Abstract
Baculoviruses infect their lepidopteran hosts via the midgut epithelium through binding of occlusion-derived virus (ODV) and fusion between the virus envelope and microvillar membranes. To identify genes and sequences that are involved in this process, a random phage display library was screened for peptides that bound to brush border membrane vesicles (BBMV) derived from the midgut epithelium of Heliothis virescens. Seventeen peptides that bound to BBMV were recovered. Two of these, HV1 and HV2, had sequence similarity to the ODV envelope protein ODV-E66 that is found in five species of alphabaculoviruses. Chemically synthesized versions of HV1 and HV2, and two peptides (AcE66A and AcE66B) derived from similar sequences of Autographa californica multiple nucleopolyhedrovirus (AcMNPV) ODV-E66, bound to unfixed cryosections of whole midgut tissues. AcE66A, but not HV1, bound to H. virescens gut BBMV proteins on a far-Western blot. Competition assays with HV1 and purified AcMNPV ODV resulted in decreased mortality of H. virescens larvae at a dose of 1 LD50, and a significant increase in survival time at higher virus concentrations. These results suggest a role for ODV-E66 in baculovirus infection of lepidopteran larval midgut epithelium.
Introduction
Baculoviruses are large dsDNA viruses that infect insects within the orders Lepidoptera, Diptera and Hymenoptera. The family Baculoviridae contains four genera, Alphabaculovirus, Betabaculovirus, Gammabaculovirus and Deltabaculovirus (Jehle et al., 2006), with Alphabaculovirus being divided into groups I and II on the basis of molecular phylogeny and the presence or absence of the gp64 envelope-protein gene. Baculoviruses are distinct among viruses in producing two different virion phenotypes, budded virus (BV), which functions to disseminate virus infection within an infected host, and occlusion-derived virus (ODV), which serves to spread infection among host larvae. ODVs are enclosed in a highly stable crystalline protein matrix to form viral occlusions, otherwise known as polyhedra. When polyhedra are consumed by lepidopteran larvae, the highly alkaline pH of the midgut dissolves the polyhedra, releasing the ODV into the gut lumen. The ODVs cross the peritrophic membrane lining the gut and bind to the brush border microvilli of midgut columnar cells (Horton & Burand, 1993). Entry of virions into the microvilli is mediated by fusion between the viral envelope and the microvillar plasma membrane (Haas-Stapleton et al., 2004). The mechanisms involved in binding and fusion of ODV to the midgut epithelial cells are incompletely understood.
More than 30 proteins are found in ODV (Braunagel et al., 2003), and six ODV envelope proteins have been characterized as per os infectivity factors (PIFs) that facilitate oral infectivity (Sparks et al., 2011). ODV-E66 is found in the ODV envelope and is a core gene in the lepidopteran baculoviruses (Braunagel et al., 2003; Braunagel & Summers, 2007; Herniou et al., 2003; Slack & Arif, 2007; van Oers & Vlak, 2007). ODV-E66 is also found in salivary gland hypertrophy virus (Garcia-Maruniak et al., 2009) and in nudiviruses (Wang & Jehle, 2009), which suggests the existence of common mechanisms for midgut infection by viruses of invertebrates. ODV-E66 interacts with several other baculovirus proteins including ODV-E25 (Braunagel et al., 1999), PIF2 and PIF3 (Peng et al., 2010a). PIF-1, PIF-2 and PIF-3 have recently been shown to form a complex on the surface of ODV that does not include P74, but the whether the other PIFs and ODV-E66 are included is unknown (Peng et al., 2010b). ODV-E66 also has hyaluronidase activity and may function in penetrating extracellular barriers during initial infection (Vigdorovich et al., 2007). A role for ODV-E66 in primary infection of the midgut has yet to be established.
Phage display libraries have previously been found to encode peptides that compete with pathogens for binding to insect-host or -vector tissues. Screening such libraries resulted in the identification of the SM1 peptide, a mimotope (i.e. mimic of an epitope) of the coat protein of the malaria parasite. The SM1 peptide blocks entry of Plasmodium into the mosquito gut and salivary glands (Ghosh et al., 2001, 2009). The aphid gut binding peptide GBP3.1 was also identified by screening a phage display library and was recently shown to impede uptake of a plant virus, pea enation mosaic virus, by its pea aphid vector (Liu et al., 2010). Here, we describe the bio-panning of a random phage display library against H. virescens midgut brush border membrane vesicles (BBMV). The selected phage peptide sequences included two clones with similarity to the baculovirus occlusion-derived envelope protein ODV-E66. Remarkably, the majority of the 17 peptides selected had similarity to known insect gut binding pathogens, including a midgut-specific insect pathogen, as well as human pathogens of worldwide medical importance. Here we demonstrate competition of one of these peptides with AcMNPV for binding to the H. virescens gut.
Results
Phage selection and blast analysis of peptide sequences
The F88.4 filamentous phage display library was screened for peptides that bind to BBMV derived from the midgut epithelia of fourth instar H. virescens. After three rounds of bio-panning, 20 clones were selected for sequencing. Of the 20 clones, three phage clones had stop codons within the 12 aa peptide sequence. Of the 17 remaining clones, 15 encoded unique peptide sequences and two clones encoded identical peptide sequences. In addition, peptide sequences from two other clones exhibited high sequence similarity (Table 1). The 17 peptide sequences (excluding Cys2 and Cys11 which were conserved among the displayed peptides) were abundant in the amino acids proline (15 %, 26 of 170 residues), phenylalanine (10.6 %, 18 of 170 residues), leucine (10 %, 17 of of 170 residues), threonine (9.4 %, 16 of 170 residues), alanine (7.6 %, 13 of 170 residues), histidine (7 %, 12 of 170 residues) and tryptophan (6.5 %, 11 of 170 residues).
For blast alignments: +, an amino acid mismatch of the same amino acid classification; −, a mismatch with a different amino acid classification. Baculovirus species abbreviations: Mamestra configurata NPV-A and B (MacoNPV), Hyphantria cunea NPV (HycuNPV), Agrotis segetum NPV (AgseNPV), Spodoptera exigua NPV (SeNPV), Adoxophyes orana GV (AdorGV), Choristoneura occidentalis GV (ChorGV), Cadra cautella NPV (CacaNPV), Autographa californica NPV (AcNPV), Neodiprion lecontei NPV (NeleNPV), Adoxophyes honmai NPV (AdhoNPV), Culex nigripalpus NPV(CuniNPV), Helicoverpa zea SNPV (HzSNPV), Helicoverpa armigera NPV (HearNPV), Rachiplusia ou NPV (RoNPV) and Bombyx mori NPV (BmNPV).
A blastp search of the short peptides, both the complete 12 aa peptide sequence including the conserved cysteines at positions 2 and 11, and the eight internal amino acids, was carried out and the results were examined for matches to other known gut pathogens (Table 1).
The H. virescens BBMV binding peptides showed similarity to several insect-gut pathogens, some with insect vectors or insect hosts, including: Pseudomonas entomophila, a pathogenic bacterium infecting Drosophila midguts (Vodovar et al., 2006); white spot syndrome virus of shrimp (van Hulten et al., 2001); and Leishmania, Trypanosoma and Plasmodium, which have complex life cycles involving sand flies, Tsetse flies and mosquitoes, respectively, and cause leishmaniasis, sleeping sickness and Chagas disease, and malaria in humans. Similarities were also seen to Sindbis virus and dengue virus, which infect the midgut epithelia of Aedes aegypti (Lambrechts & Scott, 2009; Pierro et al., 2008). The proteins in this diverse range of insect gut-associated microorganisms that showed similarity to the H. virescens gut BBMV binding peptides were primarily proteins of unknown function.
Four of the H. virescens gut BBMV binding peptides had similarity to five different baculovirus proteins derived from predicted ORFs of unknown function, including homologues of Ac75, which is a conserved lepidopteran baculovirus protein. They also showed similarity to the VP1054 capsid protein (another core baculovirus protein) from two species of Betabaculovirus, and homologues of Ac25, a core lepidopteran protein involved in binding to ssDNA.
ODV-E66 alignment
Two of the H. virescens gut binding peptides had similarity to the occlusion-derived envelope protein ODV-E66. These peptides were aligned with the lepidopteran ODV-E66 baculovirus sequences identified from blastp using Autographa californica multiple nucleopolyhedrovirus (AcMNPV) as the consensus sequence (Fig. 1). HV1 and HV2 showed similarity to both Group I and Group II nucleopolyhedrovirus (NPV) ODV-E66 sequences (Fig. 1). Group I alphabaculovirus ODV-E66 sequences are shown above the solid line and Group II sequences are below the line. The phage peptides localized to the same two regions within ODV-E66 across five different species of baculovirus. The regions encompassed amino acids 84–96 (region 1) and 179–192 (region 2) of AcMNPV ODV-E66 (Fig. 1), with region 2 being more conserved across baculovirus species at the amino acid level. The peptide SCWAVFSATLCT, designated HV1, had similarity to region 2 of Agrotis segetum NPV (AgseNPV) ODV-E66 and to region 2 of both ODV-E66 homologues (ORF 77/78 and ORF 143/144) in both species of Mamestra configurata NPV, A and B (MacoNPV-A and MacoNPV-B). The second peptide, HCSVWHVFAQCT, designated HV2, exhibited sequence similarity to several baculoviruses at two different locations: first, in region 1 of Hyphantria cunea NPV (HycuNPV); second, a 3–4 aa identity match (WHVF) in region 2 of AgseNPV, MacoNPV-A and MacoNPV-B, and also Spodoptera exigua MNPV (SeMNPV) ORF 114 (Fig. 1).
blastp alignment of peptides (HV1, HV2, AcE66A and AcE66B) with two separate regions of baculovirus ODV-E66. AcE66A and AcE66B peptides were created based on the alignment of the phage peptides HV1 and HV2 to the AcMNPV sequence. The line represents the division between Group I (above) and Group II (below) alphabaculoviruses. Both the A and B strains of MacoNPV and SeMNPV contain two copies of ODV-E66. Amino acids in bold type show the AcMNPV ODV-E66-derived sequence in the synthesized peptides AcE66A and AcE66B, and the similarities of HV1 and HV2 with ODV-E66 amino acid sequences from other baculoviruses. Sequences included in each comparison are indicated by brackets. Underlined amino acids indicate residues absent from HV1 and HV2, but which are shared with AcMNPV ODV-E66. ~, Spaces introduced for clarity of presentation of this figure.
Binding specificity of selected peptides
To assess whether the selected peptides HV1 and HV2, and peptides derived from the homologous regions of AcMNPV ODV-E66 (AcE66A and AcE66B; Fig. 1) bound to the H. virescens midgut, the peptides were synthesized with a spacer and a 5-FAM (carboxyfluoroscein) label. Two additional peptides were also synthesized: the aphid gut binding peptide GBP3.1 and a negative control peptide, C6 (Liu et al., 2010). Cryosections of H. virescens gut were incubated with the peptides and representative images from six insects using the same exposure times are shown in Fig. 2(a). The HV1 peptide (similar to region 2 of Group II ODV-E66) and AcE66A peptide (region 1 of AcMNPV ODV-E66, Group I) exhibited the strongest binding to gut sections, with consistently weaker binding seen for HV2 and peptide AcE66B (region 2 of AcMNPV ODV-E66; Fig. 1). The aphid gut binding peptide GBP3.1 also bound weakly to H. virescens midgut sections, while little to no binding was observed with the negative-control peptide C6. No fluorescence was seen in the PBS control (not shown).
Representative images of peptide binding to insect midguts. (a) Comparison of peptide binding to cryosections of H. virescens midgut epithelium. The fluorescently labelled peptides were incubated on cryosections of midgut tissues in sodium carbonate buffer (pH 9.0) for 1 h and viewed using a FITC filter. Exposure, 1 s. (b) Peptide binding to the midguts of H. virescens and the aphids M. persicae and A. pisum. Bright-field images are shown for reference. Exposure, 154 ms. Bars, 100 µm.
On finding that the aphid gut binding peptide GBP3.1 bound sections of the H. virescens gut, we also tested for binding of HV1 to the gut of two other aphid species, the pea aphid, Acyrthosiphon pisum, and the green peach aphid, Myzus persicae. HV1 bound to the guts of these species at levels similar to that seen for GBP3.1 (Fig. 2b).
Peptide binding to midgut BBMV
The ability of the synthetic peptides AcE66A, C6 and HV1 to bind to midgut BBMV proteins from fourth instar H. virescens was assessed by far-Western blotting (Fig. 3). In addition to BBMV, crude homogenates of H. virescens gut and samples from the low-speed centrifugation that pellets larger cellular particles in the BBMV protocol were included for analysis. While binding to multiple proteins in all three samples was seen for AcE66A, no binding was detected for HV1 or the negative control peptide, C6 (Fig. 3).
AcE66A binding to H. virescens midgut proteins. AcE66A, C6 and HV1 were assessed for binding to proteins from BBMV, crude homogenate (H), and the low-speed pellet (P) of cellular debris from homogenized midguts of fourth instar H. virescens. Arrows indicate the strongest binding of AcE66A to H. virescens midgut proteins. No binding was detected for the C6 or HV1 peptides in three independent experiments.
Peptide competition assays with baculovirus
To address whether the H. virescens gut binding peptides compete with baculovirus virions for binding to the H. virescens midgut epithelium, we co-fed newly moulted fourth instar larvae with a molar excess of peptide (2 µg) and 1 µg purified ODV of the recombinant virus AcIE1TV3.EGFP (Harrison et al., 2010). Of the five peptides tested (HV1, HV2, AcE66A, AcE66B and GBP3.1), the virus-plus-peptide treatment resulted in 100 % mortality, as did the virus-only treatment (Table 2). Insects fed peptide alone displayed no mortality and were indistinguishable from larvae fed PBS alone. Interestingly, insects which received the virus-plus-HV1 treatment had a significant increase in the number of larvae surviving to the fifth instar that was not observed in any of the other virus-plus-peptide treatments (Table 2). HV1 was then assessed for effects at multiple doses of ODV. Insects fed HV1 peptide alone displayed no detectable differences in insect growth or vitality compared to the negative controls treated with PBS alone. At high viral doses, all insects receiving virus alone succumbed to polyhedrosis, as did all insects fed a mixture of virus and HV1. At a dose of 1 LD50 (0.006 µg) ODV, a significant decrease in mortality was observed in the presence of HV1 compared with the virus-only treatment (P<0.05, anova, Student’s t-test; Fig. 4a). In addition, increased survival of larvae to the fifth instar was observed in the presence of HV1 at viral doses of 0.05, 0.5 and 1 µg ODV, compared with the virus-only treatments (P<0.05, χ2 test; Fig. 4b). This effect was not seen at the highest dose of 1.5 µg ODV.
A χ2 test was used to determine significant differences in survival time. Insects were starved overnight and then droplet fed with 1 µg ODV (AcIE1TV3.EGFP) and 2 µg peptide. These results are each from three independent replicates of 10 to 16 insects.
Reduced viral infection and increased survival time in H. virescens co-inoculated with the HV1 peptide and ODV. (a) ODV competition bioassay with the HV1 peptide. Mortality resulting from different doses of ODV alone (striped bars) and ODV with HV1 (solid bars) is indicated. No mortality was observed following feeding larvae with HV1 alone. Error bars show SEM. *, Statistically significant reduction in mortality in the presence of peptide (P<0.05 by anova and Student’s t-test). (b) Proportions of H. virescens larvae dying from NPV infection at fourth and fifth instars when co-inoculated with HV1 and ODV. Grey bars, insects fed ODV alone; black bars, insects fed ODV and HV1; chequered grey bars, H. virescens fed ODV alone that survived to fifth instar; striped black, ODV- and HV1-fed larvae that survived to fifth instar. *, Statistically significant increases in survival to fifth instar compared with inoculation with ODV alone (P<0.05 by χ2 analysis).
Discussion
Screening of a phage display library for peptides that bound H. virescens midgut BBMV resulted in the isolation of 17 peptides with sequence similarities to proteins from a variety of microorganisms associated, in various ways, with the guts of insects. Taken together, phenylalanine, tryptophan and histidine were abundant in the 17 gut binding peptides, and these amino acids are known to be over-represented in the binding sites of carbohydrate-binding proteins (Boraston et al., 2004; Cheng et al., 2009). Hence it is possible that peptide binding is mediated in part by sugar groups; indeed, there is mounting evidence for the involvement of glycans in the binding and uptake of insect-vector pathogens (Dinglasan et al., 2007; Dinglasan & Jacobs-Lorena, 2005). The peptide sequence similarity data also suggest that the baculovirus proteins encoded by ac75, vp1054 and ac25 warrant further investigation for potential roles in interacting with the host midgut epithelium. Of particular interest was alignment of two of the H. virescens BBMV-binding peptides with two separate regions of ODV-E66 across five different species of alphabaculoviruses. This result strongly implicates these two regions of ODV-E66 in ODV binding to midgut epithelial cells.
Peptides selected for gut binding in the lepidopteran H. virescens were able to bind to the guts of aphids, and vice versa. Peptide binding to the guts of insects with such disparate gut physiology probably reflects conserved protein and/or common glycan components of gut epithelia across insect taxa. This result highlights the potential for identifying common features shared across insect species, which could be useful for the development of tools for the management of insect pests of both medical and agricultural importance.
Although HV1 and AcE66A bound most strongly to cryosections of H. virescens gut (Fig. 2a), binding of AcE66A, but not HV1, to BBMV was seen in far-Western blots (Fig. 3). The fact that the sequences of HV1 and AcE66A align with different regions of ODV-E66 might suggest that the gut membrane protein component that associates with ODV-E66 amino acids 179–192 (similar to HV1), but not 84–96 (similar to AcE66A), loses its conformation under the denaturing conditions used for far-Western blotting. Similar results were obtained for GBP3.1 binding to aphid gut BBMV (not shown) highlighting that in vitro binding to BBMV may differ from in vivo binding events.
Co-feeding of HV1 with baculovirus ODV revealed competition between the peptide and virus as observed through decreased mortality at a dose of 1 LD50 in the presence of HV1 compared with virus alone, and increased survival of larvae to the fifth instar at higher doses. A greater proportion of larvae moulted to fifth instar prior to succumbing to NPV infection when ODV was administered with HV1, suggesting that the presence of HV1 at certain doses delayed the infection process at the midgut epithelium. The extent of this impediment was not sufficient to significantly affect the number of larvae that ultimately died from infection at these higher viral doses. The similarity of the HV1 sequence to that of ODV-E66, and decreased mortality on co-feeding HV1 and ODV at a dose of 1 LD50, suggests that HV1 interferes with ODV-E66 interaction with the H. virescens midgut epithelium. However, given the strong binding to the midgut in multiple species, HV1 could be blocking the interaction with other viral envelope proteins as well.
In terms of practical application, it may be feasible to use a peptide that competes for virus binding to the insect gut for protecting beneficial insects or laboratory colonies from virus infection. A prime candidate for such an approach is protection against tsetse fly, Glossina pallipides, cultures from infection with salivary gland hypertrophy virus (family Hytrosaviridae) (Kariithi et al., 2010), which also has an ODV-E66 homologue (Garcia-Maruniak et al., 2009).
In summary, we present evidence that the H. viresecens gut binding peptide HV1 impedes infection of the gut epithelium by AcMNPV ODV, most probably by competing for binding sites required for infection. Identification of the ligands bound by HV1 and ODV-E66 will provide insight into the mechanism of baculovirus infection. In addition, this study highlights the potential for the use of phage display libraries for the investigation of insect–pathogen molecular interactions.
Methods
Insects and brush border membrane vesicle (BBMV) preparation.
H. virescens (BioServe) were reared in a growth chamber on a 12 h light–12 h dark cycle at a constant 28 °C on an artificial diet (Southlands). For BBMV preparations, we used a modified method of Wolfersberger et al. (1987) where 25–35 midguts were homogenized in 9 ml MET buffer (300 mM mannitol, 2 mM EDTA, 17 mM Tris, pH 8.0), as described in Sparks et al. (2011). BBMV preparations were stored at −80 °C for up to 3 months. The quality of the BBMV was determined by aminopeptidase (APN; EC 3.4.11.2) activity assay using the Sigma protocol based on the method of Pfleiderer (1970). The BBMV protein concentration was quantified by the Bradford method using a Bio-Rad protein assay (Bio-Rad). APN activities were normalized for protein concentration, and BBMV with APN activity eightfold greater than the crude homogenate were used for further study.
Pea aphids, A. pisum Harris, were obtained from Berkshire Biological Supply Company (Westhampton, MA, USA) and reared on broad bean, Vicia faba. Green peach aphids, M. persicae (Sulzer), were reared on Chinese cabbage, Brassica rapa. All aphid colonies were maintained in growth chambers at 24 °C with a 12 h light–12 h dark cycle.
Phage selection.
The f88.4-LX8 phage display library (Bonnycastle et al., 1996) was used for bio-panning against the H. virescens gut. The phage library displays 12 amino acid peptides (XCX8CX, where X is any amino acid and C is cysteine). Phage were amplified and titrated using K-91 Escherichia coli (Invitrogen), and precipitated with 20 % polyethylene glycol as previously described (Smith & Scott, 1993). Twenty microlitres of phage stock (~1×1014 phage) was incubated with 100 µg of BBMV preparation in a total volume of 100 µl of sodium phosphate buffer (pH 9.0) at room temperature for 1 h. Unbound phage were then separated using a modified BRASIL method (Giordano et al., 2001). Briefly, the phage/BBMV mixture was overlaid on a single organic-phase mixture of 9 : 1 di-butyl phthalate : cyclohexane, and spun at 10 000 g for 10 min at 4 °C. Unbound phage remained in suspension, while the bound phage pelleted with the BBMV tissues. Immediately after centrifugation, the tubes were flash frozen in liquid nitrogen and then the bottom of the tube excised with a razor blade. This tube tip was placed in a sterile microcentrifuge tube with 300 µl of 50 mM glycine-HCl (pH 2.2) containing 1 mg BSA ml−1, and the tissue gently resuspended by inversion and placed on a rocker for 10 min to elute the phage. The reaction was neutralized with 8 µl 2 M Tris (pH 9.1). Fifteen microlitres of this mixture was used for titration and the remainder immediately amplified as before. After amplification, this enriched pool of phages was titrated and used for the next round of bio-panning. After three rounds of successive binding–amplification, the eluted phage were incubated with K-91 E. coli and plated onto NZY plates containing 20 µg ml−1 tetracycline. Individual bacterial clones were picked and amplified for isolation of genomic DNA using standard methods (Sambrook & Russell, 2001). The sequencing primer (5′-CTGAAGAGAGTCAAAAGC-3′) was used to determine the sequence of the random peptide.
Sequence analysis.
The short peptides identified from the phage bio-panning were analysed using blastp (Altschul et al., 1997) optimized for short read sequences. The sequence of the internal 8 aa, as well as the complete 12 aa sequence (including the loop-forming cysteines at positions 2 and 11), was analysed. Following the discovery of phage similarity to the baculovirus protein ODV-E66, those peptide sequences were aligned with all of the available ODV-E66 sequences using clustal w and BioEdit (Hall, 1999; Thompson et al., 1994).
Peptide synthesis and binding.
Peptides were synthesized by Neo-Peptide (NeoBioPharma), with N-terminal 5-FAM tags with a 6-aminohexanoic acid (ahx) spacer. The sequences of the peptides were as follows: aphid gut binding peptide GBP3.1, TCSKKYPRSPCM-OH (Liu et al., 2010); negative-control aphid peptide C6, FCRTADVIDACT-OH (Liu et al., 2010); H. virescens gut binding peptides HV1, SCWAVFSATLCT-OH and HV2, HCSVWHVFAQCT-OH; AcMNPV ODV-E66 homologous peptides AcE66A, TCLSYSFSQKCA and AcE66B, ACDWYHFTITCT-OH. Peptides were resuspended in PBS (pH 7.4) to a concentration of 14 mg ml−1 and stored in aliquots at −20 °C until use.
H. virescens midguts were dissected in PBS buffer (pH 7.4) and rinsed free of the peritrophic matrix which lines the gut cavity. Whole midguts of the green peach aphid (M. persicae), and the pea aphid (A. pisum) were dissected in PBS. The guts were then immediately placed in Tissue-Tek Optimal Cutting Temperature (O.C.T.) mounting medium (Ted Pella) and frozen. Cryosections, approximately 20 µM in depth, were made using a Universal cryostat on to ProbeOn Plus slides (Fisher Scientific). Slides were air-dried for 2 h then washed twice with PBS (pH 7.4) for 5 min to remove residual O.C.T. The sections were circled with a PAP pen (Ted Pella) to minimize the volume of reagents needed in the following steps. The sections were incubated with 71 or 142 µl of peptide solution at a concentration of 14 ng µl−1 (1000 or 2000 ng of peptide) in sodium carbonate buffer at pH 9.0 (to mimic the approximate pH of the gut) for 1 h at room temperature in a dark chamber with a humidity source. Slides were again washed twice in PBS for 5 min. Slides were mounted using Fluoro-Gel (Electron Microscopy Services) and the coverslips sealed with clear fingernail polish. Slides were stored at 4 °C in the dark until viewed using a Zeiss Axioplan II fluorescence microscope using the FITC filter. Images were captured within 24 h by using Axiovision imaging software (Carl Zeiss). Samples were exposed for 145 ms or 1 s at a magnification of ×20. For H. virescens, the experiment was performed using three to six cryosections from two insects in triplicate; the cryosections of the aphids were from two to three midguts in duplicate.
Viruses.
AcIE1TV3.EGFP (Harrison et al., 2010), which expresses EGFP under the baculovirus immediate–early 1 promoter was used in bioassays. This virus is not significantly different from wild-type AcMNPV in virulence or morphology, as indicated by bioassay and EM analysis (Harrison et al., 2010; Sparks et al., 2011). Virus was amplified in vivo using standard methods (O'Reilly et al., 1992). ODV were isolated from the polyhedra as described in Sparks et al. (2011). Briefly, polyhedra were treated with a dilute alkaline solution for 10 min and then neutralized. After 1 h on an orbital shaker to release the ODV from the calyx, samples were centrifuged at low speed to pellet the calyces and loaded onto sucrose step gradients for ultracentrifugation. The ODV bands were collected, washed in water and resuspended in a small volume of PBS (pH 7.4).
Far-Western blotting.
BBMV proteins were separated using SDS-PAGE using a Mini-PROTEAN 3 cell (Bio-Rad) following the manufacturer’s instructions. Approximately 20 µg of fourth instar BBMV proteins were loaded onto 12 % acrylamide SDS-PAGE gels along with ECL DualVue Western blotting Markers (GE Healthcare). The proteins were transferred to Hybond PVDF membrane (Amersham) using a Bio-Rad Protean III Trans–Blot. The membranes were then blocked overnight in blocking buffer (GE Healthcare) and washed with TBS (pH 7.5)/0.1 % Tween 20 three times for 5 min each with agitation. The membranes were then incubated in the dark with 140 µg of each peptide in 0.1M sodium carbonate (pH 9.0) at room temperature, and washed three times in TBS/0.1 % Tween 20. Peptide binding was visualized using a Typhoon imager (GE Healthcare) with an excitation wavelength of 457 nm and emission filter wavelength of 532 nm.
Bioassays.
Third instar H. virescens larvae showing head capsule slippage were allowed to moult overnight in the absence of diet. Insects (12–15 per treatment per replicate) were initially droplet fed purified ODV (1 µg) with or without peptide (2 µg) in a 1–2 µl volume of PBS with all six peptides tested (HV1, HV2, Ac66A, Ac66B, GBP3.1 and C6). The bioassay was repeated in triplicate. In subsequent experiments, larvae were fed a range of doses of ODV (0.006–5 µg of ODV: 1 LD50 to excess lethal doses) with or without the peptide HV1 (2 µg) in a 1–2 µl volume of PBS. Control larvae were fed peptide alone or PBS alone. Only insects that consumed the entire dose of virus and peptide were used for analysis. Insects were returned to diet and maintained in an incubator. Larvae were monitored twice daily for mortality until pupation of the survivors and control larvae. The instar at death was noted. For any questionable deaths, larvae were checked by light microscopy to confirm polyhedrosis. The bioassay was conducted with three independent replicates. Data for 36–45 larvae per treatment were analysed by ANOVA and Student’s t-test. The effects of peptide on the proportion of larvae surviving to fifth instar were evaluated using χ 2 analysis.
Acknowledgements
The authors wish to thank Dr Robert L. Harrison for helpful discussions and critical reading of the manuscript, Dr J. K. Scott, Simon Fraser University, BC, Canada, for providing the f88.4 LX8 phage library and the K-91 E. coli strain, Jessica Hayward, an undergraduate in the Women in Science and Engineering internship program, and Evelyn Chen for their assistance with insect rearing. This work was supported by Hatch Act and State of Iowa funds.