Journal of General Virology

The complete sequence of the Cydia pomonella granulovirus genome

Teresa Luque,¹ Ruth Finch,² Norman Crook,² David R. O'Reilly¹ and Doreen Winstanley²

¹ Department of Biology, Imperial College of Science, Technology and Medicine, Imperial College Road, London SW7 2AZ, UK
² Horticulture Research International, Wellesbourne, Warwick CV35 9EF, UK

The nucleotide sequence of the DNA genome of Cydia pomonella granulovirus (CpGV) was determined and analysed. The genome is composed of 123500 bp and has a G+C content of 45.2 %. It contains 143 ORFs of 150 nucleotides or more that show minimal overlap. One-hundred-and-eighteen (82.5 %) of these putative genes are homologous to genes previously identified in other baculoviruses. Among them, 73 are homologous to genes of Autographa californica nucleopolyhedrovirus (AcMNPV), whereas 108 and 98 are homologous to genes of Xestia c-nigrum GV (XcGV) and Plutella xylostella GV (PxGV), respectively. These homologues show on average 37.4 % overall amino acid sequence identity to those from AcMNPV and 45 % to those from XcGV and PxGV. The CpGV gene content was compared to that of other baculoviruses. Several genes reported to have major roles in baculovirus biology were not found in the CpGV genome, such as gp64, the major budded virus glycoprotein gene in some nucleopolyhedroviruses, and lef-7, involved in DNA replication. However, the CpGV genome encodes the large and small subunits of ribonucleotide reductase, three inhibitor of apoptosis (iap) homologues and two protein tyrosine phosphatases. The CpGV, PxGV and XcGV genomes present a noticeably high level of conservation of gene order and orientation. A striking feature of the CpGV genome is the absence of typical homologous repeat sequences. However, it contains one major repeat region and 13 copies of a single 73–77 bp imperfect palindrome.

The Baculoviridae are a family of invertebrate viruses with large, double-stranded DNA genomes. To date, most baculoviruses have been isolated from lepidopteran insects. The Baculoviridae are subdivided into two genera, Nucleopolyhedrovirus (NPV) and Granulovirus (GV) (Murphy et al., 1995). NPVs form large polyhedral occlusion bodies that contain multiple virus particles, while GVs form smaller, ovoid occlusion bodies called granules that generally contain a single virion. NPVs and GVs present major differences in infection cycle pathology (Crook, 1991; Federici, 1997). Unlike NPVs, in GV infections the nuclear membrane breaks down before occlusion body formation. Little is known about the molecular causes of these differences. Several NPVs have been completely sequenced, namely Autographa californica (Ac) MNPV (Ayres et al., 1994), Orgyia pseudotsugata (Op) MNPV (Ahrens et al., 1997), Lymantria dispar (Ld) MNPV (Kuzio et al., 1999), Bombyx mori (Bm) NPV (Gomi et al., 1999), Spodoptera exigua (Se) MNPV (IJkel et al., 1999) and Helicoverpa armigera (Ha) SNPV (Chen et al., 2001). In contrast, only two GVs, Xestia c-nigrum (Xc) GV and Plutella xylostella (Px) GV, have been sequenced so far (Hashimoto et al., 2000; Hayakawa et al., 1999).

GVs can be subdivided into two classes, the 'slow' and 'fast' GVs (Winstanley & O'Reilly, 1999). XcGV is a slow GV, whereas PxGV is a fast GV. The type species of the genus, Cydia pomonella GV (CpGV), is also a fast GV, i.e. the host typically dies in the same instar in which it was infected. CpGV is highly pathogenic for the codling moth, C. pomonella, a worldwide pest of apples, pears and walnuts (Glen & Payne, 1984). Several strains of the virus have been isolated, although most field trials have used the Mexican isolate CpGV-M (Crook et al., 1985). The molecular biology of this virus is poorly understood. Although its genome has been mapped and estimated to contain about 125.6 kbp (Crook et al., 1985, 1997), only a small number of CpGV genes have been identified. Here, we describe the complete genome sequence of CpGV and compare it to other baculovirus genomes.

Cloning and sequence analysis. The SalI cosmid library of CpGV-M1, an in vivo cloned genotype from the Mexican isolate, has been described (Crook et al., 1997). DNA fragments were subcloned from the cosmids into plasmid vectors for sequencing. Cosmid DNA or PCR products were also used as templates. M13 single-stranded DNA templates were sequenced using Sequenase (US Biochemicals) according to the manufacturer's protocol. Double-stranded DNA was sequenced using ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction kits (PE Biosystems) and ABI 373 and 377 automated sequencers. Sequence data were assembled and analysed using the Wisconsin Genetics Computer Group programmes (Devereux et al., 1984) and DNAStar (Lasergene). ORFs of 50 codons or more were considered for further analysis; smaller ORFs were included when similarity to a known gene was demonstrated. Overlapping ORFs were considered where the non-overlapping portion of the ORF exceeded 50 amino acids. The possibilities of splicing or alternative start sites were not considered. Homology searches were carried out using BLAST and FASTA programmes (Altschul et al., 1990; Pearson & Lipman, 1988). Pairwise sequence alignments were performed with the GCG Gap programme using a gap creation penalty of 4 and an extension penalty of 2. Multiple alignments were performed with ClustalW (Thompson et al., 1994) and phylogenetic analyses with PAUP*4.0b4a (Swofford, 2000). The MacVector program was used to identify repeats by comparing the genome to itself with a window size of 30 and minimum similarity of 65 %. The conservation of gene order between baculovirus genomes was visualized by generating gene parity plots essentially as described by Hu et al. (1998).

Sequence analysis of the CpGV genome

The CpGV genome consists of 123500 bp, in good agreement with the previous estimate of 125.6 kbp based on restriction mapping (Crook et al., 1997). This compares to 178733 bp and 100999 bp for XcGV and PxGV (Hashimoto et al., 2000; Hayakawa et al., 1999). The A of the granulin start codon was designated nt 1 and the sequence numbered in the direction of granulin transcription. The G+C content of the CpGV genome is 45.2 %, slightly higher than that of XcGV, PxGV, AcMNPV, BmNPV, SeMNPV and HaSNPV (39–44 %), but lower than that of OpMNPV and LdMNPV (55–58 %). Our criteria for selecting ORFs for further study were that they should be methionine-initiated ORFs of at least 50 codons having minimal overlap with other ORFs. However, following comparisons with other baculoviruses, some exceptions to these criteria were allowed. ORF86 (Cp86), encoding the core protein P6.9 homologue, and Cp53, homologous to Ac110, are 49 and 48 codons long, respectively. Cp5, Cp77 and Cp122 are all contained within larger ORFs. However, these three ORFs were considered more likely to be biologically meaningful because they have homologues in other baculoviruses. Thus, Cp5, Cp77 and Cp122 are included in Table 1, rather than the larger ORFs they lie within. Finally, Cp37 and 38, Cp51 and 52, Cp82 and 83, Cp87 and 88, and Cp137 and 138 overlap significantly (>45 codons). Ultimately, 143 ORFs were selected for further study (Table 1 and Fig. 1) and they are numbered from granulin in a clockwise direction. The number of ORFs described for other completely sequenced baculoviruses ranges from 120 (PxGV) to 181 (XcGV).

Table 1. CpGV ORFs

The positions and orientations of 143 putative ORFs in the CpGV genome are shown and compared to homologues in selected other completely sequenced baculoviruses. Red, ORFs unique to CpGV; blue, ORFs present in all GVs and absent in NPV; green, ORFs conserved in all nine completely sequenced baculoviruses. A Cp79 homologue is found in all baculoviruses included in this table but is absent from HaSNPV. The presence of baculovirus early (E) and late (L) promoter elements, located within 120 nt of the ATG, is indicated in the 'Prom.' column. 'E' indicates a TATA sequence with a CAC/GT start site sequence 20–40 nt downstream. 'L' indicates the presence of an A/T/GTAAG motif. Homologous ORFs in the genomes of AcMNPV (Ayres et al., 1994), OpMNPV (Ahrens et al., 1997), LdMNPV (Kuzio et al., 1999), SeMNPV (IJkel et al., 1999), XcGV (Hayakawa et al., 1999) and PxGV (Hashimoto et al., 2000) are shown with the percentage amino acid sequence identity to the homologous CpGV ORF. Pairwise alignments were carried out using the GCG Gap algorithm (Devereux et al., 1984).

Click here to see Table 1 (in a new browser window)

Fig. 1. Circular map of the CpGV genome. The inner circles indicate the positions of cleavage sites for the following enzymes: inner circle, SalI; middle circle, EcoRI; and outer circle, BamHI. CpGV ORFs discussed in the text are indicated outside these circles, with the arrow indicating the direction of transcription. The positions of repeat sequences are also shown, with the major repeat region from 20 to 21 kbp indicated by the bar. A scale in bp is provided in the centre of the figure.

Similar to other baculoviruses, CpGV ORFs have minimal intergenic distances and no obvious preferred orientation (51.7 % clockwise, 48.3 % anticlockwise) or clustering according to function or expression. Coding sequences represent 88.4 % of the genome, similar to XcGV and PxGV (88 and 89 %, respectively). Fifty-three ORFs overlap an adjacent ORF. Apart from the five pairs noted above, these overlaps are less than 36 codons. Thirty-eight CpGV ORFs possess a consensus early promoter motif (TATA box with a CAC/GT motif 20–40 bp downstream) within 120 bp of the initiation codon (Table 1). Nineteen of these also have a late promoter motif (A/T/GTAAG) within 120 bp of the initiation codon and 52 ORFs only have a late promoter motif. Fifty-three ORFs lack any recognized consensus promoter motif within 120 nt of the ATG.

Comparison of CpGV ORFs to those of other baculoviruses

One-hundred-and-eighteen CpGV ORFs have homologues in other baculoviruses, while 25 are so far unique to CpGV. None of these showed significant similarity to sequences in GenBank (Table 1). Sixty-three ORFs are conserved among all nine baculoviruses sequenced so far. Seventy-three CpGV ORFs have homologues in AcMNPV while 108 and 98 CpGV ORFs have homologues in XcGV and PxGV, respectively. Ninety-five CpGV ORFs have homologues in both XcGV and PxGV, of which 27 are so far unique to GVs. Seven CpGV ORFs have homologues in NPVs but not in XcGV or PxGV. Finally, 64 XcGV ORFs and 21 PxGV ORFs lack CpGV homologues (Table 2).

Table 2. Granulovirus ORFs without homologues in CpGV

The average amino acid sequence identity between CpGV and XcGV, PxGV and AcMNPV homologues is 44.6, 45.4 and 37.4 %, respectively. The most conserved ORFs are ubiquitin (Cp54), polyhedrin/granulin (Cp1), p6.9 (Cp86), sod (Cp59), lef-9 (Cp117) and lef-8 (Cp131) (Table 1). In contrast, 39K (Cp57) and lef-3 (Cp113) are among the least conserved.

Genes specific to GVs

Twenty-seven ORFs are present in all three sequenced GV genomes and absent in NPVs (Table 1). These GV-specific genes show 35.3 % mean amino acid sequence identity. The most conserved GV-specific ORFs are related to the previously described orf16L (Kang et al., 1997), whereas Cp2 is among the least conserved GV-specific ORFs. CpGV has two genes in the orf16L family, Cp20 and Cp23. They are approximately 35 % identical to each other. However, each is approximately 55 % identical to its orthologue in XcGV and PxGV, suggesting there are two independent lineages of genes in this family. The functions of most GV-specific ORFs are unknown. Cp116 encodes a putative novel inhibitor of apoptosis that seems to be GV-specific (see below). Cp46 (homologous to Xc40 and Px35) is likely to be a member of the stromelysin family within the matrix metalloproteinase superfamily (Hashimoto et al., 2000; Hayakawa et al., 1999). Ko et al. (2000) recently confirmed that Xc40 encodes an active proteinase. They suggested it is retained within infected cells until death, when it is released into the body of the insect, causing the proteolysis of host tissues. Cp46 is about 70 amino acids longer than Xc40 at the 5´ end and may have a signal sequence.

Genes involved in DNA replication and expression

Many genes implicated in DNA replication and expression are present in all sequenced baculovirus genomes, presumably reflecting their critical roles in virus replication. Early baculovirus genes are transcribed by the host cell machinery, but this is often modulated by viral transcription regulators, such as ie-0, ie-1, ie-2 and pe38 (Friesen, 1997). Both ie-0 and ie-2 are absent from CpGV, whereas ie-1 (Cp7) and pe38 (Cp24) are present, but poorly conserved. These genes are not well conserved among baculoviruses in general (about 26 % mean amino acid identity). Both ie-2 and pe38 are absent from XcGV, PxGV, LdMNPV, SeMNPV and HaSNPV.

Six genes are reported to be essential for DNA replication in AcMNPV and OpMNPV: lef-1, lef-2, lef-3, dnapol, helicase and ie-1 (Lu et al., 1997). Homologues of all are present in CpGV. With the exception of lef-3 and ie-1, they are moderately well conserved (Table 1). AcMNPV and BmNPV lef-7 stimulates DNA replication in transient assays (Gomi et al., 1997; Morris et al., 1994). However, in AcMNPV it acts in a cell-specific manner. CpGV lacks lef-7, similar to XcGV, PxGV, LdMNPV, SeMNPV and HaSNPV. In common with LdMNPV, XcGV and PxGV, CpGV encodes a DNA ligase (Cp120) and a second helicase (Cp126). The LdMNPV DNA ligase displays catalytic properties of a type III DNA ligase (Pearson & Rohrmann, 1998). The helicase-2 gene shows similarities to a yeast mitochondrial helicase called pif-1 (Pearson & Rohrmann, 1998). In LdMNPV, neither the helicase-2 nor the DNA ligase gene stimulates DNA replication in transient assays. It has been suggested that they may be involved in DNA repair (Kuzio et al., 1999).

Cp127 and Cp128 encode the large (rr1) and small (rr2) subunits of ribonucleotide reductase. OpMNPV, LdMNPV and SeMNPV also encode ribonucleotide reductase subunits (Ahrens et al., 1997; IJkel et al., 1999; Kuzio et al., 1999). The latter viruses also encode a dUTPase, but no dUTPase homologue is present in CpGV. These enzymes are involved in nucleotide metabolism and may facilitate virus replication in non-dividing cells, in which dNTP pathways are inactive. There appear to be two separate lineages of baculovirus ribonucleotide reductase genes. Those from OpMNPV and two from LdMNPV appear to be part of a novel lineage only found in some baculoviruses, whereas those from SeMNPV are part of a eukaryotic lineage (Jordan & Reichard, 1998). The CpGV rr1 gene appears to fall in the novel baculovirus lineage and shows 51 and 52 % amino acid identity to its homologues in OpMNPV and LdMNPV respectively, but only 31 % identity to SeMNPV homologues. CpGV rr2 is most similar to OpMNPV rr2 and LdMNPV rr2a (Ld147) and shows little or no similarity to the LdMNPV or SeMNPV rr2b genes (Ld120 and Se45). Op31 appears to be a fusion of two ORFs, an N-terminal one of unknown function and a C-terminal ORF encoding dUTPase (Fig. 2). However, these ORFs are separated in LdMNPV (Ld116 and Ld138) and SeMNPV (Se55 and Se54). Cp16 is homologous to the N-terminal part of Op31, Ld138 and Se54. The C-terminal part of Se54 is homologous to Ac33 but no homologue of this gene is present in CpGV. XcGV, PxGV, AcMNPV, BmNPV and HaSNPV do not encode ribonucleotide reductase subunits or dUTPase (Hayakawa et al., 1999; Hashimoto et al., 2000; Ayres et al., 1994; Gomi et al., 1999; Chen et al., 2001).

Fig. 2. Relationships of dutpase genes and associated ORFs from several baculoviruses. The solid black box indicates the 5´ region of Op31 and its homologous ORFs. The diagonally striped box represents dutpase. Ac33 homologous sequences are indicated by a horizontally striped box, whereas the 3´ portion of Ld138 (not related to other ORFs) is denoted by the stippled box. Comparisons were based on gap analysis. The length of each ORF is indicated on the right-hand side.

Many genes required for late gene transcription have been described, including lef 4–6, lef 8–11, 39K, p47 and vlf-1 (Lu & Miller, 1997). All of these are found in CpGV, with the possible exception of lef-6. Generally, these genes are more conserved than the early transcription activators (IJkel et al., 1999). This is also the case for CpGV (42 % mean amino acid identity). Of these, lef-8 (Cp131) and lef-9 (Cp117) are the most conserved (about 58 % mean amino acid identity). In contrast lef-6, if present at all, is very poorly conserved. Hayakawa et al. (1999) reported that it is absent from XcGV. However, Hashimoto et al. (2000) reported that Px60, which is homologous to Xc88, shows some similarity to NPV lef-6 genes. Cp80 is a clear homologue of Px60 and Xc88 (Table 1). These three genes are smaller than the NPV lef-6 genes (86–101 amino acids vs 138–187 amino acids). Functional studies will be necessary to determine whether Cp80, Xc88 and Px60 are true lef-6 homologues.

CpGV structural genes

The most conserved baculovirus structural protein is polyhedrin/granulin (66.5 % mean amino acid identity), the major component of occlusion bodies (Rohrmann, 1992). Other conserved CpGV structural genes are p6.9 (Cp86) and odv-e25 (Cp91) (64.2 and 56 % mean amino acid identity respectively; Table 1). CpGV lacks homologues of some structural genes, such as calyx/pep and p80/p87-capsid, both of which are also absent from XcGV and PxGV. Hayakawa et al. (1999) reported that Xc2 encodes a homologue of ORF1629 (p78/83), thought to be an essential virion component. Xc2 shows only low similarity with the NPV ORF1629 genes (e.g. 24.1 % identity over 162 aa to Op2) and is substantially shorter (231 amino acids compared to 462 to 555 amino acids). Cp2, Xc2 and Px5 are clearly homologues of each other (Table 1) and are in the same genomic position as ORF1629. However, Cp2 is smaller again than the NPV ORF1629 genes (174 amino acids) and shows no significant similarity to any of them. Similarly Px5 was not identified as an ORF1629 homologue. Cp2/Xc2/Px5 may represent GV-specific genes (Table 1). Thus, the analysis suggests that GVs may not possess an ORF1629 homologue.

The absence of a p10 homologue in CpGV is noteworthy. In NPV-infected cells, P10 forms fibrillar structures in the nucleus and cytoplasm. The protein is implicated in occlusion body morphogenesis and disintegration of the nuclear matrix, thereby disseminating the polyhedra (van Oers & Vlak, 1997). Three XcGV ORFs (Xc5, Xc19 and Xc83) present similarities to p10. Xc5 is most similar and was named p10, although it is poorly conserved. Homologues of these three ORFs are present in PxGV (Px2, 21 and 50) and Hashimoto et al. (2000) suggested they are all p10 homologues. No Xc5, Xc83 or Px2 homologues are present in CpGV. We have previously described ORF17R (Cp22), which is 56 % identical to Xc19 and 39 % identical to Px21 (Kang et al., 1997). Cp22 shares a number of motifs with P10, including a proline-rich domain and a heptad repeat sequence. It is 30 % identical to AcMNPV P10. However, it is significantly larger (329 vs 137 amino acids) and much of the sequence identity is between sequences of low complexity. A similar situation occurs with Cp62, which is homologous to Px50. Thus, although Cp22 and Cp62 may be functionally analogous to p10, we consider it unlikely they are true homologues in the evolutionary sense.

Similar to XcGV, PxGV, LdMNPV, SeMNPV and HaSNPV, CpGV does not encode the envelope glycoprotein GP64, the major envelope fusion protein of AcMNPV, BmNPV and OpMNPV (Monsma et al., 1996). However, recent evidence suggests gp64 is unique to group I NPVs (Pearson et al., 2000). In LdMNPV, the envelope fusion protein is the product of the Ld130 gene. This protein contains a furin-like proprotein convertase cleavage site also conserved in its SeMNPV homologue (IJkel et al., 2000). Ld130 homologues are present in all baculoviruses that have been completely sequenced, including species that contain gp64. The role of the Ld130 homologue in the latter species is unclear (Pearson et al., 2000). CpGV encodes an Ld130 homologue, Cp31, which shows 30 % amino acid identity to Ld130.

Auxiliary genes

Aside from structural genes or those directly implicated in DNA replication and transcription, baculoviruses have other genes that reduce their dependence on the host or enhance their fitness in other ways (O'Reilly, 1997). Among them, ubiquitin is the most conserved and is present in all sequenced genomes. The main function of cellular ubiquitin is to signal protein degradation (Haas et al., 1996). Viral ubiquitin is non-essential and its role is unclear (Reilly & Guarino, 1996). Superoxide dismutase (sod) is also well conserved and present in all sequenced baculovirus genomes. This is presumably involved in the removal of free radicals, but is non-essential (Tomalski et al., 1991) and its role in the virus life-cycle is not known. As previously reported, chitinase (Cp10) and cathepsin (Cp11) are also well conserved (Kang et al., 1998). These genes function together to promote liquefaction of the host. Other genes in CpGV include gp37 (Cp13) and fibroblast growth factor, fgf (Cp123). Hashimoto et al. (2000) report that PxGV has two fgf homologues, Px56 and Px104. Cp123 is a homologue of Px104. CpGV also has a homologue of Px56 (Cp76) but this does not show significant similarity to fgf. No enhancin homologue is present in CpGV or PxGV. In contrast, four enhancin homologues are present in XcGV and two in LdMNPV. Enhancin is a metalloproteinase and evidence suggests that it may digest components of the insect peritrophic membrane, facilitating the initiation of infection (Derksen & Granados, 1988; Wang & Granados, 1998). As noted, CpGV encodes a GP37 homologue (Cp13). GP37 (spindlin) is related to the fusolins of entomopoxviruses, which also act as enhancing factors (Yuen et al., 1990). Both CpGV and PxGV lack a conotoxin-like, ctl, homologue (Eldridge et al., 1992). Such a gene is present in XcGV but lacking in BmNPV, SeMNPV and HaSNPV. Its biological role is unknown. Cp9 and Cp79 are homologous to Ac145 and Ac150. Members of this gene family, which contain a six-cysteine motif similar to chitin-binding proteins, are also found in entomopoxviruses (Dall et al., 2001).

Genes implicated in phosphorylation and dephosphorylation

CpGV possesses genes encoding a protein kinase (PK) (Cp3) and two protein tyrosine phosphatases (PTP) (Cp66 and Cp98). Phosphorylation is a common mechanism for regulating protein activity and several baculovirus proteins, such as IE-1 and P78/83, are known to be phosphorylated (Choi & Guarino, 1995; Vialard & Richardson, 1993). CpGV PK is homologous to AcMNPV PK-1, which is present in many other baculoviruses. No homologue of AcMNPV PK-2 was found in CpGV. This gene has only been described in AcMNPV and BmNPV and is non-essential (Ahrens et al., 1997; Ayres et al., 1994). Two lineages of dual specificity PTPs (dsPTPs) have been identified in baculoviruses. OpMNPV encodes one copy of each whereas other baculoviruses encode one or none. CpGV-M1 is unique in encoding two homologues of dsPTP-2. However, one of these, Cp66, is truncated and only encodes the C-terminal end of a dsPTP. It is unclear whether this encodes a functional dsPTP, although it does include the catalytic loop [HCXXGXXR(S/T)]. Cp66 is more similar to Cp98 (54 % amino acid identity) than it is to any other dsPTP, suggesting they may derive from a duplication in CpGV. No dsPTP-1 is present in CpGV. In contrast, no dsPTP is present in XcGV, PxGV, LdMNPV or HaSNPV (Chen et al., 2001; Hashimoto et al., 2000; Hayakawa et al., 1999; Kuzio et al., 1999).

Inhibitors of apoptosis

Programmed cell death is triggered early in baculovirus infection and, to counter this, baculoviruses encode proteins that inhibit apoptosis, such as P35 and IAP (inhibitors of apoptosis). P35 homologues have only been described in AcMNPV, BmNPV and Spodoptera littoralis NPV, whereas iap genes are present in all sequenced baculovirus genomes and in many other baculoviruses (Clem, 1997). The first iap, now referred to as iap-3, was identified in CpGV by complementation of AcMNPV p35 mutants (Crook et al., 1993). IAP homologues generally contain two baculovirus IAP repeats (BIR) (Birnbaum et al., 1994), which are associated with binding to apoptosis-inducing proteins (Vucic et al., 1997), and a C-terminal zinc finger-like (RING) Cys/His motif. Two iaps, iap-1 and iap-2, are present in AcMNPV and BmNPV but their functions are poorly understood (Griffiths et al., 1999). Epiphyas postvittana MNPV encodes four iap genes. Two of these, iap-1 and iap-2, were shown to possess anti-apoptotic activity, whereas no anti-apoptotic activity was demonstrated for iap-3 or iap-4 (Maguire et al., 2000). In addition to iap-3 (Cp17), our sequence analysis has now revealed two additional CpGV iap genes (Cp94 and Cp116). However, Cp94 has a single BIR motif. The relationships between baculovirus iaps were explored by phylogenetic analyses based on the BIR and RING motifs (Fig. 3). Insect IAPs from Spodoptera frugiperda and Trichoplusia ni were included in these analyses. These showed most similarity to the IAP-3 group. CpGV IAPs did not group together, suggesting they do not derive from duplications in CpGV. Cp-IAP-3 clearly grouped with the IAP-3 group. However, Cp94 was ambiguous in these trees, generally grouping either with IAP-1 or IAP-3 sequences, but never with IAP-2s. Thus, Cp94 could not be unambiguously included in any particular iap group. The same was true for Ld-IAP-3, which, although numbered IAP-3, does not clearly belong to any group. This could be due to the fact that Ld-IAP-3 is truncated, and thus difficult to classify. In contrast, Cp116 grouped strongly with Xc-IAP, Px-IAP and a T. ni GV IAP. These sequences form a well supported clade that does not group strongly with any previously recognized iap group. Thus, we propose to designate this the iap-5 group. Furthermore, Cp116, Xc-iap and Px-iap are located in homologous positions on their respective genomes and thus, it is highly likely they are true homologues.

Fig. 3. Phylogenetic analysis of the baculovirus IAP gene family. The figure shows a strict consensus tree of the two most parsimonious trees found using 174 characters from the BIR and RING-finger domains, of which 150 were parsimony-informative. The tree has 1647 steps, a consistency index of 0.574 and a retention index of 0.678 after removing the uninformative sites. The tree search strategy adopted involved a heuristic algorithm with 10 random addition-sequence replicates, TBR branch swapping, and used the protpars matrix. Gaps were considered as missing data. Bootstrap support (%, n=1000) is indicated above the nodes. Bootstrap values lower than 50 % are omitted. Ac, AcMNPV; Bm, BmNPV; Busu, Buzura suppressaria NPV; Cf, Choristoneura fumiferana NPV; Cp, CpGV; Eppo, Epiphyas postvittana NPV; Ha, HaSNPV; Ld, LdMNPV; Op, OpMNPV; Px, PxGV; Se, SeMNPV; Xc, XcGV. The major clades of IAP genes are indicated on the right-hand side.

Baculovirus repeated ORFs (bro genes)

A striking feature of many baculovirus genomes is the presence of the bro gene family (baculovirus repeated ORFs) (Kuzio et al., 1999). This family appears to be widespread, highly repetitive and highly conserved. The number of copies varies from none in PxGV to 16 in LdMNPV. Most bro genes have a related core in their N-terminal region but show differing degrees of similarity elsewhere. Kuzio et al. (1999) divided the family into four groups based on the variable regions. Comparison of the bro genes suggests that they are the result of several gene duplication events (Kuzio et al., 1999). Ac-bro, Bm-bro-d and Ld-bro-n are the most highly conserved genes between these viruses (82 % amino acid sequence identity). The function of these genes is not yet clear, although it has been shown that they bind to DNA (Zemskov et al., 2000). BmNPV bro genes have been characterized at the transcriptional and translational levels (Kang et al., 1999). They are transcribed early and are localized in both the nucleus and cytoplasm (Zemskov et al., 2000). A single bro-related ORF (Cp63) was identified in CpGV. This is a truncated gene encoding a 55 amino acid protein that contains most of the core described by Kuzio et al. (1999). Cp-bro is most similar to the group III bro genes, whereas it does not show significant similarity to the OpMNPV bro genes. It is interesting to note that further similarities to the core sequence can be identified upstream of the predicted Cp63 start codon but there is no in-frame Met codon preceding these sequences (Fig. 4). The sequence of this region has been thoroughly verified and we are confident this is not an artefact. We can find no homology to either the N terminus or the C terminus of bro genes elsewhere in the genome, arguing against the possibility that Cp63 represents an exon of a spliced bro homologue. Consistent with this, Cp63 is not flanked by recognizable consensus splice sites. It seems most likely that Cp63 represents a bro homologue that has been truncated at both the N- and C-termini. It remains to be seen whether it is functional.

Fig. 4. Alignment of BRO core sequences. The figure shows a ClustalW alignment (using default parameters) of BRO core sequences, including eight residues upstream from the start codon of Cp63 (indicated by the asterisk). Conserved residues are shaded in black (identical amino acids) or grey (similar amino acids). Amino acid residue numbers are given on the sides of the figure.

Repeated sequences

Repeated sequence elements, known as hrs, have been identified in all baculoviruses sequenced to date (Hayakawa et al., 2000). These sequences are thought to function as origins of DNA replication (ori) and as enhancers of early gene transcription (Cochran & Faulkner, 1983; Guarino & Summers, 1986; Kool et al., 1995; Theilmann & Stewart, 1992; Xie et al., 1995). An individual NPV hr typically comprises repeats of a 60–80 bp unit centred around a palindrome. The putative hrs identified recently in HaSNPV (Chen et al., 2001) deviate from this pattern somewhat in that they are thought to include two distinct repeat motifs. Nonetheless, like all classical hrs, each HaSNPV hr element comprises multiple tandem repeats of these repeat units. NPVs have several hrs dispersed around the genome, with variable numbers of the repeat unit. A second type of replication origin, a non-hr ori, has been identified in some NPVs (Habib & Hasnain, 2000; Heldens et al., 1997; Kool et al., 1994). These are complex structures, comprising multiple direct and inverted repeats within a region spanning 800–4000 bp. They appear to be present only once per genome.

The situation in GVs is less clear. Hayakawa et al. (1999) identified nine putative hrs in XcGV. These are quite different from NPV hrs. They comprise three to six direct repeats of a 120 bp element that lacks a palindromic core. PxGV hrs seem more like NPV hrs, in that the repeat unit is centred around a palindrome, although this is shorter than that found in NPVs (15 bp vs approximately 30 bp). The repeat unit of approximately 105 bp is larger than that typically found in NPVs. Four large hrs are present in PxGV, containing 10 to 26 copies of the repeat unit.

To identify CpGV hrs, the entire sequence was compared to itself and its complement by dot matrix analyses, using a 30 bp window allowing up to 35 % mismatches. A major region of repeated sequences, spanning 1.13 kbp, was identified from 20.1 to 21.2 kbp. Three ORFs (Cp25–27) are within this region. None of them have known homologues and it is not known whether they are transcribed. This region includes several different classes of repeat (Fig. 5 A) and is reminiscent of the non-hr-type ori. It is not repeated elsewhere in the genome. An AT-rich section comprises six short imperfect repeats of consensus TTTTTATAATNATAATACA. This is followed by three large imperfect repeats, each of which we have subdivided into three sections (Fig. 5 A, B). In repeat 1, section B is extended by multiple repeats of parts of its 5´ end (designated section b). Within repeat 3, sections B and C are separated by 141 nt that are not repeated elsewhere.

Fig. 5. CpGV repeat sequences. (A) Graphical representation of the CpGV major repeat region. The AT-rich repeats are represented by open boxes. Sections A, B and C of the large repeats are indicate by stippled, horizontal striped and diagonal striped boxes, respectively. The multiple repetitions of part of section B within repeat 1 are indicated by the horizontal-striped box labelled b. (B) Alignment of repeats 1–3 within the CpGV major repeat region. Sections A, b, B and C are indicated. A PvuII restriction site in section C is indicated. (C) ClustalW alignment (using default parameters) of CpGV dispersed repeat sequences. Nucleotides conserved in at least 10 of the sequences are shaded.

Surprisingly, these analyses did not identify any other regions of the genome containing multiple tandemly repeated sequences. We were readily able to identify hrs in other baculoviruses using this approach (data not shown). Indeed, the criteria we used (up to 35 % mismatches permitted) are considerably less stringent than those used to identify hrs in other baculoviruses. Thus, we conclude that CpGV does not possess hr elements similar to those found in other baculoviruses.

As noted above, an individual hr element normally comprises several tandem repeats of a 60–120 bp unit. However, there are examples of potential hr elements with a single copy of the repeat unit. Furthermore, it has been shown that a single repeat element is sufficient for ori function (Leisy et al., 1995). We considered the possibility that all CpGV hrs only contain a single copy of the repeat unit. We therefore examined the genome for all instances where a sequence of 60 bp or greater was repeated elsewhere in the genome, permitting up to 30 % mismatches. This identified a repeated imperfect palindrome of approximately 75 bp. FASTA searches of this sequence against the complete genome revealed that it is present in the genome 13 times (Fig. 5 C). It is never present as multiple tandem repeats like a typical hr element. However, repeats 1 and 2 are separated by 144 bp and repeats 3 and 4 by only 90 bp (Fig. 1). All other copies of this sequence element are widely distributed on the genome. Six of these repeats (2, 5, 7, 9, 10 and 13) are within predicted ORFs. It remains to be seen whether these elements actually function as replication origins or as transcription enhancers.

Fig. 6. Comparison of the CpGV gene organization versus PxGV (A), XcGV (B) and AcMNPV (C). Homologues are plotted based on their relative location in the genome. ORFs with no homologues are aligned on the vertical and horizontal axes. In (A) and (B), homologues that have been separated from their expected neighbours are named. A bar indicates the largest contiguous stretch of CpGV unique ORFs.

Organization of the CpGV genome

Comparison of gene arrangements between species can give valuable information about baculovirus evolution. In general, gene order is highly conserved between CpGV, PxGV and XcGV, with only a small number of rearrangements (Fig. 6). In contrast to the NPVs, there is no large genome inversion among the GVs. On the other hand, genome organization is poorly conserved between CpGV and AcMNPV. Similar observations have been made comparing XcGV and PxGV with NPVs (Hashimoto et al., 2000; Hayakawa et al., 1999). In CpGV, 96 of 98 PxGV homologues and 100 of 108 XcGV homologues are in a conserved position. The homologues group in discontinuous clusters with a major gap on the CpGV genome corresponding to Cp24–28. Cp24 is homologous to pe-38 while Cp25 to Cp28 have no homologues in other baculoviruses, and represent the largest contiguous stretch of ORFs unique to CpGV. However, as noted, these ORFs span a large repeat sequence and may not be transcribed.

Four small regions were identified in XcGV that show a relatively conserved gene arrangement with AcMNPV and OpMNPV. Three of these are also present in PxGV (Hashimoto et al., 2000; Hayakawa et al., 1999). In CpGV, four main regions can be identified showing a moderately conserved gene arrangement with AcMNPV. These include three of the regions described in XcGV and PxGV and contain Cp7–9, Cp83–96, Cp101–108 and Cp111–114, corresponding to Ac147–145, Ac103–89, Ac83–75 and Ac65–68, respectively (Fig. 6). However, gene order within these regions is not exactly conserved. For instance, the Ac79 homologue, which one would expect to be around Cp105, is actually Cp65. A similar translocation is observed for its XcGV homologue (Xc75). The gene arrangement within these regions is also well conserved in the NPVs.

The complete sequence of CpGV has revealed many differences in gene content and arrangement compared to NPVs. Not surprisingly, it shows a greater similarity to XcGV and PxGV in terms of both gene arrangement and conservation of homologous genes. A striking feature of the CpGV genome is the lack of typical hrs, which are present in all other baculoviruses sequenced to date. This suggests it may be necessary to re-evaluate what constitutes an hr element, and/or the role they play in the virus life-cycle.

We are grateful to Salvador Carranza for help with the sequence analysis, and to Elisabeth Herniou for extensive help with the analysis of the data and critical comments on the manuscript. We thank John Kuzio for the production of Fig. 1, and George Rohrmann, Julie Olszewski and Sally Wormleaton for helpful comments on the manuscript. This work was supported in part by a grant from the BBSRC (G01317) to D.R.O'R., and by BBSRC core support to D.W.

The GenBank accession number of the CpGV genome sequence reported in this paper is U53466.

Ahrens, C. H., Russell, R. L. Q., Funk, C. J., Evans, J. T., Harwood, S. H. & Rohrmann, G. F. (1997). The sequence of the Orgyia pseudotsugata multicapsid nuclear polyhedrosis virus genome. Virology 229, 381–399.

Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990). Basic local alignment search tool. Journal of Molecular Biology 215, 403–410.

Ayres, M. D., Howard, S. C., Kuzio, J., Lopez-Ferber, M. & Possee, R. D. (1994). The complete DNA sequence of Autographa californica nuclear polyhedrosis virus. Virology 202, 586–605.

Birnbaum, M. J., Clem, R. J. & Miller, L. K. (1994). An apoptosis-inhibiting gene from a nuclear polyhedrosis virus encoding a polypeptide with Cys/His sequence motifs. Journal of Virology 68, 2521–2528.

Chen, X., IJkel, W. F. J., Tarchini, R., Sun, X., Sandbrink, H., Wang, H., Peters, S., Zuidema, D., Lankhorst, R. K., Vlak, J. M. & Hu, Z. (2001). The sequence of the Helicoverpa armigera single nucleocapsid nucleopolyhedrovirus genome. Journal of General Virology 82, 241–257.

Choi, J. & Guarino, L. A. (1995). Expression of the IE1 transactivator of Autographa californica nuclear polyhedrosis virus during viral infection. Virology 209, 99–107.

Clem, R. (1997). Regulation of programmed cell death by baculoviruses. In The Baculoviruses, pp. 237–266. Edited by L. K. Miller. New York: Plenum Press.

Cochran, M. A. & Faulkner, P. (1983). Location of homologous DNA sequences interspersed at 5 regions in the baculovirus Autographa californica nuclear polyhedrosis virus genome. Journal of Virology 45, 961–970.

Crook, N. E. (1991). Baculoviridae: Subgroup B: Comparative aspects of granulosis viruses. In Viruses of Invertebrates, pp. 73–110. Edited by E. Kurstak. New York: Marcel Dekker.

Crook, N. E., Spencer, R. A., Payne, C. C. & Leisy, D. J. (1985). Variation in Cydia pomonella granulosis virus isolates and physical maps of the DNA from three variants. Journal of General Virology 66, 2423–2430.

Crook, N. E., Clem, R. J. & Miller, L. K. (1993). An apoptosis-inhibiting baculovirus gene with a zinc-finger like motif. Journal of Virology 67, 2168–2174.

Crook, N. E., James, J. D., Smith, I. R. L. & Winstanley, D. (1997). Comprehensive physical map of the Cydia pomonella granulovirus genome and sequence analysis of the granulin gene region. Journal of General Virology 78, 965–974.

Dall, D., Luque, T. & O'Reilly, D. (2001). Insect–virus relationships: sifting by informatics. Bioessays 23, 184–193.

Derksen, A. C. G. & Granados, R. R. (1988). Alteration of a lepidopteran peritrophic membrane by baculoviruses and enhancement of viral infectivity. Virology 167, 242–250.

Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Research 12, 387–395.

Eldridge, R., Li, Y. & Miller, L. K. (1992). Characterization of a baculovirus gene encoding a small conotoxin-like polypeptide. Journal of Virology 66, 6563–6571.

Federici, B. (1997). Baculovirus pathogenesis. In The Baculoviruses, pp. 33–60. Edited by L. K. Miller. New York: Plenum Press.

Friesen, P. D. (1997). Regulation of baculovirus early gene expression. In The Baculoviruses, pp. 141–170. Edited by L. K. Miller. New York: Plenum Press.

Glen, D. M. & Payne, C. C. (1984). Production and field evaluation of codling moth granulosis virus for control of Cydia pomonella in the United Kingdom. Annals of Applied Biology 104, 87–98.

Gomi, S., Zhou, C. E., Yih, W., Majima, K. & Maeda, S. (1997). Deletion analysis of four of eighteen late gene expression factor gene homologues of the baculovirus, BmNPV. Virology 230, 35–47.

Gomi, S., Majima, K. & Maeda, S. (1999). Sequence analysis of the genome of Bombyx mori nucleopolyhedrovirus. Journal of General Virology 80, 1323–1337.

Griffiths, C. M., Barnett, A. L., Ayres, M. D., Windass, J., King, L. A. & Possee, R. D. (1999). In vitro host range of Autographa californica nucleopolyhedrovirus recombinants lacking functional p35, iap1 or iap2. Journal of General Virology 80, 1055–1066.

Guarino, L. A. & Summers, M. D. (1986). Interspersed homologous DNA of Autographa californica nuclear polyhedrosis virus enhances delayed-early gene expression. Journal of Virology 60, 215–223.

Haas, A. L., Katzung, D. J., Reback, P. M. & Guarino, L. M. (1996). Functional characterization of the ubiquitin variant encoded by the baculovirus Autographa californica. Biochemistry 35, 5385–5394.

Habib, S. & Hasnain, S. E. (2000). Differential activity of two non-hr origins during replication of the baculovirus Autographa californica nuclear polyhedrosis virus genome. Journal of Virology 74, 5182–5189.

Hashimoto, Y., Hayakawa, T., Ueno, Y., Fujita, T., Sano, Y. & Matsumoto, T. (2000). Sequence analysis of the Plutella xylostella granulovirus genome. Virology 275, 358–372.

Hayakawa, T., Ko, R., Okano, K., Seong, S.-I., Goto, C. & Maeda, S. (1999). Sequence analysis of the Xestia c-nigrum granulovirus genome. Virology 262, 277–297.

Hayakawa, T., Rohrmann, G. F. & Hashimoto, Y. (2000). Patterns of genome organisation and content in lepidopteran baculoviruses. Virology 278, 1–12.

Heldens, J. G. M., Broer, R., Zuidema, D., Goldbach, R. W. & Vlak, J. M. (1997). Identification and functional analysis of a non-hr origin of DNA replication in the genome of Spodoptera exigua multicapsid nucleopolyhedrovirus. Journal of General Virology 78, 1497–1506.

Hu, Z. H., Arif, B. M., Jin, F., Martens, J. W. M., Chen, X. W., Sun, J. S., Zuidema, D., Goldbach, R. W. & Vlak, J. M. (1998). Distinct gene arrangement in the Buzura suppressaria single-nucleocapsid nucleopolyhedrovirus genome. Journal of General Virology 79, 2841–2851.

IJkel, W. F. J., van Strien, E. A., Heldens, J. G. M., Broer, R., Zuidema, D., Goldbach, R. W. & Vlak, J. M. (1999). Sequence and organization of the Spodoptera exigua multicapsid nucleopolyhedrovirus genome. Journal of General Virology 80, 3289–3304.

IJkel, W. F. J., Westenberg, M., Goldbach, R. W., Blissard, G. W., Vlak, J. M. & Zuidema, D. (2000). A novel baculovirus envelope fusion protein with a proprotein convertase cleavage site. Virology 275, 30–41.

Jordan, A. & Reichard, P. (1998). Ribonucleotide reductases. Annual Review of Biochemistry 67, 71–98.

Kang, W., Crook, N. E., Winstanley, D. & O'Reilly, D. R. (1997). Complete sequence and transposon mutagenesis of the BamHI J fragment of Cydia pomonella granulosis virus. Virus Genes 14, 131–136.

Kang, W., Tristem, M., Maeda, S., Crook, N. E. & O'Reilly, D. R. (1998). Identification and characterization of the Cydia pomonella granulovirus cathepsin and chitinase genes. Journal of General Virology 79, 2283–2292.

Kang, W., Suzuki, M., Zemskov, E., Okano, K. & Maeda, S. (1999). Characterisation of baculovirus repeated open reading frames (bro) in Bombyx mori nucleopolyhedrovirus. Journal of Virology 73, 10339–10345.

Ko, R., Okano, K. & Maeda, S. (2000). Structural and functional analysis of the Xestia c-nigrum granulovirus matrix metalloproteinase. Journal of Virology 74, 11240–11246.

Kool, M., Goldbach, R. W. & Vlak, J. M. (1994). A putative non-hr origin of DNA replication in the HindIII-K fragment of Autographa californica multiple nucleocapsid nuclear polyhedrosis virus. Journal of General Virology 75, 3345–3352.

Kool, M., Ahrens, C. H., Vlak, J. M. & Rohrmann, G. F. (1995). Replication of baculovirus DNA. Journal of General Virology 76, 2103–2118.

Kuzio, J., Pearson, M. N., Harwood, S. H., Funk, C. J., Evans, J. T., Slavicek, J. M. & Rohrmann, G. F. (1999). Sequence and analysis of the genome of a baculovirus pathogenic for Lymantria dispar. Virology 253, 17–34.

Leisy, D. J., Rasmussen, C. H.-T. K. & Rohrmann, G. F. (1995). The Autographa californica nuclear polyhedrosis virus homologous region 1a: identical sequences are essential for DNA replication activity and transcriptional enhancer function. Virology 208, 742–752.

Lu, A. & Miller, L. K. (1997). Regulation of baculovirus late and very late gene expression. In The Baculoviruses, pp. 193–216. Edited by L. K. Miller. New York: Plenum Press.

Lu, A., Krell, P. J., Vlak, J. & Rohrmann, G. (1997). Baculovirus DNA replication. In The Baculoviruses, pp. 171–192. Edited by L. K. Miller. New York: Plenum Press.

Maguire, T., Harrison, P., Hyink, O., Kalmakoff, J. & Ward, V. K. (2000). The inhibitors of apoptosis of Epiphyas postvittana nucleopolyhedrovirus. Journal of General Virology 81, 2803–2811.

Monsma, S. A., Oomens, A. G. P. & Blissard, G. W. (1996). The GP64 envelope fusion protein is an essential baculovirus protein required for cell to cell transmission of infection. Journal of Virology 70, 4607–4616.

Morris, T. D., Todd, J. W., Fisher, B. & Miller, L. K. (1994). Identification of lef-7: a baculovirus gene affecting late gene expression. Virology 200, 360–369.

Murphy, F. A., Fauquet, C. M., Bishop, D. H. L., Ghabrial, S. A., Jarvis, A. W., Martelli, G. P., Mayo, M. A. & Summers, M. D. (editors) (1995). Virus Taxonomy. Sixth Report of the International Committee on Taxonomy of Viruses. Vienna & New York: Springer-Verlag.

O'Reilly, D. R. (1997). Auxiliary genes of baculoviruses. In The Baculoviruses, pp. 267–300. Edited by L. K. Miller. New York: Plenum Press.

Pearson, W. & Lipman, D. (1988). Improved tools for biological sequence comparison. Proceedings of the National Academy of Sciences, USA 85, 2444–2448.

Pearson, M. N. & Rohrmann, G. F. (1998). Characterisation of a baculovirus-encoded ATP-dependent DNA ligase. Journal of Virology 72, 9142–9149.

Pearson, M. N., Groten, C. & Rohrmann, G. F. (2000). Identification of the Lymantria dispar nucleopolyhedrovirus envelope fusion protein provides evidence for a phylogenetic division of the Baculoviridae. Journal of Virology 74, 6126–6131.

Reilly, L. M. & Guarino, L. A. (1996). The viral ubiquitin gene of Autographa californica nuclear polyhedrosis virus is not essential for viral replication. Virology 218, 243–247.

Rohrmann, G. F. (1992). Baculovirus structural proteins. Journal of General Virology 73, 749–761.

Swofford, D. L. (2000). PAUP*. Phylogenetic analysis using Parsimony (*and other methods), 4.0 edn. Sunderland, MA: Sinauer Associates.

Theilmann, D. & Stewart, S. (1992). Tandemly repeated sequence at the 3´ end of the IE-2 gene of the baculovirus Orgyia pseudotsugata multicapsid nuclear polyhedrosis virus is an enhancer element. Virology 187, 97–106.

Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22, 4673–4680.

Tomalski, M. D., Eldridge, R. & Miller, L. K. (1991). A baculovirus homolog of a Cu/Zn superoxide dismutase gene. Virology 184, 149–161.

van Oers, M. M. & Vlak, J. M. (1997). The baculovirus 10 kilodalton protein. Journal of Invertebrate Pathology 70, 1–17.

Vialard, J. E. & Richardson, C. D. (1993). The 1,629-nucleotide open reading frame located downstream of the Autographa californica nuclear polyhedrosis virus polyhedrin gene encodes a nucleocapsid-associated phosphoprotein. Journal of Virology 67, 5859–5866.

Vucic, D., Kaiser, W. J., Harvey, A. J. & Miller, L. K. (1997). Inhibition of reaper-induced apoptosis by interaction with inhibitor of apoptosis proteins (IAPs). Proceedings of the National Academy of Sciences, USA 94, 10183–10188.

Wang, P. & Granados, R. R. (1998). Observations on the presence of the peritrophic membrane in larval Trichoplusia ni and its role in limiting baculovirus infection. Journal of Invertebrate Pathology 72, 57–62.

Winstanley, D. & O'Reilly, D. R. (1999). Granuloviruses. In The Encyclopedia of Virology, 2nd edn, vol. 1, pp. 140–146. Edited by R. Webster & A. Granoff. London: Academic Press.

Xie, W. D., Arif, B., Dobos, P. & Krell, P. J. (1995). Identification and analysis of a putative origin of DNA replication in the Choristoneura fumiferana multinucleocapsid nuclear polyhedrosis virus genome. Virology 209, 409–419.

Yuen, L., Dionne, J., Arif, B. & Richardson, C. (1990). Identification and sequencing of the spheroidin gene of Choristoneura biennis entomopoxvirus. Virology 175, 427–433.

Zemskov, E. A., Kang, W. & Maeda, S. (2000). Evidence for nucleic acid binding ability and nucleosome association of Bombyx mori nucleopolyhedrovirus BRO proteins. Journal of Virology 74, 6784–6789.

JGV Direct table of contents

© 2001 SGM

This article is now available in the October 2001 print issue of JGV (vol. 82, 2531–2547). The complete issue of the journal may be seen in electronic form on JGV Online.