INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Foamy viruses (FVs) are unique among retroviruses in their establishment of life-long persistent infections without any accompanying pathologies. Infection is characterized by the presence of viral DNA in a large number of organs (9, 42), without detectable levels of viral RNA or protein expression (6, 9, 42, 44). Indeed, viral transcription has been detected only in the oral mucosa of a single infected animal (9). However, virus can be recovered readily by coculturing of infected tissues, peripheral blood, or throat swab specimens with susceptible cell lines (6, 18, 42, 44, 46, 49). Thus, in most locations in vivo, FV replication is latent; however, when the virus is removed from such a context, replication can proceed. In contrast to the in vivo situation, FV replication in vitro can result in either lytic or persistent infection (13, 41, 53). Infection of many cell types in vitro is often accompanied by cytopathic effects (CPE) and rapid cell killing. Since such infections do not mimic the in vivo situation, we sought to develop a tissue culture system in which there is little viral replication. For these studies we used the prototypic human FV (HFV) clone HFV13 (29). HFV has recently been renamed SFVcpz(hu) to more clearly indicate that the original HFV isolate is a chimpanzee FV isolated from a human-derived cell culture (17). It has been previously shown that several human hematopoietic cell-derived lines can be infected with SFVcpz(hu), but with low levels of viral replication and no CPE (33, 53). We examined the role of viral transcription in regulating virus production in these cell lines.

Several characteristics of FV transcription may allow for different types of viral replication, such as lytic and persistent infections. One factor which could be involved in regulating viral replication is the presence of an internal promoter (IP) (27) in addition to the conventional long terminal repeat (LTR) promoter. The IP has low basal activity and drives the expression of the requisite transcriptional transactivator, tas (25). Tas is a DNA binding protein which transactivates both the IP and the LTR promoter (15, 25-27). Interestingly, the Tas protein binds to distinct sequences in the LTR promoter and the IP which share no homology, indicating that Tas may transactivate the two promoters via different mechanisms (8, 20, 23). In addition, Tas has a higher affinity for the IP than for the LTR promoter (20). These facts, coupled with the lack of basal LTR promoter transcription in the absence of Tas, provide a number of possible ways to regulate FV replication. It is generally thought that after infection and integration into a new host cell, the low basal activity of the IP drives the expression of Tas which, due to its higher affinity for its own promoter, drives the expression of additional Tas via a positive-feedback loop (25, 26, 28). Once sufficient levels of Tas are attained, LTR promoter transcription can proceed and viral replication can commence. This bimodal, temporal pattern of transcription could be regulated at (i) Tas-independent, basal transcription of the IP, (ii) Tas-dependent transactivation of the IP, or (iii) Tas-dependent transactivation of the LTR promoter. A better understanding of how FVs achieve persistence in vitro may provide a better understanding of how they persist in vivo.

We have examined the relationship between the SFVcpz(hu) LTR promoter and the IP in a variety of lytic and persistent infections in vitro. We have shown that the SFVcpz(hu) IP is more efficiently transactivated in persistently infected cells, while the LTR promoter is more efficiently transactivated in lytically infected cells. Activation of persistently infected cells with the phorbol ester phorbol 12-myristate 13-acetate (PMA) resulted in greatly enhanced LTR promoter transcription and viral replication. However, mutation of three consensus AP-1 binding sites in the SFVcpz(hu) LTR promoter had little or no effect on lytic replication or PMA-induced viral replication in persistently infected cells. Our findings suggest that the regulation of the two FV promoters is important in determining the outcome of FV infection.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cells and viruses. Virus titers were determined using the previously described FAB indicator cell line (52). Diploid human embryonic lung (HEL) cells (ATCC CCL-137), baby hamster kidney (BHK-21) cells (ATCC CCL-10), and FAB cells were grown in Dulbecco's modified Eagle medium (DMEM) containing 10% fetal bovine serum (FBS) and antibiotics. Human erythroleukemia (H92) cells (ATCC TIB-180), Raji cells (ATCC CRL-2367), U937 cells (ATCC CRL 1593.2), and Jurkat cells (ATCC TIB-152) were grown in RPMI-1640 medium supplemented with 10% FBS. Infection of H92, Raji, Jurkat, and U937 cells by coculturing was performed as previously described (33). Green fluorescent protein (GFP) indicator lines H92-5Lg and Jurkat-5Lg were transduced with retroviral vector LN-5Lg (see below) pseudotyped with vesicular stomatitis virus glycoprotein (VSV-g) as previously described (4). H92 and Jurkat cells were selected in 600 and 1,200 µg of G418 (Life Technologies, Inc.)/ml, respectively.

Plasmids. The infectious molecular clone pSFVcpz(hu)13 was provided by R. Flugel (29). A cytomegalovirus (CMV)-driven SFVcpz(hu) vector, pC-SFVcpz(hu), was generated such that the 5' end of the viral RNA begins at the same nucleotide as in wild-type SFVcpz(hu). The U3 region of the SFVcpz(hu) LTR promoter was excised from pSub1 (3) by digestion with EagI and partial digestion with XbaI. A linker containing EagI and XbaI ends was cloned into pSub1, yielding pCMVsub1. The CMV immediate-early (CMV-IE) promoter was amplified from vector pCR3.0 (Invitrogen) using the forward primer CMVEagI (5'GATCGATCGGCCGGCGCGCGTTGADATTGATTATTG3') and the reverse primer CMVXbaI (5'ATCTAGACTCGAAGGCTTATATAGACCTCCCACCGTACACG3') (positions of introduced restriction enzyme sites EagI and XbaI are underlined). The PCR product was digested with EagI and XbaI and cloned into pCMVsub1. pC-SFVcpz(hu) was generated by subcloning an EagI/SwaI fragment from pCMVsub1 into pSFVcpz(hu)13.

Luciferase reporter assays. BHK-21 cells were seeded at 2 × 10⁴ cells per well in 48-well plates. The following day, cells were transfected with 1 µg of DNA and 2 µl of Fugene (Roche) per well according to the manufacturer's instructions. Each reaction contained 0.3 µg of reporter construct, 0.3 µg of pUC19, and 0.1 µg of CMV--galactosidase vector to monitor transfection efficiency; 0.3 µg of pCMV-tas or an additional 0.3 µg of pUC19 was added to appropriate reactions. After 48 h, lysates were prepared with 250 µl of passive lysis buffer (Promega). Ten microliters of cleared lysate was analyzed with the firefly luciferase system (Promega) and Autolumat LB 953 instrumentation (Berthold). -Galactosidase expression was measured at 420 nm using o-nitrophenyl--D-galactopyranoside (2). Nonadherent cell lines were transfected with DMRIE-C reagent (Life Technologies). For each reaction, 4 µl of DMRIE-C and 250 µl of DMEM were mixed with 2 µl of DNA in 100 µl of DMEM and incubated at 37°C for 30 min. Cells were counted, washed, and resuspended in DMEM at 2 × 10⁶ cells/ml, and 250 µl of cells was added to the DNA-DMRIE-C mixture and incubated for 4 h at 37°C. Then, 500 µl of RPMI-1640 medium supplemented with 22% FBS was added. In some cases, Jurkat cells were treated with 50 nM PMA at 20 h posttransfection. At 48 h posttransfection, cells were pelleted and lysates were prepared with 250 µl of passive lysis buffer. A 25-µl portion of cleared lysate was analyzed for luciferase activity. A second 25-µl portion of lysate was monitored for -galactosidase activity using a Galacto-Light Plus system according to the manufacturer's instructions (Tropix, Inc.).

Western blotting. Western blot analysis was performed essentially as previously described (33). Briefly, Jurkat cells persistently infected with SFVcpz(hu) were plated in T-75 flasks (Falcon) at 5 × 10⁵ cells/ml and treated with 50 nM PMA. At various times, 6 ml of cells was harvested and pelleted by low-speed centrifugation. Lysates were prepared with 250 µl of Ab buffer, and genomic DNA was sheared by passage through a 23-gauge needle. Lysates were cleared by high-speed centrifugation before loading on sodium dodecyl sulfate (SDS)-10% polyacrylamide gels and detection using enhanced chemiluminescence (Amersham).

RIPA. BHK-21 cells (2 × 10⁶) were plated on 10-cm dishes, infected with SFVcpz(hu) at a multiplicity of infection (MOI) of 0.5, and grown until extensive syncytium formation was observed (about 40 h). Growth medium was removed and replaced with 5 ml of DMEM lacking cysteine and methionine but containing 400 µCi of 35-S Express protein label (NEN). Persistently infected H92, Raji, Jurkat, and U937 cells (5 × 10⁶) were harvested and labeled as described above. After 4 h, cells were harvested in 1 ml of Ab buffer containing 2 µg of aprotinin/ml, 2 µg of leupeptin/ml, 1 µg of pepstatin A/ml, 0.57 mM phenylmethylsulfonyl fluoride (Sigma), and 1 µg of Pefablock (Roche)/ml (protease inhibitors). Genomic DNA was sheared and cleared by centrifugation. Lysates were precleared by incubation with 50 µl of protein A-Sepharose for 1 h at 4°C. BHK-21 cell lysate (100 µl) and 600 µl of H92, Raji, Jurkat, or U937 cell lysate were immunoprecipitated with 4 µl of anti-Tas antiserum and 2 µl of anti-Gag antiserum overnight at 4°C in a total volume of 1 ml. The following day, 100 µl of protein A-Sepharose (75 mg/ml in phosphate-buffered saline) was added to each reaction and mixed for 4 h at 4°C. Reactions were washed twice with radioimmunoprecipitation (RIPA) buffer (10 mM Tris, 150 mM NaCl, 1% Nonidet P-40, 1% deoxycholic acid, 0.1% SDS, 0.5% aprotinin [pH 7.4]), once with high-salt buffer (10 mM Tris, 2 M NaCl, 1% Nonidet P-40, 1% deoxycholic acid [pH 7.4]), and again with RIPA buffer. Samples were then separated on SDS-10% polyacrylamide gels and visualized by autoradiography. Quantification was done by phosphorimaging with ImageQuant software.

RPA. An RNase protection assay (RPA) was performed using a Direct Protect kit (Ambion). Jurkat cells were treated as in the Western blotting procedure, and total nucleic acid was isolated in 400 µl of lysis solution. A 25-µl portion of cleared lysate was used in each hybridization. pBS-IP was linearized with BamHI, and T7 RNA polymerase was used to generate a 305-bp ³²P-dUTP-labeled probe and, after RNase protection, 246- and 110-bp products indicating LTR promoter and IP transcripts, respectively. Plasmid pGAPDHBS was linearized with HindIII, and T7 runoff transcripts produced an unprotected 590-bp product and, after RNase protection, a protected 546-bp product.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The viral LTR promoter and IP are differentially regulated in lytic and persistent infections in vitro. We hypothesized that differential regulation of the LTR promoter and IP may play a role in determining lytic or persistent infection in vitro. Lytic SFVcpz(hu) infection is generally observed in fibroblast-derived cells, such as BHK-21 cells and HEL cells, and is characterized by titers of >10⁵/ml, CPE, extensive cytoplasmic vacuolation, and cell death. In contrast, persistent infection by SFVcpz(hu) generally occurs in but is not limited to cells of human hematopoietic lineages, in which there are few or no adverse effects on cell replication. Titers vary widely from >10³/ml in the human erythroleukemia cell line H92 and the Burkitt's lymphoma-derived Raji cell line to <10²/ml in the monocytic lymphoma-derived U937 cell line and the T-cell lymphoma-derived Jurkat cell line (Table 1). In a previous study (53), Raji cells were unable to be productively infected, but in the current study, using coculture methods, we were able to efficiently infect these cells.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the SFVcpz(hu) DNA genome, LTR promoter, and IP. Open arrows, LTR promoter and IP; closed arrows, AP-1 consensus site; vertical bands, Tas binding site; cross-hatched area, Tas-responsive element. (A) 11,955-bp SFVcpz(hu) DNA genome and known ORFs. (B) SFVcpz(hu) 3'LTR promoter. U3, R, and U5 regions, TATAA box, region of bel2 which overlaps U3, and locations of the three consensus AP-1 binding sites are indicated. (C) IP. The env and overlapping tas ORFs are shown. IP splice donor (SD) and splice acceptor (SA) sites, membrane-spanning domain (MSD), and locations of PCR primers (black arrows) used to amplify the IP are noted.

View larger version (37K): [in this window] [in a new window]   FIG. 2.   Transactivation of the LTR promoter and IP by transient transfection. (A) Schematic diagrams of constructs used in transfections. (B to E) Transactivation results. Control, promoterless control luciferase vector; LTR, LTR-luc; LTR+Tas, LTR-luc plus CMV-tas; IP, IP-luc; IP+Tas, IP-luc plus CMV-tas. The fold change in luciferase units (LU) relative to the value for the LTR promoter is shown below each column. (B) Jurkat cells, 1 LU = 131 raw LU (RLU). (C) Raji cells, 1 LU = 914 RLU. (D) H92 cells, 1 LU = 843 RLU. (E) BHK-21 cells, 1 LU = 39,656 RLU. All values are based on at least three independent experiments and are reported as the mean and standard error of the mean.

View larger version (65K): [in this window] [in a new window]   FIG. 3.   Promoter activity in lytic and persistent infections. RIPA was performed with the indicated uninfected and infected cell types using a mixture of anti-Gag and anti-Bet polyclonal antisera. Black arrows indicate 74- and 70-kDa Gag proteins. The grey arrow indicates the 52-kDa Bet protein. Locations of molecular mass markers are shown on the right. Immunoprecipitated Gag and Bet proteins were quantified by phosphorimaging after subtraction of the background from adjacent uninfected-cell lanes. The resulting values were then normalized to total 35S incorporated, and the ratio of Bet to Gag proteins was calculated (shown below the lanes). Values derived for Bet are considered IP activity, while values for Gag are considered LTR promoter activity. Inf., infection.

Fusion of HEL cells with persistently infected cells allows for LTR promoter transcription. To address whether LTR promoters integrated in persistently infected cells are capable of efficient transcription given the appropriate cellular environment, fusion experiments were performed with either H92 or Jurkat cells and infected HEL cells. Uninfected H92 and Jurkat cells were transduced with the murine leukemia virus-based vector LN-5Lg, which expresses GFP under the control of the SFVcpz(hu) LTR promoter (Fig. 4A). G418-resistant populations were obtained and named H92-5Lg and Jurkat-5Lg, respectively. These cells were then infected by coculturing with lytically infected HEL cells. During coculturing with infected HEL cells, interaction of the SFVcpz(hu) receptor on the H92-5Lg or Jurkat-5Lg cells with the SFVcpz(hu) envelope expressed on the surface of the HEL cells permits fusion between the two cell types. During this process, we observed that the infected HEL cells expressed large amounts of GFP, while very few H92-5Lg and Jurkat-5Lg cells expressed GFP (Fig. 4F and G). This observation indicates that upon fusion with infected HEL cells, the genome of H92-5Lg or Jurkat-5Lg cells enters an environment suitable for LTR promoter transcription, resulting in GFP expression.

View larger version (112K): [in this window] [in a new window]   FIG. 4.   GFP expression of H92-5Lg or Jurkat-5Lg cells upon coculturing with lytically infected HEL cells (FV-HEL cells). PHACO, phase-contrast light microscopy. GFP, fluorescent imaging of GFP protein expression. All images in panels B to G were captured at a magnification of ×200. (A) Schematic diagram of the LN-5Lg vector used to transduce H92 and Jurkat cells. (B and E) PHACO and GFP images, respectively, of FV-HEL cells cocultured with naive H92 cells. (C and F) PHACO and GFP images, respectively, of FV-HEL cells cocultured with H92-5Lg cells. (D and G) PHACO and GFP images, respectively, of FV-HEL cells cocultured with Jurkat-5Lg cells. egfp, enhanced GFP.

PMA treatment enhances viral replication in persistently infected cells. We have shown that in persistently infected cells, despite significant IP activity, there is an unexpected lack of LTR promoter activity. Based on these data, we hypothesized that factors which were necessary for efficient LTR promoter transactivation in persistently infected cells were missing. To examine this hypothesis, we treated persistently infected cells with a variety of cell activators in an effort to supply whatever factors might be necessary for efficient LTR promoter activity. We found that of the cell activators used, PMA had the most profound effect on viral titers (data not shown). Jurkat cells were significantly more responsive to PMA stimulation than the other cell types (Table 2). In Jurkat cells, PMA treatment resulted in syncytium formation and cell death, indicating a switch from persistent to lytic infection (Fig. 5A to C). The molecular clone of SFVcpz(hu) used in all of our experiments has a 646-bp deletion in U3 relative to the original HFV isolate or SFVcpz (17, 45). PMA treatment of Jurkat cells infected with SFVcpz resulted in increases in titers similar to those seen with SFVcpz(hu) (data not shown), indicating that the deleted region of the SFVcpz(hu) LTR promoter has no effect on PMA induction in Jurkat cells.

View larger version (52K): [in this window] [in a new window]   FIG. 5.   PMA induction of SFVcpz(hu) in Jurkat cells. (A) Light microscopy of untreated SFVcpz(hu)-infected Jurkat cells. (B) SFVcpz(hu)-infected Jurkat cells treated with 50 nM PMA. Arrows indicate multinucleated syncytia. (C) Uninfected Jurkat cells treated with 50 nM PMA. (D) PMA induction of the SFVcpz(hu) LTR promoter and IP in Jurkat cells, as measured by transient transfection. The same constructs as those shown in Fig. 2A were used. Values shown below the white columns represent the fold change in luciferase units (LU) relative to the value for the LTR promoter with no PMA stimulation; 1 LU = 131 raw LU (RLU). Values shown below the grey columns represent the fold change in LU relative to the value for the LTR promoter with PMA stimulation; 1 LU = 102 RLU. All values are based on at least three independent experiments and are reported as the mean and standard error of the mean.

View larger version (36K): [in this window] [in a new window]   FIG. 6.   Characterization of PMA induction in SFVcpz(hu)-infected Jurkat cells. (A and C) Naive or persistently infected Jurkat cells were treated with 50 ng of PMA/ml, and cells were harvested at the indicated times (h.p.i., hours postinfection; h.p.t., hours posttreatment. (A) Western blot analysis of protein expression using anti-Gag (top) or anti-Tas or -Bet (bottom) antisera. Black arrows, Gag; grey arrow, Bet; white arrow, Tas. (B) Schematic diagram showing the IP region and the RNase protection probe used to distinguish between LTR promoter and IP transcripts. (C) Unprotected probes are shown in lanes 1 and 2. Lanes 5 to 14 show RPA analysis at the indicated times after PMA treatment. White arrow, protected 546-bp GAPDH transcripts; black arrow, protected 246-bp LTR promoter transcripts; grey arrow, protected ~110-bp IP transcripts. (D) Relative RNA levels determined by phosphorimaging quantitation of LTR promoter and IP transcripts in panel C, normalized to GAPDH expression. The x axis shows hours posttreatment.

Consensus AP-1 binding sites are not required for LTR promoter transactivation or lytic viral replication. The presence of three consensus AP-1 binding sites in the SFVcpz(hu) LTR (Fig. 1B) suggests a possible role for AP-1 in modulating LTR promoter-based transcription. The three AP-1 binding sites found in the SFVcpz(hu) LTR promoter were previously shown by Maurer et al. to specifically bind recombinant c-Jun-v-Fos complexes (32). These authors also demonstrated that HeLa and BHK-21 cell extracts, both of which contain high levels of AP-1 family members (30, 54), were able to bind an LTR promoter fragment containing AP-1 binding sequences but not an LTR promoter fragment with the AP-1 binding sequences mutated. Given that cells such as BHK-21, which undergo lytic infection, express high levels of AP-1 (54) and that cells such as Jurkat express little or no AP-1 (21, 30), we were interested in the role of the three AP-1 binding sites in regulating lytic and persistent infections. Vector pC-SFVcpz(hu)-AP1 is expressed from the CMV-IE promoter, which was used to replace the U3 region of the 5' SFVcpz(hu) LTR promoter (Fig. 7A). The 3' LTR promoter contains specific mutations in all three AP-1 binding sites without disturbing the overlapping bel2 ORF. The CMV-IE promoter directs the initial expression of the viral genome upon transfection, but after a single round of reverse transcription, the 3' LTR promoter containing the mutated AP-1 sites is copied to the 5' end of the provirus. Thus, subsequent rounds of viral expression are mediated by the mutated LTR promoter. The use of the CMV-IE promoter also obviates the possibility that recombination will result in viruses with unmutated LTR promoters.

View larger version (24K): [in this window] [in a new window]   FIG. 7.   Replication kinetics and LTR promoter transactivation of SFVcpz(hu)-AP1 in BHK-21 cells. (A) pC-SFVcpz(hu)-AP1 vector. The U3 region of the 5' LTR promoter is replaced with the CMV-IE promoter (grey box). Three AP-1 binding sites were mutated in the 3' LTR promoter. See Materials and Methods for details. (B) BHK-21 cells were infected at day 0 at an MOI of 0.1. Titers of SFVcpz(hu) and SFVcpz(hu)-AP1 were measured at the indicated times postinfection (p.i.). Results are reported as the mean and standard deviation. (C) Transactivation of wild-type SFVcpz(hu) and mutated SFVcpz(hu) LTR promoters by Tas. Luciferase reporter constructs were constructed, and experiments were performed as described in the legend to Fig. 2. LTR, wild-type SFVcpz(hu) LTR promoter; LTR1, SFVcpz(hu) LTR promoter with the first AP-1 binding site mutated; LTR23, SFVcpz(hu) LTR promoter with the second and third AP-1 binding sites mutated; LTR123, SFVcpz(hu) LTR promoter with all three AP-1 binding sites mutated.

AP-1 binding sites are not required for PMA-mediated induction of viral replication in persistently infected Jurkat cells. In contrast to the situation in lytically infected cells, virus production in persistently infected Jurkat cells is greatly augmented upon treatment with PMA (Table 2) as well as after cross-linking with anti-CD3 and anti-CD28 monoclonal antibodies (data not shown). We speculated that the presence of the three AP-1 binding sites in the SFVcpz(hu) LTR promoter may be important for viral replication only in specific situations, such as T-cell activation. Many proteins bind AP-1 consensus sequences; these include family members such as c-Jun and c-Fos as well as AP-1-related proteins, such as CREB. In resting T cells, c-Jun is constitutively expressed, but the other AP-1 family members, such as c-Fos, are absent (19). Since phorbol ester treatment and T-cell activation by anti-CD3 and anti-CD28 monoclonal antibodies are known to increase the levels of AP-1 family members dramatically (19, 21), we wanted to determine whether AP-1 binding sites are necessary for the increased viral replication observed in such circumstances.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the current study, we examined promoter activity in a number of persistently infected cell lines because persistent infections more closely approximate FV replication in vivo than do lytic infections of fibroblast-derived cell lines. Viral replication is significantly lower in these cells, particularly Jurkat T cells, than in lytically infected cells. We demonstrate here that LTR promoter transcription is more robust in lytically infected cells, while IP transcription is more robust in persistently infected cells. We show that SFVcpz(hu) replication and transcription in Jurkat cells can be enhanced greatly by PMA treatment. We also demonstrate that mutation of three consensus AP-1 binding sites in the viral LTR promoter has little effect on lytic, persistent, or PMA-enhanced replication. Taken together, our data suggest that the lower replication in persistently infected cells is due to a lack of an LTR promoter-specific transcription factor(s) in these cells.

Natural, experimental, and zoonotic FV infections are life-long, persistent infections without any accompanying pathologies. Although in vitro FV infection can lead to the accumulation of large numbers of integrated proviruses in individual cells (33), no FV-associated cancers have ever been reported. The lack of enhancer elements in the SFVcpz(hu) LTR promoter and the absence of basal LTR promoter transcription in persistently infected cells could account for this observation. One defining characteristic of in vivo FV infection is the lack of detectable viral replication despite the ability to recover virus by coculturing of infected tissues with susceptible cells in vitro (6, 7, 9, 11, 38, 41, 42, 53). The reason for this characteristic is unknown but is likely to be complex. Viral replication may be limited by aspects of the innate immune response, such as an interferon response, which is known to dramatically inhibit FV replication in vitro (10, 39, 40). Gamma interferon produced from phorbol ester-activated peripheral blood mononuclear cells strongly inhibits FV replication (10). Thus, it is interesting that PMA treatment of infected Jurkat cells dramatically induced viral replication (Table 2), despite the fact that such treatment also induced gamma interferon expression (14, 48). Furthermore, inhibition of gamma interferon in PMA-treated cells did not result in increased viral titers (data not shown). FV infection is also known to elicit a robust humoral response (1, 16, 34, 42, 46, 47), but its role in limiting viral replication is unknown. There are no published reports on the role of a cell-mediated immune response in limiting FV replication in vivo.

Apart from host-specific factors which could limit FV replication in vivo, several studies have suggested a number of mechanisms for achieving persistent infection in vitro. In most investigations, the Bet protein was thought to be a key player. One potential role for Bet in the regulation of viral infection was suggested by Bock et al., who found that overexpression of Bet could prevent infection by SFVcpz(hu) (5). Bet arises from a spliced message comprised of the first 88 amino acids of Tas and the entire bel2 ORF (35). Normally, the spliced bet message is derived from the IP (27), but at some frequency, LTR promoter transcripts are apparently spliced using the bet splice donor and acceptor sites, resulting in full-length SFVcpz(hu) genomes which can transcribe only bet and not tas (43). These defective genomes, termed SFVcpz(hu)tas, have been suggested to play a role in mediating persistent infection in cells which normally undergo cytolytic infection (41). Cells harboring SFVcpz(hu)tas genomes produce large quantities of Bet upon infection with wild-type virus, providing an interesting link between Bet expression and the maintenance of persistent infection. It has been suggested that SFVcpz(hu)tas acts as a defective interfering virus (22, 41, 43). The presence of large amounts of this form of FV in infected animals (9) and the accumulation of this viral form during experimental infection of rabbits (42) indicate that SFVcpz(hu)tas may play an important role in FV persistence in vivo. Deletion forms of FV are also abundant during persistent infection of Dami megakaryocytic cells (51). Interestingly, in that study, FV replication could be stimulated by treatment of cells with 5-iodo-2'-deoxyuridine, indicating a possible role for promoter methylation in persistent infection. We were unable to induce FV replication in infected Jurkat cells by treatment with either 5-azacytidine or trichostatin A (data not shown). This result indicates that at least in our system, promoter inactivation by methylation or histone acetylation is not a factor in persistent infection. Furthermore, in contrast to work with systems which support lytic replication, it was previously found that in infected H92 cells, there was no evidence for a role of SFVcpz(hu)tas in persistent infection (33, 53).

Our work suggests that differential regulation of the two FV promoters may help explain persistent infection in vitro. Transcription factors which act in concert with Tas to mediate LTR promoter and IP transcription have not been identified, but detailed analyses of the SFVcpz(hu) LTR promoter and IP have identified Tas binding sites and a number of Tas-responsive elements within these promoters (Fig. 1B and C) (8, 20, 23). The only putative transcription factor binding sites within the SFVcpz(hu) and SFVcpz LTR promoters are two Ets-1 sites and three perfect consensus AP-1 binding sites (Fig. 1B) (8, 31, 37, 45). DNase footprint analysis clearly showed that these three AP-1 sites were occupied when cell extracts from HeLa and BHK-21 cells were used, but mutation of these three sites showed that they are dispensable for Tas-mediated transactivation in cells which normally undergo lytic infection (23, 32). The authors also noted a small, two- to threefold increase in phorbol ester-mediated LTR promoter transactivation that was obviated by mutation of the AP-1 sites (32). We observed similar levels of phorbol ester stimulation in our lytically infected cells but much greater effects in cells which support persistent infection (Table 2). Interestingly, our data indicate that a virus lacking all three AP-1 sites replicates like the wild-type virus in BHK-21 cells and is induced by PMA in Jurkat cells. Although the AP-1 sites in the SFVcpz(hu) LTR promoter may not be important in PMA-mediated induction, additional transcription factors upregulated by PMA treatment may augment the increases in transcription that we observed in this study.

Experiments with protein synthesis inhibitors indicated that de novo protein synthesis is required for LTR promoter transcription following PMA treatment (data not shown). Further evidence that transcription factors other than AP-1 mediate PMA induction of the SFVcpz(hu) LTR promoter arises from the observation that the kinetics of Fos and Jun induction in PMA-treated Jurkat cells are not consistent with the kinetics of LTR promoter transcription. Fos and Jun are transcribed within 1 h after PMA treatment (21), but RPA analysis showed that LTR promoter transcription is not observed until 8 to 12 h posttreatment. RPA analysis also showed substantial PMA induction of the IP, although there are no consensus AP-1 sites near the IP. This result does not exclude the possibility that the AP-1 sites in the LTR promoter enhance transcription from the IP. Lochelt et al. demonstrated that the activity of the IP is greater when the SFVcpz(hu) LTR promoter is present in cis (25).

The idea that different transcription factors may mediate LTR promoter and IP transcription is directly supported by experiments with LTR promoter and IP expression constructs from SFVagm (36). In that study, Renne et al. (36) demonstrated through promoter competition experiments that the two promoters in SFVagm are regulated by different mechanisms. In Fig. 8 we propose a similar model to explain the differences in transcription in persistent and lytic infections. In this model, LTR promoter- and IP-specific transcription factors (Tf-LTR and Tf-IP, respectively) are abundant in lytically infected or PMA-treated cells, allowing for transcription from both promoters (Fig. 8A). In contrast, in persistently infected cells, only Tf-IP are present, while Tf-LTR are limiting (Fig. 8B). The missing factors required for LTR promoter transcription in persistently infected cells may be supplied when a cell is activated in vivo. T-cell receptor stimulation of SFVcpz(hu)-infected Jurkat cells with anti-CD3 and anti-CD28 monoclonal antibodies resulted in titer increases comparable to those observed with PMA treatment (data not shown). Because T cells are likely targets for FV infection (7, 50), activation of latently infected T cells provides an intriguing model for the maintenance of persistence in vivo. Activation of latently infected T cells may provide a small burst of viral replication, thereby permitting infection of adjacent resting T cells and further dissemination throughout the host. Such small, transient episodes of viral replication could account for the inability to detect viral replication in infected hosts.

View larger version (28K): [in this window] [in a new window]   FIG. 8.   Models of persistent and lytic SFVcpz(hu) infection in vitro. (A) Lytically infected or PMA-treated cells are characterized by basal IP transcription and Tas production. Newly synthesized Tas, in combination with unknown Tf-IP, drives the expression of additional Tas. LTR-promoter based transcription is present due to the availability of a distinct set of Tf-LTR. (B) Persistent infection. IP-based transcription mirrors that in panel A, but LTR promoter transcription is absent due to the unavailability of Tf-LTR.

In summary, we propose that distinct sets of transcription factors mediate the relative strengths of the two SFVcpz(hu) promoters in lytic and persistent infections. Our current work clearly shows that the relative level of Bet expression is higher in persistent infections than in lytic infections. Bet was the only protein detectable in infected Jurkat cells. Thus, while in lytic infection SFVcpz(hu)tas may be critical in skewing expression toward excess Bet, in persistent infection this can be accomplished by differential promoter regulation. We are currently interested in the identification of the sets of transcription factors which regulate LTR promoter and IP transcription.