Latency-Associated Nuclear Antigen of Kaposi's Sarcoma-Associated Herpesvirus Recruits Uracil DNA Glycosylase 2 at the Terminal Repeats and Is Important for Latent Persistence of the Virus^▿

Cells, plasmids, and antibodies.

BC-3 and BCBL-1, KSHV-positive primary effusion lymphoma (PEL) cell lines, and BJAB and DG75, KSHV-negative cell lines, were cultured in RPMI supplemented with 10% fetal bovine serum, 2 mM l-glutamine, and penicillin-streptomycin (5 U/ml and 5 μg/ml, respectively). Human embryonic kidney 293 (HEK293) and HEK293T cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum, 2 mM l-glutamine, and penicillin-streptomycin (5 U/ml and 5 μg/ml, respectively). Primary UNG^−/− and UNG^+/+ mouse embryo fibroblast (MEF) cells (F11.1) were a gift from Hans E. Krokan (Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway) and were established from transformed clones arising spontaneously after repeated passage in culture (2). MEFs were cultured in Dulbecco's modified Eagle's medium high in glucose with glutamine-Ham's F12 with sodium pyruvate (1:1), 10% fetal bovine serum, 1× nonessential amino acid solution, and penicillin-streptomycin. pBSpuroTR containing an entire (801-bp) TR unit was generated by cloning TR at the NotI site of pBS SKII+ (Stratagene, Inc., La Jolla, CA) as described previously (71). Expression vectors for full-length and truncated mutant LANA tagged with a myc epitope at its carboxyl terminus were described previously (70). pDsRed-LANA was described previously (64). UNG2 fused to green fluorescent protein, pUNG2EGFP, and rabbit UNG2-specific antibody (PU101) was a kind gift from Hans E. Krokan (Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway). cDNA encoding the UNG2 open reading frame was PCR amplified with pUNG2EGFP as the template and cloned into pGEX2TK and pCDNA3.1HA to make a glutathione S-transferase (GST) fusion protein and a hemagglutinin (HA)-tagged protein, respectively. Rabbit anti-LANA polyclonal antibody was a kind gift from Bala Chandran (Rosalind Franklin University of Medicine and Science, North Chicago, Illinois). RTA was detected with a mouse monoclonal antibody as described previously (45).

DNA affinity column.

Preparation of the DNA affinity column used was described previously (71). Briefly, an oligonucleotide containing the LBS (17 bp, in italics and underlined) and an additional three nucleotides at the 5′ end and six nucleotides at the 3′ end (italics) (GATCCGCCTCCCGCCCGGGCATGGGGCCGCGG) with overlapping BamHI sites (bold) was synthesized and annealed with its complementary strand. Multimerized double-stranded DNA was ligated to cyanogen bromide-activated Sepharose beads and packed onto a Bio-Rad column. Nuclear extracts (NE) from BC-3 (KSHV-positive) and BJAB (KSHV-negative) cells were prepared as described earlier and bound to the LBS affinity column as described previously (71). The eluted fractions were aliquoted and frozen at −80°C.

A control DNA column was prepared by synthesizing a similar-length DNA sequence with a scrambled (SCR) LBS (5′-GATCCGCCATCGTAGATCTGTACTGTACGCGG-3′). Similar amounts of total NE were bound to the DNA columns with the multimerized LBS and the SCR LBS. Proteins eluted after extensive washing of the column were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subjected to either Western blotting or sequencing of the band after gel excision.

Matrix-assisted laser desorption ionization-time of flight and liquid chromatography quadrupole (LCQ) mass spectrometry (MS) analysis.

Distinct protein bands from KSHV-positive BC-3 cell NE were excised from a Coomassie G-250-stained gel (see Fig. 1A). These bands were subjected to matrix-assisted laser desorption ionization-time of flight and LCQ mass spectrometry proteomic analysis at the Proteomics Core Facility at the University of Pennsylvania School of Medicine. Proteins for each band with LCQ scores greater than 20 were reported.

Western blot analysis.

An aliquot of the fraction eluted with 500 mM NaCl-containing elution buffer was resolved by SDS-PAGE, followed by detection with a UNG2-specific antibody (rabbit polyclonal antibody PU101). The signals were detected with Alexa Fluor 680 (Molecular Probes, Carlsbad, CA)- and 800 (Rockland, Gilbertsville, PA)-conjugated antibodies by Odyssey infrared scanning technology (LiCor, Lincoln, NE). myc- and HA-tagged proteins were detected with 9E10 and 12CA5 hybridomas, respectively.

In vitro binding of UNG2 to LANA.

Full-length LANA (amino acids 1 to 1162) and different truncation mutant forms of LANA were in vitro translated by a coupled in vitro transcription-translation system (TNT; Promega, Inc., Madison, WI) according to the manufacturer's instructions in the presence of [³⁵S]methionine-cysteine (Perkin-Elmer, Inc., Boston, MA). pGEX-UNG2 expressing GST-fused UNG2 protein was expressed in Escherichia coli as described previously (71). Approximately 10 μg of GST-UNG2 protein was used for binding with in vitro-translated LANA and its derivatives. In vitro-translated products were precleared with glutathione-Sepharose beads in binding buffer (1× phosphate-buffered saline [PBS], 0.1% NP-40, 0.5 mM dithiothreitol, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF], 2 μg of aprotinin per ml) for 30 min. Precleared LANA was then incubated with either GST or GST-UNG2 fusion protein in binding buffer. Binding was performed overnight at 4°C with constant rotation, followed by collection of the beads through centrifugation. The beads were washed three times with 1 ml of binding buffer and then resuspended in SDS lysis buffer and resolved by SDS-PAGE. The bound fraction was analyzed after the gel was dried and exposed to a PhosphorImager plate (Molecular Dynamics, Inc.).

myc-tagged LANA expressed in HEK293T cells was also tested for binding to GST-UNG2. HEK293T cells expressing myc-tagged LANA were lysed in RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% NP-40, 1 mM EDTA, pH 8.0) with protease inhibitors (1 mM PMSF, 10 μg/ml pepstatin, 10 μg/ml leupeptin, and 10 μg/ml aprotinin) and centrifuged at 15,000 rpm and 4°C to remove cell debris. GST-UNG2 fusion protein was added to the cell lysate and subjected to overnight binding in RIPA buffer. The bound complex was washed five times to remove loosely bound proteins and resolved by SDS-PAGE. LANA was detected with anti-myc antibody by Western blotting.

Immunolocalization of LANA and UNG2.

HEK293 cells were cotransfected with pA3M LANA and pUNG2EGFP. BC-3 cells were only transfected with pUNG2EGFP, and endogenous LANA was detected with rabbit anti-LANA serum. At 24 h posttransfection, cells were spread on glass slides and followed by fixation in acetone-methanol (1:1) for 10 min at −20°C. Slides were air dried and incubated with 20% normal goat serum in 1× PBS to block the nonspecific binding sites. LANA was detected with rabbit anti-LANA polyclonal serum at room temperature in a humidified chamber, followed by washing three times for 5 min in PBS. The LANA signal was detected with Alexa Fluor 594 (Molecular Probes, Carlsbad, CA). UNG2 was visualized by fluorescence from the enhanced green fluorescent protein fusion. Slides were then washed four times with 1× PBS, mounted with Paramount G, and visualized with an Olympus confocal laser scanning microscope.

Detection of UNG activity.

To detect the activity of UNG2 when it is bound to LANA, we used a previously described (11) PCR-based assay (see Fig. 4A). The use of dUTP instead of dTTP for PCR amplification results in an amplicon which contains uracil residues. Incubation of the amplicon with UNG2 results in the excision of uracil residues from dU-containing PCR products. Heat treatment and alkaline pH conditions result in degradation of the abasic polynucleotide of the dU-containing PCR product, which is further blocked from subsequent PCR reamplification. Any DNA fragment can be used, regardless of its origin, for the UNG assay; the only important criterion is that it should be amplifiable by PCR. In this assay, we used the KSHV K1 gene as template DNA and amplified it with primers S-K1 (5′-ATGTTCCTGTATGTTGTCTGCAG-3′) and AS-K1 (5′-TCAGTACCAATCCACTGGTTGC-3′) with either dUTP or dTTP in the mixtures of deoxynucleoside triphosphates, leading to a PCR product of 870 bp, designated the dU DNA template or the dT DNA template, respectively. Subsequent PCR amplification with the dU DNA or dT DNA template treated with UNG2-LANA was then performed for up to 30 amplification cycles under identical conditions. Briefly, 1 μl of either dU or dT template DNA was assayed in a final volume of 50 μl in the presence of 1.5 mM MgCl₂, 200 nM each amplification primer, 200 nM each deoxynucleoside triphosphate, and 1.5 U of Taq polymerase. PCR products were resolved on a 1% agarose gel.

UNG2 was coimmunoprecipitated with LANA from HEK293T cells as the source of UNG activity. LANA by itself was also used to test for any UNG activity after immunoprecipitation.

Coimmunoprecipitation of LANA and UNG2.

myc-tagged LANA and HA-tagged UNG2 were transfected either separately or together into HEK293T cells. Cells were harvested 36 h posttransfection and lysed in RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% NP-40, 1 mM EDTA, pH 8.0) with protease inhibitors (1 mM PMSF, 10 μg/ml pepstatin, 10 μg/ml leupeptin, and 10 μg/ml aprotinin). Lysates were centrifuged to remove cell debris and precleared. Lysates were then incubated with anti-myc antibody (myc ascites) overnight at 4°C with rotation, followed by incubation with protein A+G-Sepharose beads at 4°C for 1 h. The resulting immunoprecipitates were collected by centrifugation at 2,000 × g for 2 min at 4°C, and the pellets were washed four times with 1 ml of ice-cold RIPA butter. Half of the immune complex was resuspended in 30 μl of 2× SDS protein sample buffer (62.5 mM Tris, pH 6.8, 40 mM dithiothreitol, 2% SDS, 0.025% bromophenol blue, 10% glycerol) for Western blotting. UNG2 was detected with anti-HA antibody, and then the membrane was reprobed with anti-myc antibody for LANA detection.

Short-term replication assay.

pBSpuroA3 (three copies of TR) was cotransfected into either UNG^−/− or UNG^+/+ MEFs with a LANA expression vector. At 96 h posttransfection, DNA was extracted by a modified form of Hirt's procedure as described previously (31). Extracted DNA was digested with EcoRI (to linearize it) or EcoRI-DpnI with sufficient enzyme overnight. Digested DNA was resolved on a 0.8% agarose gel and transferred to a nylon membrane. DpnI-resistant copies of TR plasmid were detected by hybridization with a ³²P-labeled TR probe with PhosphorImager plates (Molecular Dynamics, Inc.).

BrdU labeling and immunoprecipitation of replicated DNA.

BrdU (5-bromo-2-deoxyuridine, a thymidine analog) was administered at 72 h posttransfection to MEFs and pulsed overnight to label the replicating DNA. DNA was extracted as described previously and digested with DpnI overnight, followed by immunoprecipitation of BrdU-labeled DNA from 90% of the digested product with an anti-BrdU mouse monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) (29). The remaining 10% of the digested DNA was used as input. One microliter of the above-described immunoprecipitated DNA was used for detection of incorporated BrdU and DpnI-resistant copies of pBSpuroA3 after replication with primers described previously (71).

Purification of KSHV virions.

Virions were purified and virion proteins were fractioned and detected as described previously (45, 71). UNG2 was detected with rabbit anti-UNG2 antibody in proteins from pelleted and gradient-purified virions.

Infection of MEF cells with KSHV and induction for virus production.

UNG^−/− and UNG^+/+ MEF cells grown to 50 to 70% confluence in 100-mm-diameter tissue culture dishes were infected with the same amount of concentrated virus from an identical number of BCBL-1 cells. Virus obtained from approximately 20 million BCBL-1 cells were used to infect a single 100-mm culture dish in the presence of Polybrene (Sigma, St. Louis, MO). Two days after infection, cells were passaged and replated after thorough washing to remove any virion particles adhering to the cell surface. Efficiency of infection of these MEFs by BCBL-1-derived virus was determined by immunofluorescence assay with rabbit anti-LANA antibody.

Equal numbers of UNG^−/− and UNG^+/+ MEFs (infected with BCBL-1 cell-derived virus) were plated in a 100-mm culture dish for induction to produce virus. Cells were induced with 20 ng of 12-O-tetradecanoylphorbol-13-acetate (TPA) per ml and 1.5 mM sodium butyrate (Sigma, St. Louis, MO) for 5 days, followed by supernatant collection and clearing by centrifugation at 2,000 rpm for 15 min to remove cells and cell debris. The supernatant was then filtered through 0.45-μm-pore-size filters, and virions were pelleted at 20,000 rpm for 2 h. Virion DNA was isolated by disrupting the virus at 60°C for 2 h in lysis buffer (10 mM Tris, pH 8.5, 1 mM EDTA, pH 8.0, 1% Sarcosyl, 0.1 mg/ml proteinase K). Lysates were phenol extracted and then CHCl₃-indole-3-acetic acid extracted. The relative number of virions produced was quantified by real-time quantitative PCR (qPCR) with K1 gene amplification as described previously (72).

shRNA treatment and quantitation of KSHV episomal copies.

The UNG2-specific shRNA used was described previously (56), and the sequence is 5′-AUCGGCCAGAAGACGCUCUdTdT-3′ (sense, corresponding to the first exon [nucleotides 713 to 731] of the UNG genomic region). Firefly luciferase gene shRNA was used as a control shRNA (BD Clontech, Mountain View, CA). shRNA for UNG2 (5′-GATCCGATCGGCCAGAAGACGCTCTTTCAAGAGAAGAGCGTCTTCTGGCCGATCTTTTTTGATATCG-3′) was cloned into pSIRENRetroP (BD Clontech) after annealing with its complementary sequence. Fifteen micrograms of either pSIRENRetroP UNG2 shRNA or pSIRENRetroP Luc shRNA was transfected into 10 million BC-3 and BCBL-1 cells. Transfected cells were selected with puromycin (3 μg/ml) for 3 weeks. Puromycin-resistant colonies were assayed for UNG2 protein in a Western blot assay with anti-UNG2-specific antibodies.

Total DNA was extracted from these selected colonies by lysing of cells in 5 mM EDTA-1% sarcosyl, followed by proteinase K digestion (1). Relative numbers of KSHV episomal copies were calculated by real-time qPCR amplification of the K1 gene.

Host cellular protein binds to multimerized LBS affinity column.

Multiple studies have demonstrated a role for LANA in KSHV genome replication (28, 32, 71). However, LANA does not have any enzymatic activity associated with DNA replication, including DNA unwinding or ATPase activity, which is required for DNA replication. LANA is also important for persistence of the KSHV genome and replication of TR-containing plasmids (4, 16). Recombinant KSHV with LANA deleted as well as LANA transcript depletion by shRNA reduced the number of KSHV genomic copies, demonstrating that tethering is important for long-term persistence (26, 74). LANA has been shown to bind to cellular proteins to support tethering and replication of the KSHV genome (7, 16, 40, 63, 71). In order to identify cellular proteins associated with LANA when LANA is bound to its cognate sequence, we performed an LBS affinity binding assay (71). Fractions from BC-3 and BJAB cells (KSHV positive and negative, respectively) eluted at 500 mM NaCl were resolved by 10% SDS-PAGE. Coomassie staining of the gel detected unique bands in eluates from the BC-3 LBS affinity column (Fig. 1A). There were also a number of unique bands in BJAB NE; however, we decided to identify the bands from BC-3 NE as the immediate focus was to identify the LANA-associated proteins. These unique bands from BC-3 NE were designated LAP1 to -11 (LANA-associating proteins 1 to 11).

Identification of LAPs.

Coomassie-stained bands were excised from the gel and submitted for identification by liquid chromatography-tandem MS analysis at the Proteomics Core Facility of the University of Pennsylvania. The Mascot software, which uses a statistical scoring algorithm for identification of protein bands, was used (53). The MS spectra were also manually checked to ensure that each identification was of high confidence. Matching of the theoretical molecular weights of the identified bands also supported the MS results and the level of confidence. The identities of these bands are listed in Table 1. LAP11 was identified as a UNG (accession number NP_550433). UNG2 has been shown to remove uracil residues from replicating DNA through a BER pathway (41).

UNG2 preferentially binds to a DNA affinity column conjugated with the LBS.

In order to confirm the specificity of LANA and UNG2 binding to the LBS affinity column, we incubated BC-3 NE with LBS and SCR LBS DNA columns under similar condition. After thorough washing of the columns, proteins were eluted from both of the columns and resolved by 10% SDS-PAGE. Western blot analysis for detection of UNG2 in eluates from these two columns revealed that UNG2 predominantly associated with the LBS column rather than the SCR LBS column, although a faint signal was seen in the SCR eluate (Fig. 1B). LANA was not detected in the SCR LBS column, confirming the inability of LANA to bind to the SCR sequence (data not shown).

UNG2 forms complexes with LANA in vitro and in human cells expressing LANA.

We confirmed the above-described findings on UNG2 binding to LANA by in vitro binding assays with GST-UNG2 as bait. In vitro-translated LANA incubated with GST-UNG2 fusion protein showed precipitation of LANA (Fig. 1C, lane 3), suggesting that these proteins can interact in vitro. GST alone showed no obvious or visible precipitation of LANA (Fig. 1C, lane 2), and GST-UNG2 did not show binding to the control protein, luciferase (Fig. 1C, lane 6), suggesting that the interaction of GST-UNG2 with LANA is likely to be specific in the cells.

We also used LANA heterologously expressed in HEK293T cells to test its ability to bind GST-UNG2. HEK293T cell lysate containing LANA was precleared with glutathione-Sepharose beads and incubated with either GST-UNG2 or GST (Fig. 1D). Bound LANA was detected by anti-myc antibody in the GST-UNG2 lane (Fig. 1D, lane 3) but not in the GST lane (Fig. 1D, lane 6), corroborating the above-described results showing binding of these two proteins.

KSHV-infected cells express LANA during latent infection; therefore, we wanted to determine whether endogenous LANA expressed from the KSHV genome is also capable of interacting with UNG2. PEL (BCBL-1) cells transfected with either pA3M (myc vector) or pA3MUNG2 (myc-UNG2) were lysed and immunoprecipitated for UNG2 with myc antibody. Detection of LANA in pA3MUNG2-transfected BCBL-1 cells (Fig. 1E, lane 3) was due to coimmunoprecipitation of LANA with UNG2, supporting the association of these two proteins in KSHV-positive cells. The absence of LANA in the myc immunoprecipitation and pA3M transfection lane supports the specificity of the association in this assay (Fig. 1E, lane 6).

UNG2 colocalizes with LANA in transiently transfected and KSHV-infected cells.

HEK293 cells were transfected with pDSRed-LANA and pEGFPUNG2 and plated on glass slides. At 12 h posttransfection, cells were stained with 4′,6′-diamidino-2-phenylindole (DAPI) and analyzed by confocal laser microscopy. Detection of LANA and UNG2 showed a punctate nuclear pattern and that most of the UNG2 signal was colocalized with red fluorescent protein-LANA, again corroborating the above-described finding that these two proteins interact in vivo, as they were in same nuclear compartments (Fig. 2). We further detected colocalization of LANA and UNG2 in the KSHV-infected PEL cell line BC-3 (Fig. 2). We transfected pEGFPUNG2 into BC-3 cells and detected LANA with rabbit anti-LANA antibody, followed by staining with goat anti-rabbit Alexa Fluor 594 (red). The merged image of these two proteins showed distinct colocalization of LANA and UNG2 in BC-3 cells.

UNG2 binds to the C terminus of LANA.

Different domains of LANA have been shown to be involved in various functions, including genome tethering and transcriptional modulation (73). Therefore, we wanted to determine the domain of LANA important for the recruitment of UNG2. Mutant forms of LANA with specific domains deleted were in vitro translated and assayed for binding to GST-UNG2 with 10% of the total translated product as the input (Fig. 3). As indicated above, GST did not show any significant binding to the deletion mutant forms of LANA. Additionally, mutant forms of LANA lacking the C terminus did not show any binding to GST-UNG2 (Fig. 3, LANA amino acids 1 to 435, 1 to 756, and 1 to 950). LANA deletion mutant forms which contained the extreme C terminus (amino acids 946 to 1162) showed significant binding to UNG2, indicating that UNG2 can bind to LANA in close proximity to the DNA binding domain of LANA (amino acids 936 to 1139) (38).

Cellular UNG2 retains its activity when complexed with LANA.

We wanted to determine whether UNG2, when complexed with LANA, retains its enzymatic activity. We used an indirect PCR based assay (11) to detect UNG2 activity as described in Materials and Methods and the schematic in Fig. 4A. If the template DNA containing dU is incubated with purified UNG2 and heated at alkaline pH, UNG2 excises the uracil residues from the template and thus generates nicks throughout its entire length. This prevents reamplification of the template DNA. In contrast to this, DNA containing dT remains intact and can be reamplified. We used purified UNG2 expressed in E. coli (data not shown) and immunoprecipitated UNG2 expressed in HEK293 cells for the UNG assay. Results of the UNG2 assay of dT- and dU-containing K1 gene amplicons are shown in Fig. 4C. Immunoprecipitates from HEK293 cells transfected with the vector-alone control did not affect reamplification of the K1 gene in both the dT- and dU-containing templates. As expected, immunoprecipitates from HA-UNG2-expressing cells blocked the reamplification of dU-containing templates but not dT-containing templates, demonstrating the presence of UNG activity in the complex.

In order to assay the activity of UNG2 in a complex with LANA, we used UNG2 coimmunoprecipitated with LANA by anti-myc antibody from LANA-myc- and HA-UNG2-transfected HEK293 cells as the source of UNG2. We confirmed the coimmunoprecipitation of UNG2 by anti-HA Western blotting (Fig. 4B, lanes 3 and 4). Immunoprecipitates from cells transfected with LANA-myc and HA-UNG2 separately were also analyzed by Western blotting with anti-myc and anti-HA antibodies (Fig. 4B, lanes 1 and 2 and lanes 5 and 6, respectively). For the UNG assay, dT- and dU-containing K1 amplicons were incubated with the complex immunoprecipitated from LANA-myc-, LANA-myc-HA-UNG2-, and HA-UNG2-transfected cells. Templates from these combinations were subjected to reamplification, and the results are shown in Fig. 4D. Immunoprecipitates from LANA-myc-transfected cells did not block reamplification of the target DNA, suggesting that LANA by its own does not have any UNG activity (Fig. 4D, lanes 1 and 2). Incubation of dU residue-containing template DNA with the immunoprecipitate from LANA-myc-HA-UNG2-transfected cells blocked reamplification, suggesting the presence of UNG activity in the immunoprecipitated complex (Fig. 4D, lanes 3 and 4). Immunoprecipitates obtained with anti-myc antibody from HA-UNG2-transfected cells did not block reamplification of K1 template DNA, ruling out the possibility of nonspecific precipitation of UNG2 with anti-myc antibody (Fig. 4D, lanes 5 and 6). This suggested that the UNG2 activity associated with the LANA-UNG2 complex is mostly due to the UNG2 bound to LANA. Importantly, this assay is not a quantitative assay but rather a qualitative assay and detects the presence or absence of DNA glycosylase activity. The reduction in reamplification of dU-containing templates treated with anti-myc immunoprecipitates from LANA-myc and HA-UNG2 could be due to the unstable nature of DNA containing the RNA base uracil. Additionally, endogenous UNG2 immunoprecipitated with LANA-myc may also have contributed to the UNG2 activity detected.

Replication of TR-containing plasmids was similar in UNG^−/− and UNG^+/+ (F11.1) MEFs.

Since UNG2 binds to the carboxyl terminus of LANA, which is important for replication, we wanted to determine whether UNG2 is capable of modulating TR-mediated replication. We performed a DpnI sensitivity assay to determine the replication of TR-containing plasmids in UNG-deficient MEFs. This DpnI assay relies on the fact that DpnI digests DNA only if the adenine (A) of its recognition site (GATC) is methylated. Since eukaryotic cells do not possess dam methylase, plasmid DNA replicated in these cells lacks methylation at adenine and thus becomes resistant to DpnI digestion. We used TR-containing plasmid DNA prepared in E. coli (dam⁺) for the replication assay by transfecting it into MEFs and determining the number of DpnI-resistant copies by Southern blotting and by PCR. pBSpuroA3 (a three-TR-containing plasmid) was cotransfected with a LANA-expressing vector into UNG-deficient (UNG^−/−) and wild-type (UNG^+/+) MEFs. Episomal DNA extracted by a modified form of Hirt's procedure was digested with either EcoRI (to linearize) or EcoRI and DpnI overnight and transferred to a GeneScreen membrane, followed by detection with a ³²P-labeled TR probe. DpnI-resistant bands were quantified, and the relative density is plotted with normalization of the input as 10%. Relative densities of DpnI-resistant bands in UNG^−/− and UNG^+/+ cells did not show much difference, suggesting that UNG2 does not directly affect DNA replication (Fig. 5A). Western blotting shows the absence of UNG2 in UNG^−/− MEFs (Fig. 5B).

We further supported the above-described results by BrdU incorporation in newly replicated DNA. UNG^−/− and UNG^+/+ cells cotransfected with pBSpuroA3 and pA3MLANA were pulsed with BrdU for 12 h, followed by extraction of episomal DNA. Ninety percent of the DpnI-digested Hirt DNA was immunoprecipitated by anti-BrdU antibody. A region of the vector backbone from the input, as well as immunoprecipitated DNA, was amplified, and the relative densities of the amplicons were determined with 10% as the input. Relative densities of the amplicons from UNG^−/− and UNG^+/+ cells did not show much difference, and thus supported the above-described result (Fig. 5C).

UNG2 is packaged into virion particles.

We and others have previously shown that virions carry cellular, as well as viral, proteins, most probably for rapid establishment of latent infection following a short burst of lytic reactivation (39, 45). Therefore, we wanted to determine whether cellular UNG2 is packaged into virion particles. We determined the presence of UNG2 in pelleted and gradient-purified virions in a Western blot assay. A protein preparation used previously for detection of RTA and LANA ORCs was used for detection of UNG2. Specifically, Western blot analysis showed the presence of UNG2 predominantly in the core of KSHV virion particles, although there was some signal associated with the tegument (Fig. 6).

KSHV virions obtained from BCBL-1 cells are capable of infecting UNG^−/− MEF cells.

KSHV virions produced after TPA and sodium butyrate induction were concentrated and purified for infection assays. Equal amounts of virion-containing suspension were added to UNG^−/− and UNG^+/+ MEFs. Infected cells were stained with anti-LANA antibody after one passage to detect infection and establishment of latent infection. LANA staining was indistinguishable in UNG^−/− and UNG^+/+ cells, and representative staining is shown in Fig. 7A. Since UNG2 is a base excision repair enzyme responsible for removal of any misincorporated uracil residues, it may not have reflected a difference in initial infection. Therefore, we passaged MEFs (UNG^−/− and UNG^+/+) infected with KSHV for several rounds in order to accumulate mutations due to misincorporation of uracil and deamination of cytosine in the genome. If the inserted mutation occurs in any regulatory gene, that would most likely affect the production of virions after induction. In fact, our virion quantitation data showed an approximately threefold reduction in the number of virions obtained from UNG^−/− cells versus the number of virions obtained from UNG^+/+ cells (Fig. 7B).

We further wanted to determine the levels of RTA in these MEFs, as RTA is essential in activating the expression of KSHV lytic genes (69). As expected, we found reduced levels of RTA (Fig. 7C) in UNG^−/− MEFs, explaining in part the reduction in the number of virions produced. We next examined the infectivity of the virions produced by and purified from these UNG^−/− and UNG^+/+ MEFs. We used equivalent amounts of virion particles (quantified with virion DNA by qPCR) from UNG^−/− and UNG^+/+ cells for infection of HEK293 cells. Infectivity was detected by immunostaining for LANA in these cells. LANA staining suggested that the infectivity of the virus progeny obtained from the MEFs was highly retarded, with very little or no infectivity of virions from UNG^−/− cells. However, few cells infected with virions from UNG^+/+ cells were detected by LANA staining (Fig. 7D).

PEL cells treated with UNG2 shRNA show a reduced number of episomal copies.

BC-3 and BCBL-1 cells selected with UNG2 shRNA for 3 weeks were assayed for UNG2 expression levels. UNG2 shRNA-treated PEL cells, BC-3 cells, and BCBL-1 cells showed efficient depletion of UNG2, i.e., close to an 85% reduction compared to control luciferase shRNA-treated cells (Fig. 8A). We further determined the number of latently persisting copies of episomal DNA in these shRNA-treated cells by real-time qPCR assay. The relative number of KSHV episomal copies was significantly reduced in UNG2 shRNA-treated PEL cells (Fig. 8B). Upon induction of these cells with TPA and sodium butyrate, the number of virions produced was further reduced, which may be a consequence of increased accumulation of mutations in the KSHV genome (Fig. 8C). This result corroborates the data obtained from UNG2-deficient MEFs. In order to detect the infectivity of virions from UNG2 shRNA-treated PEL cells, we infected HEK293 cells with equivalent numbers of virions as quantified by qPCR and detected LANA expression, which is a hallmark of latent infection. LANA staining suggested that virions from UNG2 shRNA-treated cells were also able to infect HEK293 cells but with significantly reduced efficiency compared to that of control shRNA-treated cells (Fig. 8D). The bar diagrams at the bottom of Fig. 8D represent the number of LANA-positive cells in a total of approximately 100 cells in the optical field. Representations of LANA-staining cells are shown in Fig. 8D.