Promoter- and Cell-Specific Transcriptional Transactivation by the Kaposi's Sarcoma-Associated Herpesvirus ORF57/Mta Protein^▿

Plasmids and bacmids.

All plasmids were propagated and purified as described previously (55). The plasmids described previously were the following: pcDNA3-FLg50, pcDNA3.1-ORF57-Hygro (genomic ORF57), pGL3-nut1 (−1467), pGL3-B3 (kaposin), pGL3-TK, pcDNA3.1-lacZ (57), pcDNA3-FLc50, pGem3-FLc50, pGem3-50ΔSTAD, pORF57-GL3, pK-bZIP-GL3 (56), pGL3-nut1 (−706; wild type [WT] and mutant) and pcDNA3-FLc57 (45), the ori-Lyt (L) reporter plasmid 13F (101), and pGem3-50ΔLR (10).

Reporter plasmids pNut-1 (−131)-GL3 and pNut-1 (−73)-GL3 contain nut-1/PAN promoter sequences spanning positions −131 to +24 and −73 to +24, respectively. Both were subcloned as PCR products synthesized with Phusion Polymerase (New England Biotechnology), a common reverse primer (5′-GCGAGATCTAGCCAAGGTGACT-3′), and either −131 forward primer (5′-GCGAGATCTGGATTATTAAGGGT-3′) or −73 forward primer (5′-GATCTAAAATGGGTGGCTAACCTGTCCAAAAA-3′). The WT or mutant Nut-1-GL3 (−706) was used as a template to generate the respective WT or mutant versions of the nut-1/PAN reporter plasmids. The PCR products were digested with BglII and cloned into pGL3-basic (Promega) that had been digested with BglII (for the −131 reporter) or SmaI/BglII (for the −73 reporter).

Reporter plasmid pNut-1 (−1467)-GL3-mutant was constructed by subcloning the insert from pNut-1 (−706)-GL3-mutant to pNut-1 (−1467)-GL3.

Reporter plasmid Pnut46-hspluc was constructed as described previously (55) by annealing two oligonucleotides (5′-GATCTTTCCAAAAATGGGTGGCTAACCTGTCCAAAATATGGGAACACTGGAA-3′ and 5′-GATCTTCCAGTGTTCCCATATTTTGGACAGGTTAGCCACCCATT TTTGGAAA-3′) and then cloning them into the plasmid hsp-luc that had been digested with BglII.

Reporter plasmid p50-1-GL3 was constructed by PCR amplification of nucleotide positions −910 to +50 of the ORF50 promoter, introducing an NcoI site at +50. The PCR product and pGL3-basic were digested with NcoI and joined by ligation.

Reporter plasmid pCMV-Luc was constructed by transferring the Nru I/XhoI fragment containing the cytomegalovirus (CMV) promoter from pcDNA3 (Invitrogen) to pGL3-basic (Promega) that had been digested with SmaI/XhoI.

pGem3-FL50ΔNA was constructed in multiple steps. pGem3-FLc50 was digested with PstI to eliminate extraneous restriction sites in the polylinker downstream of ORF50. The resulting plasmid was digested with NcoI and AvaI to remove the Rta cDNA sequence encoding amino acids (aa) 274 to 483. The deletion was repaired by ligating a double-stranded oligonucleotide linker containing an in-frame fusion between NcoI and AvaI sites. The resulting plasmid expresses the Rta cDNA containing aa 1 to 273 fused in frame to aa 484 to 691.

pGex5x-3-Mta was constructed by PCR amplification of the ORF57 cDNA template using primers that introduced 5′-EcoRI and 3′-XhoI restriction sites. The resulting product was cloned into pGex5x-3 (GE Biosciences) following digestion of the insert and vector with those enzymes.

pGex5x-3-MtaΔA/T was constructed by ligating two PCR amplicons. Amplicon 1 corresponded to Mta ORF nucleotides 463 to 1371 with a 3′ AgeI restriction site introduced by one primer. Amplicon 2 corresponded to Mta nucleotides 1 to 335, with a 5′ AgeI site introduced by one primer. Amplicon 1 was digested with SphI/Age I, amplicon 2 was digested with Age I/KpnI, and both were ligated with pGex5x-3-Mta that had been digested with SphI/KpnI.

pcDNA3-c57ΔA/T was constructed by generating a composite PCR amplicon of AgeI-digested amplicons 1 and 2 (see above), facilitated by the complementary AgeI overhangs in each amplicon. The resulting amplicon was digested with PshAI/XhoI (the XhoI site was introduced by one outside primer) and cloned into pcDNA3-FLc57 that had been digested with the same enzymes.

pET28a-Mta was constructed by PCR amplification of the full-length Mta cDNA using primers that introduced EcoRI/XhoI restriction sites, followed by cloning it into those sites in the pET28a vector (Novagen).

pET28a-MtaΔA/T was constructed by cloning the EcoRI/XhoI fragment of pGex5x-3-MtaΔA/T into pET28b that had been digested with the same enzymes.

pRSET A-RBP-Jk was constructed by PCR amplification of the complete human RBP-Jk cDNA (48), using primers that introduced a 5′ BamHI restriction site and a 3′ PstI restriction site. The PCR product and pRSET A (Invitrogen) were digested with those enzymes and joined by ligation.

pGem-7SK (71) encodes the 7SK cDNA (a gift of Hua Zhu).

pCR2.1-nut-1 (+) contains the TA-cloned nut-1/PAN cDNA product of PCR amplification.

BAC36 is the full-length clone of KSHV described previously (112).

Cell lines and transfections.

HH-B2 cells, a gift of George Miller, were propagated as described previously (32). HH-B2 and BCBL-1 cells were transfected by electroporation, as previously described (57), at 200 V and 150 V, respectively.

CV-1 cells were a gift of Harvey Ozer and were propagated in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 290 μg/ml glutamine, and antibiotics (penicillin/streptomycin [10 U/ml and 10 μg/ml, respectively] or 50 μg/ml gentamicin). CV-1 cells were transfected using TransIt LT-1 (Mirus) according to the manufacturer's suggestions, with a total of 2.5 μg of DNA, which included 0.25 μg each of luciferase reporter and pcDNA3.1-lacZ for quantitating transfection efficiency.

293 cells (a gift of Abraham Pinter) were cultured as described previously (12) and transfected by the calcium phosphate method. The cells were plated at a density of 5 × 10⁴ per well of a six-well plate. One hundred microliters of transfection buffer (consisting of 10 μl of 0.5 M HEPES, 81 μl distilled H₂O [dH₂O], 9 μl 2 M NaCl, 0.2 μl 1 M Na₂HPO₄) was mixed with 100 μl plasmid mixture (5 μg total DNA, 10 μl 10× NTE [150 mM NaCl, 10 mM Tris-HCl, pH 7.4, 1 mM EDTA, pH 8.0], 12.5 μl 2 M CaCl₂, and dH₂O to a final volume of 100 μl). The mixture was incubated at room temperature for 15 to 30 min and then added to the cells, and the plates were returned to the incubator. Twelve to 15 h later, the media were changed and the plates were returned to the incubator for 48 h.

All KSHV-negative Burkitt's and plasma cell lymphoma cell lines (BJAB, BL-41, RPMI8226, U266, and Akata-31) were electroporated at a density of 1 × 10⁷ cells/electroporation, with 23 μg total DNA, which included 5 μg of luciferase reporter plasmid, 3 μg of pcDNA3.1-lacZ for quantitating electroporation efficiency, and titrations of Rta or Mta expression vectors. All were cultured identically to BJAB cells, as described previously (56). BJAB cells were electroporated at 150 V/0.975 μF. BL-41 cells (a gift of Y. Yuan [University of Pennsylvania]) were electroporated at 224 V/0.975 μF.

RPMI8226 and U266 (American Type Culture Collection) were electroporated at 200 V/0.975 μF and 175 V/0.975 μF, respectively. Akata-31 cells, an EBV-negative subclone of Akata cells, were a gift of Paul Farrell (40). They were electroporated at 200 V/0.975 μF.

Luciferase assays.

Luciferase assays were performed exactly as described previously (55). Luciferase activity was normalized for each transfection using β- galactosidase activity as an internal control for all cell lines except 293 (in which Mta transactivated the β-galactosidase plasmid). For all luciferase assays, corresponding transfections were performed in triplicate at least twice.

Luciferase RNA stability.

293 cells were cotransfected with 3.0 μg of the Mta expression vector or empty vector (as indicated) and the reporter plasmid pCMV-Luc or pGL3-nut1 (−706). On the following day, the cells were washed with 1× phosphate-buffered saline, half were treated with actinomycin D (ActD) (5 μg/ml; Sigma), and the cells were returned to the incubator for 18 additional hours. Total RNA was isolated using RNABee (Tel-Test), and 15 μg RNA from each transfected dish was analyzed by Northern blotting and quantitated with a phosphorimager.

Viral reactivation.

HH-B2 cells (10⁷) were stimulated with tetradecanoyl phorbol acetate (TPA) and/or transfected with the plasmids indicated. After growth in 10 ml complete RPMI for 6 days, virions were harvested from the supernatant, and encapsidated viral DNA was quantitated by Southern blotting as described previously (44). The probe was derived from KSHV ORF6 (56). Alternatively, BCBL-1 cells were transfected with expression vectors for ORF50/Rta and/or ORF57/Mta and analyzed for reactivation by immunofluorescence, as described previously (10). At least 500 Rta-positive cells were screened.

Northern blotting.

Northern blotting was performed as described previously (56). Double-stranded probes were the EcoRI fragment of pCR2.1-nut-1 (+) (for nut-1/PAN), the AccI/KpnI fragment of pGem-7SK (7SK), and the NcoI/XbaI fragment of pGL3-basic (luciferase).

Nuclear run-on assays.

293 cells were seeded in six-well dishes and transfected with pNut-1 (−1467)-GL3 alone or together with pcDNA3.1-ORF57-Hygro. Forty-eight hours following transfection, the cells were washed twice with 1× PBS, scraped into 15-ml conical tubes, and pelleted by centrifugation for 5 min at 1,000 rpm. The cells were resuspended in three times the pellet volume of 10 mM Tris-HCl, pH 7.5, 5 mM MgCl₂, 10 mM NaCl, 0.1%NP-40, 0.5 mM dithiothreitol and vortexed gently for 5 to 10 seconds. Nuclei were pelleted at 1,000 × g and frozen in 27 mM Tris, pH 7.5, 13 mM MgCl₂, 160 mM KCl, and 27% glycerol. ATP, CTP, GTP (0.5 mM final concentration [each]), and digoxigenin (DIG)-11-UTP (6 μl; Roche Applied Science) were added to the thawed nuclei, and transcription was continued by incubation at 37°C for 20 min. Ten microliters of RQ1 DNase (Promega) and 1 μl of 100 mM CaCl₂ were added to the nuclei and incubated at 37°C for 30 min. One milliliter of RNABee (Tel-Test, Inc.) and 100 μl of chloroform were added to lyse the cells. The tubes were kept on ice for 5 min and then microcentrifuged at 10,000 rpm for 20 min. RNA was precipitated with an equal volume of isopropanol. Six micrograms of pGL3-basic or pGem-7SK plasmid DNA was denatured and suctioned onto a Hybond XL membrane (GE Biosciences) in a dot-blotting apparatus. Following UV cross-linking, the blot was prehybridized for 2 h in 20 mM TES [N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid], pH 7.4, 400 mM NaCl, 4 mM EDTA, 200 μg/ml yeast RNA, 2× Denhardt's solution, and 0.1% sodium dodecyl sulfate (SDS). The total RNA that had been purified from the nuclei of each transfected cell sample was added and hybridized overnight at 65°C. The following day, the blot was developed using the DIG Luminescent Detection Kit for Nucleic Acids (Roche Applied Science).

Protein expression and purification.

Glutathione-S-transferase (GST)-Mta protein was expressed by inoculating 50 ml of LB (containing 50 μg/ml ampicillin) with a glycerol stock of Rosetta Escherichia coli (Novagen) containing pGex5x-3-Mta. Following 12 to 15 h of growth at 37°C with agitation, the culture was diluted into 450 ml of fresh LB (supplemented as described above), and growth was continued at 37°C until the culture reached an optical density of 0.6. The flask was transferred to ice for 20 min, after which 0.9 mM isopropylthio-β-d-galactosidase (IPTG) was added. The culture was grown at 16°C for 18 additional hours with agitation. Bacteria were pelleted by centrifugation and resuspended in 10 ml 1× NETN+ (20 mM Tris buffer, pH 8.0, 150 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 10% glycerol, 5 mM β-mercaptoethanol, 0.1% Triton X-100, and protease inhibitors [Sigma]). The bacteria were lysed for 10 cycles by sonication using a large tip and 15% power for 8 seconds, and the supernatant was clarified by centrifugation for 30 min at 7,000 rpm. The supernatant was aliquoted and flash frozen in a dry-ice/ethanol bath and then transferred to −80°C storage for future use. At the time of use, aliquots were thawed slowly on ice, and protein was purified by the addition of 40 μl glutathione-Sepharose (Sigma; 50% [vol/vol] in NETN+)/ml, followed by nutation for 1 hour at 4°C and washing with 7 column volumes of fresh NETN+. The GST moiety alone was expressed and purified similarly using the plasmid pGex5x-3.

His₆-Mta and His₆-RBP-Jk proteins were purified as described above with the following modifications: inocula were glycerol stocks of E. coli BL21(DE3)-RIL (Stratagene) containing the respective plasmids. Ten milliliters of saturated culture was diluted into 500 ml of fresh LB (supplemented as described above), and protein was induced for 18 to 48 h. The bacterial pellet was resuspended in NTN+ (10 mM Tris buffer, pH 7.5, 100 mM NaCl, 0.5% Nonidet P-40, and protease inhibitors [Sigma]) and purified by liquid chromatography over a Ni-nitrilotriacetic acid-agarose column (QIAGEN; 1-ml bed volume, preequilibrated in NTN+). The beads were washed with 10 ml modified NTN+ and eluted stepwise in five fractions of 0.5 M imidizole in NTN+. Fractions containing either of the His₆ proteins (identified by Western blots) were combined and dialyzed versus 1× DNA binding buffer (55).

Generation of anti-Mta serum.

Anti-Mta serum was generated exactly as described previously (57), using GST-Mta as an antigen.

GST pull-down assays.

Equal amounts of GST-Mta or GST protein, respectively (determined by Bradford analysis), were purified on GST-Sepharose (as described above). Beads were resuspended in 0.5 ml NETN+ and combined with equal volumes of rabbit reticulocyte lysate (RRL) (TNT Quick Coupled Translation kit [Promega]) that had been programmed in the presence of [l-³⁵S]methionine with plasmids expressing Rta or mutants. Following incubation at 4°C with nutation, the beads were washed twice with 0.5 ml NETN+. The beads were mixed with 30 μl of 2× Laemmli buffer, boiled, and separated by SDS-polyacrylamide gel electrophoresis (PAGE). The gels were fixed in 10% acetic acid-50% methanol-40% dH₂O, the signals were amplified by incubation in 0.1 M salicylic acid, the gels were dried, and the proteins were visualized by a PhosphorImager (Molecular Dynamics).

Immunoprecipitations and Western blotting.

BCBL-1 cells were induced with TPA (20 ng/ml) for 12 to 15 h or transfected with the indicated plasmids, and immunoprecipitations were performed as described previously (61). Anti-Mta and -Rta antisera were conjugated to horseradish peroxidase using EZ-Link Plus Activated Peroxidase (Pierce) prior to incubation with the indicated blots. Anti-alpha-actinin antibody was purchased from Sigma.

EMSAs.

Purified proteins were mixed with 1x DNA binding buffer (55) and incubated on ice for 15 min. All electrophoretic mobility shift assay (EMSA) reaction mixtures contained 10 ng of unlabeled salmon sperm DNA (50-fold molar excess over labeled DNA probe; Sigma) as a nonspecific competitor. Probe (1 × 10⁵ cpm per reaction) labeled as described previously (12) was added and incubated for 15 min at 16°C. DNA-protein complexes were resolved by electrophoresis on 7.5% polyacrylamide-0.5× Tris-borate-EDTA gels at 4°C, as described previously (12). The gels were transferred to Whatman paper, dried, and exposed to Bio-Max MS autoradiography film (Kodak).

ChIP.

Chromatin immunoprecipitation (ChIP) was performed as described previously (12), with the exception that chromatin was cross-linked at 40 h after TPA addition. DNA was resuspended in 100 μl of double-distilled H₂O, and 5 μl was used in real-time PCRs.

Real-time PCR.

Real-time PCR was performed using AmpliTaq Gold polymerase (Applied Biosystems) and a Corbett RotorGene 3000 instrument according to the manufacturer's suggestions. Cycling parameters were 5 min at 95°C and then 40 to 60 cycles of 20 s at 95°C/30 s at 55°C/40 s at 72°C.

Primer sequences (5′ to 3′) were as follows: Nut-1/PAN (forward, GTTTTCTTATGGATTATTAAGGGTC, and reverse, AGGTGAAGCGGCAGCCAAGGTGAC) and K6 (forward, CGCCTAATAGCTGCTGCTACGG, and reverse, TGCATCAGCTGCCTAACCCAG).

The ΔΔC_t method was used (51) for quantitation, with the following modifications: ΔC_t was calculated individually for each primer pair, for cells treated with TPA, and for cells left untreated, using the following formula: ΔC_t = C_t (IP) − C_t (input chromatin), in which Ct is “threshold cycle” and IP is immunoprecipitated chromatin. Next, ΔΔC_t was calculated using the following formula: ΔΔC_t = ΔC_t (for untreated cells) − ΔC_t (for TPA-treated cells). Enrichment (n-fold) was thus calculated as 2^−ΔΔCt.

Design and synthesis of molecular beacons.

Design and synthesis of molecular beacons were performed as described previously (12). Molecular-beacon sequences were as follows (lowercase letters denote stems, and uppercase letters denote regions complementary to the genomic sequences of the indicated genes): nut-1/PAN (cgctcgGTTAATGACATAAAGGGGCGTGGcgagcg) and K6 (cccctccCACCCACCGCCCGTCCAAATTCggagggg).

Immunofluorescence.

Immunofluorescence was determined exactly as described previously (57). Secondary antibodies raised against rabbit immunoglobulin (Ig) were tetramethyl rhodamine isothiocyanate (TRITC) conjugated, and those against mouse Ig were fluorescein isothiocyanate (FITC) conjugated (ICN).

ORF57/Mta and ORF50/Rta synergize to reactivate KSHV from latency.

Using firefly luciferase as a reporter gene in transiently transfected CV-1 cells, we previously reported that ORF57/Mta and ORF50/Rta synergized to transactivate the KSHV Nut-1/PAN and kaposin promoters (45). To determine whether Rta and Mta also synergized to reactivate KSHV from latency, we measured reactivation by ectopically expressing both proteins in latently infected PEL cells. First, we harvested virions from the media of transfected HH-B2 cells and measured encapsidated viral DNA using Southern dot blots. As shown in Fig. 1A, virus released from cells transfected with the empty expression vector was barely detectable, similar to untransfected HH-B2 cells. Mta expressed alone resulted in a slight increase in virus detected in the supernatant. Rta expressed alone, as expected, resulted in about a fivefold increase in virus induction relative to the empty vector. Cotransfection of the Mta and Rta expression vectors, however, yielded a synergistic induction (approximately 18-fold) of productive reactivation of KSHV from latency.

We previously used a single-cell reactivation assay to demonstrate that Rta was necessary and sufficient to reactivate KSHV from latency (56, 57). In those assays, our marker for induction of reactivation was the K8.1 protein, which is expressed in PEL cells with true late kinetics (i.e., K8.1 expression requires viral DNA replication) (54, 57, 107). To test Mta and Rta for synergistic induction of the KSHV lytic cycle, we scored reactivation similarly 48 h after electroporating BCBL-1 cells. Cells were scored positive if Rta expression was detected by visual inspection by fluorescence microscopy; for quantitation, we counted the percentage of Rta-positive cells that were also K8.1 positive (Rta⁺/K8.1⁺ divided by Rta⁺). At least 500 doubly positive cells were counted for each. Cells transfected with empty expression vector were scored similarly. Results from the empty-vector transfections (spontaneous reactivation) were subtracted from those in which the Mta or Rta vectors were transfected, and the results were plotted as relative reactivation. As shown in Fig. 1B, ectopic expression of Rta alone resulted in a slight increase in the number of doubly positive cells relative to vector-transfected cells. However, cotransfection of the Mta and Rta vectors resulted in a dramatic induction of viral reactivation. Ectopic expression of Mta alone did not produce a significant increase in Rta/K8.1-positive cells, suggesting that Mta alone is unable to induce Rta expression or complete viral reactivation in PEL cells. When these cells were scored by immunofluorescence for Mta/K8.1 doubly positive cells, the data also showed no difference in progression to K8.1 expression, comparing transfected vector to Mta-expressing cells (not shown). This single-cell assay agreed with Fig. 1A in showing that ectopic expression of Mta alone has little effect on viral reactivation but potently synergizes with ectopically expressed Rta.

Induction of nut-1/Pan expression by Rta has been used previously to measure viral reactivation (92). We repeated the approach shown in Fig. 1A, but we isolated total RNA from transfected HH-B2 cells 24 h after electroporation. As shown in Fig. 1C, vector-transfected cells showed little nut-1/PAN expression, while treatment with sodium butyrate robustly induced nut-1 expression (approximately 17-fold [Fig. 1D]). Ectopic expression of Rta alone resulted in about a threefold induction of nut-1 expression. Cotransfection of the same amount of Rta vector with increasing amounts of Mta vector demonstrated a dramatic, dose-responsive synergy of the two proteins, from 12- to 24-fold relative to vector-transfected cells. To confirm that ectopic Mta was not simply transactivating ectopic Rta expression, we coelectroporated BCBL-1 cells with the Mta vector and an expression vector for Rta fused to the V5 epitope so that we could distinguish ectopic Rta from Rta expressed from the endogenous virus. As shown in Fig. 1E, Mta did not increase ectopic Rta expression under the conditions used for Fig. 1A to D. Taken together, Fig. 1 demonstrates, using three independent assays, that Mta and Rta synergized posttranslationally to reactivate KSHV from latency.

Mta and Rta proteins are expressed with similar kinetics during KSHV reactivation in PEL cells.

Many studies have examined the kinetics of ORF57/Mta transcript expression during KSHV reactivation (22, 54, 56, 75), but relatively little is known regarding the kinetics of ORF57/Mta protein expression. To address this question, cellular protein extracts were prepared from KSHV-infected BCBL-1 cells at various times during viral reactivation following addition of TPA to the cultures. As seen in the upper part of Fig. 2A, Mta protein was detectable as a 49-kDa major species within 4 h following TPA treatment, slightly smaller than the predicted molecular mass for ORF57/Mta of 51.1 kDa (ExPASY). In other experiments, we have detected Mta protein as early as 1 h following TPA treatment (data not shown). Two minor species migrating with smaller apparent masses than the major Mta protein were also detected. Since Mta protein was not expressed in untreated, latently infected cells, these data agreed with previous classifications of Mta as a lytic viral protein. Rta protein was detectable prior to Mta protein, but both proteins increased in abundance until 24 h after TPA addition (Fig. 2A). By 48 h post-TPA, both Mta and Rta proteins were barely detectable by Western blotting (data not shown). Mta and Rta proteins were thus expressed with kinetically similar patterns during reactivation, but Rta protein expression preceded that of Mta, similar to expression of their transcripts (56).

Figure 2B, left, shows that the Mta-specific antiserum detected Mta exclusively in the nucleus in reactivating BCBL-1 cells (compare to Fig. 2B, right, DNA labeled with DAPI [4′,6′-diamidino-2-phenylindole]). Figure 2B, center, serves as a reference for cytoplasmic localization of the K8.1 glycoprotein.

To evaluate single-cell expression of Mta and Rta, we performed indirect immunofluorescence assays of tandem populations of BCBL-1 cells untreated or induced to reactivate with TPA. Approximately 1,000 cells were scored, in duplicate or triplicate, and the percentage of the total cells expressing each protein was determined. As shown in Fig. 2C, both Mta and Rta were detected in less than 1% of latently infected cells, representing the small population of cells that spontaneously reactivated the virus. At 24 h post-TPA addition, the percentages of Mta- and Rta-expressing cells remained identical. From 24 to 72 h post-TPA addition, the percentages of cells expressing Rta and Mta continued to increase, although at different rates. The percentage of Rta-positive cells increased linearly up to 72 h post-TPA in nearly 20% of reactivating cells. However, over the same time course, the rate of increase of Mta-expressing cells steadily declined and reached a plateau of about 12% of the population by 72 h post-TPA. Importantly, the total number of cells per ml of culture did not change significantly over the time course of this experiment. Since Western blotting showed that expression of both proteins was barely detectable at 48 h (data not shown), total Mta or Rta protein was not proportional to the total number of cells expressing each protein.

Expression of Mta is tightly associated with complete viral reactivation.

In Fig. 1, we demonstrated that ectopic coexpression of Mta and Rta resulted in synergistic induction of KSHV reactivation; we quantitated expression of the true late protein K8.1 as one measurement of synergistic reactivation. To compare coexpression of K8.1 with Mta or Rta when expressed in single cells from the endogenous viral genome, we performed indirect immunofluorescence assays in BCBL-1 cells uninduced or induced by TPA treatment. As shown in Fig. 2D, about 50% of untreated cells that expressed either Rta or Mta, representing the spontaneously reactivating population, also expressed K8.1. At 72 h post-TPA treatment, nearly 80% of Mta-expressing cells also expressed K8.1, while never more than about 20% of Rta-expressing cells expressed K8.1. Although Rta is necessary and sufficient to reactivate KSHV, these data suggest that expression of Mta predicted successful, full progression through the lytic gene cascade. Therefore, the expression of Mta seems to be required to commit a reactivating cell to the entire lytic cascade.

The RRE is necessary but not sufficient for Mta to synergize with Rta in transactivation of the Nut-1/PAN promoter.

To determine the function of Mta, we initially cotransfected CV-1 cells with increasing amounts of Mta or Rta, alone or together, and a plasmid containing the KSHV nut-1/PAN (−1463) promoter driving luciferase expression as a reporter (Fig. 3A shows a promoter schematic). The magnitude of transactivation for each was determined by comparison to transfection of the reporter with empty expression vector. Rta transactivated the nut-1 promoter to about 22-fold, but Mta had very little effect on the promoter alone (Fig. 3B). These effects of Rta and Mta alone were very similar to their effects on viral reactivation, as shown in Fig. 1. As we showed previously (45), when Rta and Mta were coexpressed, the two proteins dramatically synergized to transactivate the nut-1 promoter (Fig. 3B). In this experiment, the synergistic activation reached nearly 3,700-fold over basal promoter activity (and 143-fold greater than Rta transactivation alone). The optimal amounts of each expression vector (0.5 μg Rta expression vector and 3 μg of Mta expression vector) were identical to that which we previously published (45).

We and others have previously described an RRE in the nut-1/PAN promoter (15, 45, 91). Mutation of that element by replacing 9 of its 13 bp severely impaired activation of the nut-1/PAN promoter by Rta, and coexpression of Mta had no additional effect (45). The 13-bp element is conserved in the kaposin/ori-Lyt (R) and ori-Lyt (L) promoters (5, 15, 89, 101); Fig. 3C shows alignment of the three elements with the mutant nut-1/PAN element.

To characterize the cis requirements for Mta/Rta synergy on the nut-1/PAN promoter, we tested a series of promoter deletion mutants in transfected CV-1 cells. Figure 3D shows that deletion of the nut-1 promoter from −1467 to −706 resulted in an enhancement of Mta/Rta synergy, from 143-fold to 243-fold. Additional deletions of the promoter to −131 and −73 resulted in sequential reduction of synergy (71- and 22-fold, respectively). The deletion to −73 truncates the promoter just upstream of the RRE (Fig. 3A). These reductions in synergy were independent of the magnitude of Rta transactivation (data not shown). Thus, synergy was generally proportional to the size of the promoter, except for deletion of the sequences between −1467 and −706, which enhanced synergy.

As previously demonstrated, mutation of the 13-bp RRE (Fig. 3C) completely eliminated Rta activation and Mta-Rta synergy (Fig. 3D, Nut-1-1467 mut), affirming that the RRE is necessary for Mta-Rta synergy. To determine whether the RRE was sufficient for Mta-Rta synergy, we cloned the 46-bp DNA that included the RRE, the putative TATA box, and the intervening sequences (Fig. 3A) upstream of the heterologous hsp70 promoter. Although the resulting promoter was well activated by Rta alone (27-fold) (not shown), Mta did not synergize with Rta in this context. Therefore, the nut-1/PAN RRE was necessary but not sufficient for Mta-Rta synergy. Instead, sequences upstream and downstream of the RRE seemed to be required for optimal synergy of Mta with Rta.

Mta protein transcriptionally cooperates with Rta in a promoter-specific manner.

We extended these studies in CV-1 cells by testing six other viral promoters in a fashion similar to that for Fig. 3 (shown in Fig. 4). Rta, but not Mta, activated each of these promoters alone in CV-1 cells (data not shown). Figure 4 shows the level of additional activation over Rta alone when Mta was coexpressed. Among these promoters, those from ori-Lyt (L) and kaposin/ori-Lyt (R) were synergized to the highest magnitudes, 38- and 24-fold, respectively. Synergistic activations of the other promoters, in declining amounts, were 13.8 for ORF57, 13.3 for TK, 6.8 for K-bZIP, and 2.8 for ORF50. Mta-Rta synergy thus differs by promoter. Synergistic activation of the promoters shown in Fig. 4 never reached the magnitude of synergy of the nut-1/PAN −706 promoter (243-fold) (Fig. 3B). However, the two promoters that were preferred for Mta/Rta synergy were those that share the 13-bp homology with nut-1/PAN [i.e., ori-Lyt (L) and kaposin] (Fig. 3C).

Independent transactivation by Mta, and synergy with Rta, is cell line specific.

Vero cells support productive reactivation of KSHV (37, 96), so it is likely that CV-1 cells would also be permissive. However, to determine Mta's function in an uninfected cell line that is more similar to human B cells, the natural KSHV reservoir, we repeated the CV-1 experimental strategy (as in Fig. 3 and 4) in a series of human cell lines. Each was transfected with a range of plasmid amounts, and the maximal value for Rta alone was quantitated. Maximal synergy was then quantitated by cotransfecting the optimum amount of Rta expression vector with a range of Mta expression vector.

The top row of Table 1 shows the values for independent and synergistic transactivation of the nut-1/PAN promoter by Mta and Rta in CV-1 cells that were graphed in Fig. 3B. The next two rows show results from two Burkitt's lymphoma cell lines, BL-41 and BJAB. Similar to CV-1 cells, Rta transactivated the nut-1 promoter in both cell lines, while Mta did not. However, as shown in the right column, Mta synergized with Rta only in BL-41, but not BJAB, cells. The synergistic effect in BL-41 cells was quantitatively similar to that in CV-1 cells. Clearly, there was a cell-specific difference in synergy in B cells.

A number of publications have compared the cellular transcriptome of PEL cells to that of a series of lymphoid cell lines representing different stages of B-cell differentiation. These studies found that the PEL transcriptome clusters most closely with that of plasma cell lymphomas or multiple myelomas (42, 47). Therefore, we tested Mta-Rta synergy in the transfected plasma cell lymphoma lines RPMI 8266 and U266. As shown in Table 1, both cell lines reflected our CV-1 data: Rta activated the nut-1 promoter alone, Mta did not, and both proteins synergized. However, the levels of synergy of Rta and Mta were lower than in CV-1 or BL-41 cells. Independent transactivation by Rta was also clearly more robust in U266 cells.

Finally, in Akata-31 and 293 cells, Mta robustly transactivated the nut-1/PAN promoter independently of Rta (Table 1). Rta also independently transactivated the nut-1/PAN promoter in both cell lines. Coexpression of Mta with Rta also enhanced Rta transactivation alone, but the magnitude of cooperation was not synergistic (i.e., it did not exceed the level of transactivation calculated by multiplying the values of independent transactivation by Mta and Rta). Akata-31 cells are a Burkitt's lymphoma line that is a subclone of Akata cells that has been cured of latent EBV (40). 293 cells are human embryonic kidney epithelial cells that have been used extensively for genetic studies of KSHV replication. Together, these data demonstrate that transactivation of the nut-1/PAN promoter by Mta and synergy between Rta and Mta are cell line specific in human cells.

To confirm that Mta was expressed regardless of its ability to synergize with Rta, we compared its expression in transfected BJAB and BL-41 cells by immunofluorescence. As shown in Fig. 5A, Mta was well expressed in the nuclei of both cell lines. Furthermore, Mta did not increase expression of Rta in BL-41 cells transfected under the conditions used for Table 1 (Fig. 5B), demonstrating that Mta-Rta synergy was not the result of Mta increasing the intracellular Rta concentration. Finally, Mta expressed in transfected 293 cells (Fig. 5C) migrated with an apparent mass similar to that in infected PEL cells (Fig. 2A).

Independent transactivation by Mta is promoter specific.

One potential mechanism of Mta-Rta synergy is that Rta stimulates an “intrinsic” transactivation function of Mta that is otherwise not observed when Mta is expressed independently. In Akata-31 and 293 cells, an endogenous protein might provide an Rta-like function for Mta to transactivate independently of Rta. Since Mta-Rta synergy was promoter specific (Fig. 3 and 4), we reasoned that our hypothesis would be supported if Rta-independent transactivation by Mta shared the same promoter specificity as Mta-Rta synergy.

Therefore, we tested Mta's promoter specificity for independent transactivation in transfected 293 cells. Figure 6A shows that Mta transactivated the full-length nut-1/PAN promoter and each of its deletion mutants with a preference that mirrored the relative differences in Mta-Rta synergy between each two promoters (i.e., as in Fig. 3D): nut-1 (−706) was activated more strongly by Mta than nut-1 (−1476), but deletion to −131 and −73 caused a gradual decrease in Mta-directed transactivation (Fig. 6A). However, the range of relative Mta transactivation among the four promoters was less pronounced than the range in Mta-Rta synergy (Fig. 3D).

In Fig. 3D, we confirmed our earlier observation (45) that the 13-bp RRE was required for Mta-Rta synergy. To test the effect of the RRE mutation on Rta-independent transactivation by Mta, we compared Mta transactivation of the WT and mutated versions of the full-length nut-1/PAN promoter and two of its truncated versions. As shown in Fig. 6B, the RRE mutation did not completely eliminate Mta transactivation in 293 cells but reduced it from about 33% to 67% in a deletion-specific manner.

Figure 6C shows a similar analysis of Mta transactivation of the six additional KSHV promoters that we tested for Mta-Rta synergy (Fig. 4). The ori-Lyt (L) and kaposin promoters were strongly activated by Mta in 293 cells, to 44- and 39-fold, respectively. The ORF57 and TK promoters were moderately activated, to about 15-fold. The K-bZIP and ORF50 promoters were very weakly activated by Mta, to less than fivefold. Together, Fig. 6 demonstrates that the preference of independent promoter activation by Mta in 293 cells closely mirrored the specificity of Mta-Rta synergy in CV-1 cells. The RRE contributed to Mta transactivation in 293 cells but was not required.

Promoter-specific transactivation by Mta is independent of basal promoter activity.

All of the promoters tested in Fig. 6A to C were cloned into identical reporter vectors (pGL3 basic), each driving expression of the luciferase gene as a reporter molecule. It was conceivable that Mta transactivated the luciferase message posttranscriptionally and that Mta's promoter-specific effect was due to differences in intrinsic (basal) transcription controlled by each promoter. However, this was not the case: the intrinsic activity of each promoter (Fig. 6D) was independent of the magnitude of Mta transactivation shown in Fig. 6A to C. For example, the Nut-1 (−1467) and Nut-1 (−73) promoters were transactivated to nearly identical magnitudes by Mta (Fig. 6A), yet the basal activities of the two promoters differed by about eightfold. Similarly, the ORF57 promoter had the highest basal activity of all of the promoters tested, and the kaposin promoter had nearly the lowest basal activity, yet Mta activated the kaposin promoter to about a threefold-greater magnitude.

Mta enhances transcriptional initiation in 293 cells.

To further distinguish between transcriptional and posttranscriptional transactivation by Mta in 293 cells, we performed nuclear run-on transcription assays in cells transfected with the nut-1/PAN reporter in the absence or presence of increasing amounts of Mta. Run-on transcription assays are the classic method to distinguish transcriptional from posttranscriptional transactivation (85). These assays measure the relative number of RNA polymerase molecules actively transcribing a gene at a defined moment. This is achieved by separating cell nuclei from the remainder of the cell, which prevents active polymerase molecules from reinitiating transcription. By incubating the isolated nuclei in the presence of buffered nucleotides and DIG-11-UTP, only nascent transcripts incorporate DIG.

Total RNA from the nuclei was hybridized to excess transcript-specific DNA cross-linked to a nitrocellulose membrane (Fig. 7A). The label was then quantitated by detection of DIG. We quantitated the luciferase transcripts made from the transfected nut-1/PAN reporter and normalized those values with endogenous 7SK transcripts in duplicate transfections (Fig. 7A and B). As shown in Fig. 7C, nuclei from 293 cells transfected with the Mta expression vector showed a dose-responsive transcriptional transactivation up to 10-fold greater than that of nuclei from cells transfected with reporter plasmid and control empty vector. Figures 7A and B also demonstrate that the cellular 7SK gene was transactivated by ORF57/Mta when 4.5 μg of the Mta vector was transfected. This effect on endogenous 7SK transcription therefore reduced the apparent transactivation of the nut-1/PAN promoter by Mta in the normalized data (Fig. 7C), which otherwise would have been 5.5-fold by 4.5 μg of expression vector. These data suggested that Mta transactivation of the nut-1/PAN promoter in 293 cells occurred in part by enhancing transcriptional initiation, perhaps by a mechanism in which Mta functioned as a bona fide transcriptional transactivator. These experiments, however, do not exclude the possibility that Mta acted indirectly by hypothetically stimulating the abundance or activity of a cellular transcription factor that regulates the nut-1/PAN promoter.

Promoter-specific Mta transactivation occurs at the levels of transcription and RNA stability.

In Fig. 6D, we showed that promoter-specific transactivation by Mta was independent of intrinsic promoter activity. However, a direct method for testing Mta's effect on reporter transcript stability is to compare transcript accumulation in the presence and absence of the transcriptional inhibitor ActD. Figure 8A shows that when luciferase was expressed from the nut-1/PAN promoter, its transcript was barely detectable when Mta was not cotransfected, regardless of ActD treatment (lanes 1 to 4). Figure 8A demonstrates that Mta transactivated the nut-1/PAN promoter in the absence of ActD (lanes 5 and 6) and maintained that level of luciferase transcript when ActD was added (lanes 7 and 8).

Comparison of Fig. 8A to B supports a promoter-specific transcriptional and posttranscriptional role of Mta transactivation. When luciferase was expressed from the potent CMV promoter in the absence of Mta, the transcript was easily detectable when ActD was omitted from the culture but became virtually undetectable when ActD was added (Fig. 8B, compare lanes 1 and 2 to 3 and 4). Coexpression of Mta had no effect on luciferase transcript in the absence of ActD (Fig. 8B, compare lanes 5 and 6 to 1 and 2) but maintained luciferase transcript levels in the presence of ActD (Fig. 8B, compare lanes 7 and 8 to 5 and 6). If Mta were transactivating luciferase RNA only by enhancing its stability, then the levels of luciferase RNA would have increased proportionally regardless of the promoter (compare lanes 1 and 2 with 5 and 6 for both Fig. 8A and B). This was not the case, as Mta increased the luciferase transcript 11.8-fold for the nut-1/PAN promoter (Fig. 8A), but not the CMV promoter (Fig. 8B), in the absence of ActD. These data suggested that Mta was indeed transactivating the nut-1/PAN promoter, but not the CMV promoter. However, the data from Fig. 8 also demonstrated that once luciferase is expressed, whether as the result of Mta-mediated promoter transactivation (nut-1/PAN) (Fig. 8A) or by the intrinsic activity of the promoter (CMV) (Fig. 8B), Mta also enhanced the stability of the luciferase RNA (Fig. 8, lanes 7 and 8).

The Rta and Mta proteins bind directly to each other.

The least complicated mechanism to explain synergy between the Rta and Mta proteins is by direct interaction with each other. To test this hypothesis, we generated total protein extracts from BCBL-1 PEL cells that had been treated with TPA for 24 h. Equal amounts of extract were incubated with our antisera that were specific for Mta or Rta or with preimmune sera. Immunoprecipitated proteins were probed by Western blotting using the anti-Mta or anti-Rta antibodies that had been cross-linked to horseradish peroxidase (to eliminate background detection of the precipitated antibodies). As shown in Fig. 9A (left), the anti-Mta serum, but not the preimmune serum, coprecipitated Rta. As shown in Fig. 9A (right), the anti-Rta serum, but not the preimmune serum, coprecipitated Mta. These data prove that Mta and Rta interact in the same complex in extracts from infected cells (a similar result has been published elsewhere [61]).

To confirm the Mta-Rta interaction in vitro, and to determine the amino acid requirements for Rta to bind to Mta, we employed a GST pull-down approach using Mta generated as an N-terminal fusion to the GST moiety. RRLs were programmed with plasmids expressing full-length Rta, or three deletion mutants (Fig. 9B), in the presence of [³⁵S]methionine. The migration of each of the input Rta proteins is demonstrated on the left side of Fig. 9C. The right side of Fig. 9C shows that full-length Rta bound to GST-Mta in vitro, confirming the coimmunoprecipitation (Fig. 9A). Only the C-terminally truncated mutant of Rta, called ORF50ΔSTAD, interacted with Mta. Deletion of the leucine-rich repeats (ΔLR) of Rta, or of Rta aa 275 to 484 (ΔNA), eliminated interaction with Mta. None of the Rta proteins bound to the GST moiety expressed alone (right side of Fig. 9C). Our data suggested that two individual regions of Rta were required for interactions with Mta.

Mta is a DNA-binding protein.

The data described above demonstrated that Mta/Rta synergy and Rta-independent Mta transactivation were promoter specific. The simplest mechanism by which Mta achieved promoter specificity was by binding DNA in a sequence-specific fashion. To determine whether Mta bound to DNA directly, we generated Mta as an N-terminal fusion to a six-histidine epitope. The full-length protein was expressed and purified to near homogeneity in E. coli (shown by GelCode Blue stain in Fig. 10A). Using EMSAs, the His₆-Mta was tested for binding to a 99-bp DNA from the nut-1 promoter. The DNA included the 13-bp RRE, the TATA box, and the transcriptional start site of nut-1/PAN (−73 to +26) (Fig. 3A shows the probe location). As shown in Fig. 10B, His₆-Mta formed at least four complexes with the DNA at the lowest concentration of protein tested. As we increased the concentration of His₆-Mta, there was a dose-responsive increase in the intensity of the protein-DNA complexes and a concomitant appearance of slower-migrating complexes. As a negative control, we repeated the EMSAs but substituted His₆-RBP-Jk for Mta, which did not result in formation of DNA-protein complexes (Fig. 10B). These data suggested that Mta protein bound directly to the nut-1/PAN promoter in the absence of Rta.

To determine whether Mta associated with the nut-1/PAN promoter in infected cells, we performed ChIP experiments. Chromatin from mock-induced or TPA-induced BC-3 cells was cross-linked with formaldehyde 40 h post-TPA treatment. Extracts were incubated with our anti-Mta serum, positive control anti-Rta serum, or negative control preimmune serum or purified rabbit IgG. We analyzed the ChIP-assayed DNA using quantitative real-time PCR with primers specific for a 178-bp fragment of the nut-1/PAN promoter spanning the TATA box and 13-bp RRE (positions −141 to +37) as described previously (64). Figure 10C shows the results of amplification of the ChIP-assayed DNA. Equal amounts of chromatin from mock-induced and TPA-induced BC-3 cells were used as input for the ChIP (“input” lanes). However, the anti-Mta serum precipitated the nut-1/PAN promoter only following TPA induction, but not in mock-induced cells. Since the rabbit IgG control showed only weak background bands, this suggested that Mta bound to the nut-1/PAN promoter in infected cells during reactivation of KSHV from latency. Note that the lanes for the anti-Rta serum (Fig. 10C) surpassed the linear range of the PCR; quantitation is shown in Table 2.

The PCRs were quantitated in real time by detecting amplification products using fluorescent, self-quenching molecular beacons. As shown in Table 2, the background level of chromatin that was immunoprecipitated using the negative control, preimmune serum, was set at 1.0× for normalization of the other ChIPs. Mta was enriched 27.4-fold on the nut-1/PAN promoter in infected cells following TPA addition. In comparison, at the same time point following reactivation, Rta was enriched 6.3-fold on the nut-1/PAN promoter. The negative control antibody, total rabbit IgG, yielded a background value of 1.4-fold.

As a negative control for gene-specific DNA binding, we used the same ChIP-assayed DNA but substituted PCR primers and a corresponding molecular beacon that was designed to specifically detect the DNA encoding the KSHV K6 ORF. As shown in Table 2, Mta and Rta were enriched on the K6 ORF by 2.9- and 1.6-fold, respectively. These values were below the value for enrichment using the negative control IgG, which was 4.5-fold. Together, these data suggest that both Mta and Rta bound to the nut-1/PAN promoter in infected cells within the same TATA-proximal span of 178 bp. It is unclear whether direct binding of Mta to Rta contributes to our ability to ChIP Mta on the nut-1/PAN promoter in vivo.

A putative A/T hook domain contributes to DNA binding and transactivation by Mta.

Analysis of Mta's primary sequence revealed a basic domain that was enriched in arginine (Fig. 11A). The software program MOTIF (http://www.genome.jp/) revealed a putative A/T hook domain within the R domain of Mta (aa 119 to 131) (Fig. 11B). A/T hooks are DNA-binding domains found in a great variety of cellular proteins that act as accessory proteins for transcription factors (1, 4). To determine whether the putative A/T hook contributed to Mta DNA binding, we deleted the domain in the full-length His₆ fusion of Mta (called MtaΔA/T) and performed EMSA with the nut-1/PAN (−73 to +26) DNA. As shown in Fig. 11C, both WT and mutant Mta formed multiple complexes with the PAN DNA; however, three of the DNA-protein complexes formed only with WT Mta but not MtaΔA/T.

To test the functional significance of the Mta DNA-binding mutation, we asked whether the mutant Mta could transactivate the nut-1/PAN promoter/reporter in transfected 293 cells. As shown in Fig. 11D, transactivation by the MtaΔA/T mutant was reduced by at least 60% at all input plasmid amounts compared to WT Mta. These data supported a role for Mta DNA binding in transactivation of the nut-1/PAN promoter.

The A/T hook overlaps a portion of Mta that has been suggested to contribute to Mta's nuclear localization (59). We found that the MtaΔA/T DNA-binding mutant was well expressed and correctly localized to the nuclei of transfected 293 cells (Fig. 11E).