Protein Interactions Targeting the Latency-Associated Nuclear Antigen of Kaposi's Sarcoma-Associated Herpesvirus to Cell Chromosomes

Expression plasmids.

Glutathione S-transferase (GST)-LANA fusions and yeast Gal4DBD and Gal4ACT constructs were previously described (28). pDY15 expresses GFP-LANA (minus the first 2 aa) in pEGFP-C3 (Clontech). The LANA mutant mtLANA (pMW4) was made by digesting pDY15 with XhoI and AscI to delete LANA aa 1 to 15, blunt ending, and religating. Green fluorescent protein (GFP)-LANA-N (pDH389) contains LANA codons 1 to 329 cloned in pEGFP-C1 (Clontech) at BglII. GFP-LANA-C (pMW2) contains LANA codons 931 to 1164 cloned in pEGFP-C1 at BglII. LANA m1 (pMF42), LANA m2 (pMF43), and LANA m3 (pMF73) contain LANA codons 1 to 329 fused at an XbaI site to codons 928 to 1108, 928 to 1043, and 928 to 985, respectively. LANA m4 (pMF40) expresses LANA aa 1 to 329. pMF constructs have an SG5-Flag vector background. Myc-DEK (pAK7) was made by using GST-DEK (19) as a template and ligating the PCR product into pJH363 at BglII. The Myc-DEK fragment was cut from pAK7 with EcoRI and BglII and ligated into pEGFP-C2 (Clontech) at EcoRI and BamHI to obtain GFP-Myc-DEK (pAK63). DEK was also cloned into the BglII site of pSG5 (Stratagene) to make pAK6.

GST affinity assay and immunoprecipitation.

The GST assay was performed as previously described (28). Briefly, GST and GST fusion proteins were made in bacteria and bound to glutathione Sepharose 4B beads (Amersham) at 4°C overnight. The beads were washed and the amount of protein bound to beads was determined by Coomassie blue staining of proteins separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Equal amounts of each GST protein were used in the affinity assays. HeLa cells were transfected with 10 μg each of Flag-MeCP2 with calcium phosphate. Cells were harvested 48 h posttransfection, resuspended in lysis buffer, and sonicated. Supernatant from transfected cells was incubated with GST fusion proteins. The beads were washed six times and proteins were run on a SDS-PAGE gel (9% polyacrylamide) and transferred to polyvinylidene fluoride (PVDF) membrane (Bio-Rad). Interacting proteins were detected by mouse anti-Flag (Sigma) antibody (1:1,500) and visualized by using the enhanced chemiluminescence (ECL) reaction (Amersham Life Sciences). Transcription and translation of DEK were done with pAK6 in the TNT Quick Coupled System (Promega). Equal amounts of GST fusion proteins were incubated with 10 μl of the ³⁵S-labeling reaction mixture in 600 μl of lysis buffer (0.2% NP-40, 150 mM NaCl, 1 mg of bovine serum albumin [BSA] per ml). Bound proteins were separated by SDS-PAGE and detected by autoradiography.

For immunoprecipitations, HeLa or Cos1 cells were transfected in 10-cm-diameter culture dishes with 10 μg of total DNA by using calcium phosphate (HeLa) or FuGENE6 (Roche) (Cos1) and harvested after 2 days. Cells were washed in 1x phosphate-buffered saline (PBS), lysed in 2 ml of lysis buffer (50 mM Tris [pH 7.9], 100 mM NaCl, 0.5 mM EDTA, 2% glycerol, 0.5 mM phenylmethylsulfonyl fluoride [PMSF], 0.2% NP-40, 2 μg of aprotinin per ml), and sonicated for 10 s. Extracts were precleared with Sephadex G-25 beads (Amersham) for 30 min at 4°C and then incubated with Flag antibody (Sigma) or control mouse immunoglobulin G (IgG) antibody (Santa Cruz) (5 μg per ml of extract) for 2 h. For direct precipitations, rat anti-LANA antibody (4 μg per 0.5 ml of extract; ABI) was incubated for 2 h. Protein G beads were added for 2 h. Beads were washed six times with lysis buffer, and samples (coimmunoprecipitation, 18 μl; direct precipitation, 5 μl; and input extract, 15 μl) were run on SDS-PAGE (9% polyacrylamide) gel. The amount of sample used for direct immunoprecipitations was one-fifth of the amount used for the coimmunoprecipitated sample. Western blot analysis was performed with rat anti-LANA or mouse anti-DEK (BD Transduction Laboratories) monoclonal antibody and peroxidase-conjugated antirat (Chemicon) or antimouse (Amersham) IgG secondary antibody.

Immunofluorescence assay.

As previously described (28), NIH 3T3 or HeLa cells were tranfected overnight in two-well slide chambers (LabTek) with FuGENE 6 with 2 μg of total DNA. Cells were washed in PBS, fixed in 3% paraformaldehyde-PBS for 20 min at 25°C, washed in PBS, permeabilized in 0.2% Triton X-100-PBS for 10 min at 25°C, and washed again in PBS. All staining was performed at 37°C for 1 h. Flag- and Myc-tagged proteins were detected with an anti-Flag or anti-Myc mouse antibody (Sigma) and rhodamine-conjugated donkey anti-mouse IgG secondary antibody (Jackson). LANA was detected with an anti-LANA rat antibody and either a fluorescein isothiocyanate (FITC)- or rhodamine-conjugated donkey anti-rat IgG secondary antibody (Jackson). All antibodies were used at a 1:200 dilution in 1% BSA-PBS. Cells were analyzed by confocal microscopy.

Chromosome spreads.

Ten-centimeter-diameter culture dishes of HeLa or NIH 3T3 cells were transfected with calcium phosphate (HeLa) or Mirus TransIT-LT1 (Panvera) (NIH 3T3). Twenty-four hours posttransfection, cells were treated with Colcemid (0.1 μg/ml) (Sigma) for 2 (HeLa) or 6 (NIH 3T3) h for metaphase arrest. BCBL1 cells were treated with Colcemid for 24 h. Cells were washed in 1× PBS and resuspended at 8 × 10⁴ cells per ml in 75 mM KCl for 15 min to swell nuclei. Cells (200 μl) were cytospun for 10 min at 2,000 rpm onto Superfrost/Plus slides (Fisher) at 25°C. Slides were incubated in KCM solution (120 mM KCl, 20 mM NaCl, 10 mM Tris-HCl [pH 7.7], 0.1% Triton X-100) for 10 min. All antibodies were diluted 1:200 in KCM and incubated for 1 h at 37°C. Cells were washed three times for 3 min each in KCM solution after primary and secondary antibody incubations, fixed in 4% paraformaldehyde-PBS for 10 min at 25°C, and washed twice for 1 min each in water. Cells were mounted in Vectashield mounting solution with DAPI (4′,6′-diamidino-2-phenylindole) (Vector Labs) and analyzed by fluorescence or confocal microscopy.

Yeast assays.

The yeast two-hybrid screen was performed as previously described (28).

Localization of LANA on human chromosomes of KSHV-infected cells.

LANA colocalizes with the KSHV genome at discrete spots on chromosomes of KSHV-infected BCBL1 cells (4, 11). To investigate the relationship between the LANA spots and centromere-associated proteins, chromosome spreads were performed on BCBL1 cells (Fig. 1). Centromeres were detected with human serum containing antibodies against centromeric proteins and rhodamine-conjugated donkey anti-human immunoglobulin secondary antibody. LANA was detected with anti-LANA rat monoclonal antibody and FITC-conjugated antirat secondary antibody. Centromeres (red) appeared as discrete doublets on chromosomes. The LANA-staining spots (green) were separate from centromeres.

LANA is targeted to sites of mouse interphase and mitotic heterochromatin by human MeCP2.

It has been observed that LANA associates with heterochromatin (11, 38, 45, 49), but the targeting mechanism has not been elucidated. Mouse genomic DNA has pericentromeric heterochromatin (PCH), which consists of transcriptionally inactive DNA regions that contain high concentrations of methylated CpGs. In studies of LANA-mediated transcriptional repression, we found that LANA interacts with the methyl CpG binding protein MeCP2 (unpublished data). We set out to determine if LANA could be targeted to heterochromatin via MeCP2. When human MeCP2 and human heterochromatin protein 1 alpha (HP1α), another marker of heterochromatin, are overexpressed in mouse cells, they each localize to PCH and appeared as nuclear spots (29, 54). Immunofluorescence assays were performed with mouse NIH 3T3 cells cotransfected with GFP-LANA and Flag-HP1α. GFP-LANA (green) and Flag-HP1α (red) did not colocalize (Fig. 2A). This indicates that LANA does not localize to mouse heterochromatin when expressed alone, and its localization is not affected by the presence of human HP1α. However, in cells cotransfected with human Flag-MeCP2, GFP-LANA (green) and Flag-MeCP2 (red) colocalized to the nuclear PCH spots (Fig. 2B). Thus, LANA is targeted to mouse PCH by human MeCP2. NIH 3T3 cells were then cotransfected with Flag-MeCP2 and GFP-LANA and then blocked with Colcemid for 6 h, and chromosome spreads were generated (Fig. 2C). Immunofluorescence assays detected Flag-MeCP2 (red) concentrated at the discrete spots of PCH on the mouse chromosome ends. GFP-LANA (green) also localized to the same regions when expressed in the presence of human MeCP2. In the absence of human MeCP2, LANA was never seen associated with NIH 3T3 chromosomes (Fig. 2D). Thus, LANA can be targeted to chromosomes by MeCP2.

MeCP2 targets LANA to mouse interphase heterochromatin and mitotic chromosomes via LANA aa 1 to 15.

LANA contains a CBS in its N terminus, which is required for LANA's association with chromosomes (38). We made a LANA mutant (mtLANA), which has had the first 15 aa of LANA deleted, disrupting the CBS while keeping the nuclear localization signal (NLS) intact (Fig. 3A). To test the requirement for the N-terminal CBS for MeCP2 interaction, immunoprecipitation assays were performed with extracts of HeLa cells cotransfected with Flag-MeCP2 and GFP-mtLANA or GFP-wild-type LANA (Fig. 3B). Immunoprecipitated proteins were analyzed by Western blotting, and mtLANA and wild-type LANA were detected with anti-LANA monoclonal antibody. Wild-type LANA (lane 3), but not mtLANA (lane 1), coprecipitated with Flag-MeCP2 in immunoprecipitates generated by using anti-Flag antibodies. Wild-type LANA and mtLANA were present in equal amounts in the transfected-cell extracts, as indicated by direct analysis of extract (lanes 5 and 6) or immunoprecipitation with anti-LANA antibody (lanes 7 and 8).

To determine the effect of the loss of the CBS site on MeCP2-directed chomosome targeting in mouse cells, NIH 3T3 cells were cotransfected with GFP-LANA-N (LANA aa 1 to 329) or GFP-mtLANA and Flag-MeCP2, and immunofluorescence assays were performed. As demonstrated in Fig. 3C, LANA-N (green) was targeted by Flag-MeCP2 (red) to the mouse PCH nuclear spots, although diffuse nuclear staining was also apparent. Full-length LANA may make additional contacts that stabilize the heterochromatin interaction. mtLANA (green) remained nuclear diffuse in the presence of human MeCP2 and was partially excluded from the PCH regions (Fig. 3D). This result indicates that mtLANA was unable to interact with MeCP2 and is not targeted to heterochromatin. Chromosome spreads were made from the same transfection (Fig. 3E). While Flag-MeCP2 (red) again formed discrete spots on mouse chromosome ends, mtLANA was unable to associate with the mouse chromosomes. This result is consistent with mtLANA's inability to be targeted to mouse PCH by MeCP2. In summary, the previously described essential CBS of LANA is also required for MeCP2 interaction and targeting to chromosomes.

Localization of LANA C terminus.

It is established that LANA has an N-terminal CBS (38). However, GFP-LANA C-terminus proteins have been expressed and show a nuclear speckled pattern in interphase nuclei (45, 49). Ballestas et al. also recognized that the LANA C terminus can independently associate with human chromosomes (M. E. Ballestas, T. Komatsu, and K. M. Kaye, 4th Int. Workshop KSHV and Related Agents, 2001). There is a cryptic NLS in the C terminus of LANA that is functional when the truncated LANA C terminus is expressed. We first demonstrated that our LANA C-terminus construction could target human chromosomes (Fig. 4). HeLa cells were transfected with GFP-LANA-C and blocked in metaphase by Colcemid for 2 h. Chromosome spreads were made, and chromosomes were viewed by immunofluorescence. LANA-C formed spots on the chromosomes, showing that LANA-C can be targeted to chromosomes in the absence of viral episomes. This confirms that a second LANA CBS and targeting mechanism exist.

LANA interacts with DEK.

We identified DEK as a LANA-interacting protein in a yeast screen in which Gal4DBD-LANA was cotransformed into yeast with a B-cell cDNA library to seek cellular binding partners for LANA. DEK associates with and paints human chromosomes (24). Interaction between Gal4DBD-LANA and Gal4ACT-DEK is illustrated in Fig. 5A, as measured by induction of β-galactosidase activity in cotransformed yeast. A known LANA-interacting protein, SAP30, was used as a positive control in this assay.

We next examined LANA's ability to interact in vitro with DEK by using a GST affinity assay. Expression of GST fusion proteins was examined by SDS-PAGE and Coomassie staining, and equal amounts of protein were used in each assay (data not shown). In Fig. 5B, in vitro-transcribed and -translated DEK was labeled with [³⁵S]methionine and incubated with the GST proteins. The 43-kDa DEK protein interacted with the GST fusions expressing LANA(940-1164) (lane 1) and LANA(341-1164) (lane 3). There was a weak and possibly indirect interaction with GST-LANA(1-340) (lane 2). No interaction was observed with control GST-EBNA2(1-58) (GST-E2) (lane 4) or with GST protein (lane 5). Extract (2 μl) was loaded in lane 6. These results mapped the DEK interaction to the C terminus of LANA. To confirm the mapping data, an immunoprecipitation assay was performed with extracts of HeLa cells cotransfected with Flag-LANA or Flag-LANA-C and Myc-DEK (Fig. 5C). Immunoprecipitated proteins were analyzed by Western blotting, and LANA proteins were detected with an anti-Flag mouse antibody. Both Flag-LANA (lane 1) and Flag-LANA-C (lane 3) coprecipitated with Myc-DEK in immunoprecipitates generated with an anti-Myc antibody, but not those generated with a control mouse Ig antibody (lanes 2 and 4).

We further defined the DEK-interacting domain of LANA by using the Flag-tagged LANA deletion mutants shown in Fig. 5D (upper). The in vitro-transcribed and -translated LANA mutants were labeled with [³⁵S]methionine and incubated with GST-DEK or control GST protein (Fig. 5D, lower panel). The LANA mutants m1 and m2 interacted with GST-DEK (lanes 1 and 4). No interaction was observed between LANA mutants m3 and m4 and GST-DEK (lanes 7 and 10) or between the LANA mutants and the control GST proteins (lanes 2, 5, 8, and 11). Two microliters of extract was loaded for each mutant (lanes 3, 6, 9, and 12). This assay indicated that LANA aa 986 to 1043 are required for interaction with DEK.

Localization of DEK on mouse chromosomes.

We investigated DEK's localization in mouse cells in relation to the localization of human MeCP2. NIH 3T3 cells were transfected with GFP-DEK and Flag-MeCP2, and an immunofluorescence assay was performed. GFP-DEK (green) was nuclear diffuse, in contrast to the Flag-MeCP2 nuclear spots (red) (Fig. 6A). Chromosome spreads of dually transfected cells revealed that GFP-DEK (green) painted mouse chromosomes, while Flag-MeCP2 (red) localized as before to PHC (Fig. 6B). Thus, DEK associates with mouse chromosomes, but is targeted in a different manner from MeCP2.

LANA is targeted to mouse and human chromosomes by DEK.

We have shown that the C terminus of LANA can interact with chromosomes and that the C terminus of LANA interacts with DEK. To investigate DEK's ability to target LANA to chromosomes, NIH 3T3 cells (Fig. 7A) or HeLa cells (Fig. 7B) were cotransfected with Myc-DEK and GFP-mtLANA lacking the N-terminal CBS and then blocked with Colcemid for 6 h, and chromosome spreads were performed. GFP-mtLANA (green) localized to mouse and human chromosomes in the presence of Myc-DEK (red). mtLANA did not associate with chromosomes when expressed alone (Fig. 7C). Taken together, the results indicate that DEK targeting through LANA aa 986 to 1043 provides a second mechanism by which LANA can bind to chromosomes. A model for LANA chromosomal tethering is presented in Fig. 8.