Variation in the Nucleotide Sequence of ... - Journal of Virology

JOURNAL OF VIROLOGY, Nov. 2000, p. 10766–10777 0022-538X/00/$04.00⫹0 Copyright © 2000, American Society for Microbiology. All Rights Reserved.

Vol. 74, No. 22

Variation in the Nucleotide Sequence of Cottontail Rabbit Papillomavirus a and b Subtypes Affects Wart Regression and Malignant Transformation and Level of Viral Replication in Domestic Rabbits ´ RO ˆ ME SALMON, MATHIEU NONNENMACHER, SANDRINE CAZE ´ , PATRICIA FLAMANT, JE ´ RARD ORTH, AND FRANC ODILE CROISSANT, GE ¸ OISE BREITBURD* Unite´ Mixte Institut Pasteur/INSERM U.190, Unite´ des Papillomavirus, Institut Pasteur, 75724 Paris Cedex 15, France Received 6 June 2000/Accepted 18 August 2000

We previously reported the partial characterization of two cottontail rabbit papillomavirus (CRPV) subtypes with strikingly divergent E6 and E7 oncoproteins. We report now the complete nucleotide sequences of these subtypes, referred to as CRPVa4 (7,868 nucleotides) and CRPVb (7,867 nucleotides). The CRPVa4 and CRPVb genomes differed at 238 (3%) nucleotide positions, whereas CRPVa4 and the prototype CRPV differed by only 5 nucleotides. The most variable region (7% nucleotide divergence) included the long regulatory region (LRR) and the E6 and E7 genes. A mutation in the stop codon resulted in an 8-amino-acid-longer CRPVb E4 protein, and a nucleotide deletion reduced the coding capacity of the E5 gene from 101 to 25 amino acids. In domestic rabbits homozygous for a specific haplotype of the DRA and DQA genes of the major histocompatibility complex, warts induced by CRPVb DNA or a chimeric genome containing the CRPVb LRR/E6/E7 region showed an early regression, whereas warts induced by CRPVa4 or a chimeric genome containing the CRPVa4 LRR/ E6/E7 region persisted and evolved into carcinomas. In contrast, most CRPVa, CRPVb, and chimeric CRPV DNA-induced warts showed no early regression in rabbits homozygous for another DRA-DQA haplotype. Little, if any, viral replication is usually observed in domestic rabbit warts. When warts induced by CRPVa and CRPVb virions and DNA were compared, the number of cells positive for viral DNA or capsid antigens was found to be greater by 1 order of magnitude for specimens induced by CRPVb. Thus, both sequence variation in the LRR/E6/E7 region and the genetic constitution of the host influence the expression of the oncogenic potential of CRPV. Furthermore, intratype variation may overcome to some extent the host restriction of CRPV replication in domestic rabbits. The Shope papillomavirus or cottontail rabbit papillomavirus (CRPV) induces cutaneous papillomas (warts) in cottontail rabbits under natural conditions and in domestic rabbits under experimental conditions (51). Systemic regression of warts occurs in a variable proportion of rabbits (4, 30, 54) as a consequence of a specific cell-mediated immune response (15, 31). Persistent warts may progress into invasive carcinomas (46, 54). The natural history of CRPV-induced warts thus mimics that of cervical intraepithelial neoplasia associated with oncogenic human papillomaviruses (HPVs) (9). The organization of the CRPV genome is unique among papillomaviruses by the greater size of the E6 open reading frame (ORF) (21). Long E6 (LE6) and short E6 (SE6) proteins are translated in the same reading frame from transcripts initiated at two distinct promoters (2, 12). The two major transcripts detected in warts and carcinomas encode the SE6 and the E7 proteins (12, 19, 35). Like DNA extracted from warts (29), cloned CRPV DNA is infectious for domestic rabbits (20, 36). This has provided a model system to demonstrate that the E5 and L2 ORFs are dispensable for papilloma induction (7, 34). Both wart evolution and the level of CRPV replication in warts have been shown or were suspected to depend on the genetic constitution of the host (3, 23, 30, 54) and on the * Corresponding author. Mailing address: Unite´ des Papillomavirus, Institut Pasteur, 25 rue du Docteur Roux, F-75724 Paris Cedex 15, France. Phone: 33 1 45 68 87 47. Fax: 33 1 45 68 89 66. E-mail: fbreit @pasteur.fr.

genetic variability of the virus (18, 44, 46, 48, 52). Wart regression occurs in less than 10% of cottontail rabbits, and its frequency varies between 10 and 70% in domestic rabbits (9, 54). Progression into carcinomas is observed in approximately 25% of cottontail rabbits and in up to 75% of domestic rabbits with persistent warts (46, 54). Both regression and malignant conversion of domestic rabbit warts have been reported to be linked to class II DRA and DQA genes of the major histocompatibility complex (MHC), pointing to immunogenetic control of wart evolution (9, 23, 24). Warts of naturally infected cottontail rabbits usually contain large amounts of virions, in spite of a great variation in virus content observed among rabbits (3). In contrast, little, if any, infectious virus is recovered from domestic rabbit warts (3, 18, 51, 52), and vegetative viral DNA replication, late transcripts encoding L1 and L2 capsid proteins, and capsid antigens are seldom detected (35, 37, 41, 62). The first evidence for a biological variation of CRPV has been the isolation of strains recoverable from domestic rabbit warts (18, 50, 52). The existence of cancer-producing variants was proposed to account for tumor progression (44) but could not be substantiated (43, 45, 53). Our recent observation that rabbits homozygous for a specific MHC class II DRA-DQA haplotype displayed an unusual mode of wart evolution (partial regression), characterized by the persistence of a few warts only, led us to suspect a genetic heterogeneity of CRPV (9, 48; J. Salmon, N. Chanteloup, F. Viard, P. Coudert, O. Croissant, G. Orth, and F. Breitburd, unpublished data). Available se-

10766

VOL. 74, 2000

CRPV GENETIC VARIABILITY AND BIOLOGICAL PROPERTIES

10767

quence data had disclosed only a low (about 0.5%) intratypic variability of CRPV (6, 20, 26, 57). Surprisingly, by cloning and partially sequencing the CRPV genomes present in our isolate obtained from pooled Kansas cottontail rabbit warts, we disclosed the existence of two CRPV strains, a variant of the CRPV prototype (CRPVa) and a CRPV subtype (CRPVb), characterized by strikingly divergent E6 and E7 oncoproteins (48). The aim of this study has been first to fully characterize the genetic divergence between the CRPVa and CRPVb strains by establishing the complete nucleotide sequences of the two cloned genomes. We next investigated whether the CRPV variability influenced wart evolution and whether this involved the most variable genomic region containing the long regulatory region (LRR) and the E6 and E7 ORFs. The approach was to study the evolution of warts induced by CRPVa, CRPVb, and chimeric genomes in rabbits homozygous for the DRA-DQA haplotype associated with partial regression. Our last goal was to find out whether the two CRPV strains differed in their levels of viral replication in domestic rabbit warts by analyzing wart sections for vegetative viral DNA replication, transcription of the late L1 gene, and synthesis of capsid proteins. MATERIALS AND METHODS Plasmids, virus, and rabbits. The CRPVa4 and CRPVb recombinant plasmids were described previously (48). The reference CRPVa1 clone (21) was a gift from M. Yaniv, Institut Pasteur, Paris, France. The pCMV-␤-gal plasmid was made available to us by E. Meurs, Institut Pasteur. The CRPVa-L1 recombinant plasmid was constructed by inserting a CRPVa fragment from nucleotides (nt) 5865 to 7347 containing the L1 ORF into a pBluescript vector (Stratagene) at the SmaI restriction site. The CRPV suspension prepared from pooled Kansas cottontail rabbit warts was described previously (48). New Zealand White rabbits homozygous for the MHC DRA.D-DQA.B (PR) haplotype or the DRA.CDQA.G (P) haplotype, as defined by a restriction fragment length polymorphism (RFLP) (23), were provided by P. Coudert, F. Viard, and N. Chanteloup (Centre de Recherche INRA, Tours-Nouzilly, France). DNA sequencing. Cloned CRPVa and CRPVb DNAs (21, 48), CRPVb DNA extracted from the viral suspension (48), and CRPV genomes present in total DNA extracted from warts associated with CRPVa or CRPVb were used to establish the sequences of the 4.5-kb region between the 3⬘ end of the E7 ORF and the 5⬘ end of the L1 ORF (E1-L2 region) and the 0.6-kb fragment between the 3⬘ end of the L1 ORF and the 5⬘ end of the E6 ORF spanning the LRR. The strategy was that previously reported for sequencing the E6, E7, and L1 ORFs of the same CRPV DNAs (48). Briefly, nine overlapping fragments encompassing the E1-L2 region and three fragments covering the LRR were amplified by the PCR technique (47) using primers deduced from the published CRPV nucleotide sequence (21). Universal forward and reverse M13 sequences were added to the 5⬘ ends of the forward and reverse primers, respectively, to allow direct sequencing of amplification products. Amplification experiments were performed with the Expand high-fidelity PCR system (Boehringer Mannheim) using an automated thermal cycler (GeneAmp 9600; Perkin-Elmer). PCR products were purified on Centricon-100 columns (Amicon) and sequenced on both strands by the dideoxynucleotide termination method (49), using fluorescent forward and reverse M13 primers and the ABI Prism Dye-Primer Cycle sequencing kit (Applied Biosystems). Labeled DNA fragments were purified in accordance with the manufacturer’s instruction, separated in a 4% acrylamide gel under denaturing conditions, and analyzed with an ABI Prism 377 automated sequence analyzer (Applied Biosystems). Chimeric CRPV genomes. The construction of chimeric CRPV genomes is depicted in Fig. 1. Briefly, a subgenomic 2.7-kb EspI-SacI fragment containing the LRR and the E6 and E7 ORFs was amplified from the cloned CRPVa4 and CRPVb DNAs (48) by using a forward primer located between nt 7027 and 7052 for CRPVa and nt 7020 and 7045 for CRPVb and designed to introduce a ClaI site in the 5⬘ position (5⬘-CCATCGATGGGCACCCAACAATCATAGATAG ATAATTGGC-3⬘). The nucleotide positions refer to the sequences reported in this paper. The reverse primer was chosen from the pBluescript plasmid sequence downstream of the SacI cloning site. Amplification experiments were carried out with the Expand high-fidelity PCR system (Boehringer Mannheim) using an automated thermal cycler (GeneAmp 2400; Perkin-Elmer). PCR products were inserted into a pBluescript plasmid between the ClaI and SacI restriction sites. The recombinant plasmids were double digested with SpeI and EcoNI endonucleases yielding a 1.9-kb fragment containing the variable LRR/E6/E7 region, extending from nt 7420 to 1439 for CRPVa (va) and from nt 7413 to 1431 for CRPVb (vb), and a 3.7-kb fragment containing the flanking CRPV E1 and L1 ORF sequences and the plasmid. The fragments were separated by agarose gel electrophoresis and purified using a Geneclean II kit (Bio 101, Inc.). Heterolo-

FIG. 1. Strategy for constructing chimeric genomes from cloned CRPVa4 and CRPVb DNAs. The subgenomic 2.7-kb fragments of CRPVa and CRPVb encompassing the LRR/E6/E7 region were amplified by PCR using a 5⬘ primer introducing a ClaI restriction site and cloned into a pBluescript SK(⫺) [pBSK(⫺)] plasmid. The LRR/E6/E7 regions of CRPVa (va) and CRPVb (vb) were double digested by SpeI and EcoNI restriction enzymes (arrowheads) and exchanged to generate chimeric subclones. Chimeric subclones were double digested by EspI and SacI restriction enzymes (arrowheads) and ligated to the 5.2-kb SacI-EspI E1-L1 fragment corresponding to the E1 and L1 ORF sequences flanking the va or vb insert. The generated 7.9-kb CRPVa-vb or CRPVb-va full-length circular recombinant genomes were digested by SacI restriction endonuclease and cloned into a pBluescript SK(⫺) plasmid. Open and shaded boxes, CRPVa and CRPVb sequences, respectively; lines, plasmid sequences. The different ORFs and the LRR are represented. The positions of the last and first nucleotides of the CRPVa and CRPVb sequences are given. The localization of the restriction enzyme cleavage sites and their nucleotide positions on the CRPVa sequence (21) are indicated. gous fragments were ligated to obtain chimeric recombinant plasmids, which were verified by sequencing. To obtain full-length chimeric CRPVa-vb and CRPVb-va genomes, the 2.7-kb EspI-SacI chimeric fragment was excised and ligated to the 5.2-kb SacI-EspI E1-L1 fragment of the CRPV subtype corresponding to the sequences flanking the va or vb insert. After insertion into a pBluescript plasmid at the SacI site, the recombinant plasmids were amplified in Escherichia coli XLI-Blue MRF⬘ (Stratagene) and purified using the Nucleobond AX kit (PC 500; Macherey-Nagel). Inoculation of CRPV DNAs and virions. Cloned CRPV and chimeric genomes were excised from recombinant plasmids by SacI digestion, and linear viral DNA was purified by sedimentation through a 5 to 21% sucrose gradient in the presence of ethidium bromide (5 ␮g/ml) (20). After butanol extraction and dialysis (Slide-A-Lyzer cassettes; Pierce), CRPV DNAs were recircularized by self-ligation overnight at 4°C, at a low concentration of DNA (5 ␮g/ml), in the presence of T4 DNA ligase (0.5 Weiss unit per ␮g) and 1 mM ATP (Pharmacia Biotech.), that is, in conditions minimizing the formation of high-molecular-

10768

SALMON ET AL.

weight oligomers. Prior to inoculation of DNAs, the backs and flanks of the rabbits were shaved and treated twice every other day with a mixture of turpentine and acetone (1:1) to enhance the sensitivity of the bioassay (17, 20). Rabbits were anesthetized by intramuscular injection of 25 mg of ketamine (Imalge`ne 500; Rho ˆne Me´rieux, Lyon, France)/kg of body weight and 2 mg of xylazine (Rompun; 2%; Bayer)/kg. In the first experiment, CRPV genomes (10 ␮g per site) were inoculated by intradermal injection followed by repeated punctures of the sites of injection (20) at four sites on the right flanks (CRPVb) and the left flanks (CRPVa) of 5 P and 5 PR rabbits. In a second experiment, CRPVa, CRPVb, and chimeric genomes were targeted to rabbit skin with the Helios gene gun system (Bio-Rad) in conditions described by Xiao and Brandsma (59). After determining the optimal helium discharge pressure of 350 lb/in2 for the intraepidermal expression of a ␤-galactosidase reporter gene driven by a cytomegalovirus promoter (pCMV-␤-gal plasmid), 1.6-␮m gold particles were coated with viral DNAs (2 ␮g of DNA/mg of particles) as indicated by the manufacturer and were delivered at 1 ␮g per site at three sites per genome to the left flanks (CRPVa, CRPVb-va) and right flanks (CRPVb, CRPVa-vb) of 8 P and 8 PR rabbits. Wart development was photographed weekly for 24 weeks and monthly thereafter, up to 1 year. Warts were also induced by application of a suspension of CRPV virions on a 20-cm2 area of shaved skin abraded with sandpaper (23) on the flanks of 4 PR rabbits (about 4 ⫻ 1010 particles per site) and 6 P rabbits (about 4 ⫻ 109 particles per site). Tissue specimens and CRPV DNA typing. Wart specimens were collected from PR rabbits (17 specimens) and P rabbits (31 specimens) infected with CRPV virions between 4 and 10 weeks after wart outgrowth, and biopsies from a single wart induced by CRPVa or by CRPVb DNA were taken from the same P rabbit 10 weeks after outgrowth. Fourteen warts showing signs of malignant conversion from nine rabbits infected with CRPVa or CRPVb-va DNA were biopsied. Biopsy specimens were fixed in buffered 10% formalin and embedded in paraffin. Serial 5- to 7-␮m-thick sections were prepared for histological examination after hematoxylin and eosin staining and for in situ hybridization and immunohistochemistry experiments. To identify the CRPV subtype in virion-induced warts, DNA was extracted from three to five tissue sections as described previously (56) and the CRPV subtype was determined by amplification of a fragment of 132 (CRPVa) or 150 bp (CRPVb) using oligonucleotide primers flanking the 18-nt insertion present in the CRPVb E6 ORF (48). In situ hybridization and immunohistochemistry. For in situ hybridization, sections mounted on pretreated slides (SuperFrost Plus; Mensel-Glaser) were incubated with proteinase K (20 to 50 ␮g/ml) in 100 mM Tris-HCl–50 mM EDTA, pH 8, for 10 min at 37°C, postfixed in 0.4% paraformaldehyde in diethyl pyrocarbonate (DEPC)-treated phosphate-buffered saline (PBS) for 20 min at 4°C, and washed in PBS containing 0.2% glycine. For DNA-DNA hybridization, excised full-length CRPVa DNA was labeled with digoxigenin-11-dUTP by random priming (High Prime DNA labeling kit; Boehringer Mannheim). Tissue sections were denatured in 2⫻ SSC (1⫻ SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–50% formamide for 10 min at 75°C, and 20 ␮l of the hybridization mixture (50% formamide, 2⫻ SSC, 250 ␮g of salmon sperm DNA/ml) containing 2 to 4 ng of the probe denatured for 5 min at 95°C was applied to the sections, which were covered with a siliconized glass coverslip. Hybridization was carried out overnight at 33.5°C (melting temperature [Tm] ⫺ 25°C). After sections were washed as described previously (5), hybrids were detected using alkaline phosphatase-conjugated antidigoxigenin Fab fragments (Boehringer Mannheim) (1:2,000 dilution). The enzyme activity was revealed by a standard procedure using 5-bromo-4-chloro-3-indolylphosphate as the substrate and nitroblue tetrazolium chloride as the chromogen at pH 9.5. Sections were lightly counterstained with hematoxylin and mounted in Kaiser’s glycerol gelatin (Merck) for examination. For RNA-RNA hybridization, preparation of probes and hybridization procedures were mainly as described by Cox et al. (10), Angerer et al. (1), and Zeltner et al. (62). Briefly, the recombinant CRPVa-L1 pBluescript plasmid was linearized by EcoRV or SacI digestion to generate L1 riboprobes by in vitro transcription from T3 (sense probe) and T7 (antisense probe) promoters in the presence of 35S-UTP (400 ␮Ci/nmol; Amersham). After DNase treatment, the probes (about 1,500 nt) were purified by chromatography on a Sephadex G-50 column (Quick Spin columns; Boehringer Mannheim), reduced in size to about 150 nt by alkaline hydrolysis, ethanol precipitated, and redissolved in DEPCtreated distilled water containing 10 mM dithiothreitol (DTT). Sections were pretreated to block nonspecific sulfur fixation sites (61) and to reduce electrostatic binding to amino groups (1). Prehybridization was performed at 52.5°C for 2 h in a solution containing 0.6 M NaCl, 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, 2.5 ⫻ Denhardt’s solution (0.05% each Ficoll 400, polyvinylpyrrolidone, and bovine serum albumin), 150 ␮g of yeast tRNA/ml, 0.1% sodium dodecyl sulfate, and 50% formamide. This solution was boiled for 3 min and cooled on ice prior to use. Sense (6.5 ⫻ 108 cpm/␮g) and antisense (4.8 ⫻ 108 cpm/␮g) riboprobes were diluted to 2.5 ⫻ 107 cpm/ml in the same solution containing 10% dextran sulfate and heat denatured for 1 min at 100°C. After addition of 10 mM DTT, 20 ␮l of the hybridization mixture was applied to each section and hybridization was carried out for 16 h at 52.5°C (Tm ⫺ 30°C). Posthybridization washes specific for 35 S-labeled probes were performed as described previously (1). Hybrids were detected by autoradiography using Kodak NTB-2 emulsion after a 7- to 10-day exposure time. Sections were counterstained with hematoxylin and observed under bright- or dark-field illumination.

J. VIROL. For immunohistochemistry, sections were deparaffinized and subjected to a microwave treatment, three times for 5 min, in 10 mM citric acid at 750 W. Sections were incubated successively for 30 min at 37°C with normal goat serum (1:100 dilution), a rabbit antiserum (NZ 2604) against disrupted CRPVa viruslike particles (1:100 dilution) (8), and an alkaline phosphatase-conjugated, affinity-purified goat anti-rabbit antibody (1:100 dilution) (Chemicon). After sections were washed in PBS, enzyme activity was revealed as described above. Sections were examined and photographed with a Zeiss Axiophot microscope. Cells positive for viral DNA replication or capsid antigens were numerated in the entire section, and the number of fields required to cover the whole section was determined. The surface of the field was 0.317 mm2 in our conditions of observation using an internal length standard, and the results were expressed as the number of positive cells per square millimeter of section. Means were calculated from the values found for a section of each of the 17 CRPVa- and 31 CRPVbassociated specimens studied. The significance of the differences found was evaluated using Fisher’s exact test and the nonparametric Mann-Whitney test. Nucleotide sequence accession numbers. The complete CRPVa4 and CRPVb nucleotide sequences have been submitted to the EMBL/GenBank/DDBJ nucleotide sequence database and have been assigned accession No. AJ404003 and AJ243287, respectively.

RESULTS Genetic divergence between CRPVa and CRPVb. We recently reported the cloning of the genomes and the nucleotide sequences of the E6/E8, E7, and L1 ORFs of a variant of the prototype CRPV (21), referred to as CRPVa, and of a highly divergent CRPV strain, referred to as CRPVb (48). The two cloned CRPV genomes were obtained from a viral suspension prepared from pooled Kansas cottontail rabbit warts that contained mostly (about 90%) CRPVb (48). We have now established the complete nucleotide sequences of these cloned CRPVa and CRPVb genomes (For accession numbers, see Materials and Methods). The data were confirmed by direct sequencing of overlapping PCR products obtained from the CRPVb DNA extracted from the viral suspension and from the CRPVa and CRPVb genomes present in the total DNA prepared from warts induced by each virus. We also sequenced the reference CRPV DNA clone (21) in view of the two errors previously found in the published sequence of the L1 ORF (48) and identified six additional errors (Table 1). Compared to the corrected sequence, our CRPVa strain showed five nucleotide substitutions, one in the E2 ORF (G42133A, Asp3683Asn) and four shared with the CRPVb strain, three localized in the E1 and L2 ORFs (Fig. 2B) and one previously reported in the 5⬘ untranslated part of the E7 ORF (48). Five sequence variants of CRPVa have been reported so far. Taking into account the chronological order of their description, they are referred to here as CRPVa1 for the sequenced CRPV prototype originating from Kansas cottontail rabbit warts (16, 21), CRPVa2 for the Washington B strain originating from naturally infected Whidbey Island cottontail rabbits (6, 15, 26, 57), CRPVa3 for the isolate harbored by the VX7 transplantable carcinoma (20, 43), CRPVa4 for the isolate further analyzed in this study (48), and CRPVa5 for the CRPV-HE isolate originating from Kansas cottontail rabbit warts (25). The CRPVb genome (7,867 bp) was found to be 1 nt shorter than the CRPVa1 and -a4 genomes (7,868 bp). The nucleotide positions of the CRPVa and CRPVb ORFs are given in Table 2; also given are the coding capacities and the numbers and percentages of variable nucleotides and amino acids, including data previously reported for the E6, E7, E8, and L1 ORFs (48). The organization of the prototype CRPV E1-L2 region showing the overlapping ORFs is depicted in Fig. 2A. The distribution of the synonymous and nonsynonymous mutations found in the CRPVb sequence and the resulting amino acid changes are given in Fig. 2. On the whole, CRPVb differed from CRPVa1 and CRPVa4 by 241 (3.1%) and 238 variable nucleotide positions, respectively; these differences warrant

VOL. 74, 2000


10769

TABLE 1. Corrections to the prototype CRPV sequencea ORF

E1 E2, E4

L1 a b c

Nucleotide position

Base changec

1531 1535 2017 3458 3515 ⫹ 1 3533

C3T T3C A3G T3G ⫹G ⫺A

6888 ⫹ 1b 6922b

⫹T ⫺T

Consequence

Pro573Leu None Asn2193Ser E2: Ile1163Ser; E4: Leu283Val Frameshift in E2 ORF (nt 3514 to nt 3534), Gly135-Ile-Phe-Ile-Ile-Gly-Asn1413Gly135-Tyr-Ile-TyrTyr-Trp-Asp141; Frameshift in E4 ORF (nt 3515 to nt 3532), Val47-Tyr-Leu-Leu-Leu-Gly523Gly47Ile-Phe-Ile-Ile-Gly52 Frameshift (nt 6887 to nt 6921), Gln354-Glu-Ile-Lys-Lys-Thr-His-Gly-Lys-Thr-Val-His3643Asp354-AlaAsn-Gln-Glu-Asp-Pro-Trp-Lys-Asn-Ser-Thr364

As established by Giri et al. (21). As previously reported by Salmon et al. (48). ⫹, addition of; ⫺, deletion of.

considering CRPVa and CRPVb as subtypes (40). The divergence between CRPVa1 and CRPVb involved 176 nt substitutions and 65 nt deletions or insertions. A high rate (80%) of nonsynonymous nucleotide changes characterized the E2 ORF (Fig. 2), as reported earlier for the E6 and E7 ORFs (78.4 and 100%, respectively) (48). A lower rate was observed for the E1

and L2 ORFs (47.6 and 45.9%, respectively), as previously found for the L1 ORF (33%) (48). A nucleotide substitution affecting the stop codon of the E4 ORF and a deletion of 1 bp generating a frameshift in the E5 ORF resulted in different sizes of the corresponding proteins (Fig. 2B), as also reported for the E6 and E7 proteins (48) (Table 2).

FIG. 2. Sequence variations in the E1-L2 regions of CRPVa and CRPVb. (A) Map of point mutations in the E1-L2 region. The organization of the different ORFs (boxes) of CRPVa1 (21) is given. Vertical dashed lines, positions of initiation codons; bars on the solid lines below the ORFs, synonymous (above) and nonsynonymous (beneath) nucleotide substitutions. The first and last nucleotide positions of each ORF are given. Nucleotide insertion (⫹2) and deletion (⫺1) (arrows) and the resulting frameshift (fs), generated (star) or suppressed (star in parentheses) stop codons, and mutations downstream of the E5b stop codon (dashed bars) are represented. (B) Nucleotide and amino acid variations in the E1-L2 region. Nucleotide positions refer to the CRPVa1 sequence (21) as corrected in this paper. Arrows, beginning of the ORFs. Nucleotide substitutions in CRPVb are given beneath the corresponding CRPVa1 nucleotides. c, mutations common to CRPVb and CRPVa4. Conserved nucleotides (dashes) are given for the regions flanking the variable E4 stop codons (shaded), the frameshift and the premature E5 stop codon (shaded) generated by the T4342 deletion, and the CA insertion in the E5-L2 overlap. Variable nucleotide positions within the same codon are underlined. Amino acid changes are given beneath the corresponding nucleotides, and stop codons suppressed or generated (shaded) by mutations are indicated (stars).

10770

SALMON ET AL.

J. VIROL.

TABLE 2. Nucleotide positions and coding capacity of CRPVa and CRPVb ORFs Nucleotide positiona of:

No. (%) of variable nucleotides

No. of amino acidsb

No. (%) of variable amino acidsb

972 949 972 949

56 (6.4)

273 276 176 179

37 (13.4)

323 291

472 440

12 (4.1)

50

3 (6.0)

979 965

1075 1061

1356 1348

21 (5.5)

94 96

12 (12.5)

a b

1356 1342

1362 1354

3167 3159

21 (1.2)

602

9 (1.5)

E2

a b

3088 3080

3112 3104

4281 4273

30 (2.5)

390

19 (4.9)

E4

a b

3332 3324

3377 3369

4012 4028

24 (3.4)

212 220

18 (8.2)

E5

a b

4265 4257

4274 4266

4576 4340

14 (4.5)

101 25

L2

a b

4303 4344

4378 4371

5853 5846

47 (3.0)

492

16 (3.3)

L1c

a b

5780 5773

5828 5821

7342 7335

12 (0.8)

505

4 (0.8)

LRRf

a b

7346 7339

153 121

54 (7.9)

ORF or LRR

CRPV

First nucleotide

First ATG

Last nucleotide

a b a b

100 68

154 122 445 413

E8c

a b

182 150

E7c

a b

E1

E6c LE6d SE6d

43 (8.1)

29 (16.2)

6 (24.0)e

a Data for CRPVa were taken from Giri et al. (21). Nucleotide positions for the CRPVb ORFs were numbered starting at the HpaI site like those for the CRPVa ORFs. b Starting from the first methionine. c Nucleotide and amino acid sequence data for CRPVb were reported previously (48). d Referring to the LE6 and SE6 proteins translated in the same reading frame from transcripts initiated at two distinct promoters (2, 12). e Taking into account only the 25 N-terminal amino acids. f LRR located between L1 stop codon and E6 ATG. One variable position is in the untranslated part of the E6 ORF (48).

The E1 ORF showed little variation (Fig. 2; Table 2). Twenty-four of the 30 variable nucleotides in the E2 ORF, which corresponded to 17 of the 19 amino acid changes in the protein, were located in the central hinge region overlapping the E4 ORF. None of the amino acid changes affected the 88 carboxy-terminal residues within the DNA-binding domain of the transactivating E2 protein (22). Twelve nucleotide substitutions downstream of the splice acceptor site at nucleotide position 3714 in the E4 ORF (12) resulted in five amino acid changes and in the addition of eight amino acid residues to the putative CRPVb E1Ê4 protein, due to the A40153T and G40383A mutations suppressing and generating a stop codon, respectively. The frameshift due to the 1-nt deletion in the E5 ORF (T4342) affected codons 23 to 25 and resulted in a premature TAA termination codon within the conserved polyadenylation signal of CRPV early transcripts (12). The coding capacity of the CRPVb E5 ORF was reduced from 101 to 25 amino acids and 6 (24%) of these residues were found to be variable compared to those of the prototype CRPV. An insertion of 2 nt occurred in the E5 ORF, 30 nt downstream of the 1-nt deletion. Both are located in the untranslated part of the overlapping L2 ORF. Of the 16 variable amino acids in the L2 protein, none was found among the 77 amino-terminal residues and the 49 carboxy-terminal residues (Table 2; Fig. 2B). The LRR beginning after the L1 ORF stop codon (at posi-

tions 7346 and 7339 for CRPVa and CRPVb, respectively) and ending before the first E6 ORF ATG (at positions 153 and 121 for CRPVa and CRPVb, respectively) was found to contain 54 (7.9%) variable nucleotide positions (Table 2, Fig. 3). Computer analysis revealed that the putative recognition sequence for the E1 protein (39) and the eight bona fide binding sites for the E2 protein found in the CRPVa1 (21) and CRPVa4 LRRs were preserved in CRPVb, as well as a number of potential binding motifs for NF1, AP1, Tef-2, and Sp1 cellular transcription factors, which are considered important for the biological properties of papillomaviruses (39). Two T3C transitions at positions 7663 and 7665 created an additional E2 binding site in the central segment of the CRPVb LRR and another transition (A323G) generated a putative binding motif for an ets-related protein (38) 21 bp upstream of the TATA box for the LE6 transcript. The second 32-bp sequence of the direct repeat starting at position 42 was found to be deleted. This reduced to 34 bp the distance between the TATA boxes for the LE6 and SE6 transcripts and eliminated the cap site of the CRPVa LE6 transcripts mapped to this region (12, 55), as well as a putative binding motif for c-myb (27). In addition, a nucleotide transition (C74843T) affected a CACACA sequence (positions 7484 to 7489), which could correspond to the putative promoter for the transcripts encoding capsid proteins (55) (Fig. 3).

VOL. 74, 2000


10771

FIG. 3. Sequence variations in the LRR. The alignment of the region extending from the stop codons of the L1 ORFs to the initiation codons of the LE6 ORFs of CRPVa1 (a) (21) and CRPVb (b) (48) is represented, except for the conserved sequence between nt 7347 and 7426. Dashes, conserved nucleotides. The two 32-bp direct repeats found in the CRPVa LRR are overlined by arrows. The TATA boxes for the LE6 and SE6 transcripts (12) and the putative promoter for the late transcripts (55) (shaded), the cap sites of the late and LE6 mRNAs (12, 55) (bent arrows), and the binding motifs for the viral E1 (dotted box) and E2 (solid boxes) proteins are indicated. Putative binding motifs for Tef-2, NF1, Sp1, AP1 (39), and c-myb (27) cellular transcription factors identified in the CRPVa LRR are overlined by brackets, and an ets-binding motif (38) found in the CRPVb LRR is underlined by a bracket. Start and stop codons are in boldface.

CRPV variation and genetic constitution of host as related to wart evolution. Our aim was to find out whether CRPV genetic variability influenced wart evolution towards regression or persistence and malignant conversion. Our approach has been to use rabbits homozygous for two MHC class II DRADQA RFLP haplotypes, characterized by a highly divergent antigen-binding ␣1 domain of DQ␣ molecules (23, 24). Rabbits homozygous for each haplotype had been found previously to show distinct modes of wart evolution when inoculated with a viral suspension containing the two CRPV subtypes (48; Salmon et al., unpublished data). The DRA.D-DQA.B haplotype was found to be preferentially associated with the early regression of most warts and the persistence of a few warts in the majority of the rabbits (partial regression [PR] haplotype), whereas the DRA.C-DQA.G haplotype was found to be preferentially linked to wart persistence (P haplotype) (9, 48; Salmon et al., unpublished data). The cloned CRPVa4 and CRPVb genomes were excised from recombinant plasmids, recircularized, and inoculated intradermally at four sites to each of five rabbits homozygous for

the PR or the P haplotype. Warts developed in all rabbits at most (50 to 87.5%) sites (Table 3), usually within 2 to 6 weeks (Fig. 4A and B). Several modes of wart evolution were observed, as previously described (8, 23). Regression occurred most often early, within 1 to 8 weeks after wart outgrowth. Warts present at 8 weeks either persisted for months, remaining unchanged or converting to malignancy, or were progressively lost. The kinetics of wart evolution for 18 weeks after inoculation is depicted in Fig. 4A and B, and the percentages of positive sites showing early regression, persistence for 6 months after inoculation, or malignant transformation are given in Table 3. In PR rabbits, all warts induced by CRPVb showed early regression (Fig. 4A; Table 3), whereas about 80 and 70% of the sites were still positive for CRPVa 18 weeks and 6 months, respectively, after inoculation (Fig. 4A; Table 3). In P rabbits, most of the sites positive for CRPVa and CRPVb showed no early regression (Table 3; Fig. 4B) and about one-half and one-third of CRPVa- and CRPVb-positive sites, respectively, were still positive 6 months after infection. Wart regression or persistence as related to the CRPV subtype

TABLE 3. Biologic properties of CRPVa, CRPVb, and chimeric genomes No. (%) of positive sitesc

No. of: Expte

Rabbit haplotypea

1

DRA.D-DQA.B (PR)

2

a

CRPV DNAb

Positive rabbits/ inoculated rabbits

Positive sites/ inoculated sites (%)

After early regression

At 26 weeks

a b

4/4d 4/4

14/16 (87.5) 10/16 (62.5)

11 (78.6) 0

10 (71.4) 0

1 (7.1) 0

DRA.C-DQA.G (P)

a b

5/5 5/5

11/20 (55) 10/20 (50)

8 (72.7) 7 (70)

6 (54.5) 3 (30)

1 (9.1) 0

DRA.D-DQA.B (PR)

a b b-va a-vb

8/8 6/8 8/8 8/8

15/24 (62.5) 11/24 (45.8) 12/24 (50) 18/24 (75)

13 (86.7) 1 (9.1) 9 (75) 0

11 (73.3) 0 9 (75) 0

6 (40) 0 5 (41.7) 0

DRA.C-DQA.G (P)

a b b-va a-vb

8/8 5/8 6/8 6/8

18/24 (75) 12/24 (50) 19/24 (79.2) 17/24 (70.8)

11 (61.1) 7 (58.3) 6 (31.6) 7 (41.2)

4 (22.2) 0 3 (15.8) 3 (17.6)

0 0 1 (5.3) 0

As determined by RFLP analysis of rabbit leukocyte antigen DRA and DQA genes (23). Chimeric genomes (b-va, a-vb) were constructed by exchanging the LRR/E6/E7 region of CRPVa (va) or CRPVb (vb). Early regression took place within 8 weeks after wart outgrowth. Cancers developed between 26 and 47 weeks after inoculation. d One rabbit died before wart outgrowth. e Experiment 1, 10 ␮g of DNA intradermally; experiment 2, 1 ␮g of DNA by gene gun. b c

With cancer

10772

SALMON ET AL.

J. VIROL.

FIG. 4. Course of the disease induced by CRPVa, CRPVb, or chimeric genomes in rabbits homozygous for two distinct MHC class II DRA-DQA RFLP haplotypes. CRPVa (—■—) and CRPVb (- - F - -) DNAs and CRPVa-vb (- - E - -) and CRPVb-va (—䊐—) chimeric genomes obtained by exchanging the LRR/E6/E7 region, as described for Figure 1, were excised from recombinant plasmids, circularized, and inoculated by intradermal injection (A and B) (Table 3, experiment 1) or by using a gene gun (C and D) (Table 3, experiment 2) into five (A) and eight (C) rabbits homozygous for the DRA.D-DQA.B (PR) haplotype and to five (B) and eight (D) rabbits homozygous for the DRA.C-DQA.G (P) haplotype, as described in Materials and Methods. The total numbers of sites with warts observed at different times after inoculation during an 18-week follow-up are represented. Both regression occurring as early as 1 week after outgrowth and delayed outgrowth of some warts account for the fluctuations in the numbers of positive sites.

and the rabbit MHC class II haplotype is illustrated in Fig. 5. Of the three PR rabbits with persistent CRPVa-induced warts, one showed a histology-proven carcinoma 34 weeks after wart outgrowth. All five P rabbits had persistent CRPVa-induced warts, and two of them had persistent CRPVb-induced warts. One CRPVa-induced wart converted to an invasive carcinoma after 21 weeks. On the whole, only CRPVa-associated cancers were observed (Table 3). These data demonstrate that wart evolution towards regression or persistence and malignant conversion depends on both the genetic variation of CRPV and the polymorphism of rabbit MHC genes. Variability of the LRR/E6/E7 region and wart evolution. The most variable region of the CRPV genome includes the LRR and the E6 and E7 ORFs (on the whole, 7% variable nucleotide positions). This variability may affect both the level of expression and the biologic and antigenic properties of the viral E6 and E7 oncoproteins, which are likely targets involved in regression (9). Therefore, we constructed chimeric CRPV genomes by exchanging the SpeI-EcoNI fragment encompassing this region (Fig. 1). The exchanged fragment did not include the invariant first 75 nt of the LRR and contained the first 84 nt of the E1 ORF. Depending on the origin of the variable (v) region, the chimeric genomes were referred to as CRPVa-vb or CRPVb-va. Recircularized chimeric and parental CRPV genomes were each delivered at three sites onto the skin of eight PR and eight P rabbits using a gene gun (57). Warts developed in most rabbits between 3 and 7 weeks (Fig. 4C and D; Table 3). Positive sites ranged from 50 to 79.2% for chimeric DNAs and from 45.8 to 75% for parental genomes. In

PR rabbits, all warts induced by CRPVa-vb and over 90% of those induced by CRPVb showed early regression, whereas most positive sites inoculated with CRPVb-va and CRPVa persisted for 6 months after transfection (Table 3; Fig. 4C). In P rabbits, the rates of early regression of warts induced by the four genomes (from 38.9 to 68.4% of the positive sites) were higher than those observed for CRPVa and CRPVb in the first experiment (Table 3; Fig. 4B and D). Differences among the four genomes were not found statistically significant. Six months after infection, all CRPVb-induced warts had regressed and warts persisted at less than 25% of the CRPVa-, CRPVa-vb-, or CRPVb-va-positive sites (Table 3). The influence of the variability of the LRR/E6/E7 region on wart evolution as related to the DRA-DQA haplotype is illustrated in Fig. 6. Cancers arose between 19 and 42 weeks after wart outgrowth at about 40% of the sites positive for CRPVa (six cancers) or CRPVb-va (five cancers) in a total of six PR rabbits (Table 3; Fig. 6) and from a persistent CRPVb-va-induced wart in one P rabbit (Table 3). Taken together, the data indicate that the LRR/E6/E7 region plays a major part in wart evolution and that both sequence variations in this region and the genetic constitution of the host influence the outcome of the disease. CRPV duality and level of expression of late viral functions. CRPV vegetative DNA replication, late transcripts, and capsid antigens are seldom detected in domestic rabbit warts (37, 41, 62). CRPV strains recoverable from domestic rabbit warts have been reported, however (18, 50, 52), suggesting a genetic heterogeneity of the virus. This prompted us to compare the

VOL. 74, 2000


10773

in the CRPVb-induced wart tissue sections analyzed were sixto eightfold greater for viral DNA (60.1 versus 7.9 cells/mm2), L1 transcripts (12.5 versus 2.2 cells/mm2), and capsid antigens (25.8 versus 3.9 cells/mm2). To confirm these data, viral DNA replication and capsid antigen synthesis were further analyzed for warts induced by a viral suspension containing the two CRPV subtypes. Warts were collected from 10 rabbits between 4 and 10 weeks after outgrowth, and the CRPV subtype was identified in wart sections by a PCR approach (48). CRPVa was detected in 17 specimens taken from 4 PR rabbits, and CRPVb was detected in 31 specimens taken from 6 P rabbits. Warts induced by CRPVb virions yielded a significantly higher proportion of specimens positive for CRPV DNA and capsid antigen (Table 4). Furthermore, the mean numbers of positive cells per square millimeter were about 10- and 17-fold greater in CRPVbpositive wart sections for viral DNA and capsid antigen, respectively. The differences were found very significant (Table 4). The highest values of positive cells per square millimeter of wart section found for CRPVb- and CRPVa-positive specimens were 44.9 and 3.4 cells/mm2 for viral DNA and 18.2 and 1.0 cells/mm2 for capsid proteins. These values were in the range of those reported above for the warts induced in the same rabbit by cloned CRPVa and CRPVb DNAs. The data indicate that the host restriction of the expression of late viral functions in warts induced by CRPVb in domestic rabbits is less stringent than that in warts induced by CRPVa. DISCUSSION FIG. 5. Evolution of warts induced by CRPVa and CRPVb DNAs in rabbits homozygous for two different MHC class II DRA-DQA RFLP haplotypes. CRPVa (a) and CRPVb (b) DNAs were excised from recombinant plasmids, recircularized, and inoculated by intradermal injection (10 ␮g/site) as described in Materials and Methods into four sites not evenly spaced on the left (CRPVa) and right (CRPVb) flanks of rabbits homozygous for the DRA.D-DQA.B (PR) or DRA.C-DQA.G (P) haplotype. Warts (one to three per site) developed at 5 or 6 weeks after infection (p.i.) at three sites (P rabbits) or four sites (PR rabbits). CRPVb warts showed early regression in PR rabbits and persisted in P rabbits, whereas CRPVa warts persisted in both P and PR rabbits. Bars, 15 mm.

expression of the late viral functions in warts induced by the two CRPV subtypes by using in situ methods. Single CRPVaand CRPVb-induced wart biopsy specimens were taken 10 weeks after wart outgrowth from the P rabbit illustrated in Fig. 5 and analyzed for viral DNA replication, transcription of the L1 ORF, and presence of capsid antigens. As illustrated for the detection of viral DNA in CRPVb- and CRPVa-induced warts (Fig. 7A and B), an unusually high level of expression of late viral functions was observed for the CRPVb specimen (Fig. 7A, C, and D). Compared to those in the CRPVa-induced specimen, the numbers of positive cells per square millimeter

The complete nucleotide sequence reported here for the genomes of the CRPVa4 and CRPVb isolates disclosed both an intratype variability unusual among papillomaviruses and the great genetic stability of the CRPVa subtype, in agreement with the conclusions drawn from our previous partial sequence data (48). Our study also provided evidence for variations in the CRPV genome affecting the interaction of the virus with the host. The CRPVa4 and CRPVb strains were found to differ in 3.1% of the nucleotide positions, whereas the CRPVa4 and the prototypical CRPVa1 strains differed by 5 (0.06%) nt only. Compared to those for CRPVa4, the sequence data available for the Washington B strain or CRPVa2 (4,136 nt in the E6-E2 region) (6, 26, 57) showed about 0.3% variable nucleotides, which further substantiates the genetic stability of the CRPVa subtype. The LRR/E6/E7 region was found to be the most divergent (Table 2). In the CRPVb LRR, a nucleotide substitution in the putative promoter sequence for the late transcripts (55), an additional binding site for the viral E2 protein, and the deletion of a 32-bp sequence containing the cap site of CRPVa LE6 transcripts (12, 55) could affect the transcription of the viral genome. The divergence in the coding sequences

TABLE 4. Level of viral replication in warts induced by CRPVb and CRPVa in domestic rabbits Proportion (%) of specimens positivea for:

Mean no. of cells/mm2 ⫾ SEM positiveb for:

Virus inducing warts

Viral DNA

Capsid antigens

Viral DNA

Capsid antigens

CRPVb CRPVa

30/31 (96.8) 12/17 (70.6) (P ⫽ 0.017)

22/31 (71) 4/17 (23.5) (P ⫽ 0.0005)

7.92 ⫾ 1.92 0.73 ⫾ 0.25 (P ⬍ 0.0001)

2.22 ⫾ 0.76 0.12 ⫾ 0.078 (P ⫽ 0.0012)

a Specimens were considered positive with at least one positive cell observed in the wart section analyzed by in situ hybridization or immunochemistry. Proportions of CRPVb- and CRPVa-positive specimens were compared using Fisher’s exact test. b Positive cells were numerated in each section as described in Materials and Methods. The numbers of positive cells per square millimeter found for each specimen were used to calculate the mean numbers of positive cells per square millimeter for CRPVa- and CRPVb-positive specimens. The two-tailed P values were calculated using the nonparametric Mann-Whitney test.

10774

SALMON ET AL.

involved point mutations only (E1, E2, E8, L2, and L1 ORFs) or point mutations and nucleotide insertions or deletions and mutations modifying stop codons (E6, E7, E4, and E5 ORFs). This resulted in amino acid changes ranging from 0.8 (L1) to 16.2% (LE6) and in variations in the sizes of the E6, E7, E4, and E5 proteins (Table 2). It is worth stressing that the putative E5 proteins encoded by CRPVa1 (21) or CRPVa4, CRPVa2 (7), CRPVa5 (25), and CRPVb vary greatly in amino acid sequences and in size (from 25 to 101 amino acids). CRPVa2 E5 mutant DNAs encoding only the 7 or 61 aminoterminal residues were found to induce warts in domestic rabbits, although at frequencies lower than that for wild-type DNA (7, 32). In our experiments, we found no significant difference between the infectivities of CRPVa4 and CRPVb genomes, which encode 101- and 25-amino-acid E5 proteins, respectively. This supports the conclusion drawn from mutant studies (7, 32) that the E5 protein is dispensable for papilloma production. The probability of wart regression and the risk for developing a cancer were found to depend on the genetic constitution of the host and on the CRPV strain, specifically, on the sequence variation of the LRR/E6/E7 region. Of 25 rabbits inoculated with both CRPVa and CRPVb genomes and, for 16 of them, also with chimeric genomes containing the LRR/ E6/E7 region of CRPVb (CRPVa-vb) or CRPVa (CRPVb-va), warts developed in all inoculated with CRPVa genomes and in most inoculated with CRPVb (80%) and chimeric genomes (87.5%). In rabbits homozygous for the MHC class II PR haplotype, warts induced by CRPVb or CRPVa-vb showed an early regression, whereas most warts induced by CRPVa or CRPVb-va persisted with similarly high rates. In contrast, in rabbits homozygous for the P haplotype, the rates of early wart regression were similar for CRPVa and CRPVb (about 35% of positive sites) and for CRPVa-vb and CRPVb-va (68.4 and 58.8% of positive sites, respectively). Due to a higher rate of late regression of CRPVb-induced warts, about 2.5-fold more CRPVa sites remained positive 6 months after inoculation. It is worth stressing that regression was higher in P rabbits for gene gun DNA delivery than for intradermal inoculation. This could be due to a direct transfection of the epidermal antigen-presenting Langerhans cells, which may have resulted in a greater stimulation of T-cell responses (42). Spontaneous regression of CRPV-induced warts precludes tumor progression and thus represents a crucial event in host control of the oncogenic potential of the virus (9). Regression is the consequence of a specific cell-mediated immune response (15, 31), and our data support the hypothesis that antigenic peptides derived from LE6, SE6, and/or E7 proteins are major targets. The different regression rates of warts induced by the four genomes, especially in PR rabbits, are likely to be due to distinct immunogenic properties of the viral oncoproteins, in view of their great amino acid variations. Different levels of expression of the LE6 and SE6 proteins could also contribute to the phenomenon. The DQA alleles defining the PR and P haplotypes differ at 22.2% of the amino acids in their ␣1 peptide-binding domains (24). This suggests that an allelic restriction of the binding of CRPVa and CRPVb E6and/or E7-derived epitopes by MHC molecules and of their presentation to specific T lymphocytes plays a major role in determining wart regression. Cancer developed in a total of nine (36%) PR or P rabbits, in eight (32%) from warts induced by CRPVa and in six (42.9%) from warts induced by CRPVb-va. It is worth stressing that cancers did not arise from persistent CRPVb- or CRPVavb-induced warts in P rabbits. Because only two P rabbits developed a CRPVa- or a CRPVb-va-associated cancer, fur-

J. VIROL.

FIG. 6. Evolution of warts induced by chimeric and parental genomes in rabbits homozygous for the PR or the P DRA-DQA RFLP haplotype. CRPVa (a), CRPVb (b), and chimeric genomes obtained by exchanging the LRR/E6/E7 region of CRPVa (va) and CRPVb (vb) were excised from recombinant plasmids, recircularized, and inoculated (1 ␮g/site) with a gene gun apparatus as described in Materials and Methods into three sites on the left (CRPVa, CRPVb-va) and right (CRPVb, CRPVa-vb) flanks of rabbits. CRPVb- and CRPVa-vb-induced warts regressed in PR rabbits by 11 weeks after inoculation (p.i.). Carcinomas that arose from CRPVa- and CRPVb-va-induced warts are also illustrated. Bars, 5 (data at 6 and 11 weeks p.i.) and 15 mm (data at 36 weeks p.i.).

ther study is required before concluding that variations in the E6 and E7 oncoproteins affect their transforming properties. It is worth mentioning that a CRPVb-induced wart biopsy specimen taken 4 months after outgrowth from a P rabbit showed histologic features of an invasive carcinoma 8.5 months later, after two passages in athymic nude mice. This indicates that CRPVb-induced warts have a potential to progress to malignancy (F. Breitburd, unpublished data). In addition to its influence on wart evolution, CRPV intratype variation was found to affect the host restriction of the expression of late viral functions in domestic rabbits. Domestic rabbit warts usually yield little, if any, infectious virus (3, 18, 51, 52). Vegetative viral DNA replication and capsid proteins were

VOL. 74, 2000


10775

FIG. 7. In situ detection of CRPV DNA, L1 transcripts, and capsid antigens in warts induced in the same rabbit by cloned CRPVa or CRPVb DNAs. Two warts were collected 10 weeks after wart outgrowth, fixed in formalin, and embedded in paraffin, and 5-␮m-thick sections were prepared for in situ hybridization and immunochemistry experiments. Representative positive fields are illustrated. (A and B) Detection of CRPV DNA in a section of the warts induced by CRPVb (A) and CRPVa (B) using a digoxigenin-labeled CRPV DNA probe. CRPV genomes were detected in the nuclei of upper terminally differentiating granular cells (gr), where vegetative viral DNA replication occurs (41), and in keratinized superficial cells (k), where viral DNA is encapsidated (11). b, basal cell layer; dp, dermal papilla. (C) Detection of L1 transcripts in an adjacent section of a CRPVb-induced wart using a 35S-labeled antisense CRPV L1 RNA probe (dark-field illumination). Late transcripts were observed in the upper granular cells but not in the nonliving keratinized cells. (D) Immunochemical detection of CRPV capsid antigens in an adjacent section of a CRPVb-induced wart using antibodies raised against disrupted CRPV virus-like particles (8) and an alkaline phosphatase-conjugated secondary antibody. Viral antigens were detected in the nuclei of some of the uppermost granular cells and in keratinized cells, as expected (37, 41). Bars, 50 ␮m.

detected in a significantly greater proportion of warts induced by CRPVb, and the mean number of positive cells (expressed per square millimeter of wart section) was found greater by 1 order of magnitude for CRPVb-positive specimens. A similar difference in the level of expression of late viral functions was found also for warts induced in the same rabbit by CRPVa and CRPVb DNAs. That virions are produced in CRPVb-induced warts has been shown by the recovery of infectious particles from a CRPVb-induced wart grafted to a nude mouse (F. Breitburd, unpublished results). The somewhat alleviated host restriction of CRPVb replication might result from a higher capacity of the E6 and E7 oncoproteins to overcome the switch-off of the host DNA replication machinery that occurs in keratinocytes committed to terminally differentiate (28). It may also involve a higher activity or level of expression of the E1 and E2 proteins that are required for viral DNA replication (28). It is worth stressing that the C3T transition affecting the putative promoter for the late transcripts in CRPVb was previously identified in CRPVa DNA in the transplantable VX7 carcinoma, a tumor derived from a wart induced by a CRPV strain recoverable in domestic rabbits (20, 43). CRPV strains obtained from cottontail rabbits have been shown to differ in their abilities to be recovered from domestic rabbit warts (3, 18). Our data strongly suggest that this old observation, unex-

plained as yet, is a consequence of an intratype variability of CRPV. An intratypic variation has been described for HPV5 associated with the skin carcinomas of epidermodysplasia verruciformis (13) and for HPV16 and HPV18 associated with the carcinomas of the uterine cervix (33). There is some evidence that the variability of HPV16 could correspond to different oncogenic potentials (14, 58, 60). The characterization of two CRPV subtypes showing a striking divergence in the regulatory region of their genomes and in their E6 and E7 oncogenes has allowed us to demonstrate unequivocally, for the first time, that intratype variation influences the evolution of potential cancer precursor lesions. It has allowed us also to show that intratype variability is able to overcome to some extent the host restriction of viral replication. By using various chimeric CRPV genomes, our current experiments are aimed at understanding the molecular basis of the distinct biologic properties of CRPV subtypes. ACKNOWLEDGMENTS Je´ro ˆme Salmon and Mathieu Nonnenmacher contributed equally to the results. J.S. and M.N. were supported by doctoral fellowships from the Ministe`re de l’Education Nationale de la Recherche et de la Tech-

10776

SALMON ET AL.

nologie (MENRT), and J.S. was the recipient of a Pasteur-Weizmann fellowship. We acknowledge D. Senlecques for expert assistance in the preparation of the figures and the manuscript and D. Jeannel for advice on relevant statistical tests. REFERENCES 1. Angerer, L. M., K. H. Cox, and R. C. Angerer. 1987. Demonstration of tissue-specific gene expression by in situ hybridization. Methods Enzymol. 152:649–661. 2. Barbosa, M. S., and F. O. Wettstein. 1987. Transcription of the cottontail rabbit papillomavirus early region and identification of two E6 polypeptides in COS-7 cells. J. Virol. 61:2938–2942. 3. Beard, J. W. 1956. The fallacy of the concept of virus “masking”: a review. Cancer Res. 16:279–291. 4. Beard, J. W., and P. Rous. 1934. A virus-induced mammalian growth with the characters of a tumor (the Shope rabbit papilloma). II. Experimental alterations of the growth of the skin: morphological considerations: the phenomena of retrogression. J. Exp. Med. 60:723–740. 5. Bergeron, C., R. Barrasso, S. Beaudenon, P. Flamant, O. Croissant, and G. Orth. 1992. Human papillomaviruses associated with cervical intraepithelial neoplasia: great diversity and distinct distribution in low- and high-grade lesions. Am. J. Surg. Pathol. 16:641–649. 6. Brandsma, J. L., Z.-H. Yang, S. W. Barthold, and E. A. Johnson. 1991. Use of a rapid, efficient inoculation method to induce papillomas by cottontail rabbit papillomavirus DNA shows that the E7 gene is required. Proc. Natl. Acad. Sci. USA 88:4816–4820. 7. Brandsma, J. L., Z.-H. Yang, D. DiMaio, S. W. Barthold, E. Johnson, and W. Xiao. 1992. The putative E5 open reading frame of cottontail rabbit papillomavirus is dispensable for papilloma formation in domestic rabbits. J. Virol. 66:6204–6207. 8. Breitburd, F., R. Kirnbauer, N. L. Hubbert, B. Nonnenmacher, C. TrinDinh-Desmarquet, G. Orth, J. T. Schiller, and D. R. Lowy. 1995. Immunization with viruslike particles from cottontail rabbit papillomavirus (CRPV) can protect against experimental CRPV infection. J. Virol. 69:3959–3963. 9. Breitburd, F., J. Salmon, and G. Orth. 1997. The rabbit viral skin papillomas and carcinomas: a model for the immunogenetics of HPV-associated carcinogenesis. Clin. Dermatol. 15:237–247. 10. Cox, K., D. V. DeLeon, L. M. Angerer, and R. C. Angerer. 1984. Detection of mRNAs in sea urchin embryos by in situ hybridization using asymmetric RNA probes. Dev. Biol. 101:485–502. 11. Croissant, O., C. Dauguet, P. Jeanteur, and G. Orth. 1972. Application de la technique d’hybridation molećulaire in situ à la mise en évidence au microscope électronique de la re´plication ve´ge´tative de l’ADN viral dans les papillomes provoque´s par le virus de Shope chez le lapin cottontail. C. R. Acad. Sci. 274:614–617. 12. Danos, O., E. Georges, G. Orth, and M. Yaniv. 1985. Fine structure of the cottontail rabbit papillomavirus mRNAs expressed in the transplantable VX2 carcinoma. J. Virol. 53:735–741. 13. Deau, M.-C., M. Favre, S. Jablonska, L.-A. Rueda, and G. Orth. 1993. Genetic heterogeneity of oncogenic human papillomavirus type 5 (HPV5) and phylogeny of HPV5 variants associated with epidermodysplasia verruciformis. J. Clin. Microbiol. 31:2918–2926. 14. Ellis, J. R. M., P. J. Keating, J. Baird, E. F. Hounsell, D. V. Renouf, M. Rowe, D. Hopkins, M. F. Duggan-Keen, J. S. Bartholomew, L. S. Young, and P. L. Stern. 1995. The association of an HPV16 oncogene variant with HLA-B7 has implications for vaccine design in cervical cancer. Nat. Med. 1:464–470. 15. Evans, C. A., and Y. Ito. 1966. Antitumor immunity in the Shope papillomacarcinoma complex of rabbits. III. Response to reinfection with viral nucleic acid. J. Natl. Cancer Inst. 36:1161–1166. 16. Favre, M., N. Jibard, and G. Orth. 1982. Restriction mapping and physical characterization of the cottontail rabbit papillomavirus genome in transplantable VX2 and VX7 domestic rabbit carcinomas. Virology 119:298–309. 17. Friedewald, W. F. 1942. Cell state as affecting susceptibility to a virus. Enhanced effectiveness of the rabbit papilloma virus on hyperplastic epidermis. J. Exp. Med. 75:197–219. 18. Friedewald, W. F., and J. G. Kidd. 1944. The recoverability of virus from papillomas produced therewith in domestic rabbits. J. Exp. Med. 79:591–605. 19. Georges, E., O. Croissant, N. Bonneaud, and G. Orth. 1984. Physical state and transcription of the cottontail rabbit papillomavirus genome in warts and transplantable VX2 and VX7 carcinomas of domestic rabbits. J. Virol. 51: 530–538. 20. Georges, E., G. Pehau-Arnaudet, and G. Orth. 1992. Molecular and biological characterization of cottontail rabbit papillomavirus variant DNA sequences integrated in the VX7 carcinoma. Virology 186:750–759. 21. Giri, I., O. Danos, and M. Yaniv. 1985. Genomic structure of the cottontail rabbit (Shope) papillomavirus. Proc. Natl. Acad. Sci. USA 82:1580–1584. 22. Giri, I., and M. Yaniv. 1988. Structural and mutational analysis of E2 transactivating proteins of papillomaviruses reveals three distinct functional domains. EMBO J. 7:2823–2829. 23. Han, R., F. Breitburd, P. N. Marche, and G. Orth. 1992. Linkage of regres-

J. VIROL.

24.

25. 26. 27. 28. 29. 30. 31. 32. 33.

34. 35.

36. 37. 38. 39.

40. 41.

42.

43. 44. 45. 46. 47. 48.

49.

sion and malignant conversion of rabbit viral papillomas to MHC class II genes. Nature 356:66–68. Han, R., F. Breitburd, P. N. Marche, and G. Orth. 1994. Analysis of the nucleotide sequence variation of the antigen-binding domain of DR␣ and DQ␣ molecules as related to the evolution of papillomavirus-induced warts in rabbits. J. Investig. Dermatol. 103:376–380. Han, R., N. M. Cladel, C. A. Reed, and N. D. Christensen. 1998. Characterization of transformation function of cottontail rabbit papillomavirus E5 and E8 genes. Virology 251:253–263. Harry, J. B., and F. O. Wettstein. 1996. Transforming properties of the cottontail rabbit papillomavirus oncoproteins LE6 and SE6 and of the E8 protein. J. Virol. 70:3355–3362. Howe, K. M., and R. J. Watson. 1991. Nucleotide preferences in sequencespecific recognition of DNA by c-myb protein. Nucleic Acids Res. 19:3913– 3919. Howley, P. M. 1996. Papillomavirinae: the viruses and their replication, p. 2045–2076. In B. N. Fields, D. M. Knipe, and P. M. Howley (ed.), Fields virology, 3rd ed, vol. 2. Lippincott-Raven Publishers, Philadelphia, Pa. Ito, Y. 1960. A tumor-producing factor extracted by phenol from papillomatous tissue (Shope) of cottontail rabbits. Virology 12:596–601. Kidd, J. G. 1938. The course of virus-induced rabbit papillomas as determined by virus, cells and host. J. Exp. Med. 67:551–573. Kreider, J. W., and G. L. Bartlett. 1981. The Shope papilloma-carcinoma complex of rabbits: a model system of neoplastic progression and spontaneous regression. Adv. Cancer Res. 35:81–110. Meyers, C., J. Harry, Y.-L. Lin, and F. O. Wettstein. 1992. Identification of three transforming proteins encoded by cottontail rabbit papillomavirus. J. Virol. 66:1655–1664. Myers, G., C. Baker, C. Wheeler, A. Halpern, A. McBride, and J. Doorbar. 1996. Human papillomaviruses 1996. A compilation and analysis of nucleic acid and amino acid sequences. Los Alamos National Laboratory, Los Alamos, N.Mex. Nasseri, M., C. Meyers, and F. O. Wettstein. 1989. Genetic analysis of CRPV pathogenesis: the L1 open reading frame is dispensable for cellular transformation but is required for papilloma formation. Virology 170:321–325. Nasseri, M., and F. O. Wettstein. 1984. Differences exist between viral transcripts in cottontail rabbit papillomavirus-induced benign and malignant tumors as well as non-virus-producing and virus-producing tumors. J. Virol. 51:706–712. Nasseri, M., and F. O. Wettstein. 1987. A variant of CRPV DNA preferentially maintained as a plasmid in NIH 3T3 cells and characterization of its transcripts in nude mouse tumors. Virology 161:541–548. Noyes, W. F., and R. C. Mellors. 1957. Fluorescent antibody detection of the antigens of the Shope papilloma virus in papillomas of the wild and domestic rabbit. J. Exp. Med. 106:555–562. Nye, J. A., J. M. Petersen, C. V. Gunther, M. D. Jonsen, and B. J. Graves. 1992. Interaction of murine ets-1 with GGA-binding sites establishes the ETS domain as a new DNA-binding motif. Genes Dev. 6:975–990. O’Connor, M., S.-Y. Chan, and H. U. Bernard. 1995. Transcription factor binding sites in the long control region of genital HPVs, p. III-21–III-40. In G. Myers, H. Delius, J. Icenogel, H.-U. Bernard, C. Baker, A. Halpern, and C. Wheeler (ed.), Human papillomaviruses 1995. A compilation and analysis of nucleic acid and amino acid sequences. Los Alamos National Laboratory, Los Alamos, N.Mex. Orth, G. 1999. Papillomaviruses—human (Papovaviridae): general features, p. 1105–1114. In R. G. Webster and A. Granoff (ed.), Encyclopedia of virology, 2nd ed., vol. 2. Academic Press Ltd., London, United Kingdom. Orth, G., P. Jeanteur, and O. Croissant. 1971. Evidence for and localization of vegetative viral DNA replication by autoradiographic detection of RNADNA hybrids in sections of tumors induced by Shope papilloma virus. Proc. Natl. Acad. Sci. USA 68:1876–1880. Porgador, A., K. R. Irvine, A. Iwasaki, B. H. Barber, N. P. Restifo, and R. N. Germain. 1998. Predominant role for directly transfected dendritic cells in antigen presentation to CD8⫹ T cells after gene gun immunization. J. Exp. Med. 188:1075–1082. Rogers, S., J. G. Kidd, and P. Rous. 1960. Relationships of the Shope papilloma virus to the cancers it determines in domestic rabbits. Acta Unio Int. Contra Cancrum 16:129–130. Rous, P. 1936. The virus tumors and the tumor problem. Am. J. Cancer 28: 233–272. Rous, P. 1943. The nearer causes of cancer. JAMA 122:573–581. Rous, P., and J. W. Beard. 1935. The progression to carcinoma of virusinduced rabbit papillomas (Shope). J. Exp. Med. 62:523–548. Saiki, R. K., D. H. Gelfand, S. Stoffel, S. J. Scharf, R. Higuchi, G. T. Horn, K. B. Mullis, and H. A. Erlich. 1988. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239:487–491. Salmon, J., N. Ramoz, P. Cassonnet, G. Orth, and F. Breitburd. 1997. A cottontail rabbit papillomavirus strain (CRPVb) with strikingly divergent E6 and E7 oncoproteins: an insight in the evolution of papillomaviruses. Virology 235:228–234. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463–5467.

VOL. 74, 2000


50. Selbie, F. R., and R. H. M. Robinson. 1948. Serial transmission of infectious papillomatosis in the domestic rabbit. Br. J. Cancer 1:371–379. 51. Shope, R. E. 1933. Infectious papillomatosis of rabbits (with a note on the histopathology by E. W. Hurst). J. Exp. Med. 58:607–624. 52. Shope, R. E. 1935. Serial transmission of virus of infectious papillomatosis in domestic rabbits. Proc. Soc. Exp. Biol. Med. 32:830–832. 53. Smith, W. E., J. G. Kidd, and P. Rous. 1952. Experiments on the cause of the rabbit carcinomas derived from virus-induced papillomas. I. Propagation of several of the cancers in sucklings with etiological tests. J. Exp. Med. 95:299– 317. 54. Syverton, J. T. 1952. The pathogenesis of the rabbit papilloma-to-carcinoma sequence. Ann. N. Y. Acad. Sci. 54:1126–1140. 55. Wettstein, F. O., M. S. Barbosa, and M. Nasseri. 1987. Identification of the major cottontail rabbit papillomavirus late RNA cap site and mapping and quantitation of an E2 and minor E6 coding mRNA in papillomas and carcinomas. Virology 159:321–328. 56. Wright, D. K., and M. M. Manos. 1990. Sample preparation from paraffinembedded tissues, p. 153–158. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White (ed.), PCR protocols: a guide to methods and applications. Academic Press Inc., San Diego, Calif.

10777

57. Wu, X., W. Xiao, and J. L. Brandsma. 1994. Papilloma formation by cottontail rabbit papillomavirus requires E1 and E2 regulatory genes in addition to E6 and E7 transforming genes. J. Virol. 68:6097–6102. 58. Xi, L. F., L. A. Koutsky, D. A. Galloway, J. Kuypers, J. P. Hughes, C. M. Wheeler, K. K. Holmes, and N. B. Kiviat. 1997. Genomic variation of human papillomavirus type 16 and risk for high grade cervical intraepithelial neoplasia. J. Natl. Cancer Inst. 89:796–802. 59. Xiao, W., and J. L. Brandsma. 1996. High efficiency, long-term clinical expression of cottontail rabbit papillomavirus (CRPV) DNA in rabbit skin following particle-mediated DNA transfer. Nucleic Acids Res. 24:2620–2622. 60. Zehbe, I., E. Wilander, H. Delius, and M. Tommasino. 1998. Human papillomavirus 16 E6 variants are more prevalent in invasive cervical carcinoma than the prototype. Cancer Res. 58:829–833. 61. Zeller, R., and M. Rogers. 2000. In situ hybridization to cellular RNA, p. 14.3.1–14.3.14. In F. M. Ausubel, R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl (ed.), Current protocols in molecular biology. John Wiley & Sons Inc., New York, N.Y. 62. Zeltner, R., L. A. Borenstein, F. O. Wettstein, and T. Iftner. 1994. Changes in RNA expression pattern during the malignant progression of cottontail rabbit papillomavirus-induced tumors in rabbits. J. Virol. 68:3620–3630.