In Vivo Rescue of a Silent tax-Deficient Bovine Leukemia Virus from a Tumor-Derived Ovine B-Cell Line by Recombination with a Retrovirally Transduced Wild-Type tax Gene

Abstract

The lack of bovine leukemia virus (BLV) expression is a consistent finding in freshly isolated ovine tumor cells and in the B-cell lines derived from these tumors. In order to gain further insight into the mechanisms of BLV silencing in these tumors, we have used the YR2 B-cell line, which was derived from the leukemic cells of a BLV-infected sheep. This cell line contains a single, monoclonally integrated, silent provirus, which cannot be reactivated either by stimulation in vitro or by in vivo injection of the tumor cells or cloned proviral DNA in sheep. Sequence analysis of the tax gene from the YR2 cell line identified two G-to-A transitions (G₇₉₂₄ to A₇₉₂₄ and G₈₁₄₉ to A₈₁₄₉) that result in E-to-K amino acid changes at positions 228 and 303 in the Tax protein. Following retroviral vector-mediated transfer of a wild-type tax gene into YR2 cells, we showed that BLV mRNA, viral proteins, and virions were produced, demonstrating that the cellular factors required for virus expression were present in the original YR2 cell line. Injection of this transduced YR2 cell line in sheep led to the rescue of replication-competent BLV proviruses. The integrated competent proviruses exhibited unique chimeric tax genes, which arose from homologous recombination between the transduced wild-type tax and the YR2-derived tax sequences. Furthermore, in one of these functional recombinant proviruses, only the A₈₁₄₉-to-G₈₁₄₉ reversion was present, providing clear evidence that the defect underlying the silent phenotype in YR2 cells results from a single C-terminal E₃₀₃-to-K₃₀₃ amino acid substitution in the BLV Tax protein. Our observations suggest that a single strategically located mutation in tax provides a mechanism for BLV inactivation in B-cell tumors.

Bovine leukemia virus (BLV) is an exogenous B-lymphotropic retrovirus that naturally provokes B-lymphoproliferative disorders in cattle and can be used to induce a related B-cell-associated pathology in experimentally infected sheep (reviewed in references 8 and 34). Infection in cattle most often manifests itself as persistent lymphocytosis, a chronic disease characterized by the polyclonal proliferation of circulating B lymphocytes with proviral DNA integrated at various locations (16, 36). Clonal lymphoid tumors develop in fewer than 5% of infected cattle after a long latency (20, 36). Experimental transmission of BLV in sheep has been extensively studied following the inoculation of naive animals with either whole blood or fresh or cultured lymphocytes from seropositive animals or by the direct injection of cloned proviral DNA (44, 74). Infected sheep develop B-cell leukemia or lymphosarcoma at a higher frequency and with a shorter latency period than are observed in naturally infected cattle, and their malignant disease is often preceded by a preneoplastic expansion of B lymphocytes (6, 58). Integrated BLV proviral DNA has been found in every B-cell tumor examined from infected cattle and sheep examined (37, 38). Our ability to establish tumor-derived B-cell lines for in vitro studies, in combination with the in vivo BLV ovine model, provides a unique system for studying B-cell leukemogenic processes.

Structurally and functionally, BLV is related to the human T-cell lymphotropic virus types 1 and 2 (HTLV-1 and HTLV-2), which are associated with adult T-cell leukemia, tropical spastic paraparesis, and hairy T-cell leukemia in humans (23, 52, 56, 59, 79, 81). These viruses have similar genomic organizations, encode gene products with biologically similar functions, and share mechanisms of transactivation (reviewed in reference 21). BLV, HTLV-1, and HTLV-2 lack known cellularly derived oncogenes and have no specific integration sites in tumor cells, suggesting that their mechanism of tumor induction differs from that of other oncoviruses. Aside from the structural genes (gag, pol, and env), the BLV provirus contains a region called X, located between the env gene and the 3′ long terminal repeat (LTR), which encodes at least four proteins: Tax, Rex, R3, and G4. Tax and Rex, respectively, are involved in transcriptional and posttranscriptional regulation of viral expression and are essential for viral infectivity in vivo (11, 12, 18, 68). Information about the R3 and G4 accessory proteins is scarce, except for studies showing that G4 has moderate oncogenic potential in vitro (33) and that both proteins may play a role in virus propagation in the infected host (71).

BLV Tax is a nuclear protein of 34 kDa with no direct DNA binding activity (11). In HTLV-1, HTLV-2, and BLV, Tax transactivates virus transcription through the Tax-responsive element located within the U3 region of the virus LTR (11, 19, 60, 68). Transactivation requires the interaction of these sequences with cellular transcription factors, including members of the cyclic AMP-responsive element binding proteins (CREB/ATF) (1, 73). Furthermore, HTLV-1 Tax has been shown to recruit transcription factors in nuclear bodies (4, 5) and to activate the transcription of numerous cellular genes controlling cellular proliferation, including the interleukin-2 and the interleukin-2 receptor alpha genes (22, 45, 55, 61). Interestingly, both HTLV-1 and BLV Tax exhibit oncogenic potential in cell culture, since these proteins have been shown to immortalize primary rat fibroblasts and cooperate with Ha-Ras in their complete transformation (51, 62, 69, 70, 78). HTLV-I Tax has also been found to stimulate the G₁- to S-phase transition in immortalized human T lymphocytes (57, 80).

Expression of the BLV genome is repressed in vivo, and even sensitive methods, such as reverse transcription-PCR (RT-PCR) and in situ hybridization, have detected only limited viral RNA expression (2, 26, 30, 37, 42, 63). However, the virus can be reactivated ex vivo, by growing untransformed BLV-infected lymphocytes from aleukemic sheep (31, 39, 42, 50). In contrast, the lack of virus expression consistently characterizes both the tumor cells isolated from BLV-infected sheep and the transformed B-cell lines derived from these tumors (54, 63, 64). Previous analysis revealed that a significant proportion of integrated proviral DNA in the tumor cells is defective (37, 63). However, even transcriptionally competent proviruses are inactive in BLV-induced ovine tumors, suggesting that host cell factors may play an important role in virus silencing (63). Together, these data suggest that in ovine B-cell tumors, Tax-mediated transactivation of cellular genes is most likely essential for the early steps of the leukemogenic process, although the continuous expression of viral proteins is not necessary to maintain the transformed phenotype.

In order to gain further insight into the mechanisms responsible for the lack of viral expression in transformed B cells, we have been studying an ovine B-cell line, YR2, which was cloned from the leukemic B cells of a BLV-infected sheep (35, 63). This cell line contains a single, monoclonally integrated silent provirus, in which we showed two E-to-K amino acid substitutions in Tax that might impair the infectious potential of the integrated provirus. Virus expression could not be induced either in vitro, following stimulation with phorbol myristate acetate or phytohemagglutinin (unpublished observations), or in vivo by injecting sheep with the YR2 cells or the proviral DNA cloned from fresh tumor cells (63, 74). In this study, a wild-type tax gene was introduced into YR2 cells by using a gibbon ape leukemia virus (GaLV)-pseudotyped retroviral gene transfer strategy (3, 76). We show here that BLV expression was activated in the transduced tumor cells, indicating that cellular factors necessary for Tax activity were present in the native YR2 cells. In vivo inoculation of sheep with transduced YR2 cells clearly induced the propagation of rescued BLV within the infected host. The functional proviruses observed in the infected sheep peripheral blood resulted from homologous recombination between the transduced wild-type tax and the YR2-derived tax sequences. Furthermore, we have provided evidence that the naturally occurring defect underlying the silent phenotype in YR2 cells results from a C-terminal E-to-K single amino acid substitution in the BLV Tax protein. Our observations indicate that mutations in tax provide a mechanism for viral inactivation in BLV-induced B-cell tumors.

MATERIALS AND METHODS

Cells and media.

PG13 (46), a retroviral vector packaging cell line which produces GaLV-pseudotyped viral particles, was maintained in Dulbecco’s modified Eagle’s medium (Gibco) supplemented with 10% fetal calf serum (FCS) (Gibco) and 1% penicillin-streptomycin (Gibco). FLK (a BLV-infected fetal lamb kidney cell line), Rat-2 (a control cell line, ATCC CRL1764), and Rat-2 derivatives (Rat-2_LTaxSN and Rat-2_NUNL) were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% FCS, 1 mM sodium pyruvate, 2 mM glutamine, nonessential amino acids, and 100 μg of kanamycin per ml. LB155, a B-lymphoid cell line derived from a precrural tumor isolated from a BLV-infected cow (35); YR2, a cloned B-lymphoid cell line established from peripheral blood lymphocytes (PBL) isolated from a BLV-infected sheep (35, 63); and YR2 derivatives (YR2_LTaxSN and YR2_NUNL) were maintained in OPTIMEM medium (Gibco) supplemented with 10% FCS. Antibiotic selections of transduced cells were performed as indicated with 1 mg of G418 per ml. All of the cell lines were cultivated at 37°C in a 5% CO₂ humidified atmosphere.

Viral vector production and transduction protocol.

pLTaxSN results from the insertion of the 1,083-bp HincII-BamHI sequence of pGEM7zfLOR1 containing the BLV tax cDNA (provided by L. Willems) into the HpaI and BamHI sites of pLXSN (47). pLTaxSN and pNuNeo-LacZ, a control retroviral vector which expresses a fused protein resulting from fusion between the bacterial beta-galactosidase gene lacZ and the neomycin resistance gene neoR, were transfected into PG13 cells by using the Lipofectin procedure (Gibco), and transfected cells were selected with 1 mg of G418 per ml for 1 week. Viral titers of G418-resistant polyclonal cell populations were estimated by using a standard titration assay performed on Rat-2 cells and are expressed as G418-resistant CFU (G418 CFU) per milliliter. Transduction was performed as previously described in presence of 4 μg of Polybrene per ml by cultivating Rat-2 cells in the presence of viral vector-containing supernatants or by cocultivating YR2 cells and producers prior to G418 selection (3). Rat-2 and YR2 cells transduced with pLTaxSN or pNuNeo-LacZ are referred as Rat-2_NUNL or YR2_NUNL and Rat-2_LTaxSN or YR2_LTaxSN, respectively.

Southern blot analysis and probes.

High-molecular-weight cellular DNA was prepared by sodium dodecyl sulfate (SDS) (0.05%) and pronase (0.2 mg/ml) disruption of cells followed by extraction with phenol-chloroform and ethanol precipitation. Genomic DNA (20 μg) was digested with restriction endonucleases, separated by electrophoresis through a 0.8% agarose gel, and blotted onto nylon membranes (Hybond-N; Amersham). Cross-linked nitrocellulose filters were prehybridized for 4 h at 65°C in 3× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–1× Denhardt medium–0.1% SDS–200 μg of salmon sperm DNA per ml and hybridized overnight at 65°C with a ³²P-labelled specific probe (4 × 10⁸ to 5 × 10⁸ cpm/ml). After hybridization, filters were washed at 65°C in 2× SSC–0.1% SDS and finally in 0.2× SSC–0.1% SDS, dried, and exposed to Kodak X-Omat AR films. The probes used for Southern blot hybridization were a 8.3-kb total BLV probe representing the SacI fragment from plasmid pV344 (63), a BLV TAX probe corresponding to the 1-kb XbaI fragment of pGEM7zfLOR1, a 990-bp BLV ENV probe obtained by PCR amplification with primers EA and EB (see below) of YR2 proviral DNA, a 1,039-bp Moloney murine leukemia virus (MoMLV) LTR probe obtained after AflII digestion of pLTaxSN, and a NEO probe represented by a 668-bp HindII-NaeI fragment of pLTaxSN.

PCR, RT-PCR, and sequencing analysis.

The sequences of the BLV primers used in the PCR, RT-PCR, and sequencing experiments were as follows (nucleotide positions according to Sagata et al. [56] are in parentheses): T1 (7321 to 7340), 5′-GATGCCTGGTGCCCCCTCTG-3′; T2 (7604 to 7623), 5′-ACCGTCGCTAGAGGCCGAGG-3′; T6 (7767 to 7786), 5′-GTCCGTCTTTGCCCCAGACA-3′; EA (4766 to 4788), 5′-TCCTGGCTACTAACCCCCCCGT-3′; EB (5756 to 5777), 5′-TCCAGTGAGCCCCACTGACAGG-3′; C2 (7314 to 7333), 5′-GGCACCAGGCATCGATGGTG-3′; C3 (7246 to 7265), 5′-CCCCAACCAACAACACTTGC-3′; U3 (8599 to 8618), 5′-GCCAGACGCCCTTGGAGCGC-3′; and B (7200 to 7219), 5′-CGGGATCCATTACCTGATAA-3′. The sequences of the vector-specific primers were as follows (nucleotide positions in pLXSN according to Miller and Rosman [47] [GenBank accession no. M28248] are in parentheses): P1 (1230 to 1249), 5′-AGACTGTTACCACTCCCTTA-3′; P2 (1675 to 1694), 5′-ACACCCTAACTGACACACAT-3′; N1 (2281 to 2300), 5′-GACGGGCGTTCCTTGCGCAG-3′; and N2 (2591 to 2610), 5′-TCGCCGTCGGGCATGCGCGC-3′.

Genomic DNA and total RNA from cultured cells were prepared by using TriPure reagent (Boehringer) according to the manufacturer’s protocol.

One microgram of genomic DNA was amplified in buffer containing 10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl₂, 50 mM KCl, 0.2 mM deoxynucleoside triphosphates, 0.5 μM each primer, and 1 U of Taq polymerase (Boehringer). The amplification sequence consisted of a 5-min step at 94°C; 36 cycles of 1 min at 94°C, 1 min at 60°C, and 2 min at 72°C; and then a 10-min step at 72°C. PCR amplification on blood samples collected from infected sheep was performed as follows. Aliquots (500 μl) of frozen blood were mixed with an equal volume of lysis buffer (0.32 M sucrose, 10 mM Tris-HCl [pH 7.5], 5 mM MgCl₂, 1% Triton X-100). Samples were centrifuged for 30 s, washed three times in 1 ml of lysis buffer, resuspended in 100 μl of PCR buffer (50 mM KCl, 2.5 mM MgCl₂, 10 mM Tris-HCl [pH 9.0], 0.1% Triton X-100), and incubated with 1 μl of proteinase K (5 mg/ml) for 1 h at 50°C. Ten-microliter aliquots were amplified by PCR in 100 μl of reaction buffer containing 0.2 mM deoxynucleoside triphosphates, 200 ng of primers, and 5 U of Taq DNA polymerase (Boehringer). Cycling conditions and postamplification analysis were performed as described above.

Amplification of the selected region of the BLV proviral DNA (from nucleotide 7200 to 8618, with primers B and U3) prior to sequencing was performed by using Pfu proofreading DNA polymerase (Stratagene) according to the manufacturer’s protocol. Briefly, 10-μl aliquots of total cellular lysate were amplified in a 100-μl reaction volume containing 20 mM Tris-HCl (pH 8.8), 2 mM MgSO₄, 10 mM KCl, 10 mM (NH₄)₂SO₄, 0.1% Triton X-100, 0.4 mM each deoxynucleoside triphosphate, 0.5 μM primers, and 5 U of Pfu DNA polymerase. The amplification was performed with 36 cycles consisting of 1 min at 94°C, 1 min at 60°C, and 3 min at 72°C, followed by 10 min at 72°C. Ten microliters of the resulting reaction mixture was used in a second PCR amplification round under the same amplification conditions. The purified amplification product was sequenced by using the Thermo Sequenase radiolabeled terminator cycle sequencing kit according to the protocol supplied by the manufacturer (Amersham).

For RT-PCR experiments, total RNA was treated for 1 h at 37°C with 40 U of DNase I (Promega) and reverse transcribed with rTth DNA polymerase (Perkin-Elmer) in the presence of downstream primer at 70°C for 15 min according to the standard protocol supplied by the manufacturer. After addition of the upstream primer, single-stranded cDNA was amplified by PCR with 36 cycles involving denaturation at 94°C for 1 min, annealing at 60°C for 1 min, and elongation at 72°C for 2 min.

Northern blot analysis.

Total RNA was prepared by using TriPure reagent (Boehringer) according to the manufacturer’s protocol. Ten micrograms of total RNA was lyophilized, resuspended in denaturing buffer (6 M deionized glyoxal, 50% dimethyl sulfoxide, 0.1 M phosphate buffer [pH 7.0]) and incubated for 10 min on ice followed by 3 min at 55°C. RNAs were separated by electrophoresis through a 1% agarose gel containing 10 mM phosphate buffer, transferred onto a nylon membrane (Amersham) for 20 h in 20× SSC, and cross-linked by UV. Membranes were first prehybridized for 3 h at 42°C in hybridization solution (50% deionized formamide, 0.1% SDS, 10× Denhardt solution, 200 μg of salmon sperm DNA per ml) and then hybridized for 20 h in fresh hybridization mixture containing a ³²P-labelled TAX, NEO, or glyceraldehyde-3-phosphate dehydrogenase probe (4 × 10⁸ to 6 × 10⁸ cpm/ml). Filters were washed three times for 20 min in 2× SSC–0.1% SDS and then three times in 0.2× SSC–0.1% SDS, dried, and exposed to Kodak X-OMAT AR films.

Flow cytometry.

The detection of intracellular BLV p24 was performed with a mix of monoclonal anti-p24 antibodies 4H6, 4′F3, 7G6, 3B1, 2′C1, 4′G9, 2B1, and 5F7 (provided by D. Portetelle). Cells (10⁶) were fixed and permeabilized for 45 min in OrthoPermeafix (Ortho Diagnostics), labelled for 1 h at 4°C in 50 μl of anti-p24 mix (dilution, 1/100), washed in phosphate-buffered saline (PBS) containing 3% FCS, and incubated for 30 min with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin G (Sigma). After final washes in PBS-FCS, cells were resuspended in 500 μl of PBS–1% formaldehyde and analyzed by flow cytometry (FACScan; Becton Dickinson).

Electron microscopy.

YR2, YR2_NUNL, and YR2_LTaxSN cells (2 × 10⁶ each) were prepared for thin-section microscopy. Equal volumes of the cell suspensions and of 5% glutaraldehyde (Merck) prediluted in PBS were mixed and incubated at room temperature. Fixed cells were collected by low-speed centrifugation, resuspended in freshly diluted 2.5% glutaraldehyde, and, after OsO₄ postfixation, processed for thin-section transmission electron microscopy by routine techniques (25). Briefly, in order to reveal the structure of the retroviral envelope glycoprotein knobs more clearly, the cells were included in agarose blocks and treated before dehydration with tannic acid (0.1%) and uranyl acetate (2%). Epon-embedded blocks were sectioned by using a diamond knife and a Leica Ultracut S ultramicrotome to a thickness of 40 to 50 nm. Thin sections were poststained with lead citrate and evaluated with a Zeiss EM 10 A transmission electron microscope.

In vivo infection of sheep.

Sheep were inoculated intradermally with cultured cells. Sheep S11 and S12 received 10⁷ YR2 cells; sheep S17, S18, S19, and S20 were inoculated with 10⁷ YR2_LTaxSN cells; and sheep S15 and S16 were inoculated with 10⁷ LB155 cells as a positive control. Blood was collected weekly. Anti-p24 antibody titers in the serum were determined with a competitive enzyme-linked immunosorbent assay as previously described (49). All animals maintained individually characteristic leukocyte profiles that fell within normal ranges for percentages and absolute numbers of lymphocytes. For PCR experiments, 10-ml samples of total blood were immediately frozen and used for amplification. These sheep were 12 months postinfection by the end of the present study. The animals were maintained under controlled conditions at the National Institute for Veterinary Research (Brussels, Belgium).

RESULTS

Sequence analysis of the BLV provirus integrated in the cloned tumor-derived YR2 ovine B-cell line.

YR2, a B-cell line established from fresh M395 leukemic cells collected from a BLV-infected sheep (35, 63, 64), was used in the present study. The BLV provirus derived from fresh M395 cells has been found to be noninfectious in vivo (74) and severely impaired in its ability to activate transcription of the LTR-chloramphenicol acetyltransferase (LTR-CAT) reporter gene in vitro (63, 74), which leads to the hypothesis that the Tax protein, the 3′ LTR, or both elements could be altered in the YR2 provirus. Proviral tax and 3′ LTR sequences were amplified from in vitro-cultured YR2 cells by using the BLV-specific B and U3 primers (Fig. 1A) and were compared to tax and 3′ LTR sequences of M344, an infectious YR2-related provirus (data not shown). Sequence alignment confirmed the presence of two G-to-A transitions in the tax coding region of the YR2 provirus at nucleotides 7924 and 8149, which resulted in E-to-K changes at amino acids 228 and 303 of the Tax protein. No modification was detected in the YR2 provirus 3′ LTR. We concluded that one or both of the G-to-A transitions in the tax coding sequence might impair the provirus infectious potential.

FIG. 1.

Schematic representation of BLV (A) and LTaxSN (B) proviruses (not drawn to scale). (A) BLV gag, pol, and env sequences encode Gag, reverse transcriptase, and envelope proteins, respectively. tax/rex/R3/G4 indicates four overlapping open reading frames which encode the Tax, Rex, R3, and G4 proteins, respectively. The Gag- and Pol-encoding full-length 8.4-kb mRNA, the env 4.4-kb singly spliced transcript, and the tax/rex 2.1-kb doubly spliced mRNA are represented as dashed lines. The positions of the tax (open box) and rex (hatched box) open reading frames are indicated on the 2.1-kb transcript, with asterisks representing the YR2 tax A₇₉₂₄ and A₈₁₄₉ mutations. Alternatively spliced R3- and G4-encoding mRNAs are not represented. (B) In LTaxSN, the BLV wild-type tax cDNA (gray box), SV40 early promoter (black box), and neomycin resistance gene (white box) are shown. Transcripts generated from the LTR (3.8 kb) and from the internal SV40 promoter (1.8 kb) are indicated. Probes used in Southern and Northern blot hybridization are represented as thick bars. Arrows show the positions and directions of primers used in PCR and RT-PCR. Amplification product sizes according to primer sets are indicated. S, SacI.

Construction of pLTaxSN and viral vector production.

tax gene delivery into YR2 cells was approached by using a retroviral vector-mediated gene transfer strategy. pLTaxSN contains an MoMLV-derived retroviral backbone, which carries a full-length tax cDNA placed under the control of the MoMLV LTR promoter, and the neoR coding sequence, which promotes resistance to G418 driven by an internal simian virus 40 (SV40)-derived promoter (Fig. 1B). The tax gene present in the LTaxSN vector was isolated from a previously sequenced variant 1 BLV provirus (56), whereas the BLV provirus in YR2 cells belongs to the variant 2 group (52). Expression of a functional transactivating Tax protein was verified by using a CAT assay on Rat-2_LTaxSN cells transfected with pLTRCAT, a plasmid that contains the Tax-inducible BLV LTR cloned upstream of the CAT gene (53) (data not shown). The control vector NuNeo-LacZ expresses a fused protein resulting from fusion between the bacterial beta-galactosidase lacZ gene and the neoR gene (3). Since only GaLV-pseudotyped retroviral particles have been found to be capable of transducing ovine cells (3), vector-producing cell lines were derived from PG13, an amphotropic packaging cell line which expresses the GaLV Env protein (46). PG13 cells were transfected with pLTaxSN or pNuNeo-LacZ and selected with G418. Viral titers of polyclonal G418-resistant populations were estimated by titration assay on Rat-2 cells, taking advantage of the G418 resistance phenotype. The LTaxSN- and NuNeo-LacZ-producing cell populations exhibited viral titers of 8 × 10⁴ and 2 × 10⁶ G418 CFU/ml, respectively. The absence of helper viruses was confirmed by using a mobilization assay. YR2 cells were transduced by cocultivating target cells and mitomycin-treated PG13-derived vector-producing cells, and transduced cells were selected according to the G418 resistance phenotype induced by the vectors.

Retroviral transfer and expression of the LTaxSN-derived tax gene in YR2_LTaxSN cells.

The overall structure and expression of LTaxSN in G418-resistant YR2 polyclonal cell populations (YR2_LTaxSN) were determined at the molecular level. Genomic DNAs from YR2 and YR2_LTaxSN were extracted and digested with SacI, a restriction enzyme which cleaves within the LTRs of both the BLV and vector proviruses. Southern blot analyses were performed with four different probes (described in Materials and Methods): the NEO and MoMLV LTR probes specifically detect LTaxSN sequences, the BLV ENV probe reveals the presence of the BLV provirus, and the TAX probe detects both the LTaxSN- and YR2-derived tax sequences (Fig. 1A). In SacI-digested genomic DNA from YR2_LTaxSN cells, a 3.8-kb fragment absent from native YR2 cells (Fig. 2A, lanes 3, 5, and 7) which corresponds to the LTaxSN provirus was detected with either the NEO, TAX, or MoMLV LTR probe (Fig. 2A, lanes 4, 6, and 8). The structures of the tax and neoR coding sequences in YR2_LTaxSN cells were analyzed by PCR amplification of the LTaxSN proviral sequences in YR2_LTaxSN cells, using four sets of LTaxSN-specific primers. Primers P1 and T2 (P1/T2), P1/P2, T1/P2, and N1/N2 yielded the expected amplification products of 823, 1,540, 1,020 and 330 bp, respectively (Fig. 2B, lanes 2, 3, 6, 7, 10, 11, 18, and 19), which hybridized to the TAX or NEO probe. These results suggested that no gross modification occurred in the LTaxSN proviral vector following transduction of the YR2 cells.

FIG. 2.

Detection of LTaxSN and BLV sequences in YR2 and YR2_LTaxSN cells. (A) Southern blot analysis of SacI-digested genomic DNA. Twenty micrograms of digested DNA prepared from YR2 (lanes 1, 3, 5, and 7) or YR2_LTaxSN (lanes 2, 4, 6, and 8) cells was electrophoresed through a 0.8% agarose gel, transferred to nylon membranes, and hybridized with ³²P-labelled ENV, NEO, MoMLV LTR, and TAX probes. (B) PCR amplification of LTaxSN and BLV sequences. Genomic DNA isolated from parental YR₂ (lanes 1, 5, 9, 13, and 17) and Rat-2 (lanes 4, 8, 12, 16, and 20) cells and from YR2_LTaxSN (lanes 2, 6, 10, 14, and 18) and Rat-2_LTaxSN (lanes 3, 7, 11, 15, and 19) cells was amplified by PCR with LTaxSN-specific primers P1/T2, P1/P2, T1/P2, and N1/N2 or BLV-specific primers T6/U3 and analyzed by Southern blot hybridization with ³²P-labelled TAX and NEO probes.

In addition, we verified that the YR2 proviral tax gene remained unchanged in YR2_LTaxSN cells maintained in culture for 1 year and confirmed that the double-mutated tax gene found in native YR2 cells was also present (data not shown).

Northern blot analysis of total RNAs from YR2_LTaxSN and Rat-2_LTaxSN cells (positive control) probed with NEO revealed both the 1.8- and 3.8-kb signals, confirming that both genomic transcripts initiated from the LTaxSN LTR and that subgenomic transcripts from the internal SV40 promoter were produced (Fig. 3A, lanes 7 and 9). Neo-specific hybridization was absent in RNAs extracted from YR2, Rat-2, and FLK cells, as expected (Fig. 3A, lanes 6, 8, and 10). In parallel, the presence and expression of the NuNeo-LacZ vector in YR2_NUNL cells were verified by both X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) staining and resistance to G418.

FIG. 3.

Detection of BLV and LTaxSN transcripts in YR2_LTaxSN cells. (A) Ten micrograms of total RNA isolated from YR2, YR2_LTaxSN, Rat-2, Rat-2_LTaxSN, or FLK cells was analyzed by Northern blot hybridization with ³²P-labelled TAX (left panel) and NEO (right panel) probes. Sizes of mRNAs are indicated. (B) RT-PCR detection of the BLV 2.1-kb doubly spliced transcript in total RNA from YR2 (lanes 1 and 2), YR2_NUNL (lanes 3 and 4), Rat-2_LTaxSN (lanes 5 and 6), YR2_LTaxSN (lanes 7 and 8), and FLK (lanes 9 and 10) cells with two sets of splice-specific primers, EA/C3 (lanes 1, 3, 5, 7, and 9) and EA/T2 (lanes 2, 4, 6, 8, and 10). Amplification products were electrophoresed through a 1% agarose gel, transferred to nylon membranes, and hybridized with a ³²P-labelled total BLV probe. Positions and molecular sizes of amplified products are indicated.

BLV transcripts and viral proteins are produced in YR2_LTaxSN cells.

The presence of BLV-specific sequences in both the YR2 and YR2_LTaxSN cell lines was confirmed by hybridization of SacI-digested genomic DNA with the BLV ENV and TAX probes (Fig. 2A, lanes 1, 2, 5, and 6), which revealed a common 8.3-kb fragment, and PCR amplification of an 852-bp product by using T6 and U3 BLV-specific primers (Fig. 2B, lanes 13 and 14). The effect of transferring and expressing the LTaxSN-derived tax gene on the quiescent state of BLV was studied by Northern blot and RT-PCR assays. The major BLV transcripts include an 8.4-kb full-length mRNA which acts both as genomic RNA and as the Gag-Pol protein-encoding mRNA, a 4.4-kb singly spliced RNA encoding the Env protein, and a 2.1-kb doubly spliced RNA encoding the Tax and Rex regulatory proteins (Fig. 1A). Minor alternatively spliced mRNAs have been detected by nested PCR (2) and were not analyzed in the present study. Northern blot hybridization of total RNA derived from YR2_LTaxSN cells or from FLK cells (a BLV-producing cell line used as a control) revealed RNA species of 8.4, 4.4, and 2.1 kb with the TAX probe, which suggested that all of the major BLV transcripts were produced in these cells (Fig. 3A, lanes 2 and 5). TAX probe hybridization with RNA from Rat-2_LTaxSN cells revealed a single 3.8-kb signal that corresponds to RNA initiated from the LTaxSN MoMLV LTR (Fig. 3A, lane 4). BLV transcripts were undetectable in either YR2 or Rat-2 cells (Fig. 3A, lanes 1 and 3). The absence of the 2.1-kb doubly spliced BLV transcript was confirmed by RT-PCR with two different sets of BLV splice-specific primers, EA/C3 and EA/T2, and hybridization with a BLV probe (Fig. 3B, lanes 1 to 6) in YR2 cells, Rat-2_LTaxSN cells, and YR2_NUNL cells (negative controls). Conversely, analysis of the YR2_LTaxSN and FLK RNAs with the EA/C3 and EA/T2 primer sets yielded two fragments of 125 and 483 bp, respectively (Fig. 3B, lanes 7 to 10). Together, these results demonstrate that transcription of the BLV genes can be specifically activated in YR2 cells after the transfer and expression of a functional tax gene. Furthermore, we determined the sequence of the doubly spliced YR2 2.1 kb tax cDNA that is expressed in YR2_LTaxSN cells, using RT-PCR with splice-specific primers. We identified both of the G-to-A transitions that characterize the defective provirus, suggesting that transduced wild-type tax transcomplemented the tax-defective YR2 provirus.

Tax protein expression was investigated by Western blotting with anti-Tax specific monoclonal antibodies, and a 34-kDa band was detected in cellular extracts of YR2_LTaxSN and LTaxSN PG13 producing cells but not in the YR2 or YR2_NUNL cells (data not shown). Functional Tax originating from LTaxSN and inactive Tax encoded by the YR2 BLV provirus could not be differentiated by this approach. Intracellular detection of p24 antigen by flow cytometry demonstrated that the YR2_LTaxSN cells produced the capsid protein, suggesting that expression of the BLV provirus was active in the vast majority of the cell population (Fig. 4). p24 expression was stable over the entire period of this study, in which cells were continuously passaged for 12 months.

FIG. 4.

BLV p24 expression in YR2_LTaxSN cells. Flow cytometry analysis of YR2_LTaxSN cells (gray histogram) and YR2_NUNL cells (empty histogram) labelled with the anti-p24 monoclonal antibody mix and stained with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin G is shown.

Reactivation of BLV expression results in production of HTLV-BLV group viral particles.

Reactivation of BLV expression in the YR2_LTaxSN cells could lead to the production of virus particles. Thin-section electron microscopy of YR2_LTaxSN cells (Fig. 5) revealed the presence of a high number of virus-producing cell profiles. The enveloped particles appear somewhat heterogeneous in diameter. Virions contain more-or-less electron-dense, concentrically located cores showing isometric to irregular-angular outlines typical for the HTLV-BLV group of retroviruses (24). No virus particles were detected in either the YR2 or the YR2_NUNL cells (data not shown).

FIG. 5.

Thin-section electron microscopy of YR2_LTaxSN cells. According to morphological criteria, the virus particles represent typical members of the HTLV-BLV group of retroviruses. Magnification, ×90,000.

Efficient in vivo transfer of BLV provirus originating from YR2_LTaxSN cells.

Sheep were intradermally injected with 10⁷ YR2_LTaxSN cells (sheep S17, S18, S19, and S20). In parallel, control animals were injected with the native YR2 cells (negative control) (sheep S11 and S12) or LB155 cells, a BLV-producing bovine cell line (positive control) (sheep S15 and S16). Blood samples were monitored weekly for BLV-specific seroconversion by using an enzyme-linked immunosorbent assay to detect anti-p24 antibodies. No seroconversion was observed in the two control YR2-injected animals. In contrast, 1 month after injection, both the LB155-infected and the YR2_LTaxSN-infected sheep seroconverted. Antibody titers remained positive over time, with the exception of sheep S19, whose antibody levels were no longer detectable at 3 months postinoculation. This animal remained seronegative thereafter. Blood samples from the infected sheep S17, S18, S19, and S20 and from the control animals S11, S12, S15, and S16, collected 4 months postinfection, were analyzed by PCR for the presence of the BLV provirus with T6 and U3 BLV specific primers. As shown in Fig. 6, an 852-bp signal was detected in genomic DNAs of seropositive animals (S15, S16, S17, S18, and S20) but not in the PBL from either sheep S11 or S12 (negative controls) or S19 (a YR2_LTaxSN-infected seronegative animal) (data not shown). This demonstrates that the BLV provirus is present in the PBL following injection of LB155 cells or YR2_LTaxSN cells. Our data indicate that reactivation of BLV expression in the YR2_LTaxSN cells leads to the infection of the PBL of these animals by a functional BLV provirus.

FIG. 6.

PCR amplification of BLV provirus tax sequences in YR2_LTaxSN-infected sheep. Blood samples from YR2_LTaxSN-infected sheep S17, S18, and S20; YR2-infected sheep S11 and S12; and LB155-infected sheep S15 were collected at 4 months postinfection. DNA was prepared as described in Materials and Methods and amplified with the T6 and U3 BLV-specific primers. PCR amplification products were separated by agarose gel electrophoresis and analyzed by Southern blot hybridization with a ³²P-labelled total BLV probe.

BLV proviruses present in peripheral blood from YR2_LTaxSN-injected sheep exhibit recombined tax sequences.

The BLV provirus tax sequences from the YR2_LTaxSN-injected seropositive sheep were isolated, using the B and U3 BLV-specific primers (Fig. 1A). A 1,419-bp sequence spanning the entire tax gene was generated by PCR amplification of genomic DNA from sheep S17, S18, and S20 PBL.

The sheep proviral tax sequences were aligned with those from YR2 BLV tax (variant 2 and defective) and LTaxSN tax (variant 1 and functional), from nucleotide 7247 to 8246 of the BLV genome (numbering according to Sagata et al. [56]) (Fig. 7). The YR2 BLV tax nucleotide sequence differs from that of LTaxSN tax by 29 nucleotides, 13 of which are associated with changes in the deduced amino acid sequence. Eleven of these amino acid substitutions in the YR2 sequence are also present in the infectious YR2-related M344 provirus (variant 2 and functional). The remaining two differences, G-to-A transitions at nucleotides 7924 and 8149, are present in neither the LTaxSN nor the M344 functional proviruses, again leading to the hypothesis that these mutations may play a role in the inactivation of the YR2 provirus.

FIG. 7.

Nucleotide sequence alignment of tax genes from the YR2 BLV provirus, the LTaxSN vector, and the BLV proviruses in YR2_LTaxSN-infected sheep S17, S18, and S20. Only the sequences between nucleotides 7247 and 8246 (from the tax splice acceptor site into the 3′ untranslated region of the provirus) are shown. The numbering is according to Sagata et al. (56). The YR2 and LTaxSN sequences differ by 29 nucleotides (shown in boldface for LTaxSN). Dashes represent areas of nucleotide identity. Each of the proviruses from sheep S17, S18, and S20 shows a unique tax sequence, with differences occurring exclusively at the same 29 positions and originating from either LTaxSN (boldface) or YR2 (lightface). The boxes in the YR2 tax sequence indicate the nucleotides (A₇₉₂₄ and A₈₁₄₉) which differ between YR2 and the infectious YR2-related provirus M344. The position of the critical A₈₁₄₉ mutation in YR2 is indicated with a star. The TGA tax stop codon is shown at position 8170.

Nucleotide sequence alignment revealed that each unique proviral tax gene in S17, S18, or S20 originated from either YR2 or LTaxSN tax sequences apparently in a random pattern, according to the 29 nucleotides which can be used to distinguish the YR2 BLV and LTaxSN tax (Fig. 7). Assuming the possibility that S17, S18, and S20 proviruses resulted from infectious BLV viruses with chimeric functional tax genes, our data suggest that varying tax sequences were derived by homologous recombination between highly conserved YR2 and LTaxSN tax genes in YR2_LTaxSN. As schematized in Fig. 8, at least four homologous recombination events between YR2 and LTaxSN occurred to produce the chimeric tax gene present in BLV proviruses originating from S17, whereas a minimum of two homologous recombinations took place to generate the BLV provirus tax genes present in S18 and S20. The number of recombination sites and their exact locations are unknown. In addition, we observed that both A₇₉₂₄ and A₈₁₄₉ had reverted to G₇₉₂₄ and G₈₁₄₉ in the proviral tax sequences from S17 and S20, whereas only the A₈₁₄₉-to-G₈₁₄₉ reversion was detected in S18. These results clearly demonstrate that a single mutation, the G₈₁₄₉-to-A₈₁₄₉ transition that leads to the E₃₀₃-to-K₃₀₃ amino acid substitution in Tax, can impair the YR2 Tax function and induce the silent phenotype observed in YR2 cells.

FIG. 8.

Schematic representation of the tax gene present in the YR2 BLV provirus, in LTaxSN, and in the BLV proviruses from sheep S17, S18, and S20. Only the nucleotides corresponding to the 29 positions which can be used to discriminate between the YR2 and LTaxSN tax sequences (see Fig. 7) are indicated. These arbitrary units are used to scale the representation. Nucleotides originating from either YR2 or LTaxSN are depicted as hatched or white boxes, respectively. These boxes represent the minimal number of DNA fragments from both origins, since high sequence homology between YR2 and LTaxSN sequences impairs the exact localization of recombination sites. The star indicates the position of the critical mutation in YR2 (G₈₁₄₉ to A₈₁₄₉).

DISCUSSION

BLV proviruses are normally transcriptionally silent in virus-induced ovine B-lymphoid tumors. Here, we report the possibility of reactivating BLV proviral gene expression in a tumor-derived B-cell line by introducing a functional tax gene. These experiments were performed with YR2, a tumor cell line derived from leukemic B cells isolated from a BLV-infected sheep, which contains a single integrated provirus with a silent phenotype. In YR2 cells, even subgenomic BLV RNAs could not be detected by RT-PCR with splice-specific primers, excluding the possibility of a nonproductive RNA pattern such as that observed in latently human immunodeficiency virus type 1 (HIV-1)-infected cell lines, where RNA is produced but selectively spliced to restrict productive virus expression (9). Previous efforts to activate virus expression in this cell line were not successful (reference 63 and unpublished data).

Several molecular mechanisms could be involved in the lack of expression observed in ovine tumor cells, including provirus silencing by negative cellular proteins, a lack of cellular factors required for Tax activity in malignant cells, interference with the host sequences that flank the provirus, hypermethylation of the provirus, or mutations present in the proviral DNA. There is no specific nucleotide integration site for the BLV provirus, although higher-order DNA-chromatin structures may affect the DNA accessibility for proviral integration and subsequent expression. Thus, identical proviruses inserted in different chromosomal locations could vary significantly in their transcriptional activities, suggesting that a major influence is exerted by the integration site (13, 14). Furthermore, LTR-driven transcription has frequently been found to be inactivated by the presence of methylated cytosines at CpG sites, and this effect is reversible by growth with the demethylating agent 5-azacytidine (27, 29). We found that hypermethylation was not essential in silencing the YR2 provirus, since growing YR2 cells with 5-azacytidine did not stimulate virus expression (63). Alternatively, analysis of the tax gene sequence from the YR2 cell line revealed two G-to-A transitions at nucleotides 7924 and 8149, which result in E-to-K changes at amino acids 228 and 303 of YR2 Tax. These mutations were thought to play a role in the silent phenotype, but while they may be critical, additional factors could also be involved in the repression of virus expression in ovine BLV-induced tumors.

To address this issue, we sought to reactivate virus expression from the tax-defective integrated provirus by introducing a transactivation-competent tax gene into YR2 cells by using a retroviral vector gene transfer strategy with GaLV-pseudotyped viral particles (46). We clearly found that following gene transfer, BLV provirus expression can be detected, in terms of both RNA and protein production in YR2_LTaxSN cells. Furthermore, we found, by thin-section electron microscopy, that HTLV-BLV group viral particles are released by these wild-type tax-transduced YR2 cells. These data suggest that transduced wild-type tax gene expression fully reactivated BLV provirus expression and led to the production of BLV viral particles, demonstrating that a Tax defect could play an important role in the silent phenotype observed in BLV-induced B-cell tumors. In addition, (i) the integrated BLV provirus in the YR2_LTaxSN cells, grown continuously in vitro for 1 year, maintained its native mutated tax gene, and (ii) sequence analysis of tax cDNA detected the presence of both the LTaxSN and the YR2 forms of tax, confirming the concomitant expression of the wild-type unspliced and the mutated doubly spliced tax mRNAs, respectively. Taken together, our data strongly suggest that transcomplementation-mediated BLV reactivation occurred in YR2_LTaxSN cells.

Experiments were designed to investigate the in vivo infectious potential of BLV viral particles rescued by LTaxSN in a second-round infection assay in sheep. Animals were injected intradermally with YR2_LTaxSN cells, and, as anticipated, after this injection of a high load of viral protein, all four sheep seroconverted after 1 month. However, while BLV-specific antibody titers rose over time in three animals, in one animal only a weak antibody response was transiently detected between 1 and 2 months postinfection. PCR analysis of blood samples from the three seropositive sheep, collected at 4 months post-infection, revealed the presence of BLV-specific sequences, thus providing evidence for the presence of replication-competent provirus. The short duration of seroconversion by the fourth sheep may have been due to the activation of a transient BLV-specific immune response to viral proteins in the inoculum. BLV-specific PCR was repeatedly negative in blood samples from this animal, strongly suggesting that it never became productively infected by a replication-competent provirus.

Sequence analysis of proviral DNA amplified from the in vivo-infected sheep, S17, S18, and S20, clearly showed that rescued virus was propagated within the animal, as a result of homologous recombination between the LTaxSN- and the YR2-derived tax sequences. Recombination between two different retroviruses occurs when their RNA genomes are packaged together in the same virus particle, creating a heterozygous virion (28). During reverse transcription, recombination occurs frequently by a copy choice mechanism that generates viral DNA that contains genetic information derived from both RNA copies. We suggest that the high sequence homology (97.1%) between the BLV variant 1 (LTaxSN)- and variant 2 (YR2)-derived tax sequences favors crossovers, resulting in the emergence of new recombinant strains. This implies that distinct RNAs (an 8.4-kb BLV-derived RNA for YR2 and a 3.8-kb MoMLV-derived RNA for LTaxSN) can be copackaged, confirming that there is little influence of RNA origin and size on the efficiency of retrovirus recombination (10, 32).

Studies on doubly HIV-1-infected cells have suggested that both the low fidelity of HIV-1 reverse transcriptase and the recombination events occurring among different virus strains play a significant role in the development of the high diversity of HIV-1 (41). In contrast, the sequence analysis of BLV recombinant provirus strains present in YR2_LTaxSN-infected animals at 4 months postinfection indicated that recombination in BLV is not an error-prone process. Furthermore, these proviral sequences were stable in vivo after 1 year postinfection, confirming that the in vivo BLV intrastrain variability is extremely low (75).

Sequence analysis of both the amplified proviral DNA and tax cDNA from YR2_LTaxSN cells, maintained in culture for 1 year, showed that these sequences were identical to the starting material. These data suggest that a major proportion, if not all, of the BLV provirus integrated in the YR2_LTaxSN cells is still defective. In vitro recombination could have occurred through reinfection of the cells with progeny virus containing a copy of each RNA. Using Southern blot analysis, we provided evidence that the original provirus integration site is uniquely present in the YR2_LTaxSN cell line, strongly suggesting that reinfection in vitro did not occur (data not shown). Furthermore, we have never observed in vitro reinfection in any of the BLV tumor-derived cell lines that we have examined (unpublished observations). In addition, there is a substantial body of data with different retroviruses showing that superinfection does not occur (17, 65, 66). Taken together, our data strongly suggest that homologous recombination between the two retroviral sequences took place in vivo. Studies using two different molecularly cloned strains of simian immunodeficiency virus injected simultaneously into separate legs of a monkey have provided clear evidence for genetic recombination between two different retroviral strains in the infected host in vivo (77). Recombination processes in retroviruses have often been described as a strategy for generating viral strains that can escape the host immune system or resist treatment by antiviral drugs. Our in vivo experiments illustrate the ease with which recombinants can appear in an infected sheep and the powerful selection pressure that can be exerted in the host, where the virus can take advantage of the B-cell machinery to survive and generate infectious virus particles. Once a single competent BLV is selected in the animal, it can persist within its host without undergoing further significant mutations.

Retroviral vectors are stably integrated and subsequently expressed in their target cells. These proviruses cannot propagate further due to a lack of viral functions required for replication. This is the basic principle of single-cycle transfer by retroviral vectors. The results presented here demonstrate that LTaxSN-mediated transactivation of the BLV LTR induces the expression of viral genes, thus producing the BLV proteins necessary for encapsidation. Our findings support the idea that vector sequences may be mobilized in their target cell even by endogenous retrovirus proteins from a different retroviral origin, which is viewed as an important safety concern in retroviral gene therapy for the treatment of human disease (43, 48).

The tax sequences of the recombined proviruses derived from YR2_LTaxSN-infected sheep differed from one another, but all three of these chimeric tax genes had reverted the A₈₁₄₉ nucleotide mutation, demonstrating that the E₃₀₃-to-K₃₀₃ amino acid substitution is the only modification required for provirus inactivation in YR2 cells. Little is known about the involvement of the C-terminal region in the functional activity of Tax. Previous studies based on amino acid deletions or insertions in the protein did not identify any specific domain involved in Tax functional activity; however, every modification abolished transactivation, suggesting that its structure is under heavy evolutionary constraints (67). The furthest C-terminal modification artificially introduced into the tax coding sequence was a 12-mer in-frame insertion at amino acid 270. Analysis of fusion proteins containing the Gal DNA binding region and portions of Tax identified a peptide segment encompassing amino acids 157 to 197 suggested to be involved in transcriptional transactivation (72). A putative zinc finger structure, located in the N-terminal region, was shown to be absolutely required for the transactivation function of Tax (69). However, some of the mutants that have lost their transactivating capacity are still capable of acting as immortalizing oncoproteins, indicating that the transactivation and immortalization functions of Tax are independent.

In this paper, we present evidence that a naturally occurring C-terminal E₃₀₃-to-K₃₀₃ amino acid substitution in the Tax protein is responsible for virus inactivation in a BLV-induced B-cell tumor, but further study is required to determine how this mutation affects Tax function. It was previously reported that while deletions are frequent in tumor-derived proviruses, they were not found in the tax gene, and this was interpreted as added evidence of the important role of Tax in B-cell transformation (37). However, our data define mutations in the transactivating protein of BLV as a mechanism for retroviral inactivation during the process of leukemogenesis. The conclusions based on our observations are further supported by recent reports on HIV tat-defective cell lines that describe a similar strategy for HIV postintegration latency (15).

The transactivation ability of the BLV proviral tax gene is critical for successful viral replication, and this has been clearly demonstrated by in vivo studies with cloned proviruses (74). Initial low-level LTR-directed provirus expression is thought to be mediated by a Tax-independent mechanism, involving cellular transcription factors such as CREB and NF-κB, early after infection (1, 7, 40, 73). If the Tax protein, normally turned on by this early mechanism, is unable to transactivate the proviral LTR, then the viral cycle will be abortive.

Tax-mediated transactivation of cellular genes is thought to be essential in the early events of a multistep leukemogenic process. However, the BLV provirus is silent in the tumor cell, suggesting that Tax expression is not required to maintain a transformed phenotype. BLV thus appears to have developed a strategy that allows it to escape the immunosurveillance of the host by switching off Tax and virus expression. This strategy can be viewed as one of the late critical steps in leukemogenesis which is exploited by the virus to provide a growth advantage to the infected B lymphocyte, permitting the ultimate progression to full malignancy.