gurevitz, and N. Chejanovsky. 1998. Baculovirus-mediated expression of a scorpion depressant toxin improves the insecticidal efficacy achieved with excitatory toxins. FEBS Letters 422: 132-136.

Granados, R. and K. Williams. 1986. In vivo infection and replication of baculoviruses. In R.R. Granados and B.A. Federici (eds.), The Biology of Baculoviruses, Vol. 1. CRC Press, Boca Raton, Florida, pp. 89-108.

Hammock, B.D., B.C. Bonning, R.D. Possee, T.N. Hanzlik, and S. Maeda. 1990. Expression and effects of the juvenile hormone esterase in a baculovirus vector. Nature 344: 458-461.

Harrison, R.L. and B.C. Bonning. 2000. Use of scorpion neuro

earlier (Roelvink et al., 1992). In the experiment on toxin protein expression by two distinct, temporally regulated viral promoters for pest control using recombinant baculoviruses, Gershburg et al. (1998) revealed advantages for the use of the very late p10 promoter over the early p35 promoter. In a further experiment, we showed that the early p-PCm promoter could be better than the very late p10 promoter for controlling insect pests. This is probably due to an early and sufficient level of LqhIT₂ scorpion depressant toxin protein expression by the former promoter resulting in an earlier cessation of feeding.

Bioassays on the 2^nd-instar larvae of Heliothis virescens (Fabricius) with recombinant virus expressing AaIT, an excitatory neurotoxin from Androctonus australis hector (Ewing) controlled by the very late p10 promoter, demonstrated a significant 30% decrease in the time to kill (LT₅₀ 88 hrs) compared to wild-type AcMNPV (LT₅₀ 125 hrs) (McCutchen et al., 1991). Compared with AaIT, the depressant neurotoxin LqhIT₂ causes qualitatively different neuro-physiological effects and different symptoms when injected into blowfly larvae (Zlotkin et al., 1993). Expression of LqhIT₂ reduced survival times of Heliothis zea (Boddie) and H. virescens to a significantly greater extent than expression of AaIT, and recombinant viruses expressing toxin from the late p6.9 promoter killed larvae faster than a recombinant virus utilizing the very late p10 promoter (Harrison and Bonning, 2000). Furthermore, better results could be achieved most recently by the combined expression of the excitatory toxin, LqhIT₁ and the depressant toxin, LqhIT₂ than if each had been applied alone using recombinant viruses (Regev et al., 2003).

In our studies, we demonstrated that the early p-PCm promoter provides for earlier and relatively effective expression. At 12 h p.i., expression of LqhIT₂ by the early p-PCm promoter was evident, and expression by the very late p10 promoter can only be observed beginning by 36 h p.i. Although the very late p10 promoter expressed significantly more proteins at the peak of toxin expression, the expression of the early p-PCm promoter was sufficient, so that early expression proved to enhance the control of pest insects. Furthermore, although the total amount of the engineered luciferase protein produced by the early p-PCm promoter in our previous study was less than that produced by the very late p10 promoter, significantly less protein degradation and higher total luciferase enzymatic activities were both found by using the early p-PCm promoter (Wu et al., 2000; Lo et al., 2002). The better quality of the protein produced by the early p-PCm promoter was further reflected in bioassay tests. Both the effective time and the lethal concentration (EC₅₀ in Table 2) data demonstrated that the control efficacy of vAPcmIT₂ was superior to that of vAP10IT₂ against middle instars of T. ni. We also found more significant differences between ET₅₀ and EC₅₀ occurred in the 3^rd -instar T. ni larvae than in the 4^th-instar (Table 2 and data not shown). Larvae always became less susceptible to baculoviruses with age for ten- to thousand-fold differences of susceptibility between younger and elder larvae (Smite and Vlak, 1988; Tuan et al., 1994; 1995; 1999). Therefore, these experiments sug

TUAN et al. — Engineered baculovirus for better plant protection

toxins to improve the insecticidal activity of Rachiplusia ou multicapsid nucleopolyhedrovirus. Biol. Control 17: 191-201.

Harrison, R.L. and B.C. Bonning. 2001. Use of proteases to improve the insecticidal activity of baculoviruses. Biol. Control 20: 199-209.

Hoover, K., C.M. Schultz, S.S. Lane, B.C. Bonning, S.S. Duffey, B.F. McCutchen, and B.D. Hammock. 1995. Reduction in damage to cotton plants by a recombinant baculovirus that knocks moribund larvae of Heliothis virescens off the plant. Biol. Control 5: 419-426.

Hughes, P.R., H.A. Wood, J.P. Breen, S.F. Simpson, A.J. Duggan, and J.A. Dybas. 1997. Enhanced bioactivity of recombinant baculoviruses expressing insect-specific spider toxins in lepidopteran crop pests. J. Invertebr. Pathol. 69: 112-118.

Ignoffo, C.M. and C. Garcia. 1996. Rate of larval lysis and yield and activity of inclusion bodies harvested from Trichoplusia ni larvae fed a wild or recombinant strain of the nuclear polyhedrosis virus of Autographa californica. J. Invertebr. Pathol. 68: 196-198.

Ignoffo, C.M. and C. Garcia. 1997. Effect of viral concentration and body weight on mortality of larvae of Trichoplusia ni (Lepidoptera: Noctuidae) exposed to wild-type or recombinant strains of the nuclear polyhedrosis virus of Autographa callifornia (Lepidoptera: Noctuidae). Biol. Control 26: 955-960.

Jarvis, D.L., L.M. Reilly, K. Hoover, C. Schulz, B.D. Hammock, L. Guarino. 1996. Construction and characterization of immediate early baculovirus pesticides. Biol. Control 7: 228-235.

Lee, J.C., H.H. Chen, and Y.C. Chao. 1998. Persistent baculovirus infection results from deletion of the apoptotic suppressor gene p35. J. Virol. 72: 9157-9165.

Lin, J.L., J.C. Lee, M.L. Li, and Y.C. Chao. 1999. Persistent Hz-1 virus infection in insect cells: evidence for viral DNA insertion in host chromosomes and viral infection in a latent status. J. Virol. 73: 128-139.

Lo, H.R., C.C. Chou, T.Y. Wu, J.P.Y. Yuen, and Y.C. Chao. 2002. Novel baculovirus DNA elements strongly stimulate activities of exogenous and endogenous promoters. J. Biol. Chem. 277: 5256-5264.

Lu, A., S. Seshagiri, and L.K. Miller. 1996. Signal sequence and promoter effects on the efficacy of toxin-expressing baculoviruses as biopesticides. Biol. Control 7: 320-332.

Maeda, S. 1989. Increased insecticidal effect by a recombinant baculovirus carring a synthetic diuretic hormone. Biochem. Biophys. Res. Commun. 165: 1177-1183.

Maeda, S., S.L. Volrath, T.N. Hanzlik, S.A. Harper, K. Majima, D.W. Maddox, B.D. Hammock, and E. Fowler. 1991. Insecticidal effects of an insect-specific neurotoxin expressed by a recombinant baculovirus. Virology 184: 777-780.

Martens, J.W.M., G. Honee, D. Zuidema, J.W.M. van Lent, B. Visser, and J.M. Vlak. 1990. Insecticidal activity of a bacterial crystal protein expressed by a recombinant baculovirusin insect cells. Appl. Environ. Microbiol. 56: 2764-2770.

Martens, J.W.M., M. Knoester, F. Weijts, S.J.A. Groffen, Z. Hu, D. Bosch, and J.M. Vlak. 1995. Characterization of baculovirus insecticides expressing tailored Bacillus thuringiensis CryIA(b) crystal proteins. J. Invertebr. Pathol.

66: 249-257.

McCutchen, B.F., P.V. Choudary, R. Crenshaw, D. Maddox, S.C. Kamita, N. Palekar, S. Volrath, E. Fowler, B.D. Hammock, and S. Maeda. 1991. Development of a recombinant baculovirus expressing and insect-selective neurotoxin: potential for pest control. Biotechnology 9: 848-852.

Milks, M.L., I. Burnstyn, and J.H. Myers. 1998. Influence of larval age on the lethal and sublethal effects of the nucleopolyhedrovirus of Trichoplusia ni in the cabbage looper. Biol. Control 12: 119-126.

Prikhod'ko, G.G., H.J.R. Popham, T.J. Felcetto, D.A. Ostlind, V.A. Warren, M.M. Smith, V.M. Garsky, J.W. Warmke, C.J. Cohen, and L.K. Miller. 1998. Effects of simultaneous expression of two sodium channel toxin genes on the properties of baculoviruses as biopesticides. Biol. Control 12: 66-78.

Regev, A., H. Rivkin, B. Inceoglu, E. Gershburg, B.D. Hammock, M. Gurevitz , and N. Chejanovsky. 2003. Further enhancement of baculovirus insecticidal efficacy with scorpion toxins that interact cooperatively. FEBS Letter 537: 106-110.

Roelvink, P.W., M.M.M. Van Meer, C.A.D. De Kort, R.D. Possee, B.D. Hammock, and J.M. Vlak. 1992. Dissimilar expression of Autographa californica multiple nucleocapsid nuclear polyhedrosis virus polyhedrin and p10 genes. J. Gen. Virol. 73: 1481-1489.

Smits, P.H. and J.M. Vlak. 1988. Biological activity of Spodoptera exigua nuclear polyhedrosis virus against S. exigua larvae. J. Invertebr. Pathol. 51: 107-114.

Steel, R.G.D. and J.H. Torrie. 1980. "Principles and Procedures of Statistics: A Biometrical Approach" 2^nd ed., McGraw-Hill, New York.

Stewart, L.M.D., M. Hirst, M.L. Ferber, A.T. Merryweather, P.J. Cayley, and R.D. Possee. 1991. Construction of an improved baculovirus insecticide containing an insect-specific toxin gene. Nature 352: 85-88.

Tinsley, T.W. and D.C. Kelly. 1985. Taxonomy and nomenclature of insect pathogenic viruses. In K. Maramorosch and K.E. Sherman (eds.), Viral Insecticides for Biological Control. Academic Press, London, pp. 3-26.

Tomalski, M.D., and L.K. Miller. 1991. Insect paralysis by baculovirus-mediated expression of a mite neurotoxin gene. Nature 352: 82-85.

Tomalski, M.D., and L.K. Miller. 1992. Expression of a paralytic neurotoxin gene to improve insect baculoviruses as pesticides. BioTechnology 10: 545-549.

Tuan, S.J., S.S. Kao, and D.J. Cheng. 1994. Histopathology and pathogenicity of Spodoptera exigua nuclear polyhedrosis virus isolated in Taiwan. Chinese J. Entomol. 14: 33-45.

Tuan, S.J., S.S. Kao, D.J. Cheng, Roger F. Hou, and Y.C. Chao. 1999. Comparison of the characterization and pathogenesis of three lepidopteran nucleopolyhedroviruses (HearNPV, SpeiNPV and SpltNPV) isolated from Taiwan. Chinese J. Entomol. 19: 167-186.

Tuan, S.J., S.S. Kao, U.L. Leu, and D.J. Cheng. 1995. Pathogenicity and propagation of Spodoptera litura nuclear polyhedrosis virus isolated in Taiwan. Chinese J. Entomol. 15: 19-33.

Tuan, S.J., S.S. Kao, Y.C. Chao, and R.F. Hou. 1997. Investigation of pathogenicity of AcMNPV to nine lepidopteran pests in Taiwan. Chinese J. Entomol. 17: 209-225.

Tuan, S.J., W.L. Chen, and S.S. Kao. 1998. In vivo mass pro

Botanical Bulletin of Academia Sinica, Vol. 46, 2005

duction and control efficacy of Spodoptera litura (Lepidoptera: Noctuidae) nucleopolyhedrovirus. Chinese J. Entomol. 18: 101-116.

van Beek, N., A. Lu, J. Presnail, D. Davis, C. Greenamoyer, K. Joraski, L. Moore, M. Pierson, R. Herrmann, L. Flexner, J. Foster, A. Van, J. Wong, D. Jarvis, G. Hollingshaus, and B. McCutchen. 2003. Effect of signal sequence and promoter on the speed of action of a genetically modified Autographa californica nucleopolyhedrovirus expressing the scorpion toxin LqhIT₂. Biol. Control 27: 53-64.

Vaughn, J.L., R.H. Goodwin, G.J. Tompkins, and P. McCawley. 1977. The establishment of two cell lines from the insect Spodoptera frugiperda (Lepidoptera: Noctuidae). In Vitro 13: 213-217.

Volkman, L.E. 1995. Virus taxonomy: the classification and nomenclature of viruses. In F.A. Murphy, C.M. Fauquet, D.H.L. Bishop, S.A. Ghabrial, A.W. Jarvis, G..P. Martelli, M.A. Mayo, and M.D. Summers (eds.), The Sixth Report of

the ICTV. Springer-Verlag Wien, Inc., New York, pp. 104-113.

Whitt, M.A. and J.S. Manning. 1988. A phosphorylated 34-kda protein and a subpopulation of polyhedrin are thiol-linked to the carbohydrate layer surrounding a baculovirus occlusion body. Virology 163: 33-42.

Wu, T.Y., D.G. Lin, S.L. Chen, C.Y. Chen, and Y.C. Chao. 2000. Expression of highly controllable genes in insect cells using a modified tetracycline-regulated gene expression system. J. Biotechnol. 80: 75-83

Zilberberg, N, E. Zlotkin, and M. Gurevitz. 1991. The cDNA sequence of a depressant insect selective neurotoxin from the scorpion Buthotus judaicus. Toxicon 29: 1155-1158.

Zlotkin, E., M. Gurevitz, E. Fowler, and M.E. Adams. 1993. Depressant insect selective neurotoxins from scorpion venom: chemistry, action, and gene cloning. Arch. Insect Biochem. Physiol. 22: 55-73.