41st Annual Meeting of the Society for Invertebrate Pathology
and 9th International Conference on Bacillus thuringiensis

Viral disease is a major component of the cyclic population dynamics of some Lepidoptera including western tent caterpillars. Epizootics of nucleopolyhedrovirus and host population subdivision provide an arena in which selection on virulence of virus and resistance of hosts could act. Theory predicts that epizootics should select for host resistance and that viral isolates should respond to this change on a population-by-population basis. Experiments provide evidence that these interactions are occurring but that patterns are weak as compared to other factors that determine the cyclic population dynamics. In addition there is no evidence for induced immunity or selection within a generation of tent caterpillars. The factors that promote the rapid development of NPV epizootics remain a mystery and are the topic of future research.

Horizontal transmission of baculovirus disease depends upon contact of a susceptible host with infectious viral particles. Theory suggests that there should be increased potential for horizontal transmission of pathogen at high host density due to the increased likelihood of contact. Gregarious species will experience high densities locally, even though host density measured regionally may be low. However, the relative importance of local versus regional population densities of lepidopteran hosts has received less attention in baculovirus research than have the impacts of overall population density. To assess the effect of spatial distribution on the transmission of viral disease within a population a number of manipulation experiments were carried out. Aggregatory configuration of either host or diet varied between treatments, whilst within-plot density was kept constant. Our host species which all feed on cabbage were Pieris brassicae (gregarious at the larval stage), Mamestra brassicae (solitary at the larval stage) and Autographa gamma.

To understand the molecular basis for differences in speed-of-kill phenotypes the genome sequence of a fast-killing egt minus genotype of a Spodoptera frugiperda multiple nucleopolyhedrovirus from the USA (SfMNPV-3AP2) was compared with that of a slower-killing egt minus SfMNPV genotype originally isolated in Nicaragua (SfMNPV-NIC). Nucleotide sequences were strongly conserved (99.5% identity) and a high degree of predicted amino acid sequence was observed between the two isolates. The SfNIC-B genome was 132,947 bp, 1,617 bp larger than that of SfMNPV-3AP2, due mainly to a deletion of 1,428 bp located between Sf26 (egt) and Sf27 in the latter. A total of 145 open reading frames (ORFs) were identified in SfNIC-B, three of which were absent in SfMNPV-3AP2. In turn, SfNIC-B lacked the SfMNPV-3AP2 ORF129 homologue. Other genes, such as odv-e66a, p26b, were also truncated in SfMNPV-3AP2 due to small deletions, but lack of these genes has no substantial effects on the biological activity of these viruses. A deletion in the homologous region 8 of SfNIC-B was also observed. Construction of recombinant viruses that will help determine the genes involved in virulence is currently being undertaken.

A high prevalence of per os defective variants occurs among a Nicaraguan (Sf-NIC) population of the multiple nucleopolyhedrovirus of Spodoptera frugiperda (SfMNPV), indicating that interactions among genotypes are important for baculovirus survival. Past results have demonstrated a positive interaction between SfNIC-B (complete genotype) and SfNIC-C/D (pif defective genotypes) indicating that dilution of pif genes in the coinfected mixture resulted in increased infectivity compared to SfNIC-B alone. To confirm these results and evaluate the persistence of pif and pif2 deleted genotypes in the population, two bacmid recombinants, SfNIC-BΔ16K (lacking the same 16.3 kb genomic region occurring in SfNIC-C) and SfNIC-BΔpifs (encompassing a 2.8 Kb deletion including only both pif genes) were constructed, mixed individually with the complete genotype SfNIC-B and injected in S. frugiperda larvae at different ratios. After four successive passages in larvae genotype ratios converged to stable populations consisting of ca. 80%:20% of SfNIC-B:defective recombinants. This proportion resembles what is found in the wild-type SfNIC population. All mixtures re-established the infectivity of natural populations after they reached the equilibrium frequency and none of the deleted recombinant genotypes disappeared from the virus population. The proportion of pif-containing genotypes in the population seems to be critical for SfMNPV survival.

Nucleopolyhedrovirus (NPV) transmission between hosts is accomplished by occlusion derived virions (ODVs) containing single (SNPVs) or multiple (MNPVs) nucleocapsids. For MNPVs, it is believed that nucleocapsids containing genetically distinct genotypes can be enveloped into a single virion prior to occlusion. To prove this, Spodoptera frugiperda larvae were infected with a 1:1 mixture of two S. frugiperda MNPV (SfMNPV) purified genotypes, SfNIC-B and SfNIC-C. The progeny virus obtained from these larvae was inoculated in new batches of insects at five different doses. Two other batches of larvae were inoculated each with a different dose of a wild-type SfMNPV isolate containing multiple genotypes. The abundance of each genotype from each insect was then estimated by semi-quantitative PCR using specific primers. At the highest doses, viral populations behave like normal mixed infections with no differences between expected and observed progeny genotypes. However, a higher number of associated genotypes than expected occurred at the lowest doses, suggesting that genetically heterogeneous ODVs were responsible for many of the primary infections. The physical association of genetically distinct nucleocapsids is the most likely explanation for these results. This trait may guarantee the transmission of NPV diversity and, hence, survival, when OBs are scarce in the environment.

Adoxophyes honmai nucleopolyhedrovirus (AdhoNPV) has a distinctive slow-killing pathology compared to most other typical NPVs, such as Autographa californica MNPV (AcMNPV). Neonate A. honmai larvae infected with AdhoNPV die after ~17 days and, regardless of the timing of inoculation, infected larvae succumb in the final instar, and do not pupate. To elucidate factors that determine the killing speed of baculoviruses, we compared the general pathology of AdhoNPV with that of AcMNPV, which also infects A. honmai larvae but kills them after ~7 days. AdhoNPV showed a similar tissue tropism to AcMNPV, with both viruses infecting fat body and tracheal epidermal cells. However, occlusion body formation in epidermal cells was slower for AdhoNPV than for AcMNPV. The number of occlusion bodies per larva was counted to assess virus production. Growth model parameters were estimated, and revealed that the maximum growth rate was significantly lower and duration of lag phase was significantly longer for AdhoNPV than those for AcMNPV. The gene encoding ecdysteroid UDP-glucosyltransferase (EGT) was transcribed early after inoculation of penultimate-instar larvae with both AdhoNPV and AcMNPV. However, hemolymph EGT activity was detectable only after AdhoNPV-infected larvae molted to the final instar, but could be detected during the penultimate instar in AcMNPV-infected larvae.

A mixture of Adoxophyes orana granulovirus (AdorGV) and A. orana nucleopolyhedrovirus (AdorNPV) was recovered from A. orana larvae in the UK. The viruses have previously been separated, sequenced and biologically characterised. AdorGV is slow-killing with an ST50 (using an LD80 dose) of 37.0 days in neonates. AdorNPV is fast-killing with an ST50 (using an LD80 dose) of 8.8 days in neonates. As the viruses were originally found together, bioassays were performed to investigate speed of kill during a mixed infection. Neonate larvae were infected with either an LD80 dose of AdorGV, an LD80 dose of AdorNPV, and 50:50, 25:75 and 75:25 mixes of LD80 doses of AdorGV:AdorNPV. Larvae were observed daily for death and cadavers collected separately. Real-time PCR primers were designed to unique areas of the genomes and standard curves generated for AdorGV and AdorNPV. Real-time PCR was performed on the 217 resulting cadavers to determine the amount of GV and NPV DNA in each larva. The results showed that the GV remained at a low level and did not affect speed of kill in the 25:75 and 50:50 GV:NPV mixes, but slowed down the speed of kill and replicated prolifically after 12 d p.i. in the 75:25 GV:NPV mix.