The gene encoding the capsid protein P82 of the ... - Semantic Scholar

Journal of General Virology (1997), 78, 2665–2673. Printed in Great Britain ...........................................................................................................................................................................................................................................................................................

The gene encoding the capsid protein P82 of the Choristoneura fumiferana multicapsid nucleopolyhedrovirus : sequencing, transcription and characterization by immunoblot analysis Xing Li,1 Anthony Pang,2 Hilary A. M. Lauzon,1 S. S. Sohi2 and Basil M. Arif1 Laboratory for Molecular Virology1, Canadian Forestry Service, Great Lakes Forestry Centre2, 1219 Queen St East, PO Box 490, Sault Ste. Marie, Ontario, Canada P6A 5M7

A gene encoding a capsid-associated viral structural protein has been identified and sequenced in the genome of the Choristoneura fumiferana multicapsid nucleopolyhedrovirus (CfMNPV). The gene has a 1872 nucleotide open reading frame (ORF) encoding 624 amino acids with a predicted molecular mass of 71±4 kDa. Transcription, which appeared to be initiated from a conserved GTAAG motif of baculovirus late genes, was detected at 12 h, reached a maximum at 48 h and declined at 72 h post-infection (p.i.). Part of the ORF was cloned in frame into a prokaryotic expression vector, pMALc2, and the fusion protein was used to generate

Introduction Choristoneura fumiferana multicapsid nucleopolyhedrovirus (CfMNPV) has a large double-stranded supercoiled DNA genome of 125 kb. Like other members in the family, CfMNPV produces two phenotypes during its infection cycle, the occlusion-derived virus (ODV) and the budded virus (BV). The former is embedded into polyhedra in the cell nucleus forming a stable occlusion body that can survive in the environment for a relatively long time. The latter phenotype buds through the cytoplasmic membrane and is infectious to neighbouring cells, resulting in secondary infection in the host. Analysis of structural proteins by SDS–PAGE revealed that baculoviruses have a complex structure with approximately 30 polypeptides in their ODV and BV structures (Summers & Smith, 1978 ; Braunagel & Summers, 1994). Several proteins have been Author for correspondence : Basil Arif. Fax ­1 705 759 5700. e-mail barif!am.glfc.forestry.ca. The GenBank accession number of the sequence reported in this paper is U65950.

0001-4620 # 1997 SGM

antibodies in rabbits. It was shown, with the aid of the polyclonal antiserum, that this viral protein was detectable at 24 h p.i. in infected cells. The protein appeared as an 82 kDa band in occlusion-derived virus and as an 82 kDa band and a 72 kDa band in budded virus. Amino acid sequence comparisons revealed that this ORF had high homology with the ORF p87 (77 % similarity) of Orgyia pseudotsugata (Op) MNPV and the ORF p80 (60 % similarity) of Autographa californica (Ac) MNPV. Immunoblots confirmed that the CfMNPV protein had antigenic similarities to the P87 protein of OpMNPV, but not to the P80 of AcMNPV.

identified as specific to ODV or BV (Rohrmann, 1992 ; Hong et al., 1994 ; Braunagel et al., 1996 a, b ; Theilmann et al., 1996). Some proteins are common to both ODV and BV structures, including a capsid protein (P39) and another putative capsid protein (P87). The P39 is a component of the capsid, which has been confirmed by immunoelectron microscopy (Russell et al., 1991). The P87 was first identified as another capsid protein in Orgyia pseudotsugata MNPV (OpMNPV) (Mu$ ller et al., 1990). This protein has a molecular mass of 87 kDa, as identified by an anti-fusion protein antiserum, which is larger than the computer-predicted mass of 70±7 kDa. The authors thought that this was caused by some unknown post-translational modifications (Mu$ ller et al., 1990). RNA analysis showed that it was a late gene and immunofluorescence microscopical studies showed that P87 was concentrated in the nucleus late in OpMNPV infection (Mu$ ller et al., 1990). Later, Lu & Carstens (1992) found a homologue of open reading frame (ORF) p87 potentially encoding a protein of 80 kDa in Autographa californica MNPV (AcMNPV). The fact that the AcMNPV BV-specific antisera reacted with a P80 fusion protein indicated that it was a component of the virus capsid

CGGF

X. Li and others

(Lu & Carstens, 1992). In this paper we report the identification and characterization of an ORF (p82) of CfMNPV which appears to be a homologue to the ORF p87 of OpMNPV and ORF p80 of AcMNPV. Our results showed that this protein was a viral structural protein that existed in both ODV and BV structures. However, it appeared as two different size bands in BV samples when analysed by SDS–PAGE, suggesting that different forms of the protein are incorporated into the BV.

Methods + Viruses and cells. The plaque-purified CfMNPV (Ireland strain) used in this study has been described previously (Arif et al., 1984). Occlusion bodies of OpMNPV and AcMNPV, which were propagated in O. leucostigma and Trichoplusia ni larvae, respectively, were kindly provided by Peter Ebling (Natural Resources Canada, Sault Ste. Marie, Ontario, Canada). A C. fumiferana midgut cell line (FPMI-CF-203), permissive to CfMNPV, was maintained in Insect-Xpress medium (Bio}Whittaker) supplemented with 10 % FBS (Gibco}BRL). + Location and sequence analysis of p82. Most of the ORF p82 was located in the CfMNPV fragment HindIII-O2 (Fig. 1). The fragment was cloned in pUC19 (pCfHindO2). The rest of the ORF p82 was found in the BamHI-D fragment (Fig. 1), which had been previously cloned into vector pBR322 (pCfBamD). The DNA fragments, HindIII-O2 and part of BamHI-D, were sequenced by primer walking on both strands. The nucleotide and the predicted protein sequences were analysed using MacVector 4.1.4 (International Biotechnologies). The CLUSTAL program (Higgins & Sharp, 1988) was used for amino acid sequence alignment.

+ Isolation of mRNA and Northern hybridization analysis. CF203 cells were infected with CfMNPV at an m.o.i. of 8. The mRNA samples were purified at various times p.i. using a mRNA purification kit (Gibco}BRL). Samples of mRNA were electrophoresed on 1 % agarose gels (Sambrook et al., 1989) and blotted onto a nylon membrane (Amersham). A 1±48 kb probe, prepared by double-digesting plasmid pCfHindO2 with EcoRI and HindIII enzymes (Figs 1 and 4), was labelled with [α-$#P]dATP using the BRL random primer labelling kit (Gibco}BRL). Northern blot hybridization was carried out at 42 °C in 50 % formamide and the membranes were washed at 65 °C in 0±1 % SSC and 0±1 % SDS. + Primer extension. A 32-mer primer (5« TTT TTA TGT CCG CGT GAC CGC GCG TCA TTA TG 3«) designated EP872 and located 45–76 nucleotides (nt) downstream of the ATG start was designed to identify the transcriptional start of the ORF p82. The primer extension experiments were carried out as described by Sambrook et al. (1989). + Construction and expression of the maltose-binding protein (MBP)–Cfp82 gene fusion. A 1796 nt EcoRI}XbaI fragment from the plasmid pCfBamD (Fig. 1) was subcloned into the prokaryotic expression vector pMAL-c2 (New England BioLabs). Expression of the MBP–Cfp82 fusion protein was induced by 0±3 mM (final concentration) of IPTG, as recommended by the manufacturer. This MBP–Cfp82 fusion protein was purified by affinity chromatography. + Production of antisera against ODV and MBP–Cfp82 fusion protein. In order to produce an antiserum against CfMNPV ODV, the occlusion bodies were dissolved in alkali and the virions were purified by sucrose gradient ultracentrifugation, as described by King & Possee (1992). New Zealand white rabbits were injected with 100 µg of the virion solution emulsified in complete Freund’s adjuvant for the first two injections. The rest of the injections were carried out with 100 µg of the

Fig. 1. Location of ORF p82 in the CfMNPV genome. The HindIII and BamHI restriction maps (according to Arif et al., 1984 ; Barrett et al., 1996) are shown to indicate location of ORF p82. The area (2199 bp) from NsiI to XbaI restriction sites has been magnified. All predicted ORFs in this region with a coding potential of more than 60 amino acids are shown. The orientation of the polyhedrin gene (Polh) is indicated.

CGGG

Capsid protein P82-encoding gene of CfMNPV

Fig. 2. Amino acid sequence alignments of the three ORFs from CfMNPV, OpMNPV and AcMNPV. Identical amino acid residues are shown by a star, conservative amino acid substitutions by a dot, and gaps by dashes. The highly conserved regions are boxed. The repeated sequences in CfMNPV (I, II, III and IV) and in OpMNPV (1 and 2) are in italics. Amino acid sequences representing potential N-linked glycosylation sites (Asn-X-Thr/Ser) are underlined.

virion solution emulsified in incomplete Freund’s adjuvant. Blood tests followed each booster injection. The antiserum was collected after the antibody titre had peaked. The same method was used to produce antiserum against the MBP–Cfp82 fusion protein. The protein concentration in both the initial and the booster injections was 500 µg for each rabbit. + Protein electrophoresis and immunoblot analysis. ODV and BV samples were prepared as described by King & Possee (1992) and analysed by SDS–PAGE (Laemmli, 1970). CF-203 cells were infected with the virus and cells were harvested by low-speed centrifugation (4000 r.p.m. ; 5 min) at various times p.i. The pellet was dissolved in electrophoresis sample buffer. The samples were analysed in 10 % SDS–polyacrylamide gels and the proteins were transferred to nitrocellulose membranes. In order to detect P82, the membranes were incubated with MBP–Cfp82 antiserum. A secondary antibody, conjugated with a peroxidase (Bio-Rad), was used to detect antigen–antibody complexes as described by Harlow & Lane (1988). A rabbit antiserum (trpE-p87) against the OpMNPV fusion protein P87 (a gift from George Rohrmann) was also used to test cross-reactions.

Results Identification and characterization of ORF p82

Sequence analysis of the HindIII-O2 and BamHI-D fragments (total 2199 nt) revealed the presence of a large ORF (ORF 1 in Fig. 1). Its orientation was the same as that of the polyhedrin gene on the physical map of the CfMNPV DNA genome (Fig. 1). Data bank searches showed that this ORF is a homologue of the ORF p87 in OpMNPV (Mu$ ller et al., 1990) and of the ORF p80 in AcMNPV (Lu & Carstens, 1992). Since this ORF encodes a protein which migrates as an 82 kDa band (see immunoblot analysis), we decided to call it p82. The nucleotide composition neighbouring the translational start site, ATAATGG, showed the presence of purine at the ®3 and ­4 positions, which is in agreement with the rule for efficient translation (Kozak, 1986). Two copies of the baculovirus late gene transcriptional start sequence, G}ATAAG,

CGGH

X. Li and others (a)

(c)

(44)

(b) Fig. 3. (a) Comparison of regions A, B and C among the proteins of CfMNPV, OpMNPV and AcMNPV. The predicted molecular masses (kDa) of the ORFs are shown at the left. The overall amino acid identities (%) and similarities (in parentheses) are shown at the right. The regional amino acid identities (top) and similarities (bottom, in parentheses) are indicated within the region blocks. (b) Hydrophilicity profiles from three ORFs. Above the axis (­) denotes hydrophilic regions and below the axis (®) indicates hydrophobic regions. (c) An alignment of the repeated amino acid sequences from CfMNPV (I, II, III, IV) and OpMNPV (two 1 and three 2). The identical amino acids are shaded.

were identified at positions ®34 and ®182 upstream of the ATG start site (Fig. 5). Primer extension experiments (see below) showed that the transcripts were initiated within the GTAAG motif at ®34 nt upstream of the ATG. The usual polyadenylation signal, AATAAA, was not found downstream of the translational termination codon. However, Hawkins (1991) described two sequence motifs that could substitute for the usual poly(A) signal. First, the sequence of CAC}TTG could be an alternative poly(A) addition signal. In ORF p82 the sequence CATTG (underlined below) was found 32 nt downstream of the TAA (bold) termination signal. Second, a sequence rich in GT and T downstream of the CAC}TTG motif was needed for efficient termination. A region rich in GT and T residues (six GT and ten T, underlined) was found 56 nt downstream of the termination signal (TAACATGACTTTAAAAGTTTTTTATTTGTAACAAACATTGTCACGATACAAATTAAACATTGTTTTAGTCTTTGTGGTCAAAGTTGAAACGTTC). The CfMNPV ORF p82 potentially encodes a protein of 624 amino acids with a predicted size of 71±4 kDa. The protein has four potential N-glycosylation sites (Asn-X-Ser}Thr) (Fig. 2). Also, in the middle part of the protein, there are four direct repeats of ten amino acids [Ser-Met-Pro(Thr)-Ser-Pro-Pro-Gln-

CGGI

Thr-Pro-Ala]. In the third and fourth repeats, threonine substitutes for proline at the third position. Alignment of the amino acid sequence of ORF p82 of CfMNPV with the homologues from OpMNPV (Mu$ ller et al., 1990) and AcMNPV (Lu & Carstens, 1992) revealed the presence of three distinct regions (Fig. 2). Conserved regions were found at the N and C termini. Although the overall amino acid identity and similarity among the three proteins were low (27 % identity and 50 % similarity), pairwise comparisons between CfMNPV}OpMNPV and CfMNPV}AcMNPV showed 77 % and 60 % similarities, respectively (Fig. 3 a). The regional comparisons (A, B and C) between CfMNPV and OpMNPV and between CfMNPV and AcMNPV are summarized in Fig. 3 (a). Hydrophilicity plots (Fig. 3 b) show that in the middle part (region B), the proteins of CfMNPV and OpMNPV are highly hydrophilic, indicating structural conservation, even though the region is hypervariable. The protein of AcMNPV has many hydrophobic areas in this region. Alignment of the four repeated sequences in CfMNPV P82 with the five repeats in P87 of OpMNPV revealed five rows of identical amino acids except for a conserved valine replacing alanine in the last row (Fig. 3 c). The existence of the repeats in these two viral proteins and the amino acid conservation

Capsid protein P82-encoding gene of CfMNPV

identify the transcriptional start site of ORF p82, a 32-mer oligonucleotide primer complementary to the sequence 45– 76 nt downstream of the ATG start was synthesized, endlabelled and then used for primer extension experiments. The longest extension product of 110 nt was found with mRNA samples 36 h and 48 h p.i. samples (Fig. 5) indicating that the mRNA started at the T residue of the GTAAG motif located ®34 nt upstream of the ATG start. Several shorter extension products were also detected near the 82 nt position (Fig. 5). At this position there was an ATAAG motif upstream of the ORF 5 ATG start complementary to p82 (Fig. 5). The nature of the shorter transcripts has not yet been ascertained. However, Lu & Carstens (1992) have observed a similar phenomenon with p80 and suggested that these transcripts could be prematurely arrested products, possibly due to the formation of duplex RNA. Immunoblot analysis

Fig. 4. Northern blot analysis of the p82 transcript : mRNA samples (1 µg/lane) were isolated from CfMNPV-infected CF-203 cells at various times (numbers at the top indicate h p.i.). The probe was a labelled DNA fragment of EcoRI/HindIII from the pCfHindO2 (underneath). The numbers on the left are the sizes (kb) of the RNA ladder mark (Gibco/BRL). The arrow on the right indicates the 2±0 kb mRNA of the p82 gene.

within the repeats add credence to the hypothesis of a conserved structure in region B of the proteins of CfMNPV and OpMNPV. This evidence also highlights the difference with the protein of AcMNPV. In this region of the viral DNA there are six other ORFs each potentially encoding a protein of more than 60 amino acids (Fig. 1) which illustrates the coding potential of this part of the genome. The partial ORF 5 shared a high homology with the ORF p48 sequences of OpMNPV and AcMNPV (data not shown). Data bank searches did not reveal significant similarities for the remaining ORFs to any known sequences. Transcriptional analysis

Expression of the ORF p82 in cells infected with CfMNPV was ascertained by extraction of mRNAs at various times p.i. followed by analysis on 1 % agarose gels. The mRNAs were blotted onto nylon membrane and hybridized to the ORF p82 DNA probe. A 2±0 kb mRNA transcript was detected at 12 h (when the film was given a long exposure, data not shown), reached a peak at 48 h and declined by 72 h p.i. (Fig. 4). To

To confirm that the ORF p82 gene expresses a viral structural protein in CfMNPV, a 1796 nt EcoRI}XbaI fragment (Fig. 1) from the ORF p82 including its putative polyadenylation signal sequence, was inserted into the prokaryotic expression vector pMAL-2. The insert was in the same translation reading frame as the vector’s malE gene encoding an MBP with a molecular mass of 40 kDa. Under the vector promoter (Ptac), a fusion protein (MBP ­ P82) with an expected molecular mass of 110 kDa was produced in induced bacteria (data not shown). The antiserum against the CfMNPV ODV reacted with this expressed fusion protein (data not shown), suggesting that it is a viral structural protein. In order to identify the protein, an antiserum against the fusion protein (MBP–Cfp82) was produced in rabbits and used in immunoblot analysis of viral proteins (Fig. 6). The antiserum reacted strongly with an 82 kDa ODV structural protein (Fig. 6 a, lane ODV). This antiserum also reacted with two BV structural proteins, an 82 kDa band and a 72 kDa band (Fig. 6 a, lane BV). The reaction was not as strong with the BV proteins as with the ODV protein. Repeated experiments with different ODV and BV preparations showed the same results. The ORF p82 amino acid sequence shares 77 % and 60 % similarities with those of OpMNPV ORF p87 and AcMNPV ORF p80, respectively (Fig. 3 a). To examine the immunological relationships among these three proteins, a heterogeneous antiserum (trpE-Op87) from OpMNPV p87 fusion protein was tested for cross-reaction with the CfMNPV structural proteins. The same 82 kDa band in the CfMNPV ODV sample had a strong cross-reaction with this heterogeneous antiserum (Fig. 6 b, lane ODV). A weak reaction occurred with the CfMNPV BV 82 kDa band, but no reaction was seen with the 72 kDa band (Fig. 6 b, lane BV), even when a higher concentration (1 : 500) of the antiserum was used. ODV samples from the three different viruses were also tested with the MBP–Cfp82 antiserum (Fig. 6 c). An 87 kDa band in the OpMNPV ODV sample had a strong cross-reaction with the antiserum (Fig. 6 c,

CGGJ

X. Li and others

Fig. 5. Primer extension analysis of the p82 gene. The primer (EP872) complementary to the sequence downstream of ATG (bold) translational start was end-labelled and annealed to the mRNA templates extracted at various times p.i. The extended products were analysed on a denaturing 6 % polyacrylamide gel. The same primer was also used for sequencing the pCfHindO2 plasmid. This sequence result was used as a ladder to identify primer extension product sizes. The numbers at the top indicate h p.i., the S lane contained the labelled DNA markers (φX174 Hinf from Promega). The arrow on the right indicates the 110 nt extension products which stopped at the transcriptional start site (indicated by a star within the motif). The second ATAAG motif (®182 to ®178 nt) is also shown from which no detectable extension products were observed. The underlined nucleotides in the sequence are for an opposite ORF (ORF 5), translation (CAT) and transcript (CTTAT) starts.

(a)

(b)

(c)

Fig. 6. Immunoblot analysis of the P82 protein from CfMNPV (a, b), OpMNPV and AcMNPV (c). ODV and BV samples were prepared (see Methods) and electrophoresed in 10 % SDS–polyacrylamide gels (8 µg/well). Immunoblots were carried out by using Cfp82 fusion protein antiserum (a and c ; 1 : 2000) and Op87 fusion protein antiserum (b ; 1 : 500). The prestained standards (Bio-Rad) were used. The arrow on the left (a) indicates the 82 kDa polypeptide in CfMNPV and the arrows on the right (c) indicate the 82 kDa and the 87 kDa polypeptides in CfMNPV and OpMNPV, respectively.

CGHA

Capsid protein P82-encoding gene of CfMNPV

Fig. 7. Immunoblot analysis of the P82 protein in CfMNPV-infected CF-203 cells. Cells (1¬105) were infected with virus and at various times p.i. they were harvested, analysed by electrophoresis on 10 % polyacrylamide gels and then blotted onto membranes. The blots were incubated with the Cfp82 fusion protein antiserum at 1 : 2000 dilution. ODV was used as a positive control. Uninfected CF203 cells were used as a negative control. S lane contains prestained molecular mass standards. The numbers above the lanes are the h p.i.

lane OpNPV), while no cross-reaction occurred with the AcMNPV ODV proteins (Fig. 6 c, lane AcNPV). The protein detected in the CfMNPV ODV sample by reaction with either the MBP–Cfp82 antiserum or the trpEOp87 antiserum had a molecular mass of 82 kDa, which is larger than the predicted size of 71±4 kDa. Glycosylation of the protein at one or more of the four potential N-glycosylation sites (Fig. 2) could account for the discrepancy with the observed size. In order to investigate this possibility, the ODV and BV samples were treated with N-glycosidase F (New England BioLabs), which can cleave N-linked glycoproteins, followed by immunoblot analysis. The mobilities in both ODV and BV samples were not changed after the treatments while the enzyme substrate (RNase B) showed a clear shift to a lower size (data not shown). We further confirmed the above results by studying the effect of tunicamycin (Sigma), which blocks the formation of protein N-glycosidic linkages, on CF-203 cells infected with CfMNPV. Tunicamycin (5 µg}ml) was added to CfMNPV-infected cells at 12 and 24 h p.i. and samples were collected after 12 h to check the size of P82. Samples were analysed by SDS–PAGE, blotted onto membranes and reacted with the antibody. We observed that P82 still migrated as an 82 kDa band clearly indicating that it was not glycosylated (data not shown). These results suggested that the viral protein may not be glycosylated and that the difference between the predicted and observed sizes could be the result of some other post-translational modifications or due to an unusual protein structure. The MBP–Cfp82 antiserum was also used to monitor the appearance of the protein in infected cells. Cells were infected with CfMNPV and at various times p.i. samples were analysed by gel electrophoresis, blotted onto membranes and hybridized with the antiserum (Fig. 7). An 82 kDa protein was first detected at 24 h and increased steadily to at least 48 h p.i. (Fig. 7, lanes 5–7). This temporal pattern is in agreement with the Northern blot results in Fig. 4. Several smaller bands also

reacted with the MBP–Cfp82 antiserum, especially a band of 50 kDa (lane 6 and 7). These smaller proteins could be degradation products of the P82 protein. Pre-immune rabbit serum did not react with any of the cellular or viral proteins (data not shown). We have also tested the cross-reactivity of MBP–Cfp82 antiserum with AcMNPV-infected SF-21 cell extracts at various times p.i. (0 h, 12 h, 24 h). There was no cross-reaction with any AcMNPV-specific proteins (data not shown). This result clearly confirmed our evidence that the AcMNPV P80 did not share antigenic similarities with the P82 of CfMNPV.

Discussion In this paper, we have described the identification in the genome of the CfMNPV a gene encoding an 82 kDa protein. This gene is a homologue of p80 and p87 genes from AcMNPV and OpMNPV, respectively. The ORF is 1872 nt, potentially encoding a protein of 71±4 kDa. However, the actual size of the protein in ODV is 82 kDa indicating that either the protein has been modified in some way or has some unusual structural features to allow it to migrate as an 82 kDa band. The gene contained two potential transcriptional start motifs, a GTAAG at ®34 nt and an ATAAG at ®182 nt relative to the ATG translational start site. Our data show that transcription is initiated at the GTAAG motif. The tetranucleotide motif, TAAG, is rare in baculovirus genomes (O’Reilly et al., 1992) and serves as a strong promoter of late genes. It is believed that the ATAAG motif allows for stronger gene expression than the GTAAG motif. The gene did not contain the usual polyadenylation signal, AATAAA, at the 3« end and, therefore, a different signal may be used to polyadenylate the mRNA. Downstream of the stop signal, the motif CATTG (­32 nt) followed by a region rich in GT and T residues could be utilized to polyadenylate the mRNA as found by Hawkins (1991). The same genes in OpMNPV and

CGHB

X. Li and others

AcMNPV contained the normal polyadenylation signal. In CfMNPV, the transcripts of p82 were detected at 12 h p.i. and reached a maximum at 48 h. This maximum was attained earlier in cells infected with OpMNPV (36 h p.i.) or with AcMNPV (24 h p.i.). The relative slowness observed with the CfMNPV may be further evidence that this virus is slower acting than AcMNPV (unpublished observations). The coding region of these gene homologues has conserved 5« and 3« sequences separated by a hypervariable part. This phenomenon is reflected in the predicted amino acid sequences of the respective proteins (Fig. 2). The N and C termini (regions A and C) are more conserved than the middle part (region B) of the protein. Lu & Carstens (1992) have suggested that an evolutionary pressure resulted in the conservation of region C, which may be involved in maintaining virion structure. This is supported by the comparison (Fig. 3 a) where region C appears to be slightly more conserved than region A among the three viral proteins. When looking at the proteins encoded by the three genes, it is clearly seen that those of CfMNPV and OpMNPV are more closely related to each other than to that of AcMNPV. For example, the size of the AcMNPV protein (691 amino acids) is larger than the other two proteins (624 amino acids). The central part of the proteins from CfMNPV and OpMNPV is highly hydrophilic while that from AcMNPV has many more hydrophobic sequences (Fig. 3 b). Within the hypervariable region, there are conserved repeated sequences in the proteins from CfMNPV and OpMNPV (Fig. 3 c). No such conserved repeats are present in the AcMNPV protein. It will be very interesting to know if P82 proteins from other NPV have such repeated sequences and how strong their conservation is. The antigenic similarity of P82 from CfMNPV and P87 from OpMNPV further solidifies the evidence for the close evolutionary relatedness of the two proteins. Anti-P82 antiserum produced with the P82 fusion protein cross-reacted with the OpMNPV protein but not with the AcMNPV protein. Antibody generated against the OpMNPV P87 fusion protein also did not cross-react with the same protein from AcMNPV (Mu$ ller et al., 1990). These data clearly show the evolutionary divergence of the AcMNPV P80 from the other two proteins. Phylogenetic analysis of the ecdysteroid UDPglucosyltransferase genes from several viruses has also shown CfMNPV and OpMNPV to be more closely related than either to AcMNPV (Barrett et al., 1995 ; Clarke et al., 1996 ; Hu et al., 1997). The discrepancy between the predicted and the observed size of P82 could not be resolved by assuming that the protein is glycosylated. We could not detect N-linked glycosylation and Mu$ ller et al. (1990) could not detect either N- or O-linked glycosylation. The protein may, therefore, have other posttranslational modifications or some unusual structural features to account for its increased size. The function and the precise location of P82 in the viral structure have not been fully elucidated. The data presented in

CGHC

this paper corroborate other published data that the protein is associated with the viral capsid and exists in both ODV and BV structures (Mu$ ller et al., 1990 ; Lu & Carstens, 1992 ; Rohrmann, 1992). Our immunoblot analyses data show that the structural form of P82 in ODV is different from that in BV because : (i) it migrated as different size bands in ODV and in BV (the protein in BV migrated as two distinct bands) ; (ii) the two viral phenotypes had different reaction intensities with the homologous antiserum (this phenomenon has been repeatedly observed with different virus preparations) ; and (iii) heterologous anti-P87 antiserum of OpMNPV also reacted differently with the proteins from the two virus phenotypes. Indeed, the heterologous antiserum reacted with the 82 kDa protein but not with the 72 kDa protein from BV. Therefore, even within the BV, the protein exists in two antigenically distinct forms, possibly due to specific post-translational modifications. Baculovirus capsid proteins usually exist in both ODV and BV phenotypes (e.g. P39). A recent study reported that in AcMNPV an ODV capsid and envelope protein, ODV-EC27, was only found in one phenotype, ODV (Braunagel et al., 1996 b). This suggests that some capsid proteins could be different in the two phenotypes. Our data imply yet other possibilities and show that a capsid protein could be incorporated into BV as different forms to those for ODV. Such a specific protein structure could result in a change in protein sizes and}or their antigenic characteristics. We are presently conducting immunoelectron microscopy to precisely locate P82 in the virus structures. We wish to thank Dr George Rohrmann (Oregon State University, USA) for his gift of trpE-p87 antiserum of OpMNPV ; Peter Ebling (CFSSSM, Ontario, Canada) for OpMNPV and AcMNPV PIBs ; Barb Mathieson (CFS-SSM) for her assistance in producing antisera ; and Guido Caputo and Barb Cook (CFS-SSM) for SF21 cells and AcMNPV inoculum. We thank Dr John Barrett and Karen Jamieson (CFS-SSM) for critical reading of earlier versions of the manuscript. This research has been supported by a grant from the Biotechnology Strategy Fund of the Canadian Forest Service.

References Arif, B. M., Kuzio, J., Faulkner, P. & Doerfler, W. (1984). The genome

of Choristoneura fumiferana nuclear polyhedrosis virus : molecular cloning and mapping of the Eco RI, Bam HI, Sma I, Xba I and Bgl II restriction sites. Virus Research 1, 605–614. Barrett, J. W., Krell, P. J. & Arif, B. M. (1995). Characterization, sequencing and phylogeny of the ecdysteroid UDP-glucosyltransferase gene from two distinct nuclear polyhedrosis viruses isolated from Choristoneura fumiferana. Journal of General Virology 76, 2447–2456. Barrett, J. W., Lauzon, H. A. M., Mercuri, P. S., Krell, P. J., Sohi, S. S. & Arif, B. M. (1996). The putative LEF-1 proteins from two distinct

Choristoneura fumiferana multiple nucleopolyhedroviruses share domain homology to eukaryotic primases. Virus Genes 13, 229–237. Braunagel, S. C. & Summers, M. D. (1994). Autographa californica nuclear polyhedrosis virus, PDV, and ECV viral envelopes and nucleocapsids : structural proteins, antigens, lipid and fatty acid profiles. Virology 202, 315–328.

Capsid protein P82-encoding gene of CfMNPV Braunagel, S. C., Elton, D. M., Ma, H. & Summers, M. D. (1996 a).

Laemmli, U. K. (1970). Cleavage of structural proteins during the

Identification and analysis of an Autographa californica nuclear polyhedrosis virus structural protein of the occlusion-derived virus envelope : ODV-E56. Virology 217, 97–110.

assembly of the head of bacteriophage T4. Nature 227, 680–685. Lu, A. & Carstens, E. B. (1992). Nucleotide sequence and transcriptional analysis of the p80 gene of Autographa californica nuclear polyhedrosis virus : a homologue of the Orgyia pseudotsugata nuclear polyhedrosis virus capsid-associated gene. Virology 190, 201–209.

Braunagel, S. C., He, H., Ramamurthy, P. & Summers, M. D. (1996 b).

Transcription, translation, and cellular localization of three Autographa californica nuclear polyhedrosis virus structural proteins : ODV-E18, ODV-E35, and ODV-EC27. Virology 222, 100–114. Clarke, E. E., Tristem, M., Cory, J. S. & O’Reilly, D. R. (1996).

Characterization of the ecdysteroid UDP-glucosyltransferase gene from Mamestra brassicae nucleopolyhedrovirus. Journal of General Virology 77, 2865–2871. Harlow, E. & Lane, D. (1988). Antibodies. A Laboratory Manual. Cold Spring Harbor, NY : Cold Spring Harbor Laboratory. Hawkins, J. D. (1991). Gene Structure and Expression, 2nd edn. Cambridge : Cambridge University Press. Higgins, D. G. & Sharp, P. M. (1988). CLUSTAL : a package for performing multiple sequence alignment on a microcomputer. Gene 73, 237–244. Hong, T., Braunagel, S. C. & Summers, M. D. (1994). Transcription, translation, and cellular localization of PDV-E66 : a structural protein of the PDV envelope of Autographa californica nuclear polyhedrosis virus. Virology 204, 210–222. Hu, Z. H., Broer, R., Westerlaken, J., Martens, J. W. M., Jin, F., Jehle, J. A., Wang, L. M. & Vlak, J. M. (1997). Characterization of the ecdysteroid

UDP-glucosyl-transferase gene of a single nucleocapsid nucleopolyhedrovirus of Buzura suppressaria. Virus Research 47, 91–97. King, L. A. & Possee, R. D. (1992). The Baculovirus Expression System. A Laboratory Guide. New York : Chapman & Hall. Kozak, M. (1986). Point mutations define a sequence flanking the AUG initiator codon that modulates translation by eukaryotic ribosomes. Cell 44, 283–292.

Mu$ ller, R., Pearson, M. N., Russell, R. L. Q. & Rohrmann, G. F. (1990).

A capsid-associated protein of the multicapsid nuclear polyhedrosis virus of Orgyia pseudotsugata : genetic location, sequence, transcriptional mapping, and immunocytochemical characterization. Virology 176, 133–144. O’Reilly, D. R., Miller, L. K. & Luckow, V. A. (1992). Baculovirus Expression Vectors. A Laboratory Manual. New York : W. H. Freeman and Company. Rohrmann, G. F. (1992). Baculovirus structural proteins. Journal of General Virology 73, 749–761. Russell, R. L. Q., Pearson, M. N. & Rohrmann, G. F. (1991). Immunoelectron microscopic examination of Orgyia pseudotsugata multicapsid nuclear polyhedrosis virus-infected Lymantria dispar cells : time course and localization of major polyhedron-associated proteins. Journal of General Virology 72, 275–283. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning : A Laboratory Manual, 2nd edn. Cold Spring Harbor, NY : Cold Spring Harbor Laboratory. Summers, M. D. & Smith, G. E. (1978). Baculovirus structural polypeptides. Virology 84, 390–402. Theilmann, D. A., Chantler, J. K., Stewart, S., Flipsen, H. T. M., Vlak, J. M. & Crook, N. E. (1996). Characterization of a highly conserved

baculovirus structural protein that is specific for occlusion-derived virions. Virology 218, 148–158. Received 2 January 1997 ; Accepted 23 May 1997

CGHD