The 131-Amino-Acid Repeat Region of the Essential 39-Kilodalton Core Protein of Fowlpox Virus FP9, Equivalent to Vaccinia Virus A4L Protein, Is Nonessential and Highly Immunogenic

Abstract

The immunodominant, 39,000-molecular weight core protein (39K protein) of fowlpox virus (FP9 strain), equivalent to the vaccinia virus A4L gene product, contains highly charged domains at each end of the protein and multiple copies of a 12-amino-acid serine-rich repeat sequence in the middle of the protein. Similar repeats were also detected in other fowlpox virus strains, suggesting that they might confer a selective advantage to the virus. The molloscum contagiosum virus homolog (MC107L) also contains repeats, unlike the vaccinia virus protein. The number of repeats in the fowlpox virus protein does not seem to be crucial, since some strains have a different number of repeats, as shown by the difference in the size of the protein in these strains. The repeat region could be deleted, indicating that it is not essential for replication in vitro. It was not possible to delete the entire 39K protein, indicating that it was essential (transcriptional control signals for the flanking genes were left intact). The repeat region is partly responsible for the immunodominance of the protein, but the C-terminal part of the protein also contains highly antigenic linear epitopes. A role for the 39K protein in immune system modulation is discussed.

An increasing number of recombinant poxvirus vaccines have been developed recently. Most of them use vaccinia virus as the vector, but fowlpox virus (FWPV) (10), the prototypic member of the avipoxviruses, has also been used extensively to develop candidate recombinant vaccines for poultry pathogens including avian influenza virus, Marek’s disease virus, Newcastle disease virus, and infectious bursal disease virus (3). Unlike vaccinia virus, which can replicate in a wide range of mammalian as well as avian cells, avipoxviruses replicate only in avian cells. In mammalian cells, their replication is blocked but gene expression occurs. In some cell types, FWPV replication is blocked during virus assembly (30). This late blockage allows sufficient gene expression to induce a protective immune response in vivo against foreign antigens cloned into vectors. This property has been exploited to develop candidate avipoxvirus recombinants as safe nonreplicating vectors for vaccination of mammals, including humans, against rabies virus, measles virus, Japanese encephalitis virus, human cytomegalovirus, and human immunodeficiency virus (19). Among these, a canarypox-rabies recombinant virus has been subjected to a phase I clinical trial (4). Recently, the potential of FWPV recombinants for immunomodulation in chickens (34) and mammals (17) and for cancer immunotherapy (33) has been explored.

FWPV has been much less intensively studied than vaccinia virus. It has a genome 30 to 60% larger (6, 24) than that of vaccinia virus (12) and shows a very different genome organization (24, 26). Blocks of sequence within which genes exist in the same relative position in both viruses have been shown, but the genomic location of those blocks differs widely between the viruses (24). Only about 30% of the FWPV genome sequence has been determined (EMBL and GenBank databases as of September 96). The reported sequences include only a few genes encoding virus structural proteins (excluding enzymes, some of which are also packaged in the complex virion), including homologs of vaccinia virus genes A10L (p4a; EMBL:A20158) (2a), F12L (26), and F13L (5). An abundant late structural 39,000-molecular-weight protein (39K protein) expressed from a very active bidirectional promoter (15) has also been described previously by our group (2).

Since there were virtually no serological reagents available for specific FWPV structural proteins that could be used as markers for the process of morphogenesis, we produced monoclonal antibodies (MAbs) against FWPV. We isolated several MAbs recognizing a protein with an observed molecular mass of 42 kDa, and we show in this paper that it corresponds to the 39K protein described previously (2). For consistency with the previous paper, we will continue to refer to it as the FWPV 39K protein. Sequence analysis of the gene encoding this protein had shown that the protein contains multiple copies of a 12-amino-acid (aa) serine-rich repeat sequence in the central part of the protein (2). The first seven copies are perfectly conserved, whereas the last four copies have diverged extensively and the ninth copy consists of only 11 aa. The protein contains neither an N-terminal signal sequence nor a C-terminal anchor sequence. Maa and Esteban (20) mapped an immunodominant vaccinia virus protein, with an observed molecular mass of 39 kDa (hereafter referred to as p39 to distinguish it from the FWPV protein), to the region between genes encoding p4a and p4b. Subsequent sequence determination of the complete vaccinia virus genome (12) and of a vaccinia virus p39 clone (9) revealed that p39 is encoded by a gene (A4L) located between those encoding p4b (A3L) and one of the RNA polymerase subunits (A5R). The same arrangement of genes was found in FWPV, and p39 is the vaccinia virus protein with the highest homology to the FWPV 39K protein. Comparison of the vaccinia virus p39 protein sequence with the FWPV 39K protein sequence revealed that the two proteins are different (8.2% aa identity) and that only a region spanning aa 156 to 169 of vaccinia virus p39 was found to have significant homology to aa 92 to 104 of the FWPV 39K protein. Demkowicz et al. (9) therefore concluded that the 39-kDa immunodominant proteins in vaccinia virus and FWPV have evolved divergently.

The genome sequence of molluscum contagiosum virus, another poxvirus infecting humans, has been published recently (29). Analysis of the sequence of its A4L homolog revealed the presence of a repeat sequence in the central part of the protein, as found in the FWPV 39K protein, whereas p39 of vaccinia virus does not contain obvious repeats.

A large number of proteins containing repeats have been found in eukaryotes, prokaryotes, and viruses. These repeats are often involved in protein-protein interactions and are often essential for the function of the protein, as is the case for the tetratrico peptide repeats (16), the ankyrin repeat motif (23), Epstein-Barr virus nuclear antigen 2 (32), and the leucine-rich repeats (27).

In several proteins, such as EBNA-1 (18) or the circumsporozoite protein of several species of Plasmodium (25), repeats have been shown to be immunodominant. However, in other cases, such as the circumsporozoite protein of Plasmodium vivax, the repeats do not seem to be immunodominant (21). We therefore attempted to determine whether the FWPV 39K protein is essential for FWPV replication and whether the repeat region is necessary for function and responsible for the immunodominant trait of the protein.

MATERIALS AND METHODS

Virus and cells.

Avipoxviruses were grown on chicken embryo fibroblasts (CEFs) in the presence of 2% newborn calf serum. The pathogenic HP-1 strain, the partially attenuated HP-200 strain, and the attenuated HP-438 strain (the last two derived from HP-1 by six passages on CEFs, two passages on chorioallantoic membranes, and 200 or 438 passages through CEFs, respectively [22]) were provided by A. Mayr. A twice-plaque-purified isolate of HP438 (FP9) was then passaged six times to constitute a stock. The vaccinal Poxine strain was provided by Duphar. The mild vaccine Websters FPV M strain and the Chick-N-Pox vaccine strain were obtained from Salsbury Laboratories, Inc. (now Solvay Animal Health), Charles City, Iowa.

Canarypox virus (V and 229 strains), pigeonpox virus (Peekham and 950 strains), sparrowpox virus (9037 strain), and turkeypox virus were provided by the Central Veterinary Laboratory, Weybridge, United Kingdom.

Monoclonal and polyclonal antibodies.

FWPV-specific monoclonal antibodies 3D9/2F9 and 8F3/2E11 were produced by immunizing BALB/c mice intraperitoneally with 60 μg of purified HP-1 as described elsewhere (7a). MAb GB9 and GG1 were produced by immunizing BALB/c mice twice intraperitonally, 3 weeks apart, with 100 μg of purified FP9. The final boost was administered 4 weeks later with the same amount of antigen.

The anti-FLAG M2 MAb specific for the flag peptide of the fusion proteins was purchased from IBI.

Chicken hyperimmune polyclonal sera were produced by intradermally inoculating 5-week-old chickens, three times at 2-week intervals, with 5 × 10⁷ PFU/50 μl of FP9 (sera B61 to B65) or the Duphar Poxine vaccinal strain (sera C66 to C70). The chickens were bled 2 weeks after the third inoculation.

Virus purification.

Confluent CEFs were infected with FP9 at 0.4 PFU per cell. At 5 days postinfection, the supernatant was harvested along with any remaining cells. This virus suspension was freeze-thawed once and clarified for 30 min at 1,000 × g. The supernatant was then centrifuged for 1 h at 40,000 × g. The pellet was resuspended in TMN buffer (10 mM Tris [pH 7.5], 1.5 mM MgCl₂, 10 mM NaCl) and centrifuged for 2 h at 160,000 × g through a 25% (wt/wt) sucrose cushion in TMN buffer. The pellet was resuspended in 1 ml of TMN buffer and sonicated before being layered over a 15 to 40% sucrose gradient and centrifuged at 30,000 × g for 50 min. Fractions of 0.5 ml were collected from the bottom of the tube, and the virus was concentrated by centrifugation at 40,000 × g.

Preparation of envelopes and cores.

Purified virus (1.5 mg/ml) was incubated for 1 h at 37°C with 1 volume of 0.2 M Tris-HCl (pH 8.6) containing 2% Triton X-100, 2% Nonidet P-40 (NP-40), or 0.2% sodium deoxycholate (DOC) with or without 10 mM dithiothreitol. The cores were then pelleted by centrifugation at 200,000 × g for 30 min at 4°C in a TLA-100.1 rotor. The envelope fractions were collected, and the pellets were resuspended in 0.1 M Tris-HCl (pH 8.6). All the samples were centrifuged and treated again under the same conditions. Each sample was fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by Western blotting with the GB9 MAb.

Western blotting.

Protein samples fractionated by SDS-PAGE were electrotransferred onto a nitrocellulose membrane (Hybond-C; Amersham) by following standard protocols. After transfer, the membranes, blocked overnight in 5% nonfat dry milk in TN buffer (10 mM Tris, 500 mM NaCl [pH 7.5]), were incubated for 90 min at room temperature with primary antibodies (diluted in TN buffer containing 2% dried milk) followed by goat anti-mouse (1:1,000), goat anti-rabbit (1:1,000), or rabbit anti-chicken (1:15,000) secondary antibodies conjugated to alkaline phosphatase (Sigma). After several washes, the immunoblots were developed with the ImmunoPure nitroblue tetrazolium–5-bromo-4-chloro-3-indolylphosphate (NBT/BCIP) substrate kit (Pierce).

Electron microscopy. (i) Transmission electron microscopy.

CEFs were infected with 1 PFU of FP9 or 1-3 R1 repeat deletion mutant per cell and fixed in situ with 2.5% 0.1 M phosphate-buffered glutaraldehyde at different times postinfection. They were then removed with a disposable cell scraper, transferred to a centrifuge tube, and left to fix overnight. They were then washed twice in 0.1 M phosphate buffer for 30 min and then postfixed in 1% 0.1 M phosphate-buffered osmium tetroxide for 90 min. The cells were then washed in distilled H₂O five times and resuspended in 1% aqueous uranyl acetate for 1 h in the dark. After further rinsing in distilled H₂O, the pellets were embedded in 0.5 ml of 1% agar. The solidified agar-cell block was cut into 2-mm³ cubes, dehydrated through a methanol gradient, and subjected to two incubations (15 min each) in propylene oxide; the cubes were then embedded in Araldite resin, with polymerization occurring for 48 h in a 60°C oven. The blocks were then stored before being sectioned on a Reichart-Jung Ultracut E ultramicrotome; the section thickness was in the region of 90 nm. Finally, the sections were collected on 3-mm-diameter copper or nickel Athene grids, stained with lead citrate for 10 min, and examined under a Philips EM 300 transmission electron microscope at an accelerating voltage of 80 kV.

(ii) Immunogold labelling.

The cells were fixed in situ for 2 h in 0.1 M phosphate-buffered 2% paraformaldehyde–0.1% glutaraldehyde mixture. They were then gently removed with a disposable cell scraper, washed twice in 0.1 M phosphate buffer, and dehydrated in a graded ethanol series to L. R. White resin (Agar Scientific Ltd.). Infiltration of resin occurred for up to 72 h; this step involved five changes of resin and constant mixing at 4°C. Polymerization of resin took place for 22 h at 55°C. Sections with a thickness of 90 nm were cut on a Reichart-Jung Ultracut E ultramicrotome and collected onto Athene nickel grids. The sections were blocked with normal goat serum (1:10 in phosphate-buffered saline [PBS]) for 15 min and then incubated for 2 h with the GB9 MAb. The samples were jet washed in PBS before being incubated for 15 min with goat serum anti-mouse immunoglobulin G conjugated to 10-nm-diameter gold particles (British Biocell International, Cardiff, United Kingdom), diluted 1:30 in PBS. The sections were then washed in PBS and in double-distilled H₂O and allowed to air dry before being counterstained with aqueous uranyl acetate (1%) and lead citrate for 30 s each. The samples were examined in a Philips EM 300 transmission electron microscope at an accelerating voltage of 80 kV.

Cloning.

The Escherichia coli pFLAG-CTC expression vector was obtained from IBI. The entire 39K protein-coding sequence or fragments of it (Fig. 1C) were amplified by PCR from purified genomic DNA. A HindIII site or a SalI site (underlined) was included in the primers to enable the insertion and ligation in the multiple-cloning site of the plasmid. The sequence of the different primers used were as follows: PR2, 5′-GGGACAAGCTTGCTGAGAACTTCCACAAAG-3′; PR4, 5′-CCCGCAAGCTTAGTACTTTAGTACCGTCTTC-3′; PR5, 5′-CCTGTGTCGACTAAAGTTCCTTGATTGCCGC-3′; PR6, 5′-GGGACGTCGACAGGAATAATAGCATCTCTGAG-3′; PR8, 5′-CCCCCAAGCTTACTCTACCCTCGACTAGC-3′; and PR9, 5′-GGGGTGTCGACCGGTACTAAAGTACTACTACTGC-3′.

FIG. 1.

(A) Schematic representation of the 39K protein homologs in FWPV, vaccinia virus, variola virus, and molluscum contagiosum virus. The terminus charged domains are hatched, whereas the shaded rectangles represent the repeats. The different intensity of shading represents the degree of conservation between the repeats. The symbol ∇ in the variola virus protein shows the position of the 10-aa deletion in the variola virus protein in comparison with the vaccinia virus protein. The protein sequences of the vaccinia virus WR strain and the Bangladesh-75 strain of variola virus are found in the SWISS-PROT database under accession no. p29191 and p33832, respectively. The amino acid sequence of the A4L homolog of molluscum contagiosum virus is found in the GenBank database under accession no. U60315. (B and C) Structure of the FWPV 39K protein (B) showing the positions of the primers used to amplify the different fragments (C). The black rectangles in panel C represent the FLAG C-terminal peptides.

Expression in E. coli.

Single colonies of E. coli XL1 Blue transformed with the pFLAG constructs were grown for 3 h at 37°C in Luria broth containing ampicillin (100 μg/ml) and induced with isopropyl-β-d-thiogalactopyranoside IPTG (final concentration, 0.5 mM) for 1 or 4 h. The bacteria were pelleted in an Eppendorf centrifuge (1 min at 13,000 × g), resuspended in 1:10 the initial volume of SDS-PAGE sample buffer, boiled, and subjected to SDS-PAGE.

PCR.

Avipoxviruses were propagated in CEFs for 2 days. The cells were then washed in PBS, scraped into 1 ml of PBS, and pelleted. The pellet was resuspended and incubated for 2 h at 55°C in 200 μl of extraction buffer (10 mM Tris HCl, 100 mM NaCl, 10 mM EDTA, 0.5% SDS, 2% β-mercaptoethanol) containing 1.25 mg of proteinase K per ml. The DNA was then phenol-chloroform extracted and ethanol precipitated. Each virus was tested by PCR with the primers described above and a primer complementary to the sequence of the conserved repeats (PR19, 5′-CTTCGTCTTCCGTTAGTGG-3′).

FWPV knockout mutants.

The transient dominant-selection method described by Falkner and Moss (11) was used to attempt generation of FWPV mutants with deletions of the entire 39K protein or the repeat region. A cassette containing the E. coli gpt gene under the control of the vaccinia virus p7.5 early/late promoter was cloned into the SmaI site of pNEB193 (New England Biolabs, Inc.) in the same orientation as the amp gene (pGNR). Primers MASH48 (5′-GCGGATCCGATTGAGTAGTTTCATCG-3′) and MASH49 (5′-GCGCAAGCTTCAAGACATCACATACG-3′) containing, respectively, BamHI and HindIII restriction sites (underlined), were used to amplify a 1,404-bp fragment including the 39K gene (see Fig. 6A) with the genomic viral DNA as a template. This fragment was inserted into pGNR previously digested with both restriction enzymes. The resulting plasmid (pGNR39K) was subsequently digested with XbaI and SpeI to delete most of the 39K gene (see Fig. 6A) and religated, with the compatible cohesive ends, to generate pDSM1. To generate a mutant with a deletion of the repeat region of the protein (pDSM2A and pDSM2B), pGNR39K was digested with BbsI (see Fig. 6A) and religated with adapter A or B. The forward and reverse sequences of adapter A were 5′-AACTAGTTCTAGTACTTT-3′ and 5′-TACTAAAGTACTAGAACT-3′, respectively. The forward and reverse sequences of adapter B, which contained BglII and PstI sites, were 5′-AACTAGATCTAGTCTGCAGTT3′ and 5′-TACTAACTGCAGACTAGATCT-3′, respectively. The sequences of the religated plasmids were checked by automated sequencing.

FIG. 6.

(A) Structure of the PCR product (amplified with primers MASH48 and MASH49 and inserted into the pGNR plasmid), including the gene encoding the 39K protein (open areas) and 5′ termini of the adjacent genes, the equivalents of A5R and A3L (hatched areas). The repeats are indicated by arrowheads, and the positions of restriction sites used to generate the deletion mutants are shown. The late promoters of the 39K and 4b proteins (on the upper strand) are represented by narrow solid boxes, whereas the early/late promoter of the A5R equivalent (lower strand) is represented by a narrow shaded box. The cap sites of the three genes are represented by asterisks and are located 2 nucleotides upstream of the respective initiator codons. The genomic structures of the deleted clones are shown below the scale bar (in base pairs). (B) 39K protein deletion mutants. The gpt-negative clones were screened by PCR with primers MASH48 and MASH49. All the gpt-negative clones contained only a copy of the wild-type gene, like the clone represented in lane 1. One gpt⁺ intermediate clone was used as a control (lane 2), as well as the wild-type virus (lane 3). (C) Repeat-region deletion mutants. The gpt-negative clones were screened by PCR with primers PR8 and PR6. Two gpt-negative clones containing adapter A (1-3 R1 and 1-4 R1 clones) (lanes 1 and 2) and one gpt-negative clone containing adapter B (8-3 R1 clone) (lane 3) contained a deleted copy of the gene. Two gpt⁺ intermediate clones (8-2 R2 and 9-2 R2 clones) (lanes 4 and 5) and the wild-type virus (lane 6) were tested as controls. (D) Western blotting analysis, with MAb GB9, of successive plaque purifications of a gpt⁺ isolate (clone 1-3 R1) in nonselective medium (lanes 1 to 7). CEFs grown in 6-cm petri dishes were infected with the different isolates. When a cytopathic effect of approximately 80% was observed, the cells were washed, covered with 1 ml of PBS, and freeze-thawed once. One volume of electrophoresis buffer was added to 1 volume of cellular lysate, boiled, and separated by SDS-PAGE (15% polyacrylamide). In lanes 1 to 3, a wild-type protein is still expressed, but in lanes 4 to 7, only the deleted protein is observed when the isolate became gpt-negative. The wild-type virus was tested as a control (lane 8). The deletion mutant obtained after seven plaque purifications (lane 7) was passaged 12 times on CEFs and is shown in lane 10 alongside the second passage run as a control (lane 9). The wild-type virus (lane 11) and the 1-3 R1 deletion mutant (lane 12), purified on a sucrose gradient, were tested in parallel, showing that the deleted protein is present in the virion.

CEFs infected with FP9 were transfected with plasmid pDSM1, pDSM2A, or pDSM2B by using Lipofectin (Gibco BRL). Recombinant clones were plaque purified three times in selective medium containing mycophenolic acid, xanthine, and hypoxanthine, and the gpt⁺ clones were further purified without selection until they became gpt-negative. CEFs infected with different gpt-negative clones were lysed with proteinase K, and the extracted viral DNA was analyzed by PCR with primers MASH48 and MASH49. CEFs infected with the repeat knockout clones were also analysed by Western blotting.

To check whether the deletion of the 39K gene was affecting the adjacent p4b promoter, plasmid pEFL29, containing the β-galactosidase gene under the control of the FWPV p4b promoter, was deleted with SpeI, the same enzyme used to create the 39K deletion mutant. The resulting plasmid, ΔpEFL29, and the original plasmid were then used in a transient-expression assay to compare the expression of the β-galactosidase as described by Binns et al. (1). Duplicates of 2.5, 5, or 10 μg of purified plasmid DNA were transfected with 20 μg of Lipofectin into CEFs grown in 25-cm² flasks and inoculated with FP9 (multiplicity of infection, 2 PFU/cell) 2 h earlier. At 22 h after transfection, the cells were washed in PBS and freeze-thawed three times in 500 μl of 0.25 M Tris-HCl (pH 7.5)–5 mM dithiothreitol. Clarified lysate (50 μl) was distributed in four wells of a microtiter plate. A mixture of 1 μl of 60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgCl₂, 50 mM 2-mercaptoethanol, and 100 μl of 2-mg/ml o-nitrophenylgalactose (ONPG) in 60 mM Na₂HPO₄–40 mM NaH₂PO₄ was added per well. After a 30-min incubation at room temperature, the absorbance at 405 nm was measured in an enzyme-linked immunosorbent assay plate reader.

Single-step growth curve experiment.

Confluent CEFs were infected in triplicate with the repeat deletion mutant 1-3 R1 or with the wild-type virus at 10 PFU per cell. The inoculum was removed 1 h later and replaced by fresh medium. At different times postinfection, the extracellular medium was collected, and the cells were overlaid with 1 ml of fresh medium and stored at −70°C. Intracellular and extracellular viruses were subjected to titer determination by a plaque assay.

RESULTS

Structure of the FWPV 39K protein.

Hyperimmune polyclonal chicken sera obtained after immunization with FWPV (FP9 and Poxine strains) recognized predominantly a protein of about 42 kDa by Western blotting (data not shown). A highly expressed and highly antigenic protein with a similar molecular mass was also recognized by several of our MAbs (3D9, 8F3, GB9, and GG1). We showed that this protein was the 39K protein described previously by our group (2) by cloning the 39K gene as a fusion protein with the FLAG epitope (Fig. 1C) and analyzing protein expressed in E. coli by Western blotting with MAbs or chicken hyperimmune sera (Fig. 2). Analysis of the sequence of this protein (2) identified a central repeat region (aa 108 to 238) containing 11 copies of a 12-aa serine-rich sequence spanning more than one-third of the length of the protein (Fig. 1). Highly charged regions are located at each end of the protein (Fig. 1A), whereas no charged residue can be found between aa 76 and 263. Of the 75 N-terminal residues, 37% are charged, as are 32% of the 60 C-terminal residues.

FIG. 2.

Western blot analysis of the fusion proteins expressed in E. coli. Extracts from bacteria (XL1 Blue), transformed with plasmids described in Fig. 1C, were transferred onto nitrocellulose paper and analyzed by Western blotting with the anti-FLAG peptide MAb (A); with three anti-39K protein MAbs, GB9 (B), 8F3 (C), and 3D9 (D); and with two hyperimmune chicken polyclonal sera, B63 (E) and B62 (F). In each panel, uninfected bacteria have been included as a negative control (lane XL) and purified FP9 has been included as a positive control (lane PV). The other lanes (N, NM, M, MC, C, and NMC) correspond to the fusion proteins in Fig. 1C.

Comparison with vaccinia virus A4L protein.

The highest scores obtained with FASTA for FWPV 39K protein were against the procollagen alpha 2 (V) chain (optimized score, 142) and circumsporozoite protein from Plasmodium berghei (optimized score, 113), both of which contain repeats partially matching the repeats of the 39K protein. The vaccinia virus protein showing the highest homology to the FWPV 39K protein was the A4L gene product, p39 (FASTA initial and optimized scores, 55; corresponding to 18% amino acid sequence identity to the FWPV 39K protein as determined by the Genetics Computer Group GAP program). Unlike the FWPV protein, vaccinia virus p39 does not contain obvious repeats. However, interestingly, the molluscum contagiosum virus protein (MC107L) homologous to vaccinia virus A4L (showing 24% amino acid sequence identity to the vaccinia virus protein and 17% amino acid sequence identity to the FWPV 39K protein) also contains repeats in the central part of the protein (Fig. 1A): eight repeats of 9 aa (except the seventh repeat, which is 10 aa long) are clustered (aa 215 to 287). The consensus sequence of the repeats, PAPATAPAC, is conserved in the first, third to sixth, and eighth repeats, whereas the second and seventh repeats have diverged (PAPATSPAC and PAPACPTPAC, respectively). A diverged copy of the repeats (PMPATAPAA) was also located further upstream (aa 104 to 112). The corresponding FWPV, vaccinia virus, variola virus, and molluscum contagiosum virus genes occupy the same position relative to adjacent genes: all of them are located between a gene coding for major core protein 4b (A3L) and a gene coding for one of the RNA polymerase subunits (A5R) (1, 14, 29). The corresponding proteins all contain highly charged regions at each end of the protein, allowing three domains to be distinguished (Fig. 1A): (i) the N-terminal highly charged domain 1 (37, 30, 30, and 32% charged residues in FWPV, vaccinia virus, variola virus, and molluscum contagiosum virus, respectively) which shows the same length in all proteins; (ii) the uncharged or weakly charged (0, 9, 9, or 8% charged residues) central domain 2 of variable length, which can contain repeats; and (iii) the C-terminal highly charged domain 3 (31, 36, 34, or 36% charged residues), which has a similar length in FWPV, vaccinia virus, and variola virus but is twice as long in molluscum contagiosum virus. Only weak repeats were detected in vaccinia virus and variola virus p39 by a dot matrix analysis with a low stringency, and they were localized mostly in domain 2. Homologies between the FWPV protein and the vaccinia virus protein were also detected in that domain (using the DNA Strider protein dot matrix program, stringency 7, window 23 [C. Marck]). A short sequence (around aa 95) in p39 matched the 39K repeats due to the large number of serine residues. A longer fragment (around aa 145-170) of p39 matched two fragments of the 39K protein adjacent to the repeat region (around aa 80 to 110 and aa 220 to 240). The first of these fragments corresponds to the homologous fragment detected with FASTA. A 10-aa stretch near the N terminus of domain 2 has been deleted in the variola virus protein in comparison with the vaccinia virus protein.

Fractionation of purified virus with detergents.

The vaccinia p39 had been described as a core protein by Maa and Esteban (20), who treated purified virus with Triton X-100 to remove the envelopes, whereas the FWPV 39K protein was described as an envelope protein by Binns et al. (2), who used DOC instead of Triton X-100 and isolated the 39K protein in the soluble fraction. We repeated this experiment with purified FWPV, using Triton X-100, NP-40, and DOC. When FWPV was treated with 2% Triton X-100 or 2% NP-40, the 39K protein remained associated with the core, but when 0.2% DOC was used, the protein was dissociated from the core (Fig. 3).

FIG. 3.

Fractionation of purified virus by detergent treatment. Purified virus was treated with 2% Triton X-100 (lanes 1 to 4) or 0.2% DOC (lanes 5 to 8) in the presence (lanes 1, 2, 5, and 6) or absence (lanes 3, 4, 7, and 8) of dithiothreitol and centrifuged for 30 min at 200,000 × g to separate the core fraction (lanes 2, 4, 6, and 8) from the solubilized envelope fraction (lanes 1, 3, 5, and 7). Western blots were revealed with the GB9 MAb.

Immunolocalization of the 39K protein.

The morphogenesis of FWPV in CEFs is similar to the maturation process of vaccinia virus, at least for the first steps. Viral factories surrounded by crescent shaped structures appear as soon as 12 h postinfection. A few immature and mature intracellular particles can also be observed. The 39K protein, detected by immunogold labelling with the GB9 MAb, was present first in the material within the viral factories and within the viroplasm of the immature particles (Fig. 4). During the maturation of these particles into mature intracellular particles, a core structure was formed by a mechanism that remains to be understood. In these mature intracellular particles, the 39K protein had lost its central localization and was present mainly between the core and the viral membranes, as described for vaccinia virus (7).

FIG. 4.

Immunolabelling of the 39K protein detected by the GB9 MAb. CEFs were infected with FP9 (A to D) or with 1-3 R1, the repeat deletion mutant (E and F), and fixed 66 h postinfection in 1% paraformaldehyde–0.5% glutaraldehyde. The viral factory is clearly labelled as well as the viroplasm of the immature particle in formation (A). In the intracellular mature virus, the intracellular enveloped virus, and the extracellular enveloped virus, the 39K protein appears on the periphery of the core (B, C, and D respectively). The deleted protein showed the same location as the wild-type protein in the immature particles (E), the intracellular mature particles (E), and the extracellular particles (F).

Immunogenicity of the repeats.

To determine which part of the FWPV 39K protein was responsible for its high antigenicity, different fragments of the protein, represented in Fig. 1C, were expressed in E. coli and analyzed by Western blotting with MAbs and several chicken hyperimmune sera. Two MAbs (8F3 and GG1) reacted with the middle fragment (M), containing the repeats, and with all combinations of this fragment with the N-terminal (N) or C-terminal (C) fragments (i.e., NM, MC, and NMC [Fig. 2C]). Two other MAbs (3D9 and GB9) reacted with fragment C alone or as part of the larger fragments MC and NMC (Fig. 2B and D). Mild digestion of the 39K protein with trypsin, which removes charged domains 1 and 3, showed that the epitope recognized by MAb GB9 was located on the trypsin-resistant N-terminal part of the C fragment (Fig. 1) whereas the epitope recognized by MAb 3D9 was lost after digestion and was therefore located on the charged C-terminal part of fragment C (data not shown).

All polyclonal chicken sera reacted with the large fragments (NM, MC, and NMC) and with fragment M (Fig. 2E), but only one (serum B62) of nine also reacted strongly with fragment C (Fig. 2F).

Repeats are also present in the 39K protein from other FWPV strains.

To determine whether the highly antigenic repeat region of the 39K protein was a common feature in avipoxviruses, several strains of FWPV as well as other avipoxviruses were tested by Western blotting with MAb 8F3, which reacts with the repeat region. All FWPV strains tested were recognized by this MAb, indicating that all strains contain similar repeats. However, the observed molecular masses of the proteins differed: 42 kDa in HP-1, HP-200, FP9, and Poxine but only 36 kDa in Websters and Chick-N-Pox (Fig. 5A). None of the other avipoxviruses were recognized by MAb 8F3, but MAb GB9 also recognized a 36-kDa protein of pigeonpox virus (data not shown). To determine which part of the protein was responsible for the difference, each strain was tested by PCR with several combinations of primers (PR 2-6, PR 2-4, PR 8-9, and PR 4-6). The results indicated that only the central part of the protein differed in size between FWPV strains (Fig. 5B). To determine whether this difference was due to the length or to the number of repeats, PCR analysis was performed with a forward primer complementary to the sequence of a conserved repeat (PR19) and a reverse primer (PR9) corresponding to the junction between the repeat region and the C terminal fragment (Fig. 1B). With HP-200, FP9, and Poxine, a ladder of seven bands was obtained, whereas with Websters and Chick-N-Pox, only four were amplified (Fig. 5C). The repeat region of these viruses was further characterized by sequencing. All the sequences were shown to be very highly conserved. The Poxine nucleotide sequence was identical to FP9. In both Chick-N-Pox and Websters, three entire conserved repeats were missing in comparison with FP9, but the rest of the sequence of the Websters repeat region, including the very diverged repeats, was identical to FP9.

FIG. 5.

(A) Variability of size of the 39K protein in different strains of FWPV determined by Western blotting with the 8F3 MAb: HP-1 (lane 1), HP-200 (lane 2), FP9 (lane 3), Poxine (lane 4), and Websters (lane 5). (B) Amplification of the repeat region by PCR with primers PR8 and PR9 in different strains of FWPV: HP-200 (lane 1), FP9 (lane 2), Poxine (lane 3), Websters (lane 4), and Chick-N-Pox (lane 5). (C) Determination of the number of conserved repeats in the 39K protein of different strains of FWPV determined by PCR with primers PR9 and PR19. The different strains of FWPV are as in panel B.

Generation of deletion mutants.

To determine whether the FWPV 39K protein is essential for virus replication and whether the repeat region is essential for the function of the protein, two kinds of deletion mutants were constructed by the transient dominant-selection method described by Falkner and Moss (11). In the first construct (pDSM1), almost all of the 39K coding sequence was deleted (Fig. 6A). Recombinant viruses containing copies of the wild-type gene, the deleted gene, and the gpt gene were selected by plaque purification in the presence of mycophenolic acid. gpt-negative clones resulting from a second crossover were screened by PCR with primers MASH48 and MASH49. The 11 gpt-negative clones obtained contained a full-length copy of the gene (Fig. 6B), and no viable deletion mutants were obtained. We needed to check whether nonviability of the deletion mutant occurred because of the defect in the 39K gene or because the expression of the flanking genes was compromised. The pEFL29 plasmid, containing the gene coding for β-galactosidase expressed under the control of the FWPV p4b promoter (sequence including 262 nucleotides upstream of the p4b ATG initiation codon), was used to check the integrity of the promoter after digestion with SpeI (which cleaves 62 nucleotides upstream of the initiation codon of the 4b protein). The two plasmids (pEFL29 and ΔpEFL29) were compared for the level of expression of β-galactosidase in a transient-expression assay. The results showed that the deleted plasmid was at least as efficient as the original plasmid. Therefore, the deletion did not inhibit the activity of the p4b promoter. This is consistent with an analysis of vaccinia virus late promoters (8), which showed that they extend no more than 40 bp upstream of the TAAAT consensus (where the first A is the start of transcription). As with many late poxvirus genes, the initiator codon of p4b overlaps the TAAAT element. The bidirectional promoter for the A5R homolog (early/late) and the 39K protein (late) has been characterized (15), and the 39K deletion does not encroach upon this element.

Mutants with the repeat region of the 39K protein deleted were generated by the same procedure with the pDSM2A and pDSM2B plasmids. From 22 gpt-negative clones, 3 deletion mutants were identified by PCR (Fig. 6C). Two contained adapter A (clones 1-4 R1 and 1-3 R1), whereas the third contained adapter B (clone 8-3 R1). To check whether the truncated protein was expressed, lysates of cells infected with the three clones were analyzed by Western blotting with MAb 8F3 directed against the central part of the protein and MAb GB9 directed against the C-terminal part of the protein. In each case, a gpt⁺ intermediate was analyzed in parallel as a control. No protein was detected with MAb 8F3 in the deletion mutants, whereas the whole 39K protein was present in the gpt⁺ positive control (data not shown). When MAb GB9 was used, however, full-length and repeat deletion proteins could be detected in the gpt⁺ controls but only deleted protein was expressed in the mutants (Fig. 6D).

We also checked whether the repeat deletion protein was present in viral particles by Western blotting with purified virus as the antigen and by immunoelectron microscopy. The repeat deletion protein was present in the virion (Fig. 6D, lane 12) and showed the same repartition between the core or the envelope fractions when purified virions were treated with Triton X-100, NP-40, and DOC as described above for the wild-type virus (data not shown). It also showed the same localization as the full-length protein in the viral factories and in the mature and immature viral particles as determined by immunogold labelling (Fig. 4).

Stability of the deletion.

The deletion mutant was passaged 12 times in CEFs. No difference in the molecular mass was observed (Fig. 6D, lane 10).

Growth curves.

Upon titer determination of the repeat deletion mutants, we observed that the plaques were slightly smaller than the plaques formed by the wild-type virus. To determine whether the virus replication was impaired by the deletion of the repeats, a single-step growth curve experiment was performed with the 1-3 R1 deletion mutant. The results in Fig. 7 show that the replication of the deletion mutant was only slightly reduced in comparison with that of the wild-type virus. The replication rate of the mutant was no more than twofold lower; the lower yield was due to a lower level of input virus.

FIG. 7.

Single-step growth curve experiment. CEFs were infected in triplicate with the repeat deletion mutant and with the wild-type virus. The supernatant was harvested, and the cells were freeze-thawed at various times postinfection. Intracellular virus (A) and extracellular viruses (B) were subjected to titer determination. The curves were drawn by using the mean of the triplicate sample values. The error bars show the standard deviation.

DISCUSSION

Genes encoding the FWPV 39K protein and vaccinia virus p39 occupy the same relative genomic position in their respective genomes. When purified FWPV was treated with Triton X-100 or NP-40, the 39K protein remained associated with the core, as shown for p39 in vaccinia virus by Maa and Esteban (20), but was dissociated from the core by the stronger detergent DOC, as previously shown by Binns et al. (2). These results suggest, therefore, that the 39K protein is a core protein that is most probably present on the surface of the core and therefore sensitive to denaturation by DOC. Cudmore et al. (7) demonstrated that vaccinia virus p39 can be detected on the surface of isolated cores by immunoelectron microscopy. Both the FWPV and vaccinia virus proteins also share the same localization during morphogenesis.

The 39K core protein of FWPV appears to be an essential protein, since we have been unable to isolate deletion mutants. In making the deletions, we were careful not to disrupt the regulatory sequences of the flanking genes, based on published work for the upstream A5R homolog and on our own promoter assays for the downstream p4b. We cannot, at this stage, exclude the possibility that an unknown, essential regulatory element was inadvertently deleted (although ongoing work to express the 39K protein inducibly, to investigate the null phenotype, should settle that question). The main feature of the 39K protein, like the molluscum contagiosum virus homolog, is the presence of a repeat region spanning more than one-third of its length. Proteins containing repeats are widespread in prokaryotes and eukaryotes. The repeat region of these proteins often plays an essential role: for instance, any mutation or deletion of the tetratrico peptide repeats always results in loss of the function of the protein (16). A number of proteins containing repeats are involved in DNA-protein interactions, protein-protein interactions, or both (13). As a core protein, the 39K protein is likely to interact with itself, with other structural proteins, or with DNA. To check whether the repeats are essential for the function of the 39K protein, we constructed and isolated deletion mutants showing that the repeats are nonessential. The central one-third of the protein (containing the repeats) can be deleted without inhibiting the function of this essential protein or significantly reducing the production of intra- or extracellular virus. The morphology of the virions, visualized by electron microscopy (Fig. 4), was not affected by the deletion. Furthermore, the repeat deletion seems to be stable, since no change in the molecular mass of the protein was observed after 12 passages. It seems, therefore, that deletion of more than one-third of the 39K protein does not impair core formation, and hence the length of the protein is not crucial. This would imply that the essential domains of the protein, probably involved in interactions with other macromolecules, would be situated at the extremities of the protein, which are highly charged domains. That the equivalent proteins in vaccinia virus, variola virus, and molluscum contagiosum virus also contain two conserved charged domains of similar size at each end (except the C-terminal domain of molluscum contagiosum virus) supports the hypothesis that these regions play a functional role. The repeat region of FWPV, which is predicted to be very flexible, could play a role as a linker between the two highly charged, terminal domains in a model where the distance between the domains, or the overall length of the protein, would not be important. This is supported by the different numbers of repeats present in different strains of FWPV. The difference in the size of the 39K protein observed in FWPV strains has also been found in orthopoxviruses (20). The molecular mass of vaccinia virus p39, which does not contain obvious repeats, can vary in size during virus persistence. Maa and Esteban (20) isolated mutants in which the protein had been reduced in molecular mass by 2 kDa after 101 passages in cell culture, as well as mutants in which it had gained about 2 kDa in molecular mass after 65 passages, and suggested that the equivalent cowpox virus protein, with a molecular mass of 41 kDa, might have occurred by mutational changes during evolution. It would be interesting to know in which part of the proteins these mutations are located. It is noteworthy that the variola virus (Bangladesh 1975) protein, with 87.9% amino acid sequence identity to the vaccinia virus (WR) protein, has, relative to the vaccinia virus protein, a 10-aa deletion within the central domain near the junction with the N-terminal highly charged domain (aa 88 to 98 of vaccinia virus p39) (Fig. 1A).

In African swine fever virus, there is a suggested correlation between the number of repeats in j13L and virulence (31). Such a correlation is not evident for the FWPV 39K protein, which has the same number of repeats in virulent HP-1 and attenuated FP9 strains.

The FWPV 39K protein is one of the most antigenic virion proteins detected by Western blotting with chicken hyperimmune sera, and we showed that the repeats are at least partly responsible for the immunodominance: the central part of the protein, containing the repeats, was recognized by two MAbs and by all polyclonal chicken sera tested. The C-terminal one-third of the protein was also recognized by two MAbs but only by one polyclonal serum, whereas the N-terminal one-third of the protein, predicted to be the most antigenic part of the protein by the Genetics Computer Group PEPTIDESTRUCTURE program, was not recognized by any MAb or polyclonal serum. These results therefore suggest that the central part of the protein contributes to the high immunogenicity of the protein but that the C-terminal one-third of the protein also contains important linear epitopes. The situation is slightly different in the Plasmodium circumsporozoite protein (21), where all MAbs raised against the sporozoite recognize the repeat epitope and most or all antisporozoite antibodies in polyclonal sera recognize the repeats (25).

The repeats are present in several FWPV strains, their sequence being very well conserved, and the number of repeats present in the pathogenic strain HP-1 has been maintained over the 400 passages performed to derive the attenuated FP9 strain, suggesting that the presence of these repeats might confer an advantage to the virus. Although the small difference in replication rate between the wild type and the repeat-deletion mutant does not apparently confer strong selection pressure on the mutant in vitro, it may do so in vivo in the face of host immune and nonimmune responses.

The repeats might also play an important role in vivo by modulating the immune response. Schofield (28) suggested that the repeats present in the circumsporozoite protein of Plasmodium could deviate the immune response from proteins that are more critical for parasite survival. This hypothesis could also be applied to the repeat region of the FWPV 39K protein, which elicits a nonneutralizing immune response (data not shown).

Little is known about the cellular immune response against FWPV. The repeat region, clearly immunodominant for the humoral response, could also modulate the cellular response and, for instance, interfere with the presentation of the major histocompatibility complex class I (MHC-I) antigens, as does EBNA-1. Levitskaya et al. (18) suggested that the presence of the repeats within EBNA-1 would generate a signal which either prevents processing or sequesters the processing products to a cellular compartment which is inaccessible to MHC- I-restricted presentation. We are currently investigating whether the 39K protein interferes with the MHC-I presentation.

The high immunogenicity of the repeat region is probably due in part to the repetition of the epitope and to its primary structure, but it also proves that this highly expressed protein, and particularly this part of the protein, is well presented to the immune system. In this context, it would be interesting to know more about the conformation of these repeats. The results could perhaps also be applicable to a number of other repeat motifs showing a compatible periodicity, such as those in the P. falciparum circumsporozoite protein (4 aa), a proline-rich protein of wheat, a similar protein of the Paramecium bursaria chlorella virus 1 (4 aa), and the leucine-rich repeats found in the human platelet receptor glycoprotein and the leucine-rich human glycoprotein (24 aa).