Saccharomyces cerevisiae Is Permissive for Replication of Bovine Papillomavirus Type 1

Abstract

We recently demonstrated that Saccharomyces cerevisiae protoplasts can take up bovine papillomavirus type 1 (BPV1) virions and that viral episomal DNA is replicated after uptake. Here we demonstrate that BPV virus-like particles are assembled in infected S. cerevisiae cultures from newly synthesized capsid proteins and also package newly synthesized DNA, including full-length and truncated viral DNA and S. cerevisiae-derived DNA. Virus particles prepared in S. cerevisiae are able to convey packaged DNA to Cos1 cells and to transform C127 cells. Infectivity was blocked by antisera to BPV1 L1 but not antisera to BPV1 E4. We conclude that S. cerevisiae is permissive for the replication of BPV1 virus.

Papillomaviruses (PVs) are exclusively epitheliotropic viruses and have evolved a unique replication strategy that depends upon the differentiation program of keratinocytes (15). Though transient viral episome replication can occur in a number of in vitro cells (37), only keratinocytes, or cells with the potential for squamous maturation, can be productively infected, since viral capsid proteins are synthesized and virions are assembled only in terminally differentiated keratinocytes. PV capsid proteins, expressed in mammalian cells (61), insect cells (22), and E. coli (20, 27), can be used to study virion assembly and DNA encapsidation (43, 44, 52, 57-60). However, there remain large gaps in the understanding of PV life cycle.

Kreider et al. (24) first reported the use of athymic mouse xenograft culture to produce infectious human PV type 11 (HPV11) in vivo. In vitro raft culture systems have allowed differentiation-specific viral amplification, late gene expression, and virion morphogenesis for HPV31 (9, 46) and other PV types (2, 34). Recently, infectious particles have been produced (2, 8, 31, 35, 40), although the viral yield is generally small compared to input virions. However, only a small number of HPV types can be successfully grown in athymic and scid mouse xenograft systems or raft culture systems (55), and propagation of large numbers of viral particles in vitro is yet to be achieved (2).

Lambert et al. (26) first used the S. cerevisiae system to study the expression and function of the bovine PV type 1 (BPV1) E2 gene. Dostatni et al. (5) used S. cerevisiae to express full-length BPV1 E2 protein and assayed in vitro its capacity to modulate transcription. Prakash et al. (41) reported that BPV1 E2 protein regulates viral transcription by binding as a dimer to the DNA sequence ACCGN4CGGT. According to previous studies of viral DNA replication in yeast (21, 42), the basic requirements for viral cis and trans elements for episome replication are similar between S. cerevisiae and mammalian cells. We have recently observed that S. cerevisiae protoplasts, which have extensive endocytotic activity (10), can take up BPV1 virions, and the BPV1 episome can replicate (56). In the present study, we have studied whether S. cerevisiae exposed to PV virions can support production of infectious virions.

MATERIALS AND METHODS

S. cerevisiae protoplast culture and virus infection.

BPV1 virions were prepared from bovine papillomas as described previously (28). S. cerevisiae protoplast culture and virus infection were carried out as described previously (56). In brief, S. cerevisiae cells were cultured to 10⁸ cells/ml in liquid medium and harvested by centrifugation. The harvested S. cerevisiae cells were incubated in an enzyme buffer at 30°C for 3 h. The enzyme-cell mixture was checked microscopically to determine when the enzyme digestion was sufficient to produce S. cerevisiae protoplasts. S. cerevisiae protoplasts were washed with STC buffer (1 M sorbitol, 10 mM CaCl₂, 10 mM Tris-HCl; pH 7.5) twice and resuspended in S. cerevisiae medium containing 0.8 M sorbitol and 0.2 M glucose, and the density was adjusted to 5 × 10⁷ cells/ml for virus infection. Virion suspensions were dialyzed against 0.15 M phosphate-buffered saline (pH 7.4) (PBS) for 30 min. The dialyzed virus was then used to infect S. cerevisiae protoplasts. Infected or uninfected S. cerevisiae cultures were placed on a shaker with gentle agitation at 28°C in the dark. Fresh medium without sorbitol was added to the S. cerevisiae cell cultures once a day to reduce the osmoticum at the beginning of culture and subsequently based on experimental requirements.

Immunofluorescence examination of BPV L1 protein in S. cerevisiae.

S. cerevisiae protoplast culture(10 ml) was fixed by the addition of 1 ml of 37% formaldehyde in PEM buffer (100 mM Na-PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid], pH 6.9; 1 mM EGTA; 1 mM MgSO₄), with gentle agitation for 1 min, and of 88 μl of 25% glutaraldehyde (Sigma, St. Louis, Mo.; electron microscopy grade at a final concentration of 0.2% [vol/vol]). Fixed S. cerevisiae cells were agitated for 90 min in a water bath, pelleted at 1,000 × g for 5 min, and washed with 2 ml of PEM buffer three times. Washed S. cerevisiae cells were resuspended at a density of 5 × 10⁷ cells/ml in PEMS (PEM, 1 M sorbitol) buffer containing 20,000 U of lyticase (Sigma)/ml to digest the cell walls at 37°C for ca. 2 to 3 h. Digested cells were resuspended after three washes with 2 ml of PEM buffer in 2 ml of PEM containing 1% of Triton X-100 and held for 1 min. Triton X-100-treated cells were washed with 2 ml of PEM buffer three times and treated with 2 ml of fresh sodium borohydride (1 mg/ml in PEM) twice for 5 min. Cells were resuspended in 0.5 ml of PEMBAL (PEM, 0.1 M lysine, 1% globulin-free bovine serum albumin, and 0.1% sodium azide) and incubated for 1 h with continuous inversion. Cells were then pelleted and incubated in 100 μl of monoclonal antibody (MAb; 1:1,000) against BPV L1 capsid protein (MC15 [58]) on a rotary inverter overnight. MAb-labeled cells were washed with 1 ml of PEMBAL three times and incubated with 100 μl of FITC conjugated anti-mouse immunoglobulin secondary antibody, diluted 1/50 in PEMBAL, on a rotary inverter overnight. Antibody-labeled cells were washed with 0.5 ml of PEMBAL three times and mounted.

VLP preparation.

Virus-like particles (VLPs) were prepared from S. cerevisiae cultures similarly to the method used for Cos1 cells (57, 59). Briefly, cells were collected by centrifugation at 3,000 rpm for 10 min, and washed with PBS containing 2-mM phenylmethylsulfonyl fluoride (PMSF). Pellets were resuspended in 20 ml of SCE buffer (1 M sorbitol, 0.1 M sodium citrate, 10 mM EDTA; pH 6.8) containing 20 mM β-mercaptoethanol and digested with 200 μl of lyticase at 50,000 U/ml for 2 to 3 h. The digested cells were pelleted and resuspended in 5 ml of PBS containing 2 mM PMSF and homogenized in a Dounce homogenizer with a tight-fitting pestle for 10 min. Released nuclei were collected by centrifugation at 3,000 rpm at 4°C for 15 min, resuspended in 10 ml of PBS with PMSF, and sonicated for 40 s. Lysate was layered over 20% sucrose and pelleted by centrifugation at 26,000 rpm for 2 h with a Beckman SW26 rotor. Pellets were resuspended in 11.5 ml of PBS containing 5.5 g of CsCl and centrifuged in a Beckman SW41 rotor at 40,000 rpm at 21°C for 20 h. From the resulting gradient, 22 0.5-ml fractions were collected.

Immunoblotting of L1 and L2 protein.

Samples (50 μl) from CsCl gradients were dialysed in PBS, precipitated with 3 volumes of acetone at −70°C for 4 h, pelleted, resuspended in 20 μl of 1× Laemmli buffer (25), and boiled for 8 min. Samples were separated on 10% (wt/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels and electrotransferred onto nitrocellulose membranes (Bio-Rad). Blots were washed with PBS for 10 min, blocked in PBS containing 5% nonfat milk for 1 h, and probed with L1 or L2 specific MAbs (58) at 4°C overnight. Blots were then incubated with anti-mouse secondary antibody conjugated with horseradish peroxidase (Silenus Australia) and developed by enhanced chemiluminescence (Amersham Australia).

Radiolabeling of VLPs.

S. cerevisiae protoplast cultures (40 ml) were incubated in Cys/Met-free medium for 5 h prior to addition of 150 μCi of [³⁵S]Cys+Met (ICN). Cultures were incubated at 28°C for 4 days, with the addition of 20-ml fresh Cys/Met-free medium every day. S. cerevisiae cells were lysed, and fractions of density corresponding to VLPs prepared as described above. Fractions (300 μl) were incubated with 0.3 μl of L1- and L2-specific MAb for 2 h and then with protein G-Sepharose (Sigma) beads at 4°C overnight. Beads were washed with radioimmunoprecipitation assay buffer (100 mM Tris, pH 7.5; 150 mM NaCl; 5 mM CaCl₂; 0.1% Triton X-100) five times and with sterile water once and then boiled in Laemmli buffer; next, supernatant was applied to an SDS-10% polyacrylamide gel. Gels were dried and exposed to film at −70°C for 48 h.

[³H]thymidine labeling.

Fifty milliliters of S. cerevisiae protoplasts (10⁸ cells/ml) was infected with 10 μg of BPV1 virus as described above, 4 h prior to the addition of 200 μCi of [³H]thymidine. At 24 h, 50 ml of fresh S. cerevisiae medium was added. After 4 days, fractions of VLP density were prepared as described above, dialyzed against PBS, and incubated with 10 U of DNase I in 10× DNase buffer at 37°C for 30 min. Fractions were mixed with 50 μl of 3 N NaOH and boiled for 5 min, and the amount of incorporated ³H was determined by liquid scintillation counting (Beckman LS 7000).

DNA packaging analysis.

Extraction of packaged DNA from individual fractions was as described previously (57, 59). Briefly, 300 μl of fraction was dialyzed against PBS, 50 μl of 10× DNase buffer containing10 U of DNase I was added, and suspensions were held at 37°C for 30 min. After the addition of 100 μl of 10% SDS, and 500-μl phenol fractions were held at 65°C for 1 h. The aqueous phase was extracted twice with equal volumes of phenol and chloroform, and DNA precipitated with 50 μl of 3 M sodium acetate and 2 volumes of 100% ethanol at −70°C for 2 h. DNA was pelleted, resuspended in 20 μl of TE buffer, and electrophoresed on a 1% agarose gel.

Characteristic analysis of packaged DNA.

DNA samples with or without prior digestion were electrophoresed on 1% agarose gel, blotted onto nylon membrane, and probed with ³²P-labeled BPV1 DNA or BPV1 L1 DNA.

Analysis of L1 protein and viral DNA delivery in VLP-infected mammalian cells.

BPV1 VLP suspension (50 μl) was dialyzed against PBS, digested with DNase, and added to Cos1 cell cultures grown in Dulbecco modified Eagle medium supplemented with 10% fetal bovine serum. For some experiments, antibody specific for BPV1 L1 (58) or E4 (Kindly provided by John Doobar) was added to particles for 1 h prior to infection of Cos1 cell. The particle-infected Cos1 cells were harvested for Hirt DNA and protein preparations after 48 h. For protein preparation, cells were resuspended in 1× Laemmli buffer and sonicated for 40 s. Protein sample (10 μg) was boiled for 8 min and applied to a SDS-10% polyacrylamide gel. Immunoblotting assay for BPV L1 was as described above. In addition, cells were resuspended in lysate buffer (10 mM Tris-HCl, pH 7.5; 10 mM EDTA; 0.2% Triton X-100). Episomal DNA was prepared by the Hirt method (12) with some modifications (57, 59), digested with BamHI, and electrophoresed on 1% agarose gel. Southern blots were hybridized with ³²P-labeled BPV DNA. Extracted DNA was also used for PCR amplification by using oligonucleotides specific for the six early and two late genes of BPV1.

Focus formation assay.

Different volume of BPV1 VLP suspension was dialyzed against PBS, digested with DNase, and added to confluent C127 cell cultures grown in 6- or 12-well plates with Dulbecco modified Eagle medium supplemented with 2% fetal bovine serum. The medium was changed every 3 days. After 4 to 5 weeks, cells were washed with PBS twice, fixed with cold methanol for 3 min, and stained with 0.5% methylene blue plus 0.25% carbol fuschin for 15 min. Foci were counted in four separate experiments.

RESULTS

BPV virus particles are present in BPV1-infected S. cerevisiae cells.

Recent studies by ourselves (56) and others (1) have demonstrated that BPV can replicate its episome after introduction of natural virions or episomal PV DNA into S. cerevisiae cells. To investigate the extent to which BPV1 can replicate in S. cerevisiae, we first examined the extent and location of the expression of BPV1 L1 capsid protein at 20 h after exposure of S. cerevisiae cells to BPV1. Immunofluorescence microscopy of BPV1-infected S. cerevisiae cells demonstrated BPV1 L1 protein within S. cerevisiae nuclei (Fig. 1A) but not within S. cerevisiae not exposed to BPV1 (Fig. 1B). A total of 30 to 40% of S. cerevisiae cells showed significant L1 staining, which was confirmed by immunoblotting of individual colonies. Infected cells were plated and individual colonies were assayed for L1 by immunoblotting. In six experiments, 34.3% ± 13.2% of colonies were positive for L1 protein (data not shown). In addition, studies of BPV1-infected S. cerevisiae cells with cycloheximide treatment by using immunoblotting assay indicated that, at 20 h, L1 protein was at least partly newly synthesized (Fig. 1D).

FIG. 1.

Detection of BPV1 L1 protein in S. cerevisiae cells infected with BPV1 virions. S. cerevisiae cells exposed to (A) or not exposed to (B) BPV1 were probed with an L1 specific MAb 20 h after infection. (C) Electron micrograph of BPV1 virus particles purified by density gradient centrifugation from BPV1-infected S. cerevisiae cultures 4 days after infection. The size bar represents 50 nm. (D) Time course analysis of BPV L1 protein in BPV1-yeast cultures without or with cycloheximide treatment (5 μg/ml) by immunoblotting assay. A set of 1 ml of yeast culture (5 × 10⁷cells/ml) was infected with 60 ng of BPV1 virus. At different time points (as shown in Fig. 1D), 100 μl of yeast culture was collected for protein preparation. The collected yeast culture, after pelleted and washed with PBS twice, was resuspended in 50 μl of 1× Laemmli buffer and sonicated for 40 s. Then, 25 μl of protein sample was boiled for 8 min and applied subjected to SDS-PAGE. A total of 100 ng of BPV1 virus was used as positive control (V). Immunoblotting assay for BPV L1 was as described in Materials and Methods section.

Newly synthesized L1 and L2 proteins assemble into VLPs in S. cerevisiae.

To determine whether the capsid proteins of PV observed in BPV1-infected S. cerevisiae were newly synthesized, we first established that virus particles could be purified from infected cells by density gradient separation (Fig. 1C). We examined the migration of L1 and L2 proteins across a cesium chloride gradient to determine whether these proteins were associated with material of density typical of empty or full BPV1 virions. The majority of the L1 protein was found in the 1.31- to 1.29-g/ml (fractions 16 to 20) and 1.34- to 1.36-g/ml (fractions 4 to 9) gradient fractions (Fig. 2A), which correspond to the expected densities of empty and full PV virions, and VLPs could be observed in this material by electron microscopy (Fig. 1C). L2 protein was also found across the gradient, although rather more was present in the L1 containing higher-density fractions (Fig. 2A). The results confirm that BPV L1/L2 proteins in BPV1-infected S. cerevisiae cultures prefer to adopt either a heavy 1.34- to 1.36-g/ml configuration or a light 1.29- to 1.31-g/ml configuration and are therefore likely to be configured as VLPs, in keeping with prior observations that the expression of HPV L1 in S. cerevisiae results in the assembly of VLPs (16)

FIG. 2.

BPV1 capsid proteins in BPV1-infected S. cerevisiae cultures. (A) BPV1-infected S. cerevisiae cells were homogenized and fractionated on a CsCl density gradient. Fractions were separated by SDS-PAGE and probed for L1 and L2 proteins with specific MAbs. Band of 55 and 77 kDa, corresponding to L1 and L2, respectively, are indicated. (B) PV-infected S. cerevisiae cells were labeled with ³⁵S-labeled methionine and cysteine and fractionated as in panel A. VLPs were immunoprecipitated, and proteins were separated by SDS-PAGE. Arrows indicate L1 and L2 proteins at 55 and 77 kDa, respectively. The numbers represent fractions of CsCl gradients from the bottom (1.45 g/ml) to the top (1.26 g/ml).

To show that L1 and L2 proteins incorporated into VLPs in BPV1-infected S. cerevisiae cells were newly synthesized, we labeled BPV1-infected cell cultures with ³⁵S-labeled methionine and cysteine and precipitated L1 and L2 from fractions of the CsCl gradient with specific MAbs (Fig. 2B). Proteins of 55 and 77 kDa, corresponding to the known molecular masses of the BPV1 L1 and L2 proteins, were detected in material precipitated by L1 and L2 specific antibodies from fractions of a CsCl gradient prepared from BPV1-infected cultures at 4 days after infection (Fig. 2B) and also at 2 days postinfection (data not shown). Other labeled proteins were also precipitated, in lesser abundance. However, the bands corresponding in size to L1 and L2 proteins that were precipitated from infected labeled cells were not precipitated from cultures not exposed to BPV1 virus (data not shown). Incorporation of ³⁵S-labeled methionine and cysteine into proteins of appropriate molecular mass and within material of the density of PV VLPs and reactive with L1 and L2 antibodies confirms that BPV1-infected S. cerevisiae produced newly synthesized BPV1 VLPs.

DNA encapsidated by VLPs is heterogeneous.

To investigate whether the BPV1 VLPs in BPV1-infected S. cerevisiae cells could incorporate DNA, we purified DNA from dense fractions (fractions 7 and 8) and light fractions (fractions 18 and 19) from a CsCl gradient, prepared from S. cerevisiae infected 4 days previously with BPV1 virions, and also from natural BPV1 virions, in each case after DNase I treatment of the starting material to remove any DNA not packaged within particles. DNA was recovered from the fractions of the CsCl gradient corresponding to dense and light VLPs, indicating that BPV1 VLPs assembled in BPV1-infected S. cerevisiae package DNA internally (Fig. 3A). However, the electrophoresis pattern of DNA prepared from natural virions was different from that of the DNA encapsidated by BPV1 particles produced in S. cerevisiae, since the DNA encapsidated by VLPs included, in addition to the species of about the same size as PV genome found in natural virions, variable amounts of an additional species of ca. 4 kb (Fig. 3A). Hybridization of DNA purified from S. cerevisiae-produced BPV1 VLPs with a BPV1 genomic DNA probe showed that this DNA was of the same mobility as DNA from BPV1 natural virions (Fig. 3A), indicating that S. cerevisiae-produced BPV1 VLPs encapsidate full-length BPV1 genomic DNA. The additional 4-kb DNA band purified from S. cerevisiae produced VLPs did not hybridize to BPV1 DNA, suggesting that BPV1 VLPs produced in S. cerevisiae may also encapsidate DNA of S. cerevisiae origin.

FIG. 3.

Characterization of DNA extracted from high (1.35-g/ml)- and low (1.30-g/ml)-density virus-like particles prepared from BPV1-infected S. cerevisiae. (A) Undigested DNase of BPV1 virions (V), dense (D, 1.35 g/ml), and light (L, 1.30 g/ml) VLPs from BPV1-infected S. cerevisiae were electrophoresed on 1% agarose (I) and subjected to Southern blot hybridization with a ³²P-labeled BPV1 genomic probe (II). (B) DNAs of natural virions and dense (1.35 g/ml) VLPs were digested with three enzymes as shown and hybridized with ³²P-labeled BPV1 L1 gene. (C) DNA of light (1.30 g/ml) VLPs was digested with eight restriction enzymes as shown and hybridized with ³²P-labeled BPV1 L1 gene. (D) Predicted hybridization pattern of BPV1 DNA with ³²P-labeled BPV1 L1. (E) Schema showing the restriction sites of eight enzymes for the BPV1 genome and the location of the L1 gene probe.

To determine whether the BPV1 DNA encapsidated by the VLPs produced in BPV1-infected S. cerevisiae represents a complete BPV1 episome, the restriction pattern of the DNA from both dense and light VLPs was compared with that from natural BPV1 virions after digestion with three enzymes (BamHI, EcoRI, and HindIII) with unique site in the BPV1 genome (Fig. 3B). A single band representing an intact BPV1 episome at 7.95 kb hybridized with a BPV L1 gene probe in both BPV1 virion and dense VLP DNA samples (Fig. 3B). However, DNA purified from light VLPs, after digestion with the three enzymes described above, produced a more complex pattern with the BPV1 genome probe, with BamHI having an extra 2.6-kb band, three EcoRI bands (at 2.4, 3.5, and 4.0 kb) and HindIII two bands (2.5 and 4.0 kb) (data not shown). The results suggest that the DNA encapsidated by light particles might be different from that found in dense VLPs. Therefore, we used eight restriction enzyme treatments to restrict further the DNA purified from light VLPs and L1 gene probe to do Southern hybridization (Fig. 3C). The predicted hybridization pattern for BPV1 with an L1 gene probe (Fig. D and E) is shown for comparison with the pattern observed from light VLPs. Uncut DNA prepared from light VLPs shows BPV1 L1 specific hybridization to three bands, suggesting that episomal forms are present. Although the predicated L1 hybridizations were generally present in digested DNA, extra bands were observed for individual enzyme restriction, particularly for the combination of BamHI and EcoRI (Fig. 3C), indicating that heterogeneous DNA species incorporating at least a portion of the BPV1 L1 gene was encapsidated by light VLPs.

Packaged DNA is newly replicated.

To determine whether the DNA packaged by BPV1 VLPs produced in S. cerevisiae cells was newly synthesized, all cultures were labeled with [³H]thymidine 4 h after BPV1 infection. Fractions of VLP density were prepared by CsCl gradient separation. Incorporation of [³H]thymidine into the extracted DNA was then examined (Fig. 4A). Cell cultures not exposed to BPV1 had no significant incorporation of ³H into DNA in the fractions of VLP density, whereas fractions from BPV1-infected cultures incorporated significant ³H. Moreover, the maximal ³H activity was seen in the 1.30- and 1.35-g/ml fractions, showing that DNase-resistant DNA packaged within VLPs included newly synthesized viral and/or cellular DNA. In separate experiments, aliquots of each fraction of the CsCl gradient were dialyzed and treated with DNase I and DNA extracted (Fig. 4B). Inclusion of PV-associated DNA within these fractions was confirmed by hybridization with a BPV1 DNA probe (Fig. 4C).

FIG. 4.

Characterization of DNA extracted from 22 fractions of a CsCl gradient prepared from S. cerevisiae cells infected with BPV1 virions labeled with (A) or without (B and C) [³H]thymidine. (A) Incorporated radioactivity was determined by scintillation counting. The data presented are the average of three independent experiments, and the standard deviations are indicated by error bars. (B) DNA within fractions is characterized by gel electrophoresis (1% agarose). (C) DNA within fractions (see panel B) was blotted and characterized by hybridization with a ³²P-labeled BPV1 genomic probe. The numbers represent CsCl gradient fractions from 1 (1.45 g/ml) to 22 (1.26 g/ml).

VLPs produced in BPV1-infected S. cerevisiae are infectious.

To determine whether the VLPs produced in BPV1-infected S. cerevisiae cells were infectious, VLPs after dialysis and DNase digestion were added to Cos1 cells in culture. After 48 h, cells were collected for protein and Hirt DNA preparations. L1 protein was detected in protein samples prepared from Cos1 cells infected with VLPs or with E4 antibody-neutralized VLPs (Fig. 5A) but not from Cos1 infected with L1 antibody-neutralized VLPs (Fig. 5A), confirming that VLP rather than DNA was taken up by mammalian cells. DNA Southern blot hybridization revealed that episomal BPV1 DNA was detected in samples prepared from Cos1 cells exposed to dense or light VLPs (Fig. 5B), suggesting that the VLPs produced in BPV1-infected S. cerevisiae cell cultures can effectively convey packaged viral DNA to mammalian cells. PCR analysis with primers specific for different BPV1 genes revealed further that all ORFs of BPV1 could be detected within Hirt DNA samples from infected Cos1 cells (Fig. 5C).

FIG. 5.

Delivery of viral DNA by VLPs prepared from BPV1-infected S. cerevisiae protoplast cultures. Cos1 cells were exposed to dense (1.35 g/ml) or light (1.30 g/ml) VLPs. (A) Assay of BPV1 L1 protein in Cos1 cells exposed to dense (1.35 g/ml) VLPs. Cells were exposed as follows: lane 1, VLPs; lane 2, VLPs pretreated with polyclonal L1 antibody (1:1,000); lane 3, VLPs pretreated with monoclonal L1 antibody (1:1,000); lane 4, VLPs pretreated with E4 antibody (1:1,000). (B) Southern hybridization of Hirt DNA from Cos1 cells exposed to VLPs as indicated and probed with BPV1 genomic DNA. Lanes: 1, λ DNA; 2, 1.35 g of VLPs/ml; 3, 1.35 g of VLPs/ml, pretreated with polyclonal L1 antibody (1:1,000); 4, 1.35 g of VLPs/ml, pretreated with monoclonal L1 antibody (1:1,000); 5, 1.35 g of VLPs/ml, pretreated with E4 antibody (1:1,000); 6, 1.30 g of VLPs/ml; 7, 1.30 g of VLPs/ml, pretreated with polyclonal L1 antibody (1:1,000); 8, 1.30 g of VLPs/ml, pretreated with monoclonal L1 antibody (1:1,000); 9, 1.30 g of VLPs/ml, pretreated with E4 antibody (1:1,000). (C) PCR analysis for BPV1 early and late genes of Hirt DNA prepared from Cos1 cells infected with dense VLPs (1.35 g/ml).

To further confirm the infectivity of BPV1 VLPs produced in S. cerevisiae, we carried out a focus formation assay with C127 cells (43). VLPs from the dense or light fractions of a CsCl gradient were added to C127 cells. After 28 days, transformation of C127 monolayers exposed to dense or light VLPs was observed (Fig. 6). Data from four focus assay experiments has shown that the number of foci produced by BPV1 virus was (3.9 ± 1.4) × 10³/μg of L1 protein, the number of foci produced by the dense VLPs was (3.1 ± 1.2) × 10³/μg of L1 protein, and the number of foci produced by the light VLPs (1.2 ± 0.4) × 10³/μg of L1 protein. The infectivity of the dense VLPs was 2.5 times higher than the light VLPs in C127 cells. Taken together, these experiments indicate that BPV1 virions can reproduce its life cycle in S. cerevisiae cells.

FIG. 6.

Focus formation assay. C127 cells were cultured in 12-well plate and exposed to different volumes (vol/vol) of BPV1 virions (V) and light (1.30 g/ml, L) and dense (1.35 g/ml, D) VLPs. A control culture was left without infection (Mock). Cells were stained with 0.5% methylene blue plus 0.25% carbol fuschin for 15 min at 4 weeks after infection.

To estimate the efficiency of BPV1 reproduction in S. cerevisiae, we calculated the yields of VLP production after BPV1-infection. L1 protein in the dense (1.35-g/ml) and light (1.30-g/ml) VLP fractions was quantified by immunoblotting and densitometric analysis, with defined quantities of purified L1 protein of natural BPV1 virions serving as a standard (Fig. 7). Based on six independent experiments, 8.9 ± 4.5 μg of VLPs were recovered in a 0.6-ml fraction at a density of 1.35 g/ml and 13.1 ± 1.9 μg of VLPs were recovered in a 0.6-ml fraction at a density of 1.30 g/ml at 4 days postinfection. In these experiments, cells were exposed to 8.6 ± 1.8 μg of natural BPV1 virions. The data, taken together with the distribution of L1 over the entire CsCl gradient (Fig. 2A), allow a minimal estimate of (i) the output of VLPs (measured as L1) at least five times higher than that of the input L1 protein and (ii) the output of infectious particles that was at least similar to the input of infectious virus.

FIG. 7.

Estimation of VLP production from BPV1-infected S. cerevisiae cells with BPV1 virions as a standard. (A) Coomassie blue staining of SDS-PAGE gel loaded with three concentrations of BPV1 virions (V) and dense (D, 1.35-g/ml) and light (L, 1.30-g/ml) fractions purified from BPV1-infected S. cerevisiae by using a CsCl gradient. (B) Immunoblot of duplicate of panel A with an MAb (MC15) to L1 protein. Densitometry scanning was used to estimate the L1 signal. The signal for 1 μg of BPV1 virions was 7,306 ± 1,846 arbitrary units. Over six experiments, the signal for 1.35 g of VLPs/ml was 5,585 ± 4,130, a value equivalent to 0.72 μg of L1/50 μl, and the signal for 1.30 g of VLPs/ml was 7,651 ± 2,556, a value equivalent to 1.09 μg of L1/50 μl.

DISCUSSION

Using an S. cerevisiae cell culture system, we demonstrated the DNA replication, RNA transcription, and L1 protein translation of BPV1 in S. cerevisiae cells infected with natural BPV1 virions (56). The capacity of a range of types of PV genomes, including BPV1, HPV6b, HPV11, HPV16, HPV18, and HPV31, to replicate in S. cerevisiae has been demonstrated by others (1). In the present study, by using immunofluorescence microscopy, we have determined that the major capsid protein (L1) of the BPV1 is localized in the nuclei of BPV1-infected S. cerevisiae cells and that VLPs can be isolated from a lysate of infected S. cerevisiae cells. VLPs of densities typical of empty and full PV virions were present in infected cells, and these were assembled from newly synthesized L1 protein in S. cerevisiae cultures. VLPs encapsidated both viral and cellular DNA and were able to deliver encapsidated BPV1 DNA and to transform susceptible cells. S. cerevisiae exposed to natural BPV1 virions can thus produce infectious BPV1 particles, and this represents a promising model for propagation of PV virions in vitro.

The host range of the PVs is narrow. In contrast, PV capsids are able to bind to a wide variety of cells derived from a diverse number of species, indicating that a specific cellular receptor is not responsible for the narrow host range (36). Multiple receptor and uptake mechanisms have been demonstrated for PVs. PVs are believed to infect basal epithelial cells via the α₆β₄ integrin receptor (7, 32). Other receptors, including heparin and glycosaminoglycans have also been reported (19). The carboxyl-terminal portion of HPV11 interacts with heparin, and that this region appears to be crucial for interaction with the cell surface (19). Glycosaminoglycans, on the cell surface, are one group of molecules able to serve as a putative receptor (62). Heidenreich and Dierich (11) reported an integrin-like protein in C. albicans. In an independent study, Edwards et al. (6) found that the MAb Mo-1 raised against human complement receptor type 3 bound specifically to C. albicans. Several integrin-like proteins identified on the cell surface have been reported in different yeast species (14). Recently, an integrin-like protein identified at 37 kDa was immunoprecipitated with antibodies to the α₅β₁ and α_vβ₃ integrins and showed 75% homology at the nucleotide sequence level to alcohol dehydrogenase of S. cerevisiae (23). It is possible that a receptor of relatively low specificity permits uptake of BPV1 by S. cerevisiae cells after cell wall digestion. However, a recent study observed that α₆ integrin is not the obligatory cell receptor for BPV4 (47). As an alternative, therefore, virus may be internalized by a endocytotic route since S. cerevisiae protoplasts are endocytotic. Exposure of S. cerevisiae protoplasts to BPV1 virions is followed by ongoing autonomous episome replication (1, 56), transcription of the viral RNA (K.-N. Zhao et al., unpublished data) and translation of L1 and L2 capsid. The expression of L1 and L2 capsid proteins lead to the assembly of virus particles and DNA encapsidation observed in the present study.

Since the early 1980s, various yeasts have been used as an expression system for the production of VLPs. Hepatitis B (53), poliovirus (18), and a range of PVs VLPs, including HPV6 (13, 38, 45, 55), cottontail rabbit PV (16), HPV16 (39, 44, 45), HPV18 (13), and HPV11 (3, 19, 30), have been produced in yeasts. Yeasts have also been used to study the assembly of PV VLPs and the efficiency of VLP production. For example, the L2 protein of HPV6 and HPV16 is not incorporated into the VLPs synthesized in S. pombe (45). Joyce et al. (19) reported that HPV11 L1 forms particulate structures resembling native virus with an average particle diameter of 50 to 60 nm. VLPs assembled in S. cerevisiae can interact with heparin and with cell surface glycosaminoglycans resembling heparin on keratinocytes and CHO cells (19). The efficiency of VLP production depends on the expression of the PV L1 gene. Neeper et al. (38) observed that few HPV11 VLPs were produced in S. cerevisiae because of a truncation of the HPV L1 mRNA in their experiments. Yeager et al. (55) generated HPV11 pseudovirions in S. cerevisiae in which VLPs are coupled to the β-lactase gene and used them to define neutralizing antibodies. Thus, PV virions produced in S. cerevisiae are held to be useful for studies of natural PV infection. However, all of the studies were carried out after transformation of S. cerevisiae with PV genes. In recent studies we introduced the authentic viral genome of BPV1 into yeast protoplasts to study the episomal replication of the BPV1 genome in S. cerevisiae (53). In the present study, we have extended the utility of S. cerevisiae production of virions to study the life cycle of BPV1. Generally, epithelial differentiation is critical for efficient PV replication (15) since cellular differentiation is also necessary for optimal replication in other virus systems, including cytomegalovirus (54), Friend virus, (17), and human immunodeficiency virus (4). However, the complete life cycle of HPV16 has recently been displayed in cultured placental trophotoblasts (29). The present study suggests also that epithelial differentiation is not an absolute requirement for BPV1 virion production. Thus, studies of the replication of PV in S. cerevisiae, in placental trophotoblasts, and in epithelial cells should contribute to our understanding of the epithelial cell specific factors assisting viral production.

Different PV protein expression systems, including vaccinia virus and baculovirus, have allowed DNA encapsidation into expressed PV VLPs (43, 57-60). Recently, Rossi et al. (44) reported that S. cerevisiae as a PV expression system also allows DNA encapsidation into the expressed VLPs of HPV16. In the present study, two types of VLPs are identified in BPV1-infected S. cerevisiae cells. Dense VLPs encapsidate an intact 8-kb BPV1 genome, and cellular DNA was also packaged. Less-dense VLPs incorporate a mixture of virus and cell-derived DNA, and not all virus DNA is intact episome. The data support previous studies that VLPs assembled with PV L1 and L2 capsid proteins can package whole-length BPV1 genome (43, 60). Previously, we observed that the DNA packaged in 1.30 g of BPV VLPs/ml is preferentially ∼5 kb in both the vaccinia virus and baculovirus expression systems (58, 59). In the present study, the 1.30-g BPV1 VLPs packaged 8-kb BPV1 genomic DNA and 4-kb cellular DNA. The DNA packaged by VLPs in the present study was DNase I resistant, in contrast to a recent report that DNA associated with light VLPs in S. cerevisiae was not DNase I resistant (44). Mechanisms of DNA packaging by BPV1 VLPs produced in virus-infected S. cerevisiae cells may be different from those for VLPs produced in a S. cerevisiae transformation system. Dense VLPs (1.35 g) assembled in BPV1-infected S. cerevisiae cells preferentially encapsidate BPV1 genome, suggesting that mechanisms for assembly of the virus particles and encapsidation of the viral genome in BPV1-infected S. cerevisiae cultures may be analogous to those in epithelial and basal cells of the papillomas.

After Kreider et al. (24) succeeded in using raft culture technique to produce HPV11 infectious particles, different raft culture systems and several keratinocyte cell lines have been developed to establish the PV life cycle and produce infectious virions of different PV types, including HPV11 (24, 49), HPV16 (8, 48), HPV31b (34), HPV18 (35), and BPV1 (31). More recently, Meyers et al. (33) reported that HPV18/16 infectious chimeric particles could also be produced in raft culture. However, no study on the output and input of PV virions in raft cultures has been reported. Low yields of VLPs were observed for several PV types in different expression systems (38, 50), which is also an issue for raft cultures because there are no models for propagation of free PV viral particles (2). Production of PV VLPs presumably reflects in part levels of L1 protein expression (51, 58). As in the baculovirus expression system (58), S. cerevisiae cultures produce significantly more BPV1 L1 protein than they receive as infectious virus, and at least half of the packaged DNA is BPV1 derived in the current culture system. Further, we have estimated the output of infectious virions achieved in the S. cerevisiae system by using a focus assay, and this was of the same magnitude after 4 days in culture as the input virus. However, viral replication continues to occur over much longer period periods of yeast culture, and we are currently assessing the efficiency of propagation of PV episomes through multiple yeast division and of the production of PV virions in longer-term yeast cultures.

In conclusion, BPV1 VLPs produced in the S. cerevisiae system are infectious, and the output yield of both L1 protein and infectious virus particles was similar to the amount of input virus in short-term (4 days) cultures. Thus, this system may be a promising model for the propagation of free virus particles of BPV1 in vitro and will allow further genetic analysis of the life cycle of BPV1.