Functional interaction between the HIV transactivator Tat and the transcriptional coactivator PC4 in T cells.

The human immunodeficiency virus (HIV) transactivator Tat is a potent activator of transcription from the HIV long terminal repeat and is essential for efficient viral gene expression and replication. Tat has been shown to interact with components of the basal transcription machinery and transcriptional activators. Here we identify the cellular coactivator PC4 as a Tat-interacting protein using the yeast two-hybrid system and confirmed this interaction both in vitro and in vivo by coimmunoprecipitation. We found that this interaction has a functional outcome in that PC4 overexpression enhanced activation of the HIV long terminal repeat in transient transfection studies in a Tat-dependent manner. The domains of PC4 and Tat required for the interaction were mapped. In vitro binding studies showed that the basic transactivation-responsive binding domain of Tat is required for the interaction with PC4. The minimum region of PC4 required for Tat binding was amino acids 22-91, whereas mutation of the lysine-rich domain between amino acids 22 and 43 prevented interaction with Tat. Tat-PC4 interactions may be controlled by phosphorylation, because phosphorylation of PC4 by casein kinase II inhibited interactions with Tat both in vivo and in vitro. We propose that PC4 may be involved in linking Tat to the basal transcription machinery.

The human immunodeficiency virus (HIV) 1 transactivator Tat is an 86 -101-amino acid protein that is a potent transcriptional activator of expression of all HIV proteins, including Tat itself. Tat expression is essential for HIV viral replication, because Tat defective viruses are unable to replicate or express viral proteins efficiently (1,2). The Tat-responsive region of the HIV LTR has been mapped to an element immediately down-stream of the transcription start site (3,4). Tat is unusual in that it is targetted to the promoter by binding to a 59-nucleotide stem-loop structure in the nascent RNA called the transactivation-responsive (TAR) element, which is found at the 5Ј end of all HIV viral RNAs (5)(6)(7)(8). Although Tat has been demonstrated to have effects on transcriptional initiation (9), it differs from conventional activators in that its primary effect appears to be on transcriptional elongation (9 -11). In the absence of Tat, short prematurely terminated RNA transcripts are produced, but Tat appears to modify RNA polymerase II to a form that elongates more efficiently (12).
Recently a mechanism has been established to explain how Tat promotes transcriptional elongation. Phosphorylation of the C-terminal domain of RNA polymerase II by cellular kinases marks the transition from an initiation complex to an efficiently elongating polymerase complex, and a primary role of Tat appears to be to recruit such kinases. Tat targets two cyclin-dependent kinases, cdk7, the kinase component of TFIIH, and cdk9, the kinase component of positive transcription elongation factor b (pTEFb), which together are proposed to hyperphosphorylate RNA polymerase II (reviewed in Refs. 13 and 14). Tat, through its activation domain, interacts with the general transcription factor, TFIIH (15), and stimulates the phosphorylation of the C-terminal domain of RNA polymerase II by its kinase component, cdk7 (16,17). Similarly, the activation domain of Tat interacts with the cyclin T component of pTEFb, forming a complex with high affinity for the TAR element and recruiting the cdk9 component of pTEFb, which can phosphorylate RNA polymerase II (18 -20). It is therefore proposed that Tat stimulates hyperphosphorylation of RNA polymerase II by recruiting cdk7 and cdk9 and thus creates a highly efficient and processive RNA polymerase II complex.
However, Tat has been shown to interact with a wide range of cellular proteins, and it is unclear how these interactions fit into the above model of Tat transactivation. For example, Tat is found complexed to components of the basal machinery such as the core RNA polymerase II itself (21,22), the TATA-binding protein (TBP) subunit of TFIID (23,24), and the TFIID associated factor, TAF II 55 (25). Replication of the HIV virus is regulated by both viral and cellular proteins, and Tat also functions in concert with cellular transcription factors, which bind to and activate the HIV LTR. For example, the core region of the HIV LTR contains three SP1-binding sites. Tat, via its basic domain, interacts with SP1 (26), and this interaction has been shown to enhance phosphorylation of SP1 by doublestranded DNA-dependent protein kinase (DNA-PK) (27). The phosphorylation of SP1 has been correlated to enhanced HIV LTR function. The enhancer region of the HIV LTR contains binding sites for transcription factors such as NF-KB, nuclear factor of activated T cells, and activator protein 1. This region includes a tandem NF-KB repeat, which is essential for viral replication and gene expression (28). Maximal activation of the HIV LTR results from the co-operative actions of Tat and the NF-KB proteins, which bind to these elements (29). In contrast the transcription factor nuclear factor of activated T cells, which has been shown to compete with NF-KB for occupancy of these sites, interacts directly with the N-terminal region of Tat and represses Tat activation of the HIV LTR (30). Furthermore, Tat via its basic domain has been shown to interact with the transcriptional coactivators p300 and CREB-binding protein (31,32), which may serve to complex Tat with the basal transcription machinery. These coactivators are also associated with histone acetylation and chromatin remodelling. Other cellular cofactors may also interact with Tat and mediate interactions with the transcription machinery or enhancer complex. In fact, there are presently several other cellular proteins of unknown function that have been isolated as Tat interactors (33)(34)(35)(36).
We used the yeast two-hybrid system to screen for cellular proteins in CD28 activated T cells that might interact with Tat and modulate its function. There is recent evidence that Tat may modify T cell gene expression through the CD28 costimulatory receptor signaling pathway (37,38), and therefore CD28 activated cells may have unique Tat-interacting proteins. We isolated a transcriptional cofactor known as PC4 (positive cofactor 4) in this screen. We have demonstrated a specific interaction between Tat and PC4 and identified the regions of the proteins involved. In addition we have shown a functional consequence of this interaction on transcription from the HIV LTR.
PC4 cDNA was released from the library vector pJG4.5 by digestion with EcoRI and XhoI and ligated into the eukaryotic expression vector pSG5 (Stratagene) to generate SG5-PC4. PC4 was amplified by PCR and cloned into pSG5 in the reverse orientation to generate SG5-␣PC4. The N-terminal 1-62 amino acids and C-terminal 63-127 amino acids of PC4 (see Fig. 1) were amplified by PCR from pJG-PC4 and ligated into RcCMV (Invitrogen) to generate the constructs CMV-PC4NT and CMV-PC4CT, respectively. Tat86 was amplified by PCR from pCDM-Tat and ligated into RcCMV in frame and N-terminal to a synthetic influenza hemagglutinin (HA) tag using HindIII and SacII sites generated by the PCR to produce the construct CMV-TatHA. PC4 was amplified from pJG-PC4 and ligated into RcCMV in frame and N-terminal to a synthetic c-Myc tag to produce the construct CMV-PC4Myc. The PC4 bacterial expression plasmid pet11a PC4WT and derivatives have been described previously (43). Lysine mutants were generated using the following oligonucleotides: K23E/K26E (GGAATTCCATATGGAC-GAAAA GTTAGA GAGGAAAA) and K24E/K28E (GGAATTCCATATG-GACAAAGAGTTAAAGAGGGAAA). The bacterial expression plasmids pBAD-PC4, 43-127, 31-127, and 1-62 (see Fig. 1) were generated by cloning the respective PCR products into pBAD/Myc-HisB (Invitrogen) N-terminal and inframe with the Myc-His tag.
The scanning PC4 mutants that replace 6 wild type amino acids with the sequence AAASAA where constructed using a PCR technique (44). Briefly, two external primers were designed complementary to the pBad-PC4 plasmid, 5Ј and 3Ј to the PC4 insert and containing a BstEII and XbaI restriction site, respectively. For each mutant two internal primers were designed. The primer for the 5Ј product contained gcg, 6 bases forming a NheI site (gctagc), ggcagc, and 15 bases complementary to the PC4 sequence. The primer for the 3Ј product contained cgc, 6 bases forming a NheI site (gctagc), gctgct and 15 bases complementary to the PC4 sequence. Using the appropriate external and internal primer, the two products were generated by PCR, cleaved with BstEII and NheI or NheI and XbaI, gel purified, and ligated together. The full-length product was then gel purified and ligated into a BstEII/XbaIdigested pBad vector. The HIV LTR reporter construct (pHIVluc) was generated by cloning Ϫ453 to ϩ80 of the HIV LTR into the HindIII and XhoI sites of the luciferase reporter pXp1.
A cDNA library was prepared by a modification of the method of Gubler and Hoffman (46). Briefly, oligo(dT) primers incorporating a XhoI restriction site were hybridized to the mRNA, and the first strand was synthesized using MMLV reverse transcriptase (Supercript II, Life Technologies, Inc.) and dNTP mix including 5-methyl-dCTP rather than dCTP. The RNA was removed and second strand generated using RNaseH and Escherichia coli DNA polymerase I. The cDNA was bluntended using Vent DNA Polymerase and EcoRI adaptors (Promega) ligated onto the blunt ends. The cDNA was digested with XhoI, with any internal XhoI sites being protected by 5-methyl-dCTP incorporation, and ligated into the vector pJG4.5 at the unique EcoRI and XhoI restriction sites.
Yeast Two-hybrid Screening-A yeast two-hybrid screen was performed as detailed elsewhere (47). Briefly, the yeast strain EGY48 (ura3 trp1 his3 3LexA-operator-LEU2) was cotransformed by the lithium acetate method, with pEG-Tat86 and the LacZ reporter pSH18-34, which contains eight LexA operator-binding sites (LexAop). The yeast were then transformed with the pJG4.5 activated Jurkat T cell cDNA library and plated onto galactose-containing minimal medium. 74 Leuϩ colonies were replica plated onto ␤-galactosidase assay plates containing galactose or dextrose as the carbon source. Of these, 17 colonies that demonstrated ␤-galactosidase activity on galactose-but not dextrosecontaining medium were further characterized. Plasmid DNA was isolated from these colonies transformed into MC1061 E. coli by electroporation, and the plasmid DNA was subsequently isolated from ampicillin-resistant colonies. Rescued library plasmids were transformed back into EGY48 with the Tat86 bait to confirm an interaction and also with unrelated baits to test for specificity. Clones that specifically interacted with pEG-Tat86 were sequenced using the ABI PRISM Dye terminator Cycle Sequencing protocol (Perkin-Elmer).
Binding Assays-BL21 E. coli were transformed with pGEX constructs. Expression of fusion proteins was induced in exponentially growing bacteria with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside for 4 h at 37°C. GST proteins were purified from cell lysates using glutathione-Sepharose (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Protein concentrations were determined by SDS-PAGE, using bovine serum albumin as a standard. BL21(DE3) E. coli were transformed with pET11a-PC4 plasmids. Expression of PC4 proteins was induced as above. Alternatively, TOP10 E. coli were transformed with pBAD-PC4 plasmids and expression of PC4 proteins induced with 0.02% arabinose. Bacterial cells were lysed in lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 10% glycerol) by sonication, and cell debris was removed by centrifugation at 13,000 ϫ g. Transfected COS-7 cells were released from flasks with 0.1 mM EDTA in phosphate-buffered saline and incubated in lysis buffer for 30 min, and cell debris was removed by centrifugation at 13,000 ϫ g.
Binding assays were performed for 1 h at 4°C in lysis buffer containing 0.1 mg/ml bovine serum albumin using 10 g of GST proteins bound to Sepharose beads and 20 l of cell lysate containing approximately 1 g of PC4 protein. The Sepharose beads were washed in lysis buffer, and bound complexes were eluted by boiling in SDS-PAGE load buffer. Proteins were resolved by SDS-PAGE in 15% polyacrylamide, transferred to nitrocellulose for 16 h and subjected to Western analysis using anti-PC4 antibodies (rabbit polyclonal) or anti-Myc antibodies (monoclonal, 9E10).
Phosphorylation Studies-Cell lysates containing approximately 1 g of PC4 were treated with 10 units alkaline phosphatase (calf intestinal alkaline phosphatase, New England Biolabs) in phosphatase buffer (New England Biolabs) for 30 min at 37°C in a 20-l reaction. Alternatively, lysates were treated with 1 unit of casein kinase II (New England Biolabs) in CKII buffer (New England Biolabs) for 30 min at 30°C.
Transfections and Luciferase Assays-Human Jurkat T cells were cultured in RPMI supplemented with 10% fetal calf serum. Jurkat T cells (4.5 ϫ 10 6 cells in 300 l of RPMI supplemented with 20% fetal calf serum) were transfected by electroporation using a Bio-Rad Gene Pulser II at 280 V and a capacitance of 975 microfarads. In all transfections 5 g of reporter plasmid was used, and varying amounts of expression constructs were equalized by addition of parent plasmids, pSG5, or RcCMV. Cells were stimulated with a final concentration of 20 ng/ml PMA and 1 M calcium ionophore at 24 h post transfection. For HIV LTR transfections, 1 ϫ 10 5 cells were aliquotted into 96-well trays before stimulation.
Luciferase assays were performed according to the previously published method (48). Light emission was measured using a Packard Top Count Luminescence Counter. COS-7 cells were cultured in RPMI supplemented with 10% fetal calf serum. COS cells were released from flasks with trypsin, washed in phosphate-buffered saline, and resuspended in 0.8 ml of phosphatebuffered saline. Cells were transfected by electroporation at 280 V and a capacitance of 500 microfarads. COS cells were transfected with a total of 20 g of expression plasmids.
Coimmunoprecipitation-Transfected COS cells (1 ϫ 10 6 ) were lysed 48 h post transfection in 0.4 ml of lysis buffer (25 mM Tris, pH 7.4, 150 mM NaCl, 10% glycerol, 1% Nonidet P-40) containing protease inhibitors (10 g/ml leupeptin, 2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin). Lysates were precleared for 1 h by incubation with 50 l of 50% Sepharose CL4B slurry. Proteins were immunoprecipitated for 16 h with 100 l of 50% anti-HA-Sepharose slurry. After washing in lysis buffer, proteins were subjected to SDS-PAGE and analyzed by Western blotting. Proteins were detected using ECL (Amersham Pharmacia Biotech) and visualized by autoradiography. Alternatively, for quantitation proteins were visualized using the Fuji luminescent image analyzer (LAS-1000 plus) and quantitated using the Fuji Image Gauge software.

RESULTS
Tat Interacts with the Coactivator PC4 -The yeast two-hybrid system was used to screen for Tat-interacting proteins expressed in activated Jurkat T cells. A Tat bait construct (pEG-Tat86) was generated in the vector pEG202 to produce a fusion protein comprising the first 86 amino acids of HIV Tat as a C-terminal fusion to a LexA DNA-binding domain. The bait was targeted to LexA DNA-binding sites upstream of the endogenous LEU2 gene in the yeast strain EGY48 (ura3 trp1 his3 3LexA-operator-LEU2) and also upstream LexA DNA-binding sites of the LacZ reporter construct, pSH18-34. Target proteins encoded by a cDNA library were expressed as fusion proteins to an "acid blob" activation domain under the control of a GAL1 promoter. Interaction of the bait and a target protein activated the endogenous LEU2 gene, detected by growth of yeast on leucine-deficient medium and activation of the LacZ reporter, which was visualized by a ␤-galactosidase assay.
A cDNA library was generated in the vector pJG4.5 from Jurkat T cells activated with PMA, calcium ionophore and an anti-CD28 antibody. This expression library was screened for proteins that interacted with the Tat86 bait. Three distinct clones were detected that interacted with the Tat86 bait but did not interact with several nonrelated baits, including a Drosophila bicoid protein bait and baits generated from human c-Rel cDNA. The cDNAs encoding the interacting proteins were recovered from the yeast and sequenced. A clone that represented a particularly strong Tat interactor as assessed by ␤-galactosidase activity was found to be a full-length cDNA clone of the previously characterized transcriptional coactivator, PC4 (Refs. 49 and 43 and Fig. 1). PC4 is a 127-amino acid bipartite protein (Fig. 1). The N-terminal region of PC4 (amino acids 1-62) contains two serine-enriched acidic (SEAC) domains (43,49), whereas the C-terminal region of the protein (amino acids 63-127) contains a single-stranded DNA-binding motif (50). A lysine-rich region, which may be involved in the nonsequence specific binding of PC4 to double-stranded DNA, is situated between the 2 SEAC domains in the N-terminal region of the protein (50). PC4 has been demonstrated to function as a coactivator for a range of transcriptional activators in in vitro transcription systems. It has been suggested that PC4 may act as an adaptor type molecule linking activators with the preinitiation complex because PC4 has been shown to interact with both transcription factors and components of the basal transcription machinery (e.g. TFIIA). Therefore, PC4 is a candidate for a Tat coactivator involved in the interaction of Tat with the basal transcription machinery.
A Specific Protein Interaction between Tat and PC4 Can Be Demonstrated in Vitro-To confirm the interaction between Tat and PC4, binding assays were performed between a GST-Tat protein expressed in bacteria and subsequently immobilized on glutathione-Sepharose beads and a bacterially expressed PC4 protein. Western analysis of E. coli lysates using an anti-PC4 antibody revealed a 16-kDa protein ( Fig. 2A, lane 1). This PC4 protein bound to the GST-Tat fusion protein ( Fig.  2A, lane 3) but did not bind to a GST protein alone or to a GST-Myb fusion protein ( Fig. 2A, lanes 2 and 4, respectively). The presence of GST fusion proteins in the binding studies was confirmed by reprobing the blot with an anti-GST antibody ( Fig. 2A, lower panel).
To determine whether Tat could also interact with endogenous cellular PC4, binding assays were performed between GST-Tat fusion protein immobilized on glutathione-Sepharose and COS cell lysates. Cell lysates were incubated with GST alone or GST-Tat protein. Western analysis of the COS cell lysate with an anti-PC4 antibody revealed a single 16-kDa protein (Fig. 2B, lane 1). This protein bound to GST-Tat (Fig.  2B, lane 3) but not to the GST control (lane 2). These experiments show that Tat can interact with both bacterially expressed and endogenous COS cell PC4 in vitro and confirm the interaction detected in the yeast two-hybrid system.
Modification of PC4 Levels Affects Tat-mediated Activation of the HIV LTR-To determine whether the PC4-Tat interaction has a functional consequence on Tat transactivation of the HIV LTR, transient transfection assays were carried out in Jurkat T cells. Human Jurkat T cells were cotransfected with an HIV LTR luciferase reporter and with increasing amounts of a PC4 expression plasmid (SG5-PC4), with or without a Tat expression plasmid (pCDM-Tat). Cells were either unstimulated or stimulated for 8 h with PMA and calcium ionophore to mimic T cell activation. PMA and calcium ionophore stimulation of the transfected Jurkat T cells resulted in a low level of activation of the HIV LTR. Transfection of increasing amounts of the PC4 expression plasmid had no activating effect on the HIV LTR in either unstimulated or stimulated cells in the absence of Tat (Fig. 3A).
Cotransfection of a Tat expression plasmid with the HIV LTR reporter lead to a large increase in transcription in either unstimulated (42-fold) or stimulated (118-fold) cells (Fig. 3A). Transfection of increasing amounts of the PC4 expression plasmid resulted in a further increase in activity in both stimulated and unstimulated cells in a dose-dependent manner, usually in the order of 2-3-fold stimulation (Fig. 3A).
To confirm the involvement of PC4 in Tat-mediated activation of the HIV LTR, the effect of expression of an antisense PC4 expression plasmid was examined. Jurkat T cells were transfected with the HIV LTR reporter and increasing amounts of the antisense PC4 expression plasmid (SG5-␣PC4), with or without Tat. Transfection of the antisense plasmid had no effect on the HIV LTR in the absence of Tat in either unstimulated or stimulated cells (Fig. 3B). This is in agreement with the lack of effect of PC4 overexpression on the HIV LTR in the absence of Tat (Fig. 3A). However, transfection of increasing amounts of antisense PC4 plasmid reduced the Tat-dependent activation of the HIV LTR promoter in a dose-dependent manner (Fig. 3B).
PC4 sense and antisense plasmids were also cotransfected into Jurkat T cells to determine whether antisense PC4 expression could reverse the effect of overexpression of PC4 on Tat activation of the HIV LTR. As before, transfection of an antisense PC4 plasmid reduced Tat activation of the HIV LTR in both stimulated and nonstimulated cells by 2-3-fold (Fig. 3C). PC4 (sense) expression enhanced Tat-mediated activation of the HIV LTR as described above, and this induction was reduced in a dose-dependent manner by cotransfection of increasing amounts of antisense PC4 plasmid, back to approximately the levels seen with Tat alone (Fig. 3C).
To confirm that transfection of the antisense PC4 plasmid is able to reduce the level of PC4 protein in cells, protein expression was examined in antisense or control transfected COS cells. Cells were transfected with CMV vector (Fig. 3D, lane 1 To control for transfection efficiency and protein loading, all cells were also cotransfected with a Tat-HA tagged expression plasmid, and the Western blot was reprobed with an anti-HA antibody (Fig. 3D, lower  panel). Protein expression was visualized using the Fuji luminescent image analyzer (LAS-1000 plus) and quantitated using the Fuji Image Gauge software. Antisense PC4 expression reduced PC4-Myc expression levels to 23% of the control lane after normalization against Tat-HA expression levels. The ability of PC4 to enhance Tat-mediated HIV LTR activity and the ability of antisense PC4 to reduce it demonstrate a functional link between Tat and PC4 and a role for PC4 in Tat-mediated activation of the HIV LTR.
Phosphorylation of PC4 Inhibits Interaction with Tat Both in Vitro and in Vivo-PC4 can be phosphorylated by casein kinase II (CKII) on a number of serine residues (51) primarily situated in the first SEAC domain in the N-terminal region of the protein. The majority of the protein in the cell appears to be in the phosphorylated form (51). The phosphorylated and unphosphorylated proteins display different mobilities under SDS-PAGE conditions (51). Upon transfection of a PC4-Myc expression construct into COS cells, two bands were detected upon Western analysis that may represent differentially phosphorylated forms of the protein. The phosphorylation status of a Myc-tagged PC4 protein expressed in COS cells and the effect of phosphorylation on interactions with Tat were therefore examined.
COS cells transfected with the PC4-Myc construct were lysed in the absence (Fig. 4A, lane 1) or presence (lane 2) of the phosphatase inhibitor sodium orthovanadate (NaV). In the absence of NaV, two proteins of 23 and 26 kDa were detected using an anti-Myc antibody, whereas in the presence of NaV only the slower mobility 26-kDa protein was detected, suggesting that this protein is a phosphorylated form of PC4. The 26-kDa form could be converted to the 23-kDa form upon incubation with alkaline phosphatase (Fig. 4A, lane 4), whereas no change was seen upon incubation with alkaline phosphatase buffer alone (compare lanes 1 and 3). Furthermore, only the 26-kDa protein was detected upon incubation of the extracts with CKII (Fig. 4A, lane 6), whereas no change was seen upon incubation with CKII buffer alone (lane 5). We also tested whether bacterially expressed PC4 could be phosphorylated by CKII. In bacterial cell lysates expressing PC4-Myc, a 23-kDa protein, was detected (Fig. 4A, lane 7), as well as smaller degradation products. This protein is unaffected by incubation in CKII buffer (Fig. 4A, lane 8), but upon incubation with CKII, a shift to the 26-kDa protein is seen (lane 9). The PC4-Myc degradation products, which are most likely N-terminal deletions (the Myc tag is on the C terminus), are not affected by incubation with CKII (Fig. 4A, lane 9). This is not unexpected because phosphorylation by CKII has been shown to be largely restricted to the N-terminal 1-22 amino acids of the PC4 protein (51). Therefore, it appears that the 23-kDa protein is an unphosphorylated form of PC4-Myc, whereas the 26-kDa protein corresponds to PC4-Myc phosphorylated by CKII. The anti-PC4 antibody used in experiments in Fig. 2 (A and B) cross-reacts only weakly with the phosphorylated protein, therefore explaining the detection of only one protein band in extracts of COS cells (data not shown). Similarly lysates used in the experiment in Fig. 3D were prepared in the presence of sodium orthovanadate, and therefore only a single band presumably representing phosphorylated PC4-Myc was detected.
Binding studies were conducted to determine whether the phosphorylation status of PC4 affects its interactions with Tat.
Next we investigated whether the phosphorylation status of PC4 affects its interaction with Tat in cells. COS cells were transfected with the PC4-Myc contruct in the presence or absence of a HA-tagged Tat (Tat-HA) expression construct. Proteins were immunoprecipitated from COS cell lysates with an anti-HA antibody (12CA5) coupled to Sepharose beads. To determine whether PC4-Myc could bind to Tat-HA in vivo, 12CA5 immunoprecipitates were probed with the anti-Myc antibody. Western analysis revealed the presence of both the unphosphorylated (23 kDa) and phosphorylated (26 kDa) forms of PC4-Myc in lysates from COS cells cotransfected with PC4-Myc and Tat-HA (Fig. 4C, lane 1) and from COS cells transfected with PC4-Myc alone (Fig. 4C, lane 2). Neither of these protein bands were present in mock transfected COS cells (data not shown). The unphosphorylated PC4-Myc protein was coimmunoprecipitated with the Tat-HA protein (Fig. 4C, lane 3) but not immunoprecipitated in the absence of Tat-HA (Fig. 4C, lane 4). However, the phosphorylated PC4-Myc protein did not coimmunoprecipitate with Tat-HA. Therefore, a specific interaction between unphosphorylated PC4-Myc and Tat-HA, but not phosphorylated PC4-Myc was detected in vivo. Analysis of the blot with anti-HA antibodies confirmed the presence of a 21-kDa Tat-HA protein in only the Tat transfected lysates and immunoprecipitated with the anti-HA-Sepharose (Fig. 4C,  lanes 1 and 3). Taken together these experiments show that unphosphorylated but not phosphorylated PC4 interacts with Tat both in vivo and in vitro.
The Basic Domain of Tat Is Involved in the Interaction with PC4 -To investigate which region of Tat is involved in the interaction with PC4, binding studies were conducted between GST-Tat mutant proteins immobilized on glutathione-Sepharose and bacterial lysates containing a PC4-Myc tagged protein. Constructs were used that express GST fusion proteins of the one exon form of Tat (GST-Tat 1-72), the activation domain of Tat (GST-Tat 1-48), and the C-terminal region of Tat containing the basic TAR binding domain (GST-Tat 49 -86). Expression of the 23-kDa PC4-Myc protein was detected by Western analysis using an anti-Myc antibody (Fig. 5A, lane 1). The PC4-Myc protein bound to Tat 1-72 (lane 3) and Tat 49 -86 (lane 5). However, the PC4 protein did not bind to Tat 1-48 (Fig. 5A, lane 4) or to Tat 1-72 with the basic domain mutated (Tat Basic, lane 6). As expected PC4 did not bind to the GST protein alone. To confirm that the GST-Tat fusion proteins were expressed and used in approximately equal amounts in the binding studies, the blot was reprobed with an anti-GST antibody (Fig. 5A, lower panel).
To confirm that the basic region of Tat was involved in the interaction with PC4, constructs were generated to express a GST fusion protein of Tat 1-57. This construct contains the activation domain (1-48), which did not interact with Tat as well as the basic TAR binding domain (49 -57). As a control, GST-Tat 1-57 basic that has the basic domain mutated to alanine residues was also examined. As expected PC4 interacted with the wild type GST-Tat 1-72 protein but not to GST alone (Fig. 5B, lanes 3 and 2, respectively). Furthermore, an interaction was detected between PC4 and GST-Tat 1-57 but not the GST-Tat 1-57 basic control protein (lanes 4 and 5). As before, Western analysis of the blot confirmed that the GST fusion proteins were present in the binding studies (Fig. 5B,  lower panel). Therefore, the interaction between PC4 and Tat involves the TAR binding basic domain of Tat.
Identification of Regions of PC4 Important for Tat Interaction-To determine which regions of PC4 are important in the interaction with Tat, binding assays were performed between GST-Tat fusion protein immobilized on glutathione-Sepharose beads and PC4 wild type or mutant proteins (Fig. 1). Expression of PC4 wild type protein (Fig. 6A, lane 1) and PC4 22-127 (lane 5) was detected by Western analysis using an anti-PC4 antibody. The anti-PC4 antibody did not cross-react efficiently with PC4 31-127 or PC4 43-127 (data not shown), so it was necessary to generate PC4-Myc fusion proteins. Expression of PC4-Myc (lane 9), , and PC4 43-127Myc (lane 17) was confirmed by Western analysis using an anti-Myc antibody. Bands corresponding to the expected size of each protein as well as smaller degradation products were seen in each case.
Bacterial cell lysates containing wild type PC4 or mutant proteins were incubated with GST alone, GST-Tat, or GST-Myb as a control. Both the wild type PC4 protein and PC4 22-127 bound to GST-Tat (Fig. 6A, lanes 3 and 7, respectively) but not to GST alone or GST-Myb. PC4-Myc bound to GST-Tat (Fig. 6A,  lane 11), but not to GST or GST-Myb (e.g. lanes 10 and 12, respectively) as expected. In contrast, PC4 31-127Myc bound to GST-Tat only very weakly (Fig. 6A, lane 15) and PC4 43-127 did not bind to . These data suggest that the first SEAC domain contained in amino acids 1-21 is not required for interactions with Tat, whereas removing the lysinerich region impairs the interaction with Tat. As in previous binding experiments, all blots were reprobed with an anti-GST antibody to confirm the presence of GST fusion proteins in the binding studies (data not shown).
To further define the Tat-interacting region of PC4, mutants were generated deleting the C-terminal region of the protein.
Taken together these results suggests that the lysine-rich motifs of PC4 between amino acids 22 and 43 are important for the interaction with Tat, whereas the first SEAC domain of PC4 contained in amino acids 1-21 and the C-terminal 91-127 amino acids are not required for the interaction.
minal region of PC4 alone interact with Tat. Therefore, these regions of the protein were used in transfection studies to test the specificity of the coactivator function of PC4 with Tat. The scanning mutants PC4 mut23-40 and PC4 mut35-40, in which the lysine motifs are mutated to the sequence AAASAA, were also examined in transfection studies because these mutants also do not interact with Tat.
Expression constructs were generated containing the N-terminal 1-62 amino acids of PC4 (PC4NT), the C-terminal 63-127 amino acids of PC4 (PC4CT), and the mutants PC4 mut23-28 and PC4 mut35-40. Jurkat T cells were cotransfected with the HIV LTR luciferase reporter and increasing amounts of the PC4CT construct with or without Tat expression plasmid. Cells were either unstimulated or stimulated with PMA and calcium ionophore.
Overexpression of PC4CT alone lead to activation of the HIV LTR in the absence of Tat both in unstimulated and stimulated cells in a dose-dependent manner (Fig. 7A). Tat activated the HIV LTR in both unstimulated and stimulated cells. However, PC4CT did not appear to co-operate with Tat activation of the HIV LTR in either unstimulated or stimulated cells. Transfection of the PC4NT expression plasmid had no effect on the HIV LTR in the absence or presence of Tat (Fig. 7B), whereas the wild type PC4 protein enhanced Tat activation of the HIV LTR as before.
Finally PC4 mut 23-28 and PC4 mut 35-40, like wild type PC4, had no activating effect on the HIV LTR in the absence of Tat (Fig. 7C). However, unlike the wild type PC4 protein, these mutants did not enhance HIV LTR activation in the presence of Tat.
Therefore, in the absence of Tat interaction, mutant PC4 proteins did not activate Tat-dependent HIV LTR function. The C-terminal region of PC4 alone is a constitutive activator of the HIV LTR but is unable to enhance Tat-mediated activation of the HIV LTR. DISCUSSION PC4 was identified in this study as a Tat-interacting protein by a yeast two-hybrid screen of an activated Jurkat T cell library. The interaction between Tat and PC4 was confirmed FIG. 7. Functional analysis of PC4 mutants. A, Jurkat T cells were cotransfected with 5 g of reporter plasmid containing the HIV LTR with increasing amounts of PC4-CT plasmid (amino acids 63-127) in the absence and presence of Tat expression plasmid as indicated. Cells were either unstimulated (Ⅺ) or stimulated with PMA and calcium ionophore for 8 h (f). Parent plasmid was added to a total of 30 g in each transfection. Luciferase activity was measured using the Packard Top Count Luminesence counter and is depicted in counts per second (CPS). Columns represent the means, and error bars represent the standard error of five replicate assays. The fold inductions are indicated below. B, Jurkat T cells were transfected with 5 g of HIV LTR reporter plasmid in the presence or absence of 5 g Tat expression plasmid. Some cells were cotransfected with 5 g of PC4 or PC4-NT (amino acids 1-62) expression plasmid as indicated. Assays were conducted as in A. C, Jurkat T cells were transfected with 5 g of HIV LTR reporter plasmid in the presence or absence of 5 g of Tat expression plasmid. Some cells were cotransfected with 5 g of PC4, PC4 mut23-28, or PC4 mut35-40 expression plasmid as indicated. Assays were conducted as in A.
both in vitro and also in vivo using GST pull-down experiments and coimmunoprecipitations. We have shown here that PC4 enhances Tat-mediated activation of the HIV LTR in vivo. PC4 has been demonstrated to interact with both components of the basal transcription machinery such as TFIIA as well as a variety of activators such as VP16 and GAL4-AH (43,49), leading to the suggestion that it may act as an adaptor-type molecule between the transcription machinery and activators. Therefore, PC4 is a candidate for a Tat coactivator facilitating interactions between Tat and the transcription machinery on the HIV LTR.
Several studies have demonstrated coactivator function of PC4 with activators, such as VP16 and GAL4-AH (43,49); however, such experiments have thus far employed in vitro transcription systems. In this study we demonstrate coactivator function for PC4 with Tat in transient transfection experiments in Jurkat T cells. In these studies overexpression of PC4 had no activating effect on the HIV LTR alone but enhanced Tat-mediated activation of the HIV LTR in a dose-dependent manner. Similarly, antisense PC4 expression had no effect on the HIV LTR in the absence of Tat but reduced Tat-mediated activation of the HIV LTR. The lack of an antisense effect in the absence of Tat serves to demonstrate the specificity of the antisense result. This was also shown by the ability of antisense PC4 to counter the effects of overexpression of PC4 on Tat-mediated activation of the HIV LTR and the ability of antisense PC4 to reduce PC4 expression at the protein level. We therefore conclude that PC4 acts as a coactivator of Tat in vivo. While this study was in progress, Kannan and Tainsky (52) also demonstrated coactivator function for PC4 in vivo. PC4 was shown to interact with the transcription factor, activator protein 2, and regulate its transcriptional activity in vivo.
In vitro binding studies showed that PC4 interacts with the basic TAR binding region of Tat. PC4 interacted with the basic region (amino acids 50 -57) when it was presented with the activation domain (i.e. as  or with the C-terminal region of Tat (i.e. as Tat 49 -86), suggesting that this domain is sufficient for binding to PC4. The basic domain of Tat is involved in binding to the TAR RNA structure (5,53) and therefore is important for targeting Tat to the promoter region of the HIV LTR. The basic region was also shown to be involved in interactions with cellular proteins, such as the activator SP1 (26) and the transcriptional coactivator CREB-binding protein/ p300 (31,32). How interactions between these cellular proteins and the basic domain of Tat affect the Tat-TAR interaction has not been addressed. Further studies are required to determine whether PC4 can interact with the Tat-TAR complex or whether the PC4-Tat and Tat-TAR complexes form independently. PC4 has been shown to interact with both doublestranded DNA and single-stranded DNA (50), but the possibility exists that PC4 can also interact with RNA.
The region of PC4 between amino acids 22 and 91 was shown to be involved in the interaction with Tat. Deletion of the first 21 amino acids does not abolish binding, but deletion of the first 43 amino acids or mutation of the lysine-rich region between amino acids 22 and 31 abolishes binding, suggesting that the integrity of amino acids 22-43 are required for interaction with Tat. On the other hand, PC4 1-62 does not bind to Tat as would be expected if the binding determinants are located between amino acids 22 and 43. There are several possible explanations for these results. First, it is possible that the truncated PC4 1-62 does not fold correctly to generate the Tat-interacting domain. Second, there may be a complex interaction region combining amino acids 22-43 and a second determinant lying between amino acids 62 and 91. Although we cannot rule out the possibility that mutation of the lysine residues may alter the structure of the protein, it is clear using several different types of mutants that the lysine motifs of PC4 are required for the interaction with Tat. Functional studies showed a correlation between the ability of PC4 proteins to interact with Tat and their ability to function as a coactivator. Specifically, PC4 mutants in which the lysine motifs were disrupted failed to display coactivator function with Tat in our transient transfection studies. Further work involving determining the structure of the N-terminal region of PC4 may be required to define the role of the lysine motifs in the interaction between PC4 and Tat.
PC4 is a bipartite protein containing a regulatory N-terminal domain and a single-stranded DNA-binding C-terminal domain (50). In our transfection assay, the N-terminal 1-62 amino acids of PC4, which cannot bind to Tat, did not retain coactivator function. This is in agreement with previous in vitro experiments using the GAL4-AH activator (50). Suprisingly, the C-terminal domain of PC4 (amino acids 63-127), to which single-stranded DNA binding activity has been attributed, acted as a constitutive activator in our in vivo system. This is in contrast to in vitro studies that have shown the C-terminal domain of PC4 to be either inactive (50) or in fact to act as an inhibitor of transcription (54). However, this region of PC4 did not co-operate with Tat in activating the HIV LTR, which is in agreement with the lack of interaction between the C-terminal domain and Tat in our binding studies. The C-terminal domain of PC4 can bind with high affinity to single-stranded DNA in vitro (50), and this may relate to the constitutive activity observed with expression of the C-terminal domain of PC4. The crystal structure of the PC4 C-terminal domain has now been determined and revealed that this domain forms dimers containing two single-stranded DNA-binding channels (55). Furthermore, PC4 C-terminal domain dimers bind with high affinity to two opposing unpaired DNA strands such as formed by internally melted duplexes and are able to destabilize dsDNA (56). This has led to the suggestion that PC4 may play a role in promoter opening during transcription. Further studies are required to determine whether the activity displayed by the C-terminal region of PC4 on the HIV LTR in vivo is due to its single-stranded DNA binding properties and if in fact this region is involved in promoter opening.
The single-stranded DNA binding properties of PC4 are displayed only by the truncated C-terminal domain of the protein (amino acids 63-127) or phosphorylated full-length PC4 (50). PC4 can be phosphorylated by CKII on serine residues situated primarily in the first SEAC domain in the N-terminal region of the protein (51), and this has been suggested to cause a conformational change in the protein exposing the C-terminal single-stranded DNA-binding region (50). When the C-terminal region of the protein is expressed alone, it is no longer controlled by phosphorylation of the N-terminal regulatory domain and would therefore be permanently able to bind singlestranded DNA. The constitutive activator function displayed by the C-terminal region of the protein in our studies may therefore be a consequence of this lack of regulation of the singlestranded DNA binding activity of PC4. Further studies are needed to determine whether such activity is also displayed by the full-length protein in which the C-terminal single-stranded DNA-binding region is exposed by phosphorylation.
Although phosphorylation regulates the single-stranded DNA binding activity of PC4, it also appears to control interactions between PC4 and a variety activators. In vitro studies have demonstrated that phosphorylation of PC4 by CKII inhibits interactions with the viral activator VP16 and also inhibits coactivator function (51). Phosphorylation of PC4 also abolishes coactivator function of PC4 with GAL4-AH in vitro (43).
Similarly, Tat interacts with only the unphosphorylated form of PC4 both in vitro and also in coimmunoprecipitation experiments. This suggests that the interaction between Tat and PC4 is similar to that demonstrated for other activators. In contrast, the transcription factor NF-Y interacts with PC4 irrespective of its phosphorylation status, via the C-terminal domain of PC4 (57).
A model for Tat transactivation of the HIV LTR has been suggested in which Tat is able to recruit the cellular kinases cdk7 (TFIIH) and cdk9 (pTEFb) to hyperphosphorylate RNA polymerase II, converting it to a highly processive form (reviewed in Refs. 13 and 14). How the interaction between Tat and PC4 fits with this model remains to be determined. However, in vitro studies have shown that TFIIH can phosphorylate PC4 within the preinitiation complex, and this is coupled with release of PC4 from the complex and promoter clearance (58). From this work, Malik et al. (58) proposed a two-step model for PC4 coactivator function in which PC4 is involved in formation of the preinitiation complex through interactions with the basal machinery and transcription factors. PC4 then requires phosphorylation and subsequenct release from the transcription complex before promoter clearance and elongation can occur efficiently (58). It could be speculated then that PC4 is involved initially in the formation of the preinitiation complex through interactions with Tat and the basal machinery and/or other transcription factors. As eluded to earlier, determining whether the Tat-PC4 complex forms co-operatively or independently to the Tat-TAR complex may shed light on the exact role PC4 plays. Subsequently, recruitment of cdk7 by Tat, which enhances the phosphorylation of the C-terminal domain of RNA polymerase II, may also promote phosphorylation of PC4 and release of PC4 from the preinitiation complex. Phosphorylation of both of these proteins may then enhance promoter clearance and elongation. It is possible that once released the phosphorylated PC4 protein may then have a role in opening of the promoter through its single-stranded DNA-binding domain. Because PC4 is a potential phosphorylation target, it may be the subject of regulation by signal transduction pathways. We can find no evidence of changes in absolute levels of PC4 in Jurkat T cells in response to PMA/calcium ionophore or CD28 activation. It is, however, possible that modification of PC4 could occur in response to these signals.
In this study we have shown that PC4 and Tat interact and that PC4 can enhance Tat-mediated activation of the HIV LTR. Furthermore, our data suggest a complex interaction between the basic domain of Tat and the region of PC4 between amino acids 22 and 91. The lysine motifs within this region of PC4 are required for the interaction. PC4 may facilitate interactions between Tat and the transcription machinery or the enhancer complex, and it is possible that these interactions are controlled by PC4 phosphorylation.