Interaction of the single-stranded DNA-binding protein Puralpha with the human polyomavirus JC virus early protein T-antigen.

Large T-antigen, the major regulatory protein encoded by polyomaviruses, including Simian Virus 40 (SV40) and JC virus (JCV), is a multifunctional phosphoprotein that is involved in many viral and cellular events. In addition to its integral role in viral replication and cellular transformation, T-antigen also regulates transcription of both viral and cellular genes. In particular, the viral late promoter has been used as a model for the analysis of T-antigen-mediated transcriptional activation. Earlier studies have demonstrated that the cellular protein Puralpha is able to attenuate the transcriptional activity of JCV T-antigen. We investigated the mechanism whereby Puralpha affects T-antigen function. Co-immunoprecipitation studies demonstrated that Puralpha and JCV T-antigen associate in vivo, and glutathione S-transferase affinity binding assays revealed that these two proteins interact in vitro. Moreover, we localized the sequences of Puralpha that are important for the interaction between Puralpha and JCV T-antigen. In addition, we demonstrated that Puralpha interacts with the SV40 T-antigen. Transient transfection studies demonstrated that Puralpha and JCV T-antigen interact functionally as well. More specifically, Puralpha and a deletion mutant that interacts with T-antigen attenuated T-antigen-mediated transcriptional activation. A Puralpha deletion mutant that is unable to interact with JCV T-antigen, however, was found to be incapable of abrogating JCV T-antigen transactivation. Taken together, these data demonstrate that Puralpha and T-antigen interact both physically and functionally and that this interaction modulates T-antigen-mediated transcriptional activation. The implication of these findings with respect to the cellular role of Puralpha is discussed.

Polyomaviruses have proven to be powerful tools in the identification of key cellular regulatory proteins and the mechanisms that underlie their biological activities. The dependence on cellular machinery for a successful infection and the alterations in cellular processes promoting optimal conditions for viral multiplication have contributed to the utility of these viruses in identifying cellular regulatory proteins and pathways (1). In particular, studies of the major regulatory protein encoded by polyomaviruses, large T-antigen (T-antigen), have provided major insights into cellular biochemical activities including regulation of viral DNA replication, transcriptional activation of both viral and cellular promoters, and cellular transformation (2).
The human polyomavirus, JC virus (JCV), 1 is the etiologic agent of the fatal subacute human neurodegenerative disease progressive multifocal leukoencephalopathy (reviewed in Ref. 3). Unlike SV40 and other polyomaviruses, JCV has a narrow tissue tropism. In immunocompromised patients, JCV lytically infects oligodendrocytes, the myelin-producing cells in the central nervous system, and in tissue culture, JCV propagates efficiently only in primary human fetal glial cells. Studies from our and other laboratories have indicated that the cell-specific activation of the JCV promoter is determined at the level of viral gene transcription and requires multiple cellular proteins present in glial cells (reviewed in Ref. 4). One such cellular factor that influences JCV is Pur␣.
Pur␣ is a 322-amino acid sequence-specific single-stranded DNA-binding protein that has been implicated in the control of both DNA replication and transcription. There are several lines of evidence supporting a role for Pur␣ in DNA replication. Pur␣ was initially characterized as a HeLa cell nuclear protein that binds a sequence element, called the PUR element, adjacent to a region of stably bent DNA 1.6 kilobases upstream of the human c-myc gene (5,6). This element is near the center of a region implicated as an initiation zone for chromosomal DNA replication. Moreover, PUR elements are present at several eukaryotic origins of DNA replication (5). Pur␣ has also been shown to interact with viral origins of DNA replication. Pur␣ has been shown to bind the JCV and bovine papillomavirus origins of replication (7,8). Although the biological activity of Pur␣ on eukaryotic and bovine papillomavirus replication is yet undescribed, Pur␣ has been shown to inhibit JCV DNA replication in glial cells (7).
Pur␣ has also been implicated in control of gene transcription involving both viral and cellular promoters. Pur␣ has been shown to activate several promoters, including the JCV early gene promoter (9), the human immunodeficiency virus type I (10), the myelin basic protein promoter (11), and the neuronspecific FE65 gene promoter (12). In addition, Pur␣ has also been implicated in the expression of the neuronal nicotinic acetylcholine receptor gene promoter (13), the single-stranded cAMP response element (14), and the vascular smooth muscle ␣-actin gene promoter (15).
In addition to its role in transcription and replication, there is evidence suggesting that Pur␣ is involved in the control of cell growth and proliferation. Johnson et al. (16) have demonstrated that Pur␣ binds the hypophosphorylated form of the human retinoblastoma tumor suppressor gene product pRb.
Although the functional consequence of the interaction between these two proteins is yet undescribed, the selective interaction between Pur␣ and the hypophosphorylated form of a protein intimately involved in cell cycle progression suggests a potentially crucial role in the cell cycle.
With respect to JCV, we have previously demonstrated that Pur␣ and T-antigen exert antagonistic effects on each other's transcriptional activity on the JCV promoter (9). Like other polyomaviruses, the genome of JCV can be divided into an early region, which is expressed prior to the onset of viral DNA replication and encodes the viral regulatory proteins (the Tantigens), and the late region, which is expressed after viral DNA replication and encodes the viral capsid proteins (VP1, VP2, and VP3) (reviewed in Ref. 17). The noncoding regulatory region, which contains the origin of viral DNA replication and the transcriptional control region, lies between the two coding regions. This transcriptional control region is composed of a bidirectional promoter/enhancer, with the JCV early promoter controlling expression of the early genes and the JCV late (JCV L ) promoter directing transcription of the late genes. Pur␣ has been shown to activate the JCV early promoter, and coexpression of JCV T-antigen decreases this Pur␣-induced increase in the level of JCV early gene transcription (9). Likewise, Pur␣ has been shown to abrogate JCV T-antigenmediated transcriptional activation of the JCV L promoter (9).
In this report, we examine the mechanism underlying the ability of Pur␣ to abrogate JCV T-antigen-mediated transactivation of JCV L . Using co-immunoprecipitation and in vitro affinity chromatography assays, we demonstrate that Pur␣ associates with JCV T-antigen. In addition, we localize the region of Pur␣ that is involved in this interaction. Transient transfection studies demonstrate that Pur␣ and a Pur␣ deletion mutant that retains the ability to interact with T-antigen abrogate T-antigen-mediated transcriptional activation of the JCV L promoter. A mutant Pur␣ that cannot interact with JCV T-antigen, however, is unable to attenuate JCV T-antigenmediated transactivation of the JCV L promoter. Taken together, these data suggest that Pur␣ and JCV T-antigen interact with one another and that this interaction modulates T-antigen-mediated transcriptional activation of the late promoter of JC virus.

MATERIALS AND METHODS
Cell Culture-HJC-15b cells (4) were derived from a JC virus-induced hamster brain tumor (18) and were grown in Dulbecco's modified Eagle's medium supplemented with 5% heat-inactivated fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin. U-87 MG cells (ATCC HTB-14), a human glioblastoma cell line, and SVG cells, human fetal astroglial cells immortalized with SV40 T-antigen (19), were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and the same antibiotics. All cells were maintained at 37°C in a humidified atmosphere containing 7% CO 2 .
The plasmid pJCV L ⌬kB-CAT contains the HindIII-PvuII fragment containing the control region of the MAD-1 strain of JCV cloned into the late orientation upstream of the chloramphenicol acetyltransferase (CAT) gene as described previously (20).
Immunoprecipitations and Immunoblot Analysis-The JCV T-antigen-expressing cell line HJC-15b was transfected with 30 g of pEBV, pEBV-Pur␣, or pEBV-LacZ. Forty-eight h after transfection, cells were washed three times in 1ϫ PBS and lysed in lysis 150 buffer (LB 150) containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, 50 mM sodium fluoride, 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, and 1 g/ml aprotinin for 20 min on ice. Lysates were scraped and collected into microcentrifuge tubes, vortexed briefly, and microcentrifuged for 20 min at 4°C. For immunoprecipitations, 100 g of lysate was incubated with normal mouse serum (preimmune), a monoclonal antibody against SV40 T-antigen (pAb416) (Oncogene Sciences), or a monoclonal antibody against the histidine epitope tag (anti-T7) (Novagen) for 1 h at 4°C. Immune complexes were precipitated with 20 l of protein A-Sepharose (Amersham Pharmacia Biotech), washed 5 times with 1 ml of LB 150, eluted by boiling in Laemmli buffer, and separated on SDS-polyacrylamide gels. Proteins were transferred to supported nitrocellulose membranes (Schleicher and Schuell) in Western transfer buffer (192 mM glycine, 25 mM Tris base, and 20% methanol). For Western blot analysis, the membranes were blocked for 30 min in 10% nonfat dry milk in PBS-T (1ϫ PBS, 0.1% Tween-20) and then incubated for 1 h with a 1:1000 dilution of anti-SV40 T-antigen or anti-T7 antibodies at room temperature. After washing, the membranes were incubated for 1 h with a horseradish peroxidase-linked goat anti-mouse antibody at room temperature. Antibody detection was achieved by ECL according to the manufacturer's recommendations (Amersham Pharmacia Biotech).
GST Affinity Chromatography Assays (GST Pulldown)-HJC-15b cells expressing JCV T-antigen and SVG cells expressing SV40 Tantigen were lysed in LB 150 buffer for 20 min on ice. Lysates were scraped and collected into microcentrifuge tubes, vortexed briefly, and microcentrifuged for 20 min at 4°C. Two hundred g of whole cell extract was incubated with 5 g of GST or GST-Pur␣ fusion proteins coupled to glutathione-Sepharose beads in 300 l of LB 150 buffer for 1 h at 4°C with continuous rocking. After the incubation, the beads were pelleted and washed five times with LB 150 buffer. Bound proteins were eluted with Laemmli sample buffer, heated to 95°C for 10 min, separated by SDS-PAGE, and analyzed by immunoblot analysis for T-antigen. Ten g of crude extracts was loaded on the gels as migration controls. Alternatively, 8 l of 35 S-labeled in vitro translated Pur␣ was incubated with 5 g of GST or GST JCV T-antigen fusion proteins immobilized on glutathione-Sepharose beads in 300 l of LB 150 buffer plus 1 g/l bovine serum albumin for 1 h at 4°C with continuous rocking. After the incubation, the beads were pelleted and washed five times with LB 150 buffer. Bound proteins were eluted with Laemmli sample buffer, heated to 95°C for 10 min, separated by SDS-PAGE, and analyzed by fluorography for the presence of Pur␣. One-twentieth of the input reaction was loaded as a migration control.
In Vitro Transcription and Translation-35 S-Labeled Pur␣ was syn-thesized in vitro from XbaI linearized pCDNA3-Pur␣ using the TNTcoupled transcription-translation wheat germ extract (Promega) following the manufacturer's instructions.
Protein Purification-GST fusion proteins were expressed and purified as described previously (21). Briefly, bacteria were grown overnight at 37°C in Luria Bertani (LB) medium supplemented with 100 mg/liter ampicillin. The following morning, the cells were diluted 1:10 in fresh LB medium, grown to an absorbance at 595 nm of 0.6 -0.7, and induced for 2 h at 37°C with 0.5 mM isopropyl-␤-D-thiogalactopyranoside. Cells were pelleted at 6,500 ϫ g at 4°C; resuspended in NETN buffer (20 mM Tris, pH 8.0, 100 mM NaCl, 1 mM EDTA, and 0.5% Nonidet P-40) containing 1 g/ml leupeptin, 1 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride; and sonicated on ice. The bacterial lysate was centrifuged at 40,000 ϫ g at 4°C to remove insoluble material. Glutathione-Sepharose beads (Amersham Pharmacia Biotech) were added to the supernatant, and binding of the GST fusion proteins was allowed to occur for 3 h at 4°C. Beads were pelleted and washed three times with 50 volumes of NETN buffer each time. The integrity and purity of the GST fusion proteins were analyzed by SDS-PAGE followed by Coomassie Blue staining.
Transfections and CAT Assays-Transient transfection assays were carried out by the calcium phosphate method as described previously (22). Briefly, 4 ϫ 10 5 cells were plated on a 60-mm plate and grown overnight. Three h prior to transfection, the cells were fed with new growth medium. Transfections were carried out with 3 g of the reporter plasmid pJCV L ⌬B-CAT, alone or in combination with 2.5 g of CMV-JCV T-antigen, and 10 g of pEBV-Pur␣, pEBV-Pur␣ (1-154), or pEBV-Pur␣ (216 -322). Experiments were designed to be promoter controlled with either pCMV-X or pEBV-His B plasmids added to equalize the total amounts of promoter in each reaction mixture. The precipitate was removed after 3 h, and a glycerol shock was applied. Forty-eight h posttransfection, the cells were harvested, and a crude protein extract was prepared by repeated cycles of freezing and thawing. Extracts were quantitated by the Bio-Rad Bradford assay, and 75 g of protein was assayed for CAT activity (23). The fold transactivation was measured by scintillation counting of the spots cut from the thin-layer chromatography plates. Each experiment was repeated three or more times with different plasmid preparations.

Pur␣ Associates with JCV T-antigen in Vivo and in Vitro-
In the first series of experiments, co-immunoprecipitation experiments were performed to determine whether Pur␣ and JCV T-antigen associate with each other in cells. To this end, the JCV T-antigen-expressing cell line HJC-15b was transfected with a control plasmid (pEBV) or a plasmid encoding a histidine epitope-tagged Pur␣ (pEBV-Pur␣). Cellular extracts obtained from these transfected HJC-15b cells were immunoprecipitated with preimmune serum (normal mouse serum) or an antibody that recognizes the histidine epitope tag (anti-T7), and immune complexes were analyzed by Western blot analysis for the presence of JCV T-antigen. In extracts from HJC-15b cells transfected with the epitope-tagged Pur␣, anti-T7 antibody co-immunoprecipitated JCV T-antigen (Fig. 1A, lane 5). This co-immunoprecipitation was specific because normal mouse serum did not immunoprecipitate JCV T-antigen (lane 4), and the anti-T7 antibody that recognizes the histidine tagged Pur␣ did not immunoprecipitate JCV T-antigen from cells that were not transfected with Pur␣ (lane 3). The expression and immunoprecipitation of Pur␣ in HJC-15b cells transfected with pEBV-Pur␣ was confirmed by immunoprecipitation followed by Western blot analysis with the anti-T7 antibody (Fig. 1B). To rule out any possible contribution of the histidine epitope tag to the interaction between Pur␣ and JCV T-antigen, HJC-15b cells were transfected with the control plasmid pEBV and a plasmid encoding a histidine epitope-tagged ␤-galactosidase protein (pEBV-LacZ). Cellular extracts obtained from these transfected HJC-15b cells were immunoprecipitated with preimmune serum or anti-T7 antibody, and immune complexes were analyzed by Western blot analysis for the presence of JCV T-antigen. In extracts from HJC-15b cells transfected with pEBV-LacZ, the anti-T7 antibody did not co-immunoprecipitate JCV T-antigen (Fig. 1C, lane 5) indicating that the interaction between Pur␣ and JCV T-antigen is not mediated via the histidine epitope tag. The expression and immunoprecipitation of ␤-galactosidase in HJC-15b cells transfected with pEBV-LacZ was confirmed by immunoprecipitation followed by Western blot analysis with the anti-T7 antibody (Fig. 1D). Taken together, these results demonstrate that Pur␣ and JCV T-antigen are able to form a complex in cells in which both proteins are present.
To provide additional evidence for the interaction between Pur␣ and JCV T-antigen, in vitro GST pulldown experiments were performed. JCV T-antigen containing HJC-15b cellular extract was incubated with Escherichia coli-purified GST or GST-Pur␣ proteins immobilized on glutathione-Sepharose beads. After washing, proteins retained on the beads were analyzed by immunoblot analysis with an antibody that recognizes JCV T-antigen. As shown in Fig. 2A Because JCV T-antigen possesses greater than 70% amino acid homology with the SV40 large T-antigen (24), the ability of Pur␣ to interact with SV40 T-antigen was assessed. To this end, GST pulldown experiments were performed with cellular extract from cells constitutively expressing SV40 T-antigen. As shown in Fig. 2C, SV40 T-antigen was specifically retained on the Sepharose column containing GST-Pur␣ (lane 3) but not on the column containing GST alone (lane 2). Thus, Pur␣ is able to interact with T-antigen produced not only by JCV but also by SV40.
Localization of Sequences in Pur␣ Important for Interaction with JCV T-antigen-Structurally, Pur␣ is composed of a several modular domains (see Fig. 4A). In particular, the central region of Pur␣ is composed of three aromatic and basic repeats (class I) interspersed with two acidic leucine-rich repeats (class II). Other notable structural features of Pur␣ include an aminoterminal glycine-rich region and an amphipathic ␣-helix and a glutamate-glutamine-rich domain near the carboxyl terminus. To identify the region(s) of Pur␣ necessary for the interaction with JCV T-antigen, a series of amino-terminal, carboxyl-terminal, and internal deletion mutants fused to GST was constructed (see Fig. 4A), and GST pulldown experiments utilizing HJC-15b cellular extract and these various deletion mutants were performed. Consistent with the data presented above, full-length Pur␣ fused to GST bound to JCV T-antigen (Fig. 3A,  lane 3), whereas GST alone did not (Fig. 3A, lane 2). An aminoterminal GST-Pur␣ deletion construct removing the aminoterminal 54 amino acids encompassing the glycine-rich region of Pur␣, GST-Pur␣ (54 -322), showed a slightly reduced ability to interact with JCV T-antigen compared with the full-length GST-Pur␣ (Fig. 3A, compare lanes 3 and 4). Removal of the amino-terminal 85 amino acids, including the glycine-rich region and the first class I repeat, GST-Pur␣ (85-322), demonstrated only weak binding activity compared with GST-Pur␣ (1-322) (Fig. 3A, lane 5). Further amino-terminal deletion mutants, GST-Pur␣ (167-322), GST-Pur␣ (216 -322), and GST Pur␣ (274 -322) were unable to interact with JCV T-antigen (Fig. 3A, compare lanes 3 and 5). These results indicate that sequences amino-terminal to amino acid 167 are involved in the interaction between Pur␣ and JCV T-antigen.
To further define the sequences within Pur␣ that are important for its interaction with JCV T-antigen, similar experiments were performed with a series of carboxyl-terminal GST-  (Fig. 3B, lanes  3-7). The smallest carboxyl-terminal deletion mutant containing only the amino-terminal 71 amino acids, GST-Pur␣ (1-71), was unable to interact with JCV T-antigen (Fig. 3B, lane 8). These results are consistent with those obtained with the amino-terminal GST-Pur␣ deletion mutants and further refine the region of Pur␣ that is involved in the interaction with JCV T-antigen to amino acids 72-123.
Two additional Pur␣ deletion mutants were tested for the ability to interact with JCV T-antigen. One of these mutants, GST-Pur␣ (⌬72-229), removes the central region of Pur␣; the other, GST-Pur␣ (85-229), contains only the central region of Pur␣. As shown in Fig. 3C, both of these GST-Pur␣ deletion constructs were able to interact with JCV T-antigen (lanes 4 and 5), albeit to a lesser extent when compared with full-length GST- Pur␣ (1-322) (lane 3). The interaction between the central region of Pur␣, GST-Pur␣ (85-229), is not surprising considering the previous demonstration that GST-Pur␣ (85-322) was able to interact with JCV T-antigen (Fig. 3A, lane 5) and that the carboxyl-terminal deletion construct GST-Pur␣ (1-123) was able to interact with JCV T-antigen (Fig. 3B, lane 7). The observation that GST-Pur␣ (⌬72-229) interacts with JCV Tantigen is somewhat surprising because neither the deletion construct containing only the first 71 amino acids, GST-Pur␣ (1-71), nor the amino-terminal deletion constructs GST-Pur␣ (167-322), GST-Pur␣ (216 -322), and GST-Pur␣ (274 -322) interact with JCV T-antigen. Upon closer evaluation of GST-Pur␣ (⌬72-229), an interesting observation can be made. This deletion construct fuses the first one-third of the first class I repeat (Ia) with the last two-thirds of the last class I repeat (Ic), essentially reconstituting a class I repeat (hybrid) (Fig. 4B). Taken together, these observations indicate that a class I repeat is involved in the interaction with JCV T-antigen. Other sequences, however, are also important for the interaction of Pur␣ with JCV T-antigen. This is supported by the observations that GST-Pur␣ (167-322), which contains an intact class I repeat, is unable to interact with JCV T-antigen and that GST-Pur␣ (85-322), which contains two intact class I repeats, interacts weakly with JCV T-antigen. Nonetheless, the above studies demonstrate that the minimal region of Pur␣, which is important in the interaction with JCV T-antigen, resides between amino acids 72 and 123. Additional studies involving smaller deletions as well as point mutations will be necessary to help define the amino acid contacts between Pur␣ and JCV T-antigen.
Pur␣ Inhibits JCV T-antigen-mediated Transcriptional Activation-Earlier studies have demonstrated that Pur␣ has the ability to interfere with the stimulatory action of JCV T-antigen on JCV late gene transcription (9). To investigate the functional significance of the physical interaction between Pur␣ and JCV T-antigen, transient transfection assays were performed. In these studies, the plasmid pJCV L ⌬B, which contains 286 base pairs of DNA sequence of the JCV control region upstream of the reporter CAT gene in the late orientation, was introduced into U-87 MG human glial cells singly or in combination with plasmids expressing JCV T-antigen, Pur␣, or mutant Pur␣ proteins.
As shown in Fig. 5, the activity of the JCV L promoter is increased when cells are co-transfected with the pJCV L ⌬B reporter plasmid and the JCV T-antigen expressing plasmid (compare lanes 1 and 2). The stimulatory effect of JCV Tantigen is independent of an increase in DNA replication as the auxiliary sequences required for JCV DNA replication located on the early side of the origin are deleted in the reporter construct used (25). In agreement with earlier studies, the T-antigen-induced transcriptional activity of the JCV L promoter is decreased when cells are co-transfected with a plasmid expressing Pur␣ (Fig. 5, compare lanes 2 and 3) (9). To determine whether the physical interaction between JCV T-antigen and Pur␣ underlies their functional interaction, the ability of Pur␣ mutants to modulate the transcriptional activity of Tantigen on the JCV L promoter was examined. One Pur␣ mutant, Pur␣ (1-154) , retains the ability to interact with JCV T-antigen, whereas the second Pur␣ mutant, Pur␣ (216 -322), is unable to interact with JCV T-antigen. Expression of these mutants was verified by Western blot analysis using the anti-T7 antibody, which recognizes the histidine tag on both pro-  2 and 4). Alternatively, Pur␣ (216 -322) has little effect on T-antigen-mediated transcriptional activation of the JCV late promoter reporter construct (compare lanes 2 and 5). Neither full-length Pur␣ nor either deletion mutant affected the basal level of transcription in the absence of JCV T-antigen (Fig. 5, lanes 6 -8) (9). Of note, expression of JCV T-antigen remains unaltered upon expression of wild-type Pur␣ or either mutant Pur␣ (data not shown) (9). Thus, the ability of Pur␣ to inhibit T-antigen-mediated transcriptional activation is correlated with its ability to physically interact with T-antigen. Taken together, these results suggest that Pur␣ specifically inhibits T-antigen-induced transcriptional activation through its physical association with JCV T-antigen. DISCUSSION Previous studies have demonstrated that the cellular protein Pur␣ plays a role in the transcriptional regulation of the JCV promoters (9). More specifically, Pur␣ stimulates transcription from the JCV early promoter. Moreover, T-antigen attenuates the Pur␣-induced level of early gene transcription. Although Pur␣ has little effect on the late promoter itself, Pur␣ is able to decrease T-antigen-mediated transcriptional transactivation of the JCV L promoter. In this study, we have investigated the mechanism responsible for the ability of Pur␣ to abrogate the transcriptional activation of the JCV L promoter by T-antigen. Results from the current studies demonstrate that Pur␣ interacts with JCV T-antigen in vivo and in vitro. Additional studies, including co-immunoprecipitation and co-localization assays, utilizing mice transgenic for JCV T-antigen (JC-91) have also demonstrated an interaction between these two proteins. 2 Of note, the human and mouse Pur␣ proteins exhibit a very high degree of conservation, differing by only two amino acids (26). Moreover, we also demonstrate the Pur␣ is able to interact with the T-antigen produced by another polyomavirus, SV40. Transient transfection assays demonstrate that a mutant Pur␣ protein that retains the ability to interact with JCV T-antigen is, like the wild-type Pur␣ protein, able to abrogate the JCV T-antigen-mediated transactivation of the JCV L promoter. A mutant Pur␣ protein that is unable to interact with JCV Tantigen, however, is not capable of attenuating the JCV Tantigen-induced transactivation of the JCV L promoter. Taken together, these data demonstrate that Pur␣ and T-antigen interact both physically and functionally and that the interaction between these two protein modulates T-antigen-mediated Several interesting observations can be made regarding the physical interaction between Pur␣ and JCV T-antigen. JCV large T-antigen present in HJC-15b cells exists as several isoforms with different electrophoretic mobilities (Figs. 1-3). Because JCV T-antigen is a phosphoprotein with several phosphorylation sites (27), these different isoforms are believed to represent different phosphorylation states of the protein. Interestingly, Pur␣ interacts with all of the isoforms of JCV T-antigen present in HJC-15b cells (Figs. 1-3). This is in contrast to the interaction between Pur␣ and another protein, which exists in different phosphorylated states, the product of the retinoblastoma gene, pRb. Johnson et al. (16) have demonstrated that Pur␣ binds selectively to the hypophosphorylated form of pRb. The phosphorylation state of JCV T-antigen, however, may not dictate the association between Pur␣ and JCV T-antigen as in vitro translated Pur␣ interacts with GST-JCV T-antigen, which, being produced in bacteria, does not contain phosphorylated residues (Fig. 3). Nonetheless, future experiments will be required to elucidate the contribution of the phosphorylation state of JCV T-antigen to the intermolecular complex formed between Pur␣ and JCV T-antigen.
The transcriptional antagonism between JCV T-antigen and Pur␣ is reminiscent of the interaction between the related SV40 large T-antigen and the tumor suppressor p53 protein.
Several studies have demonstrated that SV40 T-antigen abrogates p53-mediated transcriptional activation (28 -30). Moreover, p53 has been shown to inhibit SV40 late promoter transactivation by SV40 T-antigen (31). This is similar to the case of Pur␣ and JCV T-antigen in that T-antigen abrogates Pur␣mediated transcriptional transactivation, and Pur␣ inhibits JCV T-antigen-mediated transactivation of the JCV late promoter (Ref. 9 and this study). The interaction between SV40 T-antigen and p53 not only has functional consequences with respect to transcription, but this interaction has also been shown to regulate replication. SV40 T-antigen, when complexed with p53, has been shown to be unable to replicate an SV40 origin-containing DNA (32,33). Interestingly, p53 has been shown to inhibit JCV DNA replication by interacting with JCV T-antigen (34). Although the functional significance of the interaction between Pur␣ and T-antigen with respect to DNA replication is yet undescribed, the results reported here invite investigation into the functional relevance of the interaction between these two proteins in other well characterized functions of T-antigen.
Another well characterized function of viral oncoproteins including large T-antigen is cellular transformation. One mechanism by which these proteins are able to cause cellular transformation is via their interaction with the products of cellular tumor suppressor genes, such as p53 and pRb. Several viral oncoptroteins have been shown to interact with p53, including JCV T-antigen (34,35), SV40 T-antigen (36,37), adenovirus E1B protein (38), and the E6 protein from the human papillomavirus (39). The interaction between these proteins and p53 results in functional inactivation, as in the case of T-antigen, or increased p53 turnover, as in the case of human papillomavirus E6. Similar to p53, pRb is targeted by several different transforming viruses and is complexed with JCV Tantigen (40), SV40 T-antigen (40,41), adenovirus E1A protein (42), and human papillomavirus E7 protein (43). These observations are interesting in light of our results demonstrating an interaction between JCV and SV40 large T-antigens with the cellular protein Pur␣. This raises interesting questions regarding the cellular role of Pur␣. Interestingly, overexpression of Pur␣ causes growth inhibition in vitro. 3 This is particularly noteworthy in light of several observations regarding Pur␣. The gene encoding Pur␣, PURA, has been localized to human chromosome band 5q31 (44). Loss of heterozygosity at 5q31 is frequently associated with hematologic malignancies particularly myelodysplastic syndrome and myeloid leukemias (45,46). Moreover, the recent demonstration of PURA gene deletions in many cases of myelogenous leukemia and myelodysplastic syndrome suggest that Pur␣ may also be involved in tumor development (47).
In this report, we demonstrate that Pur␣ interacts with the large T-antigen of JCV and SV40. In addition, we localize the region of Pur␣ that is important for its interaction in JCV T-antigen. Using transient transfection assays, we also demonstrate that these two proteins interact functionally as well. Full-length Pur␣ and a deletion mutant that retains the ability to interact with JCV T-antigen abrogate T-antigen-mediated transcriptional activation. A Pur␣ deletion mutant that is unable to interact with JCV T-antigen, however, is incapable of alternating the transcriptional activation of JCV T-antigen. Taken together, these data demonstrate that a physical interaction underlies the functional antagonism between these two proteins.  4). A Pur␣ mutant, Pur␣ (216 -322), which is unable to interact with JCV T-antigen, however, has no effect on T-antigenmediated transactivation (compare lanes 2 and 5). The results shown represent the mean of three independent experiments. Bars indicate standard deviation.