Baculoviridae is a family of large enveloped viruses with double-stranded, covalently-closed, circular DNA genomes (80-180 kb) that infect arthropods exclusively and primarily insects (52). Baculoviruses are characterized by rod-shaped enveloped nucleocapsids and two phenotypes of virions are typically evident during the replication cycle in host insects. The occlusion-derived virion (ODV) is occluded in a crystalline protein matrix during the final stages of virus replication in the nucleus of infected cells, and the budded virion (BV) that is formed as nucleocapsids bud through the plasma membrane of infected cells. Budded virions are responsible for virus spread to tissues throughout the host producing a systemic infection, and ODV are responsible for horizontal transmission of baculovirus from host to host upon ingestion of the virus occlusion body (OB). There are four genera recognized based on ODV-OB morphology, the taxonomic ORDER of susceptible insect hosts and genome sequence phylogenic analysis; these include: Alpha-, Gamma- and Deltabaculovirus (formally known collectively as the genus Nucleopoly-hedrovirus [NPV]) and Betabaculovirus (formally known as the genus Granulovirus) (26). Historically, baculovirus species have been named on the basis of the host from which they were isolated. This is a system of nomenclature that is not without its problems as multiple baculovirus species have been derived from the same host species; for example, Mamestra configurata nucleopolyhedrovirus (MacoNPV) species A and B (33). In addition, the same baculovirus species can be derived from field populations of multiple host insect species; for example, Xestia c-nigrum granulovirus (XecnGV) has been isolated from at least five noctuid species (19). Species designations within Baculoviridae are currently based on host range and restriction endonuclease (REN) profiles of genomic DNA, and increasingly DNA sequence analysis is used to differentiate species (52).
Baculoviruses have long been recognized as potential biological control agents for insect pests based on their causal association with spectacular epizootics, particularly among lepidopteran and sawfly forest pests. Interest in baculoviruses as potential biopesticides naturally led to efforts to screen insect populations for new and potentially more efficacious isolates or strains. With the application of new molecular tools such as restriction endonucleases (REN) and targeted polymerase chain reaction (PCR) technology to characterize baculoviruses it became evident that multiple genotypes occur among baculovirus populations in field isolates derived from different geographic or temporal populations of a single host insect species. Genotypically distinct strains from these populations have been derived from field isolates either by plaque purification in permissive cell lines (28) or upon in vivo cloning (41, 49). Not only are genetic variants/strains common in pooled samples of host insect larvae from a geographic population of insects but substantial numbers of genetic variants of a baculovirus species can occur in single infected individuals (9). Genetic variation in field populations has typically been detected by means of restriction fragment length polymorphisms (RFLP) and purified strains from these populations often have been demonstrated to have differences in infectivity or virulence both in vivo and in vitro. Advancements in DNA sequencing technology are beginning to allow the complete sequence analysis of multiple strains of a baculovirus species and therefore detailed information on genetic variation including patterns of variation are beginning to emerge. In this short review specific examples of the characterization of baculoviruses field isolates with respect to infectivity and pathogenicity, as well as, genetic variation will be used to demonstrate some general trends. Where examples exist, the comparisons of complete genome sequence of different isolates of a baculovirus species will be reviewed in order to examine common regions of genetic variability, and potential mechanisms that maintain genetic diversity in baculovirus populations will be discussed. Finally the potential role of genetic variation in host-pathogen interactions of baculoviruses will be discussed.
There are currently several baculovirus species for which the complete genome sequence is available for more than one strain or isolate. Although these complete genome sequences by no means reflect the total diversity of genomic difference within baculovirus species they do reveal some interesting trends (Table 1).The MacoNPV-A isolates 90/2 and 90/4 were originally derived from M. configurata populations in close geographic proximity and during the same temporal pest outbreak. The complete sequence of these isolates revealed that the MacoNPV-A 90/4 strain (AF539999) had 99.5% identity to strain 90/2 (NC_003529) but also had 521 single nucleotide substitutions and numerous small insertion/deletions relative to 90/2 (33, 34). The major region of divergence between the two genomes occurs between hr1 and orf31, a region that in 90/2 contains three baculovirus repeat ORF (bro) genes and one of these, bro-a, is absent in strain 90/4. This region which accounts for only 7.7% of the genome contains 50% of the total nucleotide changes in 90/4 relative to 90/2. This along with evidence from the genome sequence of the closely related MacoNPV-B species indicates that this region and particularly the bro-a, -b and -c genes are hot spots for sequence variation and gene rearrangements in the NPVs from M. configurata (33). Despite the high level of DNA sequence identity there are 49 ORFs that have 1-12% amino acid variation between the two strains and in addition there numerous differences in promoter motifs for many genes and substantial differences in hr elements except for hr2. While genome sequence analysis has not as yet provided definitive answers as to why 90/4 is less infectious than 90/2 a number of avenues for further investigation have been identified (34). Although the MacoNPV-A 90/4 strain was plaque-purified, the genome sequence assembled from shotgun clones clearly showed 214 polymorphisms of which the majority occurred within ORFs and there are 47 amino acid polymorphisms spread through 26 genes. It is possible that such a pool of sequence polymerphisms within in a population of a virus strain is an efficient mechanism for maintaining adaptability to changing environments and therefore an evolutionary advantage.
Table 1. Comparison of Complete Genome Sequence of Variant of NPV Species
The field isolate from which 90/4 was plaque purified was a heterogeneous population as determined by RFLP analysis which showed substantial populations of submolar fragments (34). In many previous genome sequencing studies no attempts were made to determine whether a plaque purified strain was representative of the population from which it was derived. In this study, numerous clones of selected REN fragments derived from DNA preparations from the original MacoNPV field isolate were sequenced and several genotypes including that of the 90/4 strain were identified. The MacoNPV-90/4 genotype represented approximately 25% of the mixed genotype population from which it was derived. The 90/2 genotype also appeared to be present in this mixed genotypic population of MacoNPV-A.
The genomes of two HearNPV strains, C1 (NC_ 003094) and G4 (NC_002654), have also been sequenced and compared in detail (5, 58). These strains were both isolated from H. armigera populations in Hubei province, China but as stated above the isolates have different virulence traits. HearNPV-C1 is 644 bp smaller than strain G4 and has 555 bp substitutions relative to G4 with the vast majority of these occurring in coding regions and 183 amino acid variations occur in 51 ORFs. This level of variability is similar to that observed between the two MacoNPV-A strains. The major differences between HearNPV C1 and G4 were in the bro-b gene and hr1, hr4 and hr5 not unlike the situation in MacoNPV-A. Zhang et al. (58) suggest that because the hr regions have been implicated in baculovirus regulatory processes as transcriptional enhancers and potential origins of DNA replication the observed alterations and recombination in these regions may explain the differences in temporal pattern of replication of G4 versus C1 virus that has been observed in cell culture lines.
The genome sequences of two isolates of AnpeNPV which infect the Chinese oak silkworm, Antheraea pernyi, have been completed (17, 43). A field isolate, AnpeNPV-Liaoning Z strain (NC_008035), and a pick-plaque purified strain AnpeNPV-L2 derived from the Lianoning isolate (EF207986) have 99.9% DNA sequence identity. However, the AnpeNPV-LZ genome is 384 bp longer due largely to insertions in intergenic regions and higher numbers of repeats in hr regions. Some of the most significant differences between the two strains include a truncation of egt in the LZ strain (79 aa vs 132 aa for the L2 strain) and 25 bp sub-stitutions in the putative desmoplakin and DNA polymerase genes that create frame shifts and slight truncations of these proteins (17). Interestingly, a sliding window analysis of nucleotide divergences between the two genomes indicated the differences were not random (17) but the regions of highest divergence did not include the hr regions or bro genes as was the case for MacoNPV and HearNPV isolate comparisons. Unfortunately, not much is known about the biological differences between the AnpeNPV strains so the relevance of the genetic diversity is not clear (17).
Two SfMNPV strains have recently been sequenced: SfMNPV-19 (EU258200) a virulent geographic isolate from Brazil (56) and SfMNPV-3AP2 (NC_009011) a plaque-purified strain derived from a field isolate from Missouri, USA that has a fast killing phenotype (21). Analysis of SfMNPV-19 DNA sequence that was derived from two or more REN cloned fragments suggest the number of polymorphisms is very low (0.12%) in this isolate (56). The latter investigators undertook a Clustal W alignment of the genomes of the two SfMNPV strains and the results confirmed that these are variants of the same virus species. The SfMNPV-3AP2 genome is 1235 bp smaller than SfMNPV-19 largely due to a 1427 nt deletion in 3AP2 which results in the loss of a portion of the 3' end of egt and the 5' of SfMNPV-19 ORF26. In addition, the putative dUTPase gene in SfMNPV-19 is truncated compared to the homologue in 3AP2 (56). The SfMNPV-3AP2 clone was one of a series of plaque isolates derived from six individual-cadaver-isolates from which the SfMNPV DNA was purified and used to transfect Sf21 cells. One plaque was selected from each of the single-cadaver-isolates and insect bioassays indicated that all but two were significantly more virulent than the parental field isolates (21), and the remaining two strains were not infectious per os. Harrison et al. (21) sequenced the egt region of each of the plaque purified strains and all had significant deletions ranging from the 1427 nt deletion in 3AP2 to an almost 12kbp deletion in SfMNPV-4AP2 that starts in the 3' end of lef-7 and stops near the 3' end of pif 2. The deletion of a functional pif 2 gene product in two of the plaque isolates probably explains their lack of oral infectivity. Although Harrison et al. (21) indicate there is no evidence that egt deletion mutants have a selective advantage in cell culture infections, it is remarkable that all the plaque isolates analyzed have the egt deletions. The genomic alteration in these SfMNPV plaque isolates are very similar to those described earlier for nine plaque-purified strains from a mixed genotype Nicaraguan isolate of SfMNPV (47, 48). The REN profiles of all the SfMNPV-NIC strains were analyzed and eight were shown to have deletions and sequence analysis of cloned REN fragments confirmed that all lacked a functional egt and several had large deletions encompassing a region that includes pif-2 and pif-1 in the parental genome. These authors demonstrated that mixtures of SfMNPV-NIC genotype B, a plaque strain representing the entire genome, plus specific ratios of certain deletion genotypes were more virulent than genotype B strain alone (48). These results suggest that retaining the mixed genotype populations that occur in natural field isolates may have an advantage for developing bioinsecticides based on baculoviruses (48).
In contrast to the SfMNPV in vitro selected plaque isolates cloned from wild-type SfMNPV field isolates, SeMNPV variants have been cloned from the wild-type isolates SeMNPV-US (Florida, USA) and SeMNPV-SP2 (Spain) using an in vivo cloning approach (41). In this case numerous rounds of in vivo cloning were required to produce pure and stable genotypic populations. The predominant region of variation in these genotypes was mapped to the hr1 region and there were differences in the number of repeats within hr1. In contrast to the increased virulence associated with some of the deletion mutants and mixed genotype populations of SfMNPV variants, the SeMNPV variants acted as parasitic genotypes that persisted and reduced the virulence of SeMNPV isolates (42).