Transcriptional Regulation of Kaposi's Sarcoma-associated Herpesvirus-encoded Oncogene Viral Interferon Regulatory Factor by a Novel Transcriptional Silencer, Tis *

Viral interferon regulatory factor (vIRF) encoded by Kaposi's sarcoma-associated herpesvirus (KSHV) has been shown to transform NIH3T3 and Rat-1 cells, inhibit interferon signal transduction, and regulate the expression of KSHV genes. We had previously characterized the vIRF core promoter and defined a 12-O-tetradecanoylphorbol-13-acetate (TPA)-responsive region in the upstream regulatory sequence of vIRF gene. Here, we have further identified a novel transcriptional silencer, named Tis in this region. Tis represses the promoter activities of vIRF and heterologous herpes simplex virus thymidine kinase genes in both position- and orientation-independent manners. Deletion analysis has identified acis-element of 23 nucleotides that is essential for the negative regulation. Two Tis-binding protein complexes, named vR1 and vR2, were observed by electrophoretic mobility shift assays using nuclear extracts from both KSHV-negative and -positive cell lines. A sequence fragment GAGTTAATAGGTAGAG in thecis-element was shown to be required for the DNA-protein interactions as well as the repression of vIRF promoter activity. Point-mutation analysis identified TTAAT and GTTAATAG as the core sequence motifs for the binding of vR1 and vR2, respectively. These results define the function of a novel transcriptional silencer in the regulation of vIRF gene expression.


From the Departments of Pediatrics and Microbiology, University of Texas Health Science Center at San Antonio, San Antonio, Texas 78229-3900
Viral interferon regulatory factor (vIRF) encoded by Kaposi's sarcoma-associated herpesvirus (KSHV) has been shown to transform NIH3T3 and Rat-1 cells, inhibit interferon signal transduction, and regulate the expression of KSHV genes. We had previously characterized the vIRF core promoter and defined a 12-O-tetradecanoylphorbol-13-acetate (TPA)-responsive region in the upstream regulatory sequence of vIRF gene. Here, we have further identified a novel transcriptional silencer, named Tis in this region. Tis represses the promoter activities of vIRF and heterologous herpes simplex virus thymidine kinase genes in both position-and orientation-independent manners. Deletion analysis has identified a cis-element of 23 nucleotides that is essential for the negative regulation. Two Tis-binding protein complexes, named vR1 and vR2, were observed by electrophoretic mobility shift assays using nuclear extracts from both KSHV-negative and -positive cell lines. A sequence fragment GAGTTAATAGGTAGAG in the cis-element was shown to be required for the DNA-protein interactions as well as the repression of vIRF promoter activity. Point-mutation analysis identified TTAAT and GTTAATAG as the core sequence motifs for the binding of vR1 and vR2, respectively. These results define the function of a novel transcriptional silencer in the regulation of vIRF gene expression.
The life cycle of herpesviruses consists of two phases: latent infection and lytic infection. Viral latent infection is generally associated with the development of herpesviruses-related tumors. Viral lytic replication often causes cell death; however, it also produces virions that are spread to other hosts (15). In Kaposi's sarcoma tumors, the majority of KSHV-infected cells express viral latent antigens; nonetheless, a small number of these cells also undergo viral lytic replication (16). KSHV encodes a unique set of nonstructural genes targeting at cellular signaling pathways, most of which are viral lytic genes (17). It has been suggested that KSHV lytic replication in a small number of Kaposi's sarcoma tumor cells is essential for sustaining Kaposi's sarcoma lesions through a paracrine mechanism (16). Thus, delineation of the molecular mechanism controlling the expression of KSHV lytic genes could help understand the pathogenesis of KSHV-related diseases.
One of the KSHV nonstructural regulatory lytic genes is the ORF-K9 that encodes the viral interferon regulatory factor (vIRF). vIRF is a 449-amino acid protein that shares sequence homology to cellular interferon regulatory factors (IRFs) (17). IRFs are a family of transcription factors that regulate interferon signal transduction through binding to interferon-stimulated response elements in the promoter of interferon-responsive genes (18 -24). Early reports have demonstrated that vIRF causes cellular transformation of NIH3T3 and Rat-1 cells through the inhibition of interferon-and IRF-mediated signal transduction, prevents UV-induced apoptosis, and regulates the expression of KSHV genes (25)(26)(27)(28). vIRF exerts these functions through direct binding to IRF-1, IRF-3, p300, cAMP response element-binding protein, and p53 tumor suppressor (29 -32). These results indicate that vIRF is a genuine viral oncogene and a potent transcriptional regulator. Elucidation of the molecular mechanism controlling vIRF expression could lead to the understanding of its precise role in regulating the expression of cellular and KSHV genes.
Two ORF-K9-related transcripts, a major transcript of 1.7 kb and a minor transcript with additional 84 nt sequence upstream of the 5Ј-end of the major transcript have been identified (33,34). The expression of the minor transcript is weak, and can only be detected by nested-RT-PCR in uninduced KSHV-infected cells (33). We have previously demonstrated that the major transcript of vIRF gene is a viral early lytic transcript (34), whose expression can be induced by chemical stimuli such as 12-O-tetradecanoylphorbol-13-acetate (TPA) (34 -37). We have also characterized the core promoter of vIRF gene, and defined a TPA-responsive region (34). In this study, we have further identified a novel transcriptional silencer Tis in the upstream regulatory sequence of the vIRF promoter.
Construction of pBLCAT and ⌬pCAT-991 Reporter Plasmids-The CAT reporter plasmid pBLCATReg was constructed by inserting the PCR-amplified sequence from Ϫ337 to Ϫ125 upstream vIRF transcriptional start site into the HindIII/BamHI sites upstream of the HSV-TK promoter (pBLCAT2) fused to the CAT gene. pBLCATReg-rev was constructed by inserting the sequence from Ϫ337 to Ϫ125 in a reverse orientation into pBLCAT2 ( Fig. 2A).
CAT reporter plasmids pBLCATRegI, pBLCATRegII, and pBLCA-TRegIII were constructed by inserting the corresponding regulatory sequence of vIRF promoter into the HindIII/BamHI sites upstream of TK promoter in pBLCAT2. These plasmids contain the upstream regulatory sequences of vIRF promoter from Ϫ258 to Ϫ161, Ϫ323 to Ϫ234, and Ϫ499 to Ϫ334, respectively (Fig. 3, A and B).
⌬pCAT-991 was generated by deleting the RegI-2 sequence in wild type reporter plasmid pCAT-991 using QuikChange TM Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA). Briefly, sequence deletion was performed on the pCAT-991 construct using oligonucleotide primers that lack the targeted sequence. The reaction was performed for 18 cycles at 95°C for 30 s, 55°C for 1 min, and 68°C for 9 min. After temperature cycling, the parental supercoiled dsDNA was digested with DpnI restriction enzyme at 37°C for 60 min and used to transform the Epicurian Coli XL1-Blue supercompetent cells (Stratagene). The oligonucleotides used for the deletion of RegI-2 sequence were as follows (* denotes deletion position): Da oligonucleotide 5Ј-GATTTTGTATGA-TGTTT*GCTCTTTCTGACATATC-3Ј; Ds oligonucleotide 5Ј-GATATG-TCAGAAAGAGC*AAACATCATACAAAATC-3Ј.
The constructs D1pBLCATRegI-2, D2pBLCATRegI-2, and D3pBLCATRegI-2 were generated by inserting the oligomers with the target sequence motifs deleted into the SalI site upstream of the TK promoter in pBLCAT2 (Fig. 7, A and B). All constructs were confirmed by DNA sequencing with a Big Dye Terminator Cycle Sequencing Kit on an ABI 373-S sequencer (PE Bio-system, Foster City, CA).
Transient Transfection and CAT Assay-All plasmids used for transient transfection were prepared with QIAfilter Plasmid Maxi Kit (Qiagen Inc., Valencia, CA). Transfection experiments were performed with LipofectAMINE TM 2000 reagent according to the instructions of the manufacturer (Invitrogen, Carlsbad, CA). CAT assay was performed as previously reported (25). Transfection efficiencies were normalized by co-transfection with a reporter plasmid pSV-␤-galactosidase and ␤-galactosidase activity was determined following the instructions of the manufacturer (Promega). The conversion rate of the modified 14 C-labeled chloramphenicol was calculated with a Multi-Analysis Program (Bio-Rad, Hercules, CA).
Preparation of Nuclear Extracts-Nuclear extracts from HeLa, BC-1, BC-2, and BCBL-1 cells were prepared as described previously (40). Briefly, 1 ϫ 10 7 cells were collected and washed three times with phosphate-buffered saline by centrifugation. Each cell pellet was suspended in 400 l of low-salt buffer (1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol (DTT), 10 mM Hepes (pH 7.9)), and swelled on ice for 15 min. After the addition of 25 l of 10% Nonidet P-40, the cell pellet was mixed and vortexed for 10 s. Following centrifugation at 13,000 ϫ g for 30 s, the pellet containing the nuclei was resuspended by gentle stirring with a pipette tip in 50 l of extraction buffer (25% glycerol, 1.5 mM MgCl 2 , 420 mM NaCl, 1 mM DTT, and 20 mM Hepes (pH 7.9)). The mixture was then incubated with rocking for another 20 min at 4°C. After centrifugation at 16,000 ϫ g for 45 min, the supernatant was collected, diluted with equal amount of dilution buffer (5 mM Hepes (pH 7.9), 20% glycerol, and 1 mM DTT), and stored at Ϫ70°C for future use.
To determine the core sequence motif(s) of Tis, the M0 oligomer, 5Ј-GAGTTAATAGGTAGAG-3Ј (Fig. 8A), was used as wild type sequence. Sequential point mutations were introduced for each of the nucleotides to test their effects on DNA-protein interactions. The mutant oligomers were employed as competitors in EMSA with the labeled oligomer Tis as probe. The wild type and mutant oligomers were listed in Fig. 8A.

RESULTS
The Upstream Sequence of vIRF Gene Promoter Contains a cis-Element Repressing vIRF Promoter Activity-In a previous study, we have mapped the vIRF core promoter to a DNA fragment from ϩ62 to Ϫ56 upstream of vIRF transcriptional start site (34). Reporter plasmids containing DNA fragments from ϩ62 to Ϫ56 (pCAT-56) and Ϫ125 (pCAT-125) had strong CAT activities similar to pBLCAT2 (Fig. 1). In contrast, reporter plasmids containing DNA fragments from ϩ62 to Ϫ337 (pCAT-337), Ϫ499 (pCAT-499), and Ϫ991 (pCAT-991) had 88, 91, and 93% reduced CAT activities, respectively, compared with pCAT-125 (Fig. 1C). These results suggest that the 5Јflanking region of vIRF gene contains a negative regulatory cis-element repressing its promoter activity and this element is Repression of Heterologous HSV-TK Promoter Activity by the Upstream Regulatory cis-Element of vIRF Promoter-To determine whether the upstream negative regulatory cis-element of the vIRF gene promoter can exert its effect on a heterologous promoter, a single copy of the DNA fragment (Ϫ337 to Ϫ125) was inserted into the upstream of the TK promoter in pBLCAT2 ( Fig. 2A) to generate pBLCATReg plasmid. Compared with that of pBLCAT2, the TK promoter activity (pBLCATReg) was reduced 87% in HeLa cells (Fig. 2B), indicating that the cis-element also has a repression effect on the heterologous TK promoter. Similarly, the TK promoter activity was reduced 85% when the cis-element was inserted in a reverse orientation into pBLCAT2 (pBLCATReg-rev and Fig.  2B), indicating that the repression function of the cis-element is orientation-independent.
Mapping of the Negative Regulatory cis-Element in vIRF Promoter-To map the region involved in the repression of promoter activity, the sequence region from Ϫ337 to Ϫ125 was dissected into two fragments (Ϫ161 to Ϫ258 and Ϫ234 to Ϫ323), and their corresponding CAT reporter plasmids pBLCATRegI and pBLCATRegII were constructed by inserting these fragments into the upstream of TK promoter in pBLCAT2 (Fig. 3, A and B). Since the sequence region from Ϫ499 to Ϫ334 had no repression effect (Fig. 1), its corresponding construct pBLCATRegIII was used as a control. When transiently transfected into HeLa cells, pBLCATRegII showed promoter activity similar to pBLCATRegIII, suggesting the region from Ϫ323 to Ϫ234 is not responsible for the repression of promoter activity. In contrast, pBLCATRegI had 85% lower promoter activity than that of the parental plasmid pBLCAT2 (Fig. 3C), indicating that the region between Ϫ258 and Ϫ161 in the vIRF gene promoter is involved in the transcriptional repression of the promoter activity.
To further map the negative regulatory cis-element, 4 overlapping oligonucleotides within the region from Ϫ258 to Ϫ161 (RegI) were synthesized and used to construct plasmids pBLCATRegI-1, pBLCATRegI-2, pBLCATRegI-3, and pBLCA-TRegI-4 (Fig. 4, A and B). As shown in Fig. 4C, pBLCATRegI-1, pBLCATRegI-3, and pBLCATRegI-4 had CAT activities similar to that of pBLCAT2. In contrast, pBLCATRegI-2 had 85% lower CAT activities than that of the parental plasmid pBLCAT2. These results indicate that the RegI-2 sequence contains the negative regulatory cis-element, whose negative repression function is independent on its position to the TATA box. A, diagram illustrating the upstream regulatory sequence of vIRF promoter and DNA fragments used to generate CAT reporter constructs. B, 5Ј-deletion constructs of vIRF promoter were created by inserting different PCR-amplified DNA fragments upstream of the vIRF initiating ATG into pCAT-Basic vector, a promoter-less and enhancer-less vector, and used for defining the negative regulatory sequence of vIRF promoter. The constructs were named according to the 5Ј-end nucleotide number upstream of the transcriptional start site (ϩ1). C, repression of vIRF promoter activity by its upstream promoter sequences. CAT reporter plasmids were transfected into HeLa cells with LipofectAMINE TM 2000 reagent. The cells were harvested 48 h post-transfection, and CAT assay was performed to assess the CAT activity. The result is the average relative CAT activity with standard deviation from three independent experiments compared with that of pCAT-125. The pBLCAT2 containing the HSV-TK promoter fused to the CAT gene was used as a reference control.
FIG. 2. The negative regulatory cis-element in the vIRF promoter represses heterologous HSV-TK promoter activity. A, the DNA fragment from Ϫ337 to Ϫ125 upstream of vIRF transcriptional start site was inserted into the HindIII and BamHI sites upstream of HSV-TK promoter (pBLCAT2) fused to the CAT gene in a forward or a reverse orientations. B, CAT activity was assessed after transfection of the reporter plasmid into HeLa cells. The result is the average relative CAT activity with standard deviation from three independent experiments compared with that of pBLCAT2.
FIG. 3. The region from ؊258 to ؊161 in vIRF gene promoter contains the negative regulatory cis-element. A, diagram illustrating the sequences from vIRF gene initiating ATG to the upstream regulatory sequence of vIRF promoter (Ϫ991), and two different regions (Ϫ323 to Ϫ234 and Ϫ258 to Ϫ161) used for mapping the negative regulatory cis-element. The region from Ϫ499 to Ϫ334 was used as a control sequence without repression. The number refers to the nucleotide number upstream of the vIRF transcriptional start site (ϩ1). B, three different regions of the upstream regulatory sequences of vIRF promoter were inserted into the HindIII/BamHI sites upstream of the HSV-TK promoter (pBLCAT2) fused to the CAT gene to test their effect on TK promoter activity. The result is the average relative CAT activity with standard deviation from three independent experiments compared with that of pBLCAT2.
To further confirm the effect of RegI-2 on the repression of vIRF promoter activity, RegI-2 sequence was deleted in the reporter plasmid pCAT-991 by site-directed mutagenesis. As shown in Fig. 5, deletion of the RegI-2 sequence (⌬pCAT-991) relieved its repression effect on wild type pCAT-991. The CAT activity of ⌬pCAT-991 reached 78% of pCAT-125 (100%). These results demonstrated the critical role of the RegI-2 sequence on the repression of vIRF promoter activity. Taken together, the above results indicate that the negative regulatory cis-element (Ϫ241 to Ϫ219) in vIRF promoter is a transcriptional silencer, and is named Tis.
Binding of Transcriptional Regulators to Tis-To determine whether Tis has any potential binding site(s) for transcriptional repressor(s), its sequence, RegI-2 (5Ј-CTCCACGAGTTA-ATAGGTAGAGT-3Ј), was used as a probe in an EMSA. As shown in Fig. 6A, when crude nuclear extract from HeLa cells was used, two major shifted complexes were observed (lane 2). Treatment of the nuclear extracts with unlabeled RegI-2 strongly inhibited the intensities of the shifted complexes (lane 3), indicating that both shifted complexes were RegI-2-specific DNA-protein interaction complexes. The protein complexes bound to the upper and lower bands were named as vIRF regulatory sequence-binding protein complex 1 and 2 (vR1 and vR2), respectively.
To determine whether the DNA-protein binding pattern could be reproduced in KSHV-infected cell lines, the nuclear extracts from BC-1, BC-2, and BCBL-1 cells were employed in EMSA. As shown in Fig. 6B, the RegI-2-specific DNA-protein interaction complexes were also observed in BC-1, BC-2, and BCBL-1 cell lines, indicating that the RegI-2-binding proteins are present in both KSHV-negative and -positive cell lines.

Mapping of the Minimal Binding Site(s) of Transcriptional Repressor(s)-To define the binding site(s) of transcriptional repressor(s) involved in the transcriptional repression, the
MatInspector2.2 program was used to analyze the potential binding sites of transcriptional factors (41). Comparative analysis of RegI-2 sequence with the GenBank TM data base did not identify any identical binding sequence of known transcriptional factors. However, the sequence was found to contain the core sequence motifs corresponding to several potential transcriptional factor(s), including a Myc/Max motif, a Nkx2.5 motif, and an ␦EF1 motif (Fig. 7A).
FIG. 4. Identification of the region from ؊241 to ؊219 in vIRF gene promoter as the negative regulatory cis-element. A, RegI sequence (Ϫ258 to Ϫ161), and four overlapping oligomers (Ϫ258 to Ϫ230, Ϫ241 to Ϫ219, Ϫ231 to Ϫ201, and Ϫ203 to Ϫ161) from the upstream regulatory sequence of vIRF promoter used for mapping the negative regulatory cis-element. B, the constructs pBLCATRegI-1, pBLCATRegI-2, pBLCATRegI-3, and pBLCATRegI-4 were generated by cloning the above oligomers into SalI site of pBLCAT2. C, the constructs pBLCATRegI-1, pBLCATRegI-2, pBLCATRegI-3, and pBLCATRegI-4 were transfected into HeLa cells, and the CAT assay was performed. The result is the average relative CAT activity with standard deviation from three independent experiments compared with that of pBLCAT2.
FIG. 5. Deletion analysis to demonstrate the essential role of RegI-2 in the repression of vIRF promoter activity. ⌬pCAT-991 was generated by deleting the RegI-2 sequence in the wild type pCAT-991 construct. Deletion of RegI-2 in pCAT-991 strongly increased the CAT activity to the level comparable with that of the control pCAT-125, a vIRF promoter reporter construct that lacks RegI-2 sequence (Fig. 1).

FIG. 6. DNA-protein interaction of RegI-2 (Tis) sequence. A,
RegI-2 was used as a probe in an EMSA to identify its DNA-protein interaction complexes using nuclear extract from HeLa cells. The nuclear extract was added to all reactions except that of lane 1. Two major shifted complexes named vR1 and vR2 were observed (lane 2), which were competed by unlabeled RegI-2 (lane 3). B, vR1 and vR2 complexes could also be detected using nuclear extracts from KSHV-infected cell lines. EMSA with the nuclear extracts from HeLa (lanes 2 and 3), BC-1 (lanes 4 and 5), BC-2 (lanes 6 and 7), and BCBL-1 (lanes 8 and 9) using RegI-2 as a probe. Lane 1 is a control without nuclear extract; lanes 2, 4, 6, and 8 were the reactions without competitors; lanes 3, 5, 7, and 9 were the reactions competed by unlabeled RegI-2.
To further determine whether these sequence motifs are critical for the binding of vR1 and vR2, the oligonucleotides D1RegI-2, D2RegI-2, and D3RegI-2 were used as competitors of RegI-2 probe in an EMSA. As shown in Fig. 7D, D1RegI-2 with the sequence motif CACGAG deleted strongly competed with RegI-2 for the formation of both vR1 and vR2 (lane 6). This pattern is similar to that when RegI-2 itself is used as a competitor (lane 3), suggesting that the sequence motif CAC-GAG is not essential for the binding of the repressor(s). When the oligomers D2RegI-2 (lane 4) and D3RegI-2 (lane 5) were used as competitors, the DNA-protein binding pattern remained unchanged, indicating that both sequence motifs TTA-ATAGG and ATAGGTAG were essential for the DNA-protein interactions. Indeed, the oligomer D1RegI-2 produced the same shifted patterns as RegI-2 did (Fig. 7E). In contrast, no shifted complex could be detected when labeled oligomers D2RegI-2 and D3RegI-2 were used as probes (Fig. 7, F and G). Taken together, these findings suggest that both sequence motifs TTAATAGG and ATAGGTAG are required for the binding of the repressor(s).

Determination of the Core Binding Sequence Motif(s) of Transcriptional Repressor(s) and Identification of Tis as a Novel
Transcriptional Silencer-To determine the contribution of each nucleotide in the DNA-protein interaction and identify the core binding motifs of the transcriptional repressor(s), point mutations were introduced into the wild type oligomer (M0) sequence 5Ј-GAGTTAATAGGTAGAG-3Ј. A panel of sequential mutant oligonucleotides (M1 to M15) were generated and used as competitors in an EMSA using oligomer RegI-2 as a probe (Fig. 8A). As shown in Fig. 8B, the wild type oligomer M0 had strong competition for both vR1 and vR2. When used as competitors, the mutants M1, M2, and M11 to M15 had almost the same competition effect as RegI-2 for both vR1 and vR2 complexes. However, mutants M4 to M8 competed less efficiently for vR1, indicating that the TTAAT motif is required for the binding of vR1 (Fig. 8, B and C). Similarly, mutants M3 to M10 did not compete for vR2, suggesting the GTTAATAG motif is essential for the binding of vR2 (Fig. 8, B and D).
To further confirm the effect of the mutated nucleotides (M3 to M10) on the binding of the repressor(s), representative mutants M6 and M9 with mutations of critical nucleotides were labeled and used as probes in EMSA. As shown in Fig. 8E, no protein binding could be detected using labeled oligomers M6 (lane 2) and M9 (lane 5), confirming the importance of the mutated nucleotides for the binding of transcriptional repressor(s).
The above results indicate that Nkx2.5 is the most likely transcriptional factors involved in the DNA-protein interactions of vR1 and vR2. To further determine whether Myc/Max, ␦EF1, and Nkx2.5 are involved in Tis DNA-protein interactions, gel supershift assays with specific antibodies to the respective transcriptional factors were performed. As shown in Fig. 9A, antibodies to all three transcriptional factors had no effect on the gel shift pattern. Since Nkx2.5 is a heart tissuespecific transcriptional factor, it should not be expressed in cell lines used in our studies. Indeed, the Nkx2.5 transcript was not detected in HeLa, BC-1, and 293 cells by RT-PCR although it was detected in a control RNA sample from heart tissue (Fig.  9B). Together, the above results strongly indicate that transcriptional factors Myc/Max, ␦EF1, and Nkx2.5 are not involved in Tis DNA-protein interactions, and Tis is a novel transcriptional silencer. FIG. 7. Deletion analysis of the binding sites of transcriptional repressor(s). A, potential binding sites of known transcriptional factors in RegI-2 sequence. B, RegI-2 was used as a wild type in deletion analysis to define the motifs involved in Tis transcriptional repression. The sequence motifs CACGAG, TTAATAGG, and ATAGGTAG within RegI-2 were independently deleted to generate the oligonucleotides D1RegI-2, D2RegI-2, and D3RegI-2. C, the plasmids D1pBLCATRegI-2, D2pBLCATRegI-2, and D3pBLCATRegI-2 were constructed by inserting the oligomers D1RegI-2, D2RegI-2, and D3RegI-2 into the upstream of the TK promoter in pBLCAT2, respectively. The constructs were transfected into HeLa cells and the CAT assay was performed. The result is the average relative CAT activity with standard deviation from three independent experiments compared with that of pBLCAT2. D, EMSA with RegI-2 as a probe and various oligomers as competitors. Nuclear extract from HeLa cells was added to all reactions except that of lane 1. Both vR1 and vR2 complexes were observed (lane 2), which were competed by unlabeled RegI-2 (lane 3) and D1RegI-2 (lane 6), but not by D2RegI-2 and D3RegI-2 (lanes 4 and 5). All competitors were 100-fold excess of the probe. E-G, EMSA with oligomers D1RegI-2, D2RegI-2, and D3RegI-2 as probes. vR1 and vR2 complexes were observed using probe D1RegI-2 (E), but not detected using probes D2RegI-2 (F) and D3RegI-2 (G). Lanes 1, 4, and 7 are reactions without nuclear extract; lanes 2, 5, and 8 are reactions with nuclear extract; lanes 3, 6, and 9 are reactions competed by unlabeled D1RegI-2, D2RegI-2, and D3RegI-2, respectively.

DISCUSSION
We have previously observed that a region in vIRF promoter has a transcriptional repression effect (34). In this study, we have employed deletion analysis to further examine a 1.052-kb 5Ј-flanking region of the vIRF gene in an attempt to identify the functional regulatory element that is responsible for such repression. Our results demonstrated that the sequence region from Ϫ337 to Ϫ125 had a strong repression effect on the CAT reporter gene driven by both vIRF promoter and heterologous TK promoter (Figs. 1 and 2). Dissection of this region showed that RegI (Ϫ258 to Ϫ161) is required for the repression activity (Fig. 3). The repression function of this cis-element is positionand orientation-independent, indicating the presence of a transcriptional silencer within this region. Further dissection of the RegI sequence identified a 23-bp DNA fragment from Ϫ241 to Ϫ219 (RegI-2) as the transcriptional silencer Tis (Fig. 4). Deletion of Tis in wild type pCAT-991 strongly relieved its repression effect on vIRF promoter (Fig. 5). EMSA identified two major shifted complexes, vR1 and vR2, in this region, which can be competed by unlabeled RegI-2 (Fig. 6), indicating that both vR1 and vR2 were specific. These results point to the presence of binding motif(s) of transcriptional repressor(s) in Tis. Since Tis is capable of repressing the promoter activities of both vIRF gene and heterologous HSV-TK gene, the repression function of RegI-2 may be universal rather than vIRF promoter specific.
Our data demonstrated that vR1 and vR2 complexes observed in EMSA with RegI-2 as probes are specific and responsible for the transcriptional repression of vIRF promoter. Comparative analysis with the database of transcriptional factors did not match RegI-2 with any identical sequences of known transcriptional silencers. However, the RegI-2 sequence was found to contain the potential core sequence motifs corresponding to the binding sites of transcriptional factors Myc/Max, Nkx2.5, and ␦EF1 (Fig. 7). Deletion of the sequence motifs CACGAG, TTAATAGG, and ATAGGTAG corresponding to these transcriptional factors relieved the repression effect of RegI-2 on promoter by 2, 80, and 50%, respectively, indicating that the sequence motif TTAATAGGTAG is critical for the repression of the vIRF promoter activity. When D1RegI-2, D2RegI-2, and D3RegI-2 were used as competitors in EMSA, D1RegI-2 inhibited the formation of the shifted complexes, while D2RegI-2 and D3RegI-2 had no effect. Meanwhile, when D1RegI-2, D2RegI-2, and D3RegI-2 were used as probes for EMSA, D1RegI-2 had the shifted pattern similar to that of RegI-2, while both of D2RegI-2 and D3RegI-2 formed no shifted complexes (Fig. 7). Point mutation analysis demonstrated that the sequence motif TTAAT is critical for the binding of the repressor(s) in vR1, and the sequence motif GTTAATAG is critical for the binding of repressor(s) in vR2 (Fig. 8). This sequence motif is very similar to the binding motif of Nkx2.5. However, antibodies to Nkx2.5, ␦EF1, and Myc/Max did not supershift vR1 and vR2 complexes or affect gel shift patterns in EMSA, indicating that all three transcriptional factors are not involved in Tis DNA-protein interactions (Fig. 9A). Furthermore, Nkx2.5 is a known heart tissue-specific factor whose expression is absent in cell lines used in this study, thus excluding its role in the formation of vR1 and vR2 complexes (Fig. 9B).
Regulation of eukaryotic gene expression is usually involved with multiple transcriptional factors, which function to activate or repress transcription via binding to cis-regulatory elements. Transcriptional repressors, in particular, play a central role in vital biological processes, such as the development and regulation of cell growths. Our results have demonstrated that the Tis region from Ϫ241 to Ϫ219 in the vIRF gene promoter is a transcriptional silencer. This region interacts with a yet unidentified transcriptional repressor(s) that represses vIRF promoter activity. Negative regulation has been suggested to play a critical role in modulating the expression of eukaryotic genes (42)(43)(44)(45). A large number of studies performed in the last decade have led to a better understanding of the mechanism of negative regulation in gene expression (46 -54). Although there are some other potential binding sites such as AP1 and SP1 for the binding of positive regulators in the upstream regulatory sequence of the vIRF promoter, our data suggest that negative regulation is predominant in the regulation of vIRF gene expression in latent steady-state.
Transcriptional repressors are generally divided into two categories: passive repressors and active repressors. Passive repressors function by competing with activators or basal transcriptional factors for the access to their binding DNA sequences, and the recruitment of inhibitory chromatin components to the promoter. The active repressors inhibit gene transcription by directly interacting with positive regulators or transcriptional factors of basal transcriptional machinery (44,45). Previous studies have demonstrated that vIRF is a viral early gene whose expression is minimal in KSHV latent replication but could be increased to high level during KSHV lytic replication (34 -37). Based on the previous reports and our current study, we propose the following model for the role of Tis in the regulation of vIRF gene expression (Fig. 10). In KSHV latent replication, the expression of vIRF gene is repressed by the transcriptional repressor(s), vR1 and vR2, which bind to Tis and interact with the basal transcription complexes to repress vIRF gene transcription. In KSHV lytic replication or after induction with TPA, the positive regulator(s), either of viral or cellular origin, might interact with the transcriptional repressor(s) vR1 and vR2 or other components of basal transcriptional machinery to prevent the interaction of vR1 and vR2 with Tis or other transcriptional factors, thus activating the transcription of the vIRF gene. We put forward this hypothesis based on the following evidence. 1) The expression of vIRF gene is minimal in latent replication in KSHV cell lines (34 -37) and the CAT activities of the reporter constructs pCAT-991, pCAT-499, and pCAT-337 containing Tis were all strongly repressed (34). 2) Deletion of Tis relieved its repression on both vIRF and TK promoters. 3) Deletion of the sequence motifs within Tis abolished the binding of the repressor(s) to the transcriptional silencer. 4) Both the expression of vIRF gene and the CAT activity of the reporter construct pCAT-991 that contains a TPA-responsive element (TPA-RE) could be induced to higher level by TPA treatment or transfection with KSHV transactivator ORF-50 in KSHV-infected cell lines (33)(34)(35)(36). Although some other mechanisms might also be involved in the regulation of vIRF gene expression, in particular, the direct transactivation of vIRF promoter by viral or cellular positive regulators such as the product of KSHV ORF-50, Tis is likely to have a critical role in the efficient repression of vIRF gene expression in KSHV latent replication. Future studies are intended to identify the repressor(s) and other elements that are involved in the regulation of vIRF gene expression, and determine whether transcriptional silencing is a general transcriptional regulation mechanism for KSHV lytic genes during viral latent infection.