Leucine Zipper Domain Is Required for Kaposi Sarcoma-associated Herpesvirus (KSHV) K-bZIP Protein to Interact with Histone Deacetylase and Is Important for KSHV Replication*

Background: The repressive effect of K-bZIP on ORF50 and OriLyt promoters is controversial with regard to its importance in KSHV replication. Results: The leucine zipper domain is essential for K-bZIP to interact with HDAC and inhibit HDAC activity. Conclusion: Our results suggested that K-bZIP regulates gene expression through its interaction with HDAC. Significance: The finding is a first step in designing systems or therapies to control KSHV gene expression. The Kaposi sarcoma-associated herpesvirus (KSHV; or human herpesvirus-8)-encoded protein called K-bZIP (also named K8) was found to be multifunctional. In this study, we discovered that K-bZIP interacts with histone deacetylase (HDAC) 1/2 in 12-O-tetradecanoylphorbol-13-acetate-stimulated BCBL-1 lymphocyte cells. K-bZIP appears to repress HDAC activity through this interaction, which we determined to be independent of K-bZIP SUMOylation. We dissected the domains of K-bZIP and found that the leucine zipper (LZ) domain is essential for the interaction of K-bZIP and HDAC. In addition, we constructed a KSHV bacterial artificial chromosome (BAC) with LZ domain-deleted K-bZIP (KSHVdLZ) and transfected this mutated KSHV BAC DNA into HEK 293T cells. As a result, it was consistently found that K-bZIP without its LZ domain failed to interact with HDAC2. We also showed that the interaction between K-bZIP and HDAC is necessary for the inhibition of the lytic gene promoters (ORF50 and OriLyt) of KSHV by K-bZIP. Furthermore, we found that the LZ domain is also important for the interaction of K-bZIP with the promoters of ORF50 and OriLyt. Most interestingly, although it was found to have suppressive effects on the promoters of ORF50 and OriLyt, KSHVdLZ replicates at a significantly lower level than its BAC-derived revertant (KSHVdLZRev) or KSHVWT (BAC36) in HEK 293T cells. The defectiveness of KSHVdLZ replication can be partially rescued by siRNA against HDAC2. Our results suggest that the function of K-bZIP interaction with HDAC is two-layered. 1) K-bZIP inhibits HDAC activity generally so that KSHVdLZ replicates at a lower level than does KSHVWT. 2) K-bZIP can recruit HDAC to the promoters of OriLyt and ORF50 through interaction with HDAC for K-bZIP to have a temporary repressive effect on the two promoters.

Kaposi sarcoma-associated herpesvirus (KSHV) 2 is associated with Kaposi sarcoma (1,2) and several B cell malignancies, such as primary effusion lymphoma and multicentric Castleman disease (1,(3)(4)(5)(6). Cell types identified for successful KSHV infection include monocytes, endothelial/spindle cells, B cells, and epithelial cells (6 -11). After primary infection, KSHV can set up a latent infection in the host cells where KSHV genomes exist as episomes in the nucleus, and the latent infection can be reactivated to a lytic infection to produce and release infectious viral particles (10,12,13). Accumulating studies have revealed that the KSHV latency to lytic switch is important in viral pathogenesis, in secondary infection (to maintain the number of infected cells), and in tumorigenesis (14 -17). Unfortunately, the mechanism by which KSHV is reactivated from latent infection is still unclear.
The findings that histone deacetylase (HDAC) inhibitors, such as sodium butyrate, can reactivate KSHV from latent infection (18,19) suggest that procedures affecting the activities of histone acetylase and HDAC might be related to viral reactivation. HDACs are a category of enzymes with the ability to change substrates from the acetylated to the deacetylated state, which results in a tighter chromatin structure (for histones) or reduced transactivation activity of other substrates, such as gene transcription regulators like p53 (19,20). HDAC1 and -2 are the Class I proteins of the category and exist as abundant nuclear proteins in all kinds of mammalian cells (21). HDAC1 and -2 always exist together and are found in three different complexes (Sin3, NuRD/NRD/Mi2, and CoREST) (20). HDACs have been found to be associated with several herpesviruses, such as cytomegalovirus (CMV) and herpes simplex type 1 (HSV-1), interacting with viral promoters or viral proteins and negatively affecting their gene expression, leading to subsequent reduction in viral replication (22)(23)(24)(25). Several
Molecular Cloning and Site-directed Mutagenesis of KSHV BAC-To mutate the K-bZIP gene, we used overlapping PCR to produce a mutated DNA fragment (for a diagram, see Fig. 4B) that was then cloned into a full-length K-bZIP cDNA in pcDNA3 to replace wild type with mutant sequence. The mutations were verified by DNA sequencing. To generate KSHV BAC with LZ-deleted K-bZIP, we used a galK counterselection BAC system (52). Briefly, we first replaced K-bZIP gene with a galK cassette (generated by PCR using the primers in Table 1) based on BAC36 (53). The recombinant bacteria (SW102) can only grow on minimal medium with galactose as the only carbon source and yield red colonies. Then the galK cassette was replaced by a PCR product that contains LZ-deleted or BACderived revertant K-bZIP DNA (generated by PCR using the primers shown in Table 1). This is achieved by selecting against the galK cassette by resistance to 2-deoxygalactose on minimal plates with glycerol as the carbon source. 2-Deoxygalactose is harmless unless phosphorylated by functional galK. Phosphorylation by galK turns 2-deoxygalactose into 2-deoxygalactose 1-phosphate, a non-metabolizable and therefore toxic intermediate that killed non-transformants. The resultant KSHV BACs (BACdLZ and BACdLZRev) were further verified by DNA sequencing of K-bZIP gene.
Luciferase Assay-Cells were collected 24 h after co-transfection of K-bZIP-and/or HDAC2-expressing plasmids with a luciferase reporter gene directed by the promoter of ORF50 (genomic location: 70561-71598, Genbank TM accession number U75698 (54), pORF50-luc), ORF59 (genomic location: 96737-98034, accession number U75698, pORF59-luc), or OriLyt (genomic location: 24093-24342, accession number U75698, pOriLyt-luc). pRL-TK was included in each transfection system as an internal control for the normalization of the DNA transfection. The Dual-Luciferase reporter assay system (Promega) was used to examine the responsiveness of the promoters to K-bZIP and/or HDAC2. Each assay was performed in triplicate, and the luciferase activity was normalized by the amounts of total protein. Transfection efficiency was normalized with the Renilla luciferase activities (pRL-TK). The cell lysates were assayed for firefly luciferase and Renilla luciferase activities by using a TD-20/20 luminometer with a dual autoinjector (Promega, Turner Designs). The luciferase assays were carried out according to the manufacturer's instructions (Promega).
HDAC Activity Assay-HDAC activity was assessed with the HDAC activity assay kit (Upstate-Millipore, Lake Placid, NY) according to the manufacturer's instructions. Immune complexes were incubated with 20,000 cpm [ 3 H]acetyl-labeled histone H4 peptide (Upstate-Millipore, Lake Placid, NY) in 1ϫ HDAC buffer at room temperature for 24 h with rolling. Reactions were stopped by adding 50 l of 0.12 N acetic acid, 0.72 N HCl. The released acetate was extracted in 0.5 ml of ethyl acetate and mixed in 5 ml of scintillation solution, and radioactivity was measured in a scintillation counter. All assays were performed in triplicate.
Nick Translation-Double-stranded DNA probes for in situ hybridization and Southern blot assays were labeled by nick translation as described previously (55). Briefly, 1 g of plasmid DNA (BAC36; the whole KSHV genome was cloned in the BAC vector, which was a gift from Dr. S. J. Gao) (53), 10ϫ nick translation buffer, 0.05 mM dNTP (dATP, dCTP, and dGTP), 0.01 mM dTTP, 0.04 mM biotinylated UTP, 1 unit of DNA polymerase I, and appropriate concentrations of DNase I were incubated at 15°C for 50 min. Labeled fragments obtained from the protocol were 200 -500 bases long as determined on 2% agarose gels.
Immunocytochemistry and Fluorescence in Situ Hybridization-For the immunofluorescence assay of adherent cells, cells were grown on round coverslips (Corning Glass Inc., Corning, NY) in 24-well plates (Falcon, BD Biosciences). For immunofluorescence assays in BCBL-1 cells, cells were washed with PBS, fixed with 1% paraformaldehyde, and cytospun to slides for the immunofluorescence assay, using different antibodies as desired. Cells were fixed in 1% paraformaldehyde (10 min at room temperature) and permeabilized in 0.2% Triton (20 min on ice) by sequential incubation with primary and Texas Red-or fluorescein isothiocyanate (FITC)-labeled secondary antibodies (Vector Laboratories, Burlingame, CA) for 30 min each (all solutions in PBS). For simultaneous detection of viral protein and specific DNA sequences, cells were first immunostained for viral proteins and then treated for 1 h at 37°C with RNase (Roche Applied Science; 100 g/ml in PBS) for the detection of DNA. After refixing in 4% paraformaldehyde (10 min at room temperature), samples were equilibrated in 2ϫ SSC (1ϫ SSC is 0.15 M NaCl plus 0.015 M sodium citrate), dehydrated in ethanol (70,80, and 100% ethanol for 3 min each at 20°C), air-dried, and incubated overnight at 37°C with the hybridization mixture. For DNA detection, probe and cells were simultaneously heated at 94°C for 4 min to denature DNA. After hybridization, samples were washed at 37°C with 55% formamide in 2ϫ SSC (twice for 15 min each), 2ϫ SSC (10 min), and 0.25ϫ SSC (twice for 5 min each). Hybridized probes were labeled with Texas Red-avidin (Vector Laboratories; 1:500 in 4ϫ SSC plus 0.5% BSA), and signals were amplified by using biotinylated anti-avidin (Vector Laboratories; 1:250) followed by another round of Texas Red-avidin staining. Finally, cells were equilibrated in PBS, stained for DNA with Hoechst 33258 (0.5 g/ml), and mounted in Fluoromount-G (Fisher Scientific).
Immunoblot Analysis-Proteins were separated by sodium dodecyl sulfate-7.5% polyacrylamide gel electrophoresis (SDS-PAGE) (10 -20 g loaded in each lane), transferred to nitrocellulose membranes (Amersham Biosciences), and blocked with 5% nonfat milk for 60 min at room temperature. Membranes were incubated overnight at 4°C with primary antibody followed by an incubation with a horseradish peroxidase-coupled secondary antibody (Amersham Biosciences) and detection with enhanced chemiluminescence (Pierce). For the detection of protein in the immunoprecipitation, we used secondary antibodies from TrueBlot ULTRA (eBioscience; catalog number 18-8817 for mouse or 18-8816 for rabbit). Membranes were stripped with stripping buffer (100 mM ␤-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.8); washed with PBS, 0.1% Tween 20; and used to detect additional proteins.
Preparation of Nuclear Extracts-Nuclear extracts were obtained essentially as described previously (56). Briefly, monolayer cells were washed with PBS and scraped into fresh Eppendorf tubes. Cell pellets were resuspended in cold buffer A (10 mM HEPES-KOH, pH 7.9 at 4°C, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride) and incubated at 4°C for 10 min. After centrifugation, pellets were resuspended in cold buffer C (20 mM HEPES-KOH, pH 7.9, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.5 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride) by vortexing and incubated at 4°C for 30 min. Clarified extracts were transferred to fresh tubes and stored at Ϫ70°C until use.
Coimmunoprecipitation-Antibodies were coupled to protein G-Sepharose beads (Amersham Biosciences) according to the manufacturer's instructions. After a wash with PBS, 0.1% bovine serum albumin, beads were incubated overnight at 4°C with clarified extracts; washed again in PBS, 0.1% bovine serum albumin; and resuspended in a mixture of PBS and 2ϫ Laemmli buffer (20 l of each). After heating at 95°C for 5 min, beads were removed by centrifugation, and supernatants were analyzed by SDS-PAGE and immunoblotting.
Viral DNA Isolation from Cells and Southern Blotting-Viral DNA samples were isolated by Hirt's method with modification for KSHV genome-harboring cells (57). The DNA was separated by running a 1% agarose gel and transferred onto a Nylon membrane (Bio-Rad). A biotin-labeled probe made from the full-length KSHV BAC DNA by nick translation was hybridized to the membrane. After the membrane was washed according to standard protocol, the hybridization signal was detected using an Ambion detection system (Ambion Inc., Austin, TX).
Chromatin Immunoprecipitation (ChIP) Assay and Real Time PCR-HEK 293T cells harboring BACdLZ, its BAC-derived revertant, or BAC36 were stimulated with TPA (20 ng/ml) and sodium butyrate (0.5 mM) for 24 h and fixed with 1% formaldehyde. A ChIP assay was performed using the EZChIP kit (Upstate-Millipore, Billerica, MA) according to the manufacturer's protocol. The amount of the DNA ChIPed by antibodies (normal IgG, anti-K-bZIP, and anti-HDAC2) was examined by real time PCR in a 25-l reaction using the primers shown in Table 1 that were designed according to Ref. 27. All samples were analyzed in triplicate using SYBR Green 2ϫ Master Mix (Applied Biosystems, Foster City, CA) on an Applied Biosystems 7900HT Fast Real-Time PCR System. After an initial denaturation incubation of 5 min at 95°C, 45 cycles of threestep cycling were performed with an annealing temperature of 60°C. Melt curve analysis was then performed to verify product specificity. The relative ratio of Ct of each antibody-ChIPed DNA over respective IgG (rabbit IgG or mouse IgG)-ChIPed DNA was determined by examining the change in threshold cycle number (delta Ct) with standard deviation calculated to account for variance in both IgG-and antibody-ChIPed DNA signal.
Confocal Microscopy-Cells were examined with a Leica TCS SPII confocal laser-scanning system. Two or three channels were recorded simultaneously and/or sequentially and controlled for possible breakthrough between the green and red channels and between the blue and red channels. To balance the signal strength, we used Leica image enhancement software to scan the image to separate the signal from its background. Because of the variability among cells in any culture in terms of cellular size and shape, we selected the most typical cells that were photographed, and they are presented here at high magnification.

HDAC1 and -2 Were Recruited into DNA Replication
Domains with K-bZIP-Cellular proteins begin interacting with viral proteins soon after infection. Previously, K-bZIP has been shown to colocalize with ND10 proteins, such as PML, which is known to defend against herpesviral infection (58,59). We wished to explore whether the other cellular defensive proteins (especially HDAC family proteins) would interact with KSHV K-bZIP. First we examined the relationship of K-bZIP with HDAC1, HDAC2, and PML at early times after reactivation of BCBL-1 cells by TPA (Fig. 1). After 12 h of TPA treatment, the cells were stained with anti-K-bZIP and secondary antibody in red (A, D, G, and J-L) and anti-HDAC1 (B and J), anti-HDAC2 (E and K), and anti-PML (H and L) in green. We can see that upon reactivation by TPA KSHV latency (without treatment of TPA; Fig. 1, J, K, and L) was switched to the lytic stage, thus making K-bZIP detectable. K-bZIP formed small punctate dots (A, D, and G). After merging (C, F, and I), K-bZIP dots were shown to colocalize with ND10 (PML in green; G-I) but were not associated with HDAC1 (A-C) or HDAC2 (D-F) that diffuse in the nucleus. The colocalization of K-bZIP and PML was consistent with previous reports (36,60).
Previously, we also reported that K-bZIP is present in viral DNA replication domains (61). Viral DNA replication happens at the late stage after the reactivation of BCBL-1 cells by TPA. As shown in Fig. 2, G-I, we performed DNA hybridization using a probe that was made from KSHV genomic DNA by nick translation. The BCBL-1 cells were stained with anti-K-bZIP antibody in red (G), and the DNA is shown in green (H). There are two cells in Fig. 2, G-I; the small green dots in the right cell represent the episomes (around 10 per cell on average) of KSHV DNA still in latency, which is similar to that in BCBL-1 cells not treated with TPA (Fig. 2J). However, in the cell on the left, the green domain represents the viral DNA replication domain (H). At this point, K-bZIP no longer exists as dots but instead as a large domain (G) that overlaps the DNA replication domain (I).
The DNA replication of KSHV in BCBL-1 cells after treatment with TPA was also evidenced by Southern blot assay (Fig. 2K), demonstrating that no variations in KSHV DNA signal can be detected without treatment with TPA, compared with increased KSHV DNA signal at 24 h after TPA treatment.
In all cells with observable DNA replication domains, K-bZIP formed domains that overlapped the DNA replication domains (61). Therefore, K-bZIP domains can be used to show KSHV DNA replication domains. At 48 h after TPA treatments, BCBL-1 cells were fixed, and an immunofluorescence assay was performed to show HDAC1 and -2 ( Fig. 2) localization. HDAC1 and HDAC2 always form a complex together to deacetylate histones and modify DNA conformation into a tight structure, making it more difficult for transcriptional activators to contact their targets. We observed that both HDAC1 and -2 are recruited to K-bZIP domains (Fig. 2, C and F) some time between 12 and 48 h. These results suggest that cellular defensive proteins (HDAC1 and HDAC2) respond to the reactivation process of KSHV and might play a critical role in blocking lytic gene expression.
K-bZIP Interacts with HDAC and Reduces HDAC Activity-To determine the manner in which HDAC1 and HDAC2 are recruited into the DNA replication domains, we asked whether K-bZIP interacts with HDAC1 or HDAC2. Nuclear extracts were prepared from BCBL-1 cells with or without TPA treatment and incubated with protein G beads conjugated with anti-K-bZIP (mouse), anti-HDAC1 (mouse), or anti-HDAC2 (mouse) antibody or normal IgG from mouse as a control. Pulled down proteins were then detected by Western blot assays using anti-K-bZIP, -HADC1, and -HDAC2 antibodies. Because the molecular weights of HDAC1 and HDAC2 are close to that of the heavy chain of IgG, we used secondary antibodies from TrueBlot ULTRA to mask the band of IgG heavy chain. As can be seen in Fig. 3A, we discovered that K-bZIP interacted with HDAC1 and HDAC2 as they both can bind to the anti-K-bZIP-conjugated beads. Because K-bZIP was found to be related only to HDAC1/2 in DNA replication domains at the late stage ( Fig. 2) but not at the early stage ( Fig. 1; debatable, not easily detected), it is also unknown whether the interaction of K-bZIP with HDACs was mediated by KSHV DNA replication. Next, we wondered what the consequence of the HDAC/ K-bZIP interaction might be. To measure the HDAC2-specific deacetylation activity, cell lysates from BCBL-1 cells (treated or untreated with TPA) were immunoprecipitated with anti-  anti-K-bZIP, -HDAC1, and -HDAC2, and mouse IgG (mIgG) were bound to protein G beads and incubated with nuclear extracts. The eluted protein complexes were detected by Western blot using antibodies to detect the respective proteins as indicated on the right. To avoid the heavy chain of IgG, we used secondary antibodies from TrueBlot ULTRA (eBioscience; catalog number 18-8817 for mouse or 18-8816 for rabbit). B, effect of K-bZIP on HDAC2 activity. BCBL-1 cells were untreated (lanes 1 and 3) or treated (lanes 2 and 4) with TPA for 48 h. A total of 1.2 mg of cell lysates was immunoprecipitated with anti-HDAC2 (left) or anti-K-bZIP (right) antibodies. One-third of the precipitant was assayed for deacetylase activity with or without 1 M sodium butyrate. The results shown for deacetylation activity are the averages of three independent assays. IP, immunoprecipitation. The S.E. of three assays was shown as the error bar.
HDAC2 antibody (because HDAC1 and -2 exist in the same complexes, we only used anti-HDAC2 antibody) or anti-K-bZIP antibody and used for the assay. The results showed that HDAC2-associated deacetylation activity decreased slightly after treatment with TPA for 48 h (Fig. 3B). HDAC2 activity from the complex co-immunoprecipitated by anti-K-bZIP antibody was reduced significantly (Fig. 3B). However, the K-bZIP-associated HDAC activity was reduced by an HDAC inhibitor, sodium butyrate, suggesting that the K-bZIP-bound HDAC(s) may retain at least partial deacetylation activity that is sensitive to sodium butyrate.
Leucine Zipper Domain, Not K-bZIP SUMOylation, Is Important for Interaction of K-bZIP and HDAC2-To determine whether KSHV DNA replication is required for the interaction of K-bZIP and HDAC2, we transfected a K-bZIP-expressing plasmid into HEK 293 cells free of the other components of KSHV. Nuclear extracts were prepared 24 h post-transfection and incubated with anti-K-bZIP antibody (mouse)-or anti-HDAC2 antibody (rabbit)-conjugated beads, and Western blot assays were performed to check proteins in the complexes (Fig.   4A). As shown, HDAC2 can be pulled down by anti-K-bZIP antibody; K-bZIP can also be pulled down by anti-HDAC2 antibody. Therefore, K-bZIP interaction with HDAC2 was not dependent on KSHV DNA replication.
SUMOylation can affect the location of a protein in the nucleus and its interaction with other molecules. K-bZIP can be SUMOylated, so we wondered whether SUMOylation affects the interaction of K-bZIP with HDAC2. The K-bZIP protein SUMOylation motif has been mapped to a single amino acid residue (lysine 158) (43), and a SUMO interaction motif has been identified in amino acids 72-75 (41). We constructed several deletion mutations and a point mutation of K-bZIP (Fig. 4B), including pK8_K158R point mutant, pK8_dl72-75 (deleted SUMO interaction motif), pK8_dl122-132 (deleted nuclear localization signal), and pK8_dl205-219 (deleted leucine zipper domain). After co-transfection of these plasmids with pgfpSUMO-1 (SUMO-1 was N-terminally fused with GFP) into HEK 293 cells, the cell lysates were run in a Western blot assay using anti-GFP and anti-K-bZIP antibodies as indicated in Fig. 4B (right). As can be seen, K-bZIPs with mutated FIGURE 4. Interactions of K-bZIP with HDAC2. A, nuclear extracts were prepared from HEK 293 cells that were transfected with wild-type (WT) pK-bZIP. Mouse (m) anti-K-bZIP, mouse IgG, rabbit (r) anti-HDAC2, and rabbit IgG were bound to protein G beads and incubated with nuclear extracts. The eluted protein complexes were detected by Western blot using antibodies to check the respective proteins as indicated on the right. B, HEK 293 cells were co-transfected with pgfpSUMO-1 and K-bZIP-expressing plasmids (mutation diagram is shown on the left) for 24 h. Cell lysates were assayed by Western blot using antibodies as indicated on the right. C, nuclear extracts were prepared from HEK 293 cells that were transfected with wild type or mutated pK-bZIP. Mouse anti-K-bZIP, mouse IgG (mIgG), rabbit anti-HDAC2, and rabbit IgG (rIgG) were bound to protein G beads and incubated with nuclear extracts. The eluted protein complexes were detected by Western blot using antibodies to detect the respective proteins as indicated on the right. D, immunofluorescence assay to show K-bZIP with deleted LZ domain distribution in the nucleus. IP, immunoprecipitation.

FIGURE 5. Interaction of K-bZIP and HDAC2 with KSHV gene promoters and transcription regulation of OriLyt, ORF50, and ORF59 promoters by K-bZIP, ORF50, HDAC, and K-bZIP mutants.
A, HEK 293 cells were co-transfected with reporter plasmids containing the firefly luciferase gene under the control of the OriLyt (lower), ORF59 (middle), or ORF50 (upper) promoter with pcDNA3 (to normalize the input DNA amount), K-bZIP-, or ORF50-expressing plasmid. Renilla luciferase plasmid was included in each transfection as an internal control. At 24 h post-transfection, Dual-Luciferase assays were performed with the cell lysates of transfected cells. Relative luciferase activities were calculated by dividing the normalized firefly luciferase activity of each reporter by that of the pGL3 plasmid in pcDNA3-transfected cells. B, ChIP assay. HEK 293T cells harboring BAC36, BACdLZ, or its BAC-derived revertant were fixed with 1% paraformaldehyde at 24 h post-treatment with TPA (20 ng/ml) and sodium butyrate (0.5 mM). The whole cell lysates were sonicated to fragment the DNA. The DNA was then ChIPed using anti-HDAC2, anti-K-bZIP, or anti-Ac-histone H3 (AcH3) antibody. The ChIP assay was performed using an EZChIP kit (Upstate-Millipore, Billerica, MA) according to the manufacturer's protocol. The ChIPed DNAs were finally examined by real time PCR using the primers amplifying promoters of ORF50, ORF59, and OriLyt ( Table 1). The relative DNA quantities were calculated by comparing the PCR signals from normal IgG and specific antibodies with that from input. C, the same as in A but different plasmids (K-bZIP mutants) were applied for co-transfection. Results in A and C are mean Ϯ S.D. from three independent assays. SUMOylation target (K158R), deleted SUMO interaction motif, and deleted nuclear localization signal were not SUMOylated. However, K-bZIP with loss of SUMOylation can still be pulled down by anti-HADC2 antibody in a co-immunoprecipitation assay (Fig. 4C, left). Therefore, SUMOylation of K-bZIP is not important for its interaction with HDAC1/2.
To map out the domain that is essential for the interaction of K-bZIP with HDAC1/2, we deleted the leucine zipper domain, pK8_dl205-219. K-bZIP with a deleted leucine zipper domain can still be SUMOylated (Fig. 4B, right) but loses the ability to interact with HDAC2 in the co-immunoprecipitation assay (as shown in Fig. 4C, right). Therefore, the leucine zipper domain is essential for the interaction of K-bZIP with HDAC2. We found that the distribution pattern of K-bZIP with deleted LZ also changed, losing the punctate nuclear pattern and becoming diffuse (Fig. 4D).

K-bZIP Protein Represses Several KSHV Lytic Promoters via Its
Interaction with HDAC2-A comprehensive test of KSHV gene activation by ORF50 and K-bZIP proteins was conducted recently (27). However, the effects of K-bZIP on important lytic stage promoters (including ORF50, ORF59, and OriLyt) were not clear. Previous unpublished experiments performed in our laboratory have repeatedly revealed that K-bZIP has inhibitory effects on the promoters of ORF50 and OriLyt, effects that were not evident in the global detection assays (27). For that reason, we performed co-transfection of the luciferase reporter vectors using K-bZIP-or ORF50-expressing plasmids to test the effects of either K-bZIP or ORF50 on those promoters. Results showed that K-bZIP has repressive effects on the KSHV lytic promoters (ORF50 and OriLyt) but no significant effect on the ORF59 promoter (Fig. 5A).
To know whether the K-bZIP or LZ-deleted K-bZIP could bind to the OriLyt or ORF50 promoters in HEK 293T cell lines, a ChIP assay using anti-HDAC2, anti-K-bZIP, or anti-acetylated histone (Ac-histone H3) antibody was performed. The ChIPed DNAs were then examined by real time PCR using the primers amplifying promoters of ORF50, ORF59, and OriLyt ( Table 1). As shown in Fig. 5B, after comparing the PCR signals from normal IgG and specific antibodies with that from input, we found that both HDAC2 and K-bZIP bind to promoters of ORF50 and OriLyt as does Ac-histone H3 in HEK 293T cells harboring BAC36 and BACdLZRev (Fig. 5B, upper  and lower panels). Association of K-bZIP with the promoter of ORF59 was not obvious. Although the association of HDAC2 and Ac-histone H3 with all promoters is evident and was detected in HEK 293T cells harboring BACdLZ, the level of DNA ChIPed by anti-HDAC2 is significantly lower compared with that from cells with BAC36 and BACdLZRev (Fig.  5B). That the effect is prominent only in ORF50 and OriLyt suggests that the K-bZIP leucine zipper domain must be important for K-bZIP to recruit HDAC2 to the promoters of ORF50 and OriLyt.
Finally, we wanted to know whether the interaction of K-bZIP with HDAC2 is important for the repression of the two promoters by K-bZIP. We co-transfected the HEK 293 cells with the luciferase-tagged reporter plasmids with ORF50 or OriLyt with K-bZIP, non-SUMOylated pK-bZIPK158R, or LZ-impaired pK-bZIPdl205-219. As shown in Fig. 5C, pK-bZIP158KR represses ORF50 and OriLyt nearly as well as wild-type K-bZIP. However, pK-bZIPdl205-219 lost the ability to repress the two promoters. Given the result that the LZ domain is important for interaction with HDAC (Fig. 4), these results suggest that the repressive function of K-bZIP might be due to recruitment of HDAC2 to promoters of OriLyt and ORF50 for K-bZIP.
Leucine Zipper Domain Is Necessary for K-bZIP to Interact with Promoters of ORF50 and OriLyt-As reported above, K-bZIP affected KSHV promoters of ORF50 and OriLyt. We asked whether the effects of K-bZIP on the KSHV promoters require interaction with DNA and if so whether the HDAC could be brought to KSHV promoters by K-bZIP to play repressive roles on gene expressions. For that purpose, we con- Primers used for PCRs of K-bZIP, ChIP assay, BAC system, and KSHV DNA Primers were designed according to the KSHV genomic sequence (accession number U75698 ). The primers were used for amplifying K-bZIP gene, KSHV gene promoters, and ORF73 and for the BAC system.

Purpose of the primer DNA sequence K-bZIP gene
structed a KSHV BAC DNA with LZ-deleted K-bZIP gene and its BAC-derived revertant (BACdLZRev). The BAC DNAs were verified as follows. 1) XhoI digestion showed that the pattern of KSHV BACdLZ was indistinguishable from that of its BACderived revertant or BAC36 (Fig. 6A). 2) PCR of K-bZIP gene showed a slightly smaller band from KSHVdLZ than that from its BAC-derived revertant or BAC36 (Fig. 6B). 3) DNA sequencing showed that amino acids 205-219 were deleted inframe (data not shown). We then transfected the BAC DNA (BAC36, BACdLZ, or BACdLZRev) into HEK 293T cells to make HEK 293T cell lines harboring KSHV genomes: HEK 293T/BAC36, 293T/BACdLZ, and 293T/BACdLZRev. The transiently transfected cells were purified by cell sorting of a GFP marker so that nearly all cells were KSHV BAC-positive. After treatment with TPA for different times as indicated in Fig. 6C, we collected the whole cell lysate samples. KSHV proteins (K-bZIP, RTA, latency-associated nuclear antigen, and ORF45) were then analyzed by Western blot. As can be seen, HEK 293T cell lines harboring KSHV BACs can all express viral proteins that are important for viral replication, including K-bZIP. Fig. 6C shows that KSHV RTA expression is reduced during reactivation when LZ is deleted from K-bZIP in the BAC (the second row of Western blot), which could be caused by the loss of the repressive effect of K-bZIP on HDAC. Co-immunoprecipitation assays (Fig. 6D) using anti-K-bZIP and anti-HDAC2 antibodies showed that LZ-deleted K-bZIP cannot interact with HDAC2; this is consistent with the results obtained in BCBL-1 cells and the transfection system (Figs. 3 and 4).
We wished to determine whether the LZ domain of K-bZIP and HDAC2 are important for KSHV to replicate in HEK 293T cells. Cells harboring BAC36, BACdLZ, or BACdLZRev were prepared in 6-well plates and treated with TPA for different times as indicated in Fig. 7 (upper panel). Cells were collected together with medium and treated with protease K, and total viral DNA was extracted. KSHV DNA from cells was quantified by real time PCR compared with an external BAC36 standard. As can be seen, KSHV-BACdLZ replicated at a lower level than its BAC-derived revertant and KSHV-BAC36, especially at the late time point of 48 h. This might suggest that the LZ is important for late (48 h) but not early (24 h or less) actions of K-bZIP. Taken together, the leucine zipper domain is important for FIGURE 6. KSHV BACdLZ and its BAC-derived revertant. A, a seamless (galK counterselection) BAC system was used to construct KSHVdLZ, which contains an LZ-deleted K-bZIP. XhoI digestion was followed by running an agarose gel to separate the DNA bands. The gel shows that BACdLZ has the same pattern as those of BAC36 and BACdLZRev. B, PCR using the primer to amply K-bZIP gene shows that the size from BACdLZ is slightly smaller that those from BAC36 and BACdLZRev. C, HEK 293T cells harboring BAC36, BACdLZ, and BACdLZRev were treated with TPA and sodium butyrate for 0, 12, 24, and 48 h, and the whole cell lysates were applied for Western blotting assay using antibodies against latency-associated nuclear antigen (LANA), RTA, K-bZIP, and ORF45. Tubulin was used for controlling the sample loading. D, co-immunoprecipitation assays were performed to determine the interaction of K-bZIP with HDAC2 using the nuclear extracts from HEK 293T cells harboring BAC36, BACdLZ, and BACdLZRev. Results show that K-bZIP with deleted LZ fails to interact with HDAC2 (middle panel), whereas WT K-bZIP interacts with HDAC2 (left and right panels). mIgG, mouse IgG; rIgG, rabbit IgG; IP, immunoprecipitation; r, rabbit; m, mouse.
K-bZIP to interact with some KSHV promoters and for KSHV replication in HEK 293T cells. The interaction of K-bZIP with HDAC has two different functions: 1) to repress HDAC activity as shown in Fig. 3B to favor viral replication and 2) to recruit HDAC to and repress promoters of ORF50 and OriLyt. However, the general effects of K-bZIP through interaction with HDAC are apparently positive for KSHV replication because abolishing interaction with HDAC by deleting the LZ domain produced a defective phenotype of viral gene expression and DNA replication.
To further demonstrate that the effects of K-bZIP on KSHV DNA replication are connected to HDAC activity, we knocked down HDAC2 using small hairpin RNA carried by a lentivirus against HDAC2 (shRNA plasmid: sc-44262-SH) from HEK 293T cells harboring the three different BAC DNAs. The effect of the siRNA to specifically inhibit HDAC2 gene expression is shown by Western blot assay in Fig. 7, right, lower panel. Interestingly, the defectiveness of KSHVdLZ replication can be significantly rescued by the inhibition of HDAC2 gene expression as shown in Fig. 7, left, lower panel. Taken together, although the K-bZIP has a negative effect on some KSHV promoters, its general function in KSHV is enhancing viral DNA replication through inhibiting HDAC activity.

DISCUSSION
Interest in K-bZIP of KSHV originated from the fact that it is the positional homologue of Zta, which is a reactivator in EBV (6,34,39,54,62). Both Zta of EBV and K-bZIP of KSHV are related to the basic leucine zipper (bZIP) family of transcription factors; moreover, both genes are adjacent to another conserved transcription activator, RTA. However, after comparison of the functions and protein structures of K-bZIP of KSHV with Zta of EBV, it was recognized that the two proteins have very limited similarities (33). First, KSHV K-bZIP alone cannot switch KSHV from latent to lytic infection (33). Second, within the amino acid sequence, K-bZIP of KSHV and EBV Zta are not significantly homologous (33). Finally, K-bZIP of KSHV lacks a basic DNA binding region adjacent to its dimerization domain, and it has not been demonstrated to interact with DNA directly (33). Therefore, KSHV K-bZIP might have different functions than does EBV Zta.
The role of K-bZIP in gene regulation has been widely investigated, and it is believed to be essential for KSHV reactivation. However, the part it plays in KSHV reactivation is not yet fully understood. Accumulated evidence shows that K-bZIP can interact with different cellular and viral proteins to present both trans-repressive and trans-activating activities (29,40,44,63,64). K-bZIP can interact with SUMO (43), ND10 components (60,65), CBP (44,66), and CCAAT/enhancer-binding protein (67). Although it has several functions, its primary function is to arrest cell cycle progression in the G 1 phase (65, 68); this is accomplished by its interaction with cellular proteins and results in the regulation of cell cycle protein production (65,67). Its post-translational modifications (including SUMOylation) have also attracted tentative interest because they are related to the repressive effect of K-bZIP on gene regulation (43) and because in addition SUMOylation often aids protein interaction (69). K-bZIP protein was discovered to inhibit TGF-␤ signaling through interaction with CBP (44). CBP is an acetylase and modifies histone structure to loosen DNA conformation, making it more accessible for gene transcription factors.
HDAC family proteins deacetylate histones and have the opposite function of CBP. Using a co-immunoprecipitation assay, we observed that K-bZIP can interact with HDAC1 and  Table 1). The DNA copies were determined by comparison with a standard KSHV BAC DNA (known concentration of the DNA). Lower panel, HEK 293T/BAC36, /BACdLZ, and /BACdLZRev cells were infected with lentivirus carrying shRNA against HDAC2 and selected with puromycin. The puromycin-resistant cells were assessed for HDAC2 production by Western blot assay as shown on the right, and tubulin was used to control the sample loading. The three types of cells were then stimulated with TPA for different times as indicated, and the viral DNA level was determined by real time PCR as in the upper panel.
-2. And yet, both HDAC1 and HDAC2 are present in the DNA replication domain, and both colocalize with K-bZIP; for this reason, we believe that K-bZIP has additional, as yet unverified functions. The interactions of K-bZIP and HDAC were not dependent on KSHV DNA replication because K-bZIP interacts with HDAC in a co-transfection system. SUMO modification did not affect the interaction of K-bZIP with HDAC. Moreover, LZ-deleted K-bZIP failed to interact with HDAC, demonstrating that the leucine zipper domain is essential for the interaction.
However, the functions of K-bZIP on KSHV promoters are not the same. On the one hand, in a co-transfection system using a luciferase assay, we found that the interaction of K-bZIP with HDAC is required for K-bZIP to perform its inhibitory function on KSHV promoters ORF50 and OriLyt. On the other hand, K-bZIP interaction with HDAC can directly repress the activity of HDAC, and a recent comprehensive study of the function of K-bZIP showed that it can activate 21 other KSHV promoters (16). These data suggest that the repression by K-bZIP of some promoters may be important for the maintenance of latency, whereas later activation of other promoters will be part of the lytic reactivation pathway. In the established HEK 293T cells harboring KSHV BAC, after treatment with TPA, KSHV with a LZ-deleted K-bZIP had a reduced replication phenotype. This implies that the repressive effect of K-bZIP on HDAC activity is more important than its effect on some viral promoters, explaining why KSHVdLZ has a defective phenotype.
Several herpesviral proteins have been discovered to interact with HDAC and to have different functions. CMV IE1 interacts with HDAC, represses deacetylase activity, and enhances viral replication (22,56). However, IE2 of CMV interacts with HDAC2 and has repressive effects on several promoters (70). Here, we are the first to report that KSHV K-bZIP interacts with HDAC, playing an important role in enhancing viral DNA replication. Its repressive effects on some important KSHV promoters (such as ORF50 and OriLyt) seem to contradict its importance in KSHV replication. We were curious whether the repressive effects of K-bZIP on promoters require interaction with HDAC because HDAC is generally an inhibitor of gene expression. ChIP assay results (Fig. 5B) suggested that 1) HDAC2 can bind to the promoter of ORF59 that is not associated with K-bZIP, 2) the leucine zipper domain is important for the interaction of K-bZIP with the promoters of ORF50 and OriLyt that might be mediated by HDAC, and 3) the repressive effects of K-bZIP on KSHV promoters might work through a direct interaction with HDAC and promoters. The repressive effects of OriLyt and ORF50 on KSHV promoters did not result in lower KSHV replication (Fig. 7). The repressive effects of K-bZIP on HDAC activity are important to KSHV replication because the knockdown of HDAC2 partially complements KSHVdLZ replication in HEK 293T cells.
In summary, we discovered that K-bZIP interacts with HDAC1/2; this interaction might be crucial for presenting HDAC1/2 in viral DNA replication domains. We found that the leucine zipper domain is essential for the interaction of K-bZIP and HDAC2 and that this interaction is independent of SUMOylation. We also provide evidence that K-bZIP is able to repress HDAC deacetylase activity and interact with and inhibit the lytic gene promoters (ORF50 and OriLyt) of KSHV. Our results suggest that K-bZIP might regulate KSHV gene expression through interacting with HDAC. Most importantly, our observations that the leucine zipper domain of K-bZIP is important for KSHV replication in HEK 293T cells and that K-bZIP can inhibit HDAC activity suggest that this inhibition of HDAC plays an important role in viral replication. Our model for the function of K-bZIP interaction with HDAC has two layers. 1) K-bZIP can bring HDAC molecules to some promoters, thereby having a repressive effect, and 2) it can bind to HDAC and inhibit the effects of the deacetylase, which in turn causes K-bZIP to have a positive effect on KSHV replication.