Proteomic Profiling of the Human Cytomegalovirus UL35 Gene Products Reveals a Role for UL35 in the DNA Repair Response

Cell lines.

U2OS human osteosarcoma and human embryonic kidney 293A cells were maintained in Dulbecco's modified essential medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Human nasopharyngeal carcinoma cells (CNE-2Z) were maintained in α-minimal essential medium (MEM) supplemented with 10% FBS. CNE-2Z cells with the stable silencing of PML (CNE-2Z shPML) were previously described (80) and maintained in α-MEM supplemented with 10% FBS and 0.4 μg/ml puromycin (Bioshop). HeLa-Fucci cells are HeLa cells stably transformed with fluorescently tagged cdt1 (mKO2-hCDT1) and geminin (mAG-hGem) as described previously (75) and were cultured in DMEM supplemented with 10% fetal bovine serum, GlutaMax (Gibco), and penicillin-streptomycin.

Plasmids.

Plasmids encoding UL35 and UL35a (HCMV strain AD169) with a C-terminal sequential peptide affinity (SPA) tag (UL35-S) in the pMZS3F vector (101) or a triple FLAG epitope tag (UL35-F) in pCMV-3FC have been described previously (77). The SPA tag consists of a calmodulin binding peptide and a triple-FLAG epitope tag separated by a tobacco etch virus (TEV) protease cleavage site (101). The control vector (LacZ) encodes SPA-tagged β-galactosidase expressed from pMZS3F. Untagged wild-type UL35 (UL35-wt), UL35a, and UL35N (UL35 amino acids 1 to 447) in pCMV-3FC have been described previously (76). A Myc-tagged construct containing the WD40 motif and C terminus of DCAF1 was a gift from Yue Xiong (62). Full-length (FL) DCAF1 and a C-terminal deletion containing amino acids 1 to 1417 (DCAF1 1-1417), containing N-terminal tandem FLAG and hemagglutinin (HA) epitope tags, was a gift from Filippo Giancotti (51).

Primary antibodies.

Rabbit serum raised against UL35 was a gift from Bonita Biegalke (55). Mouse anti-FLAG (M2 clone) was from Sigma-Aldrich. Rabbit primary antibodies against FLAG, PML, DDB1, USP7, γ-H2Ax (A300-081A), and the DCAF1 N terminus (amino acids 25 to 75; A301-887A) and C terminus (amino acids 1457 to 1507; A301-888A) all were from Bethyl Laboratories. Sheep anti-PML was a gift from David Bazett-Jones (University of Toronto). Rabbit anti-myc was from AbCam. Antibodies against phospho-histone H3 Ser10 (8656-R) and Cdc2 (954) were from Santa Cruz, while the phospho-Cdc2 Tyr15 antibody (9111S) was from Cell Signaling. Rabbit anti-lysine 48-conjugated ubiquitin was from Millipore. 53BP1 antibody was from BD Biosciences (612522).

Immunofluorescence microscopy.

Cells were seeded into 6-well plates on glass coverslips (∼700,000 cells/well) and transfected with expression plasmids using Lipofectamine 2000 (Invitrogen) per the manufacturer's instructions, using a DNA/Lipofectamine 2000 ratio of 2 μg to 2 μl for 293A cells and 2 μg to 4 μl for U2OS and CNE2Z cells. Transfected cells were fixed at 48 h posttransfection (or at the indicated times) with 3.7% formaldehyde in phosphate-buffered saline (PBS) (20 min), permeabilized with 0.5% Triton X-100 in PBS (10 min), and blocked with 4% bovine serum albumin (BSA) in PBS (20 min) prior to incubation with primary (1 h) and secondary (45 min) antibodies in 4% BSA in PBS. Primary antibodies were detected using either goat anti-mouse or anti-rabbit FAb fragments conjugated with Alexa fluor 488 or Alexa fluor 555 (Invitrogen). In experiments using sheep anti-PML primary antibodies, donkey secondary antibodies raised against sheep, rabbit, and mouse (conjugated with Cy2, Cy3, and Cy5, respectively) were used to prevent cross-reaction between goat secondary antibodies and sheep primary antibody. Coverslips were mounted onto slides using ProLong Gold antifade fluorescent mounting medium (Invitrogen) containing 4′,6′-diamidino-2-phenylindole (DAPI) for the visualization of nuclear DNA. Images were acquired using the 63× oil objective (numeric aperture, 1.4) on a Leica DM IRE2 inverted fluorescence microscope. Images were processed using OpenLAB (ver.4.0.2) and Adobe Photoshop CS5 using only linear adjustments. For the quantification of USP7 nuclear bodies or γ-H2AX or 53BP1 foci, the foci were counted in each of 100 randomly transfected or untransfected cells in 3 to 6 separate experiments as indicated in the figure legends. Statistical analyses (Student's t test) were conducted using Microsoft Excel 2007 software.

Western blotting.

Cells were lysed in cell lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% NP-40) plus protease inhibitor cocktail (P8340; Sigma) and clarified by centrifugation at 13,000 × g for 15 min at 4°C, and the protein concentration was determined using Bradford reagent (Bio-Rad). Equal amounts of cell lysates were subjected to SDS-PAGE, and proteins were transferred to nitrocellulose membranes, blocked with 4% milk, and incubated with the indicated primary antibodies (1 h) and goat anti-mouse or goat anti-rabbit HRP-conjugated secondary antibodies (45 min; Santa Cruz). Antibodies were detected by chemiluminescence using Western Lighting chemiluminescent reagent (PerkinElmer) and exposure to photographic film (Amersham). For the detection of phosphorylated H3 and Cdc2 by Western blotting, cells were lysed in 9 M urea buffered by 10 mM Tris-HCl, pH 6.8, and 50 μg of each lysate was analyzed by SDS-PAGE and Western blotting as described above.

Coimmunoprecipitation.

293A cells in 10-cm plates were cotransfected with the indicated plasmids and, 40 h later, were harvested and lysed in cell lysis buffer as described above for Western blotting. One mg of cleared lysates was used for immunoprecipitation (IP) and diluted in cell lysis buffer to give a final concentration of 4 mg/ml. For anti-FLAG IP, cell lysates were incubated with anti-FLAG resin (20-μl bed volume; Sigma) for 2 h at 4°C while rotating. Beads were washed four times for 10 min in 1 ml of lysis buffer, and protein was eluted with 50 μl of protein sample buffer (5% SDS, 20 mM Tris, pH 8, 10% dithiothreitol [DTT], 20% glycerol). Western blotting then was performed as described above. For DCAF1 immunoprecipitations, cell lysates prepared as described above were precleared with a 20-μl bed volume of protein A/G beads (Santa Cruz) by incubation for 45 min at 4°C. Protein A/G beads were removed by centrifugation, and cell lysates then were incubated with 2 μg DCAF1 antibody overnight at 4°C with rotating. Antibody-DCAF1 complexes then were recovered by incubation with a 20-μl bed volume of protein A/G beads for 45 min at 4°C. Beads were washed four times with 1 ml cell lysis buffer and eluted in 50 μl of protein sample buffer.

DCAF1 silencing experiments.

293A cells (250,000 to 500,000 cells) seeded in 10-cm dishes were transfected with 100 pmol short interfering RNA (siRNA) against DCAF1 (sc-76898; Santa Cruz) or Allstars negative-control siRNA (1027281; Qiagen) using Lipofectamine 2000, followed by two more rounds of siRNA transfection 24 h apart. Twenty-four h after the final round of siRNA transfection, cells were transfected with 6 μg of plasmid expressing FLAG-tagged proteins using Lipofectamine 2000. Seventy-two h later, cells were lysed and Western blotting (using 30 μg of total cell lysate) and immunoprecipitations were performed as described above.

DNA content analysis by flow cytometry.

For DNA content analysis, cells were fixed overnight at −20°C in 70% ethanol, washed in PBS with 0.5% BSA, immunostained with rabbit anti-FLAG primary antibody (Bethyl) and fluorescein isothiocyanate (FITC)-conjugated anti-rabbit secondary antibody (Santa Cruz Biotechnology), treated with 100 μg/ml RNase A for 1 h at 37°C, and stained with 50 μg/ml propidium iodide. All samples were analyzed using a FACSCalibur flow cytometer (BD Biosciences), and data were collected using CellQuest software. Cell cycle analysis was performed using FlowJo software (Treestar Inc.).

Cell cycle analysis in HeLa-Fucci cells.

For microscopy, HeLa-Fucci cells were seeded into 6-well cluster plates containing coverslips and were transfected 24 h later using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions and using a DNA/Lipofectamine 2000 ratio of 2 μg to 4 μl. Cells were fixed (3.7% formaldehyde) and permeabilized (0.1% Triton X-100) 24 h after transfection. Coverslips were blocked for 1 h in PBS containing 4% BSA and then incubated for 1 h each with rabbit anti-FLAG primary antibody (Bethyl) and goat anti-rabbit Alexa Fluor 350 (Molecular Probes) secondary antibody. Coverslips were mounted onto slides with ProLong Gold antifade mounting medium (Invitrogen), and images were acquired using a Leica DM IRE2 inverted fluorescence microscope. mKO2-Cdt1 and mAG-geminin fluorescent protein expression was detected using Tx2 (BP560/40nm) and yellow fluorescent protein (YFP) (BP500/20 nm) filter cubes, respectively. Images were processed with OpenLab software (version 4.0.2). For the quantification of cell cycle phase by fluorescence-activated cell sorting (FACS), HeLa-Fucci cells were seeded into 10-cm tissue culture plates and transfected as described above with a DNA/Lipofectamine 2000 ratio of 8 μg to 16 μl. Forty-eight h posttransfection, cells were collected, fixed, and permeabilized as described above. FLAG expression was detected with rabbit anti-FLAG followed by goat anti-rabbit Alexa fluor 647 (Molecular Probes) secondary antibody. Cells were filtered to remove aggregates before being examined using a FACSCalibur flow cytometer (Becton Dickinson). FLAG expression was detected following excitation at 633 nm and collection at 661 nm (661/16 nm BP filter, FL4 channel). Cdt1 (mKO2-hCDT1) and geminin (mAG-hGem) were excited using a 488-nm laser, and fluorescent signals were collected at 585 (585/42 nm BP filter, FL2 channel) and 530 nm (530/30 nm BP filter, FL1 channel), respectively. The data were analyzed using FloJo software (Tree Star, version 9.0.1).

Mass spectrometry.

For the mass-spectrometric analysis of UL35- and UL35a-interacting proteins, five 150-cm² dishes of subconfluent (75 to 85%) transiently transfected FLAG- or SPA-tagged UL35 or UL35a-expressing cells were scraped into PBS, pooled, washed with PBS, and collected by centrifugation at 1,000 × g for 5 min at 4°C. Cell pellets were stored at −80°C. Cell pellets were weighed and resuspended with lysis buffer (50 mM HEPES-NaOH, pH 8.0, 100 mM KCl, 2 mM EDTA, 0.1% NP-40, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF], 1 mM DTT, and 1:500 protease inhibitor cocktail; Sigma, St. Louis, MO) in a 1:4 (wt/vol) ratio. Resuspended cells were incubated on ice for 10 min, subjected to one additional freeze-thaw cycle, and then centrifuged at 27,000 × g for 20 min at 4°C. The supernatant was transferred to a fresh 15-ml conical tube, and 1:1,000 benzonase nuclease (25 U/ml; Novagen, San Diego, CA) plus 30 μl packed, preequilibrated FLAG-M2 agarose beads (Sigma, St. Louis, MO) were added. The mixture was incubated for 2 h at 4°C with end-over-end rotation. Beads were pelleted by centrifugation at 1,000 × g for 1 min and transferred to a clean Eppendorf tube. The beads then were washed once with 1 ml lysis buffer and two times with 1 ml ammonium bicarbonate rinsing buffer (50 mM ammonium bicarbonate, pH 8.0, 75 mM KCl). Elution was performed by incubation with 150 μl of 125 mM ammonium hydroxide (pH 11.0). The elution step was repeated twice. The eluate was centrifuged at 1,000 × g for 1 min, transferred to a fresh centrifuge tube, and lyophilized. One μg of mass spectrometry-grade tosyl phenylalanyl chloromethyl ketone (TPCK) trypsin (Promega, Madison, WI) dissolved in 70 μl of 50 mM ammonium bicarbonate (pH 8.3) was added to the eluate and incubated at 37°C overnight. The resulting peptides were lyophilized and resuspended in buffer A (0.1% formic acid). Liquid chromatography (LC) analytical columns (75-μm inner diameter) and precolumns (100 μm) were prepared in house from fused silica capillary tubing from InnovaQuartz (Phoenix, AZ) and packed with 100-Å C₁₈-coated silica particles (Magic, Michrom Bioresources, Auburn, CA). Peptides were subjected to LC-electrospray ionization (ESI)-tandem mass spectrometry (MS/MS) using a 120-min reversed-phase LC (RPLC; 95% water-acetonitrile, 0.1% formic acid) buffer gradient running at 250 nl/min on a Proxeon EASY-nLC pump in line with a hybrid LTQ-Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA). A parent ion scan was performed in the Orbitrap using a resolving power of 30,000, and then the six most intense peaks were selected for MS/MS (minimum ion count of 1,000 for activation) using standard collision-induced dissociation (CID) fragmentation. Fragment ions were detected in the LTQ. Dynamic exclusion was activated such that MS/MS findings of the same m/z (within a −0.1 and +2.1 Thompson window; exclusion list size, 500) detected 3 times within 45 s were excluded from analysis for 60 s. For protein identification, Thermo.RAW files were converted to the .mzXML format using Proteowizard (41) and then searched using X!Tandem (15) against the human RefSeq database (version 37). X!Tandem search parameters were the following: complete modifications, none; cysteine modifications, none; potential modifications, +16@M and W, +32@M and W, +42@N terminus, +1@N, and Q. Data were analyzed using the ProHits (54) and SAINT software tools (14).

Identification of UL35 and UL35a host protein interactions by AP-MS.

To identify host proteins that interact with UL35 and UL35a, FLAG-tagged UL35 and UL35a were expressed in 293A cells by transfection, and cell lysates were subjected to affinity purification with an anti-FLAG resin. Eluates were subjected to trypsin digestion, and the resulting peptides were identified using nanospray liquid chromatography electrospray ionization-tandem mass spectrometry (nLC-ESI-MS/MS). For UL35, we conducted five FLAG immunoprecipitations (IPs; two with the SPA tag and three with the triple FLAG tag), and for UL35a four FLAG IPs (two with the SPA tag and two with the triple FLAG tag) were performed. Four control affinity purifications also were conducted using 293A cells transfected with the same plasmid containing the LacZ open reading frame. MS data were subjected to SAINT (for significance analysis of interactome) analysis (14), and putative UL35/UL35a interactors with an average SAINT probability score of greater than 0.80 (AvgP) are listed in Fig. S1 in the supplemental material. All of the hits also were cross-referenced against a large in-house AP-MS database using ProHits (54). Proteins observed in more than 20% of the AP-MS analyses in our database (i.e., associated with a project frequency of >0.2) were removed, as these proteins are likely to interact with the antibody or solid-phase resin material in a nonspecific manner. Proteins that met both the SAINT and the ProHits project frequency criteria are included in Table 1 and represent those polypeptides with a high probability of interacting specifically with UL35 and/or UL35a. Despite displaying a project frequency of greater than 20%, we also included DDB1 in Table 1, as it is known to associate with DCAF1 and DDA1 (68) and did not appear in the control samples. Also, the high number of peptides recovered for DDB1 (spectral counts) in each experiment is considerably above that seen for nonspecific interactions, showing that DDB1 was efficiently recovered with UL35. A complete list of the recovered cellular proteins and SAINT analysis is provided in Table S1 in the supplemental material.

Since UL35 and UL35a share sequence identity, it is not surprising that some cellular proteins associated with both viral proteins. These include the ubiquitin-specific protease USP7, O-linked N-acetylglucosaminyl (O-GlcNAc) transferase (OGT), and the nuclear import factor importin 4 (IPO4). USP7 is an important regulator of p53 stability (49) and also is targeted by other herpesviruses, namely, the herpes simplex virus protein ICP0 (21) and the Epstein-Barr virus (EBV) protein EBNA1 (31). O-GlcNAc is an important posttranslational modification catalyzed by OGT that regulates several important nuclear events, including the regulation of p53 and several other transcription factors that help control expression from the CMV MIEP, including NF-κB, Ying-yang 1 (YY1), and CREB (69, 85).

Four proteins interacted specifically with UL35, of which three (DCAF1, DDB1, and DDA1) are known to form an E3-ubiquitin ligase complex with Cullin 4A (Cul4A) and Roc1 that may regulate cell cycle and/or DNA repair processes (32, 68). DCAF1 (DDB1-Cul4 associated factor 1) also is known as Vpr binding protein (VprBP) and was identified as a target of the human immunodeficiency virus type 1 (HIV-1) Vpr protein (48). DDB1 (DNA damage binding protein 1) connects DCAF1 and DDA1 (DET1 and DDB1 associated 1) to Cul4A (36, 68). The fourth UL35-specific interaction identified was with SART3 (squamous cell carcinoma antigen recognized by T cells 3), an RNA binding protein involved in tumor immunology and mRNA splicing (28, 63).

Three proteins (TNPO1, G3BP2, and DDX18) were recovered with UL35a only. Transportin 1 (TNPO1) is a nuclear import factor, and G3BP2 (RasGAP SH3 binding protein 2) has been implicated in regulating NF-κB activity (72) as well as p53 and mdm2 (43). DDX18 is a DEAD-box RNA helicase that may have a role in regulating cell proliferation (18), but it is largely uncharacterized.

UL35 inhibits formation of USP7 NBs in a PML-independent manner.

We used coimmunoprecipitation to validate some of the protein interactions identified by mass spectrometry. The USP7 interaction was further explored for several reasons. First, both UL35 and UL35a were found to associate with this ubiquitin-specific protease. Second, although other herpesvirus proteins are known to interact with USP7, this is the first example of a betaherpesvirus interacting with this cellular target. Finally, USP7 is known to associate with PML (22, 80), and UL35 has dramatic effects on the organization of PML bodies in transfected cells (76, 77, 82). 293A cells were transfected with plasmids expressing FLAG-tagged β-galactosidase (LacZ), UL35, UL35a, or UL35N (76) (Fig. 1). UL35N consists of the first 447 amino acids of UL35, which do not contain the UL35a sequence (Fig. 1A). We analyzed the levels of USP7 before FLAG IP and did not observe a change in the overall levels of USP7 in the presence of UL35 or UL35a compared to those in LacZ control transfections and actin loading controls (Fig. 1) despite good expression and more than 90% transfection efficiency (data not shown). Transfected proteins were recovered by anti-FLAG immunoprecipitation and analyzed for the recovery of USP7 by Western blotting (Fig. 1B). USP7 copurified with UL35 very strongly, while UL35a was recovered less efficiently despite similar levels of expression and recovery by IP (Fig. 1B). Unlike results with UL35 and UL35a, USP7 was not recovered by UL35N to a greater degree than the β-galactosidase negative control (Fig. 1B), suggesting that USP7 interacts with the C terminus of UL35 (i.e., UL35a) but that additional UL35 sequences also contribute to the affinity or stability of the UL35-USP7 interaction.

USP7 has been shown previously to associate with PML NBs, and we have recently demonstrated that USP7 is a negative regulator of PML (80). We also have recently described the ability of UL35 to alter PML NBs by recruiting PML to UL35 NBs (76). Since USP7 associates with both PML and UL35, we tested if USP7 localization was affected by the expression of UL35. U2OS cells expressing FLAG-tagged UL35 or UL35a were immunostained for FLAG and USP7 (Fig. 2A). In UL35-transfected cells, there was no significant recruitment of USP7 to UL35 NBs (Fig. 2A, arrows). In untransfected control cells, USP7 had a pan-nuclear staining, with some cells containing a small number of NBs (Fig. 2A, arrowheads). The formation of USP7 NBs has been reported previously and in some cases reflects the association of USP7 with PML NBs (22, 80, 86). Similarly to untransfected cells, USP7 also formed NBs in UL35a-transfected cells; however, the presence of USP7 NBs was noticeably reduced in UL35-transfected cells. Thus, although USP7 did not associate with UL35 structures, UL35 did appear to affect the ability of USP7 to form NBs. We quantified the inhibition of USP7 NB formation by UL35 by determining the percentage of transfected cells containing USP7 NB (Fig. 2B). Cells transfected with the LacZ negative control showed amounts of USP7 NBs similar to those of untransfected cells (∼40%), indicating that the transfection process did not alter USP7 NB formation. Compared to untransfected control cells, both FLAG-tagged UL35 (UL35-F) and untagged UL35 (UL35-wt) caused a significant decrease (P < 0.01) in USP7 NBs, resulting in only ∼5% of transfected cells containing USP7 NBs. In contrast, untagged UL35a (P = 0.02) and FLAG-tagged UL35a (P = 0.27) had a much smaller effect on the level of USP7 NBs, while the LacZ control vector had no appreciable effect on USP7 NB formation (P = 0.57).

Since some USP7 NBs are associated with PML NBs and since UL35 alters PML NBs, we investigated whether the apparent loss of USP7 NBs is a consequence of UL35-mediated PML alterations that could prevent USP7 association with PML. To this end, we expressed UL35 in CNE-2Z cells in which all PML isoforms were silenced with shRNA (Fig. 2C) (23, 80). As expected, compared to wild-type PML-expressing CNE-2Z cells, UL35 expression inhibited USP7 NB formation (Fig. 2D) (P < 0.01) and PML association (Fig. 2C). Neighboring untransfected cells still formed USP7 NBs, which associated with PML (Fig. 2C, arrowheads). Interestingly, in untransfected cells in which PML was silenced and PML bodies therefore were absent, USP7 still formed NBs (Fig. 2C, arrowheads), indicating that PML is not required for USP7 NB formation. As with wild-type CNE-2Z cells, the expression of UL35 inhibited USP7 NB formation in the absence of PML (P < 0.01) relative to untransfected control cells, indicating that the effect of UL35 on USP7 NB formation is independent of PML. The quantification of the results in CNE-2Z cells showed that although UL35 did not inhibit USP7 NB formation as efficiently as it did in U2OS cells, the inhibition was very similar in wild-type and PML-silenced cells (P > 0.05) (Fig. 2D). Thus, the data provide evidence that UL35 interacts with USP7 in cells and can affect its subnuclear localization.

UL35 associates with components of the CUL4A-DDB1-DCAF1 E3 ubiquitin ligase complex.

In addition to USP7, UL35 was found to associate with DCAF1, DDB1, and DDA1. These three proteins all are components of a Cullin 4-based E3 ubiquitin ligase complex (Cul4^DCAF1) consisting of the scaffold protein Cullin4A, the E3 ligase Roc1, the substrate recognition protein DCAF1, and the adaptor protein components DDB1 and DDA1 (Fig. 3A) (36, 68). Interestingly, Cul4-DDB1 complexes are targeted by several different viruses that hijack the ubiquitin ligase activity of these complexes to direct specific cellular proteins for degradation that may not normally be targeted by this E3 ligase (94). In addition, CUL4-DDB1-based complexes are associated with DNA damage responses (DDR) and cell cycle control (68). Given the complex interactions between CMV and host cells, which often involve the manipulation of DDR and the cell cycle, we chose to further explore the interaction of UL35 with this important cellular complex.

First, we confirmed the association between UL35 and DCAF1 and DDB1 by coimmunoprecipitation and Western blotting (Fig. 3B). 293A cells were transfected with LacZ, UL35N, UL35, and UL35a plasmids with FLAG tags. UL35 did not affect the levels of DCAF1 or DDB1 (Fig. 3A, input lanes). Consistently with our mass spectrometry analysis, DCAF1 and DDB1 were recovered with UL35 following FLAG immunoprecipitation but not with UL35a. Similarly, UL35N did not recover detectable levels of DDB1 or DCAF1, indicating that full-length UL35 is required for association with these two cellular proteins (Fig. 3B). In addition, the immunoprecipitation of endogenous DCAF1 using an antibody directed against the N terminus of DCAF1 recovered UL35 but not UL35a (Fig. 3C).

We next examined whether the association of UL35 with DDB1 was dependent on DCAF1, as would be expected if UL35 interacted with the Cul4^DCAF1 complex through DCAF1. To this end, we depleted DCAF1 with siRNA and then expressed FLAG-tagged UL35, UL35a, or β-galactosidase (LacZ) followed by immunoprecipitation with anti-FLAG antibody (Fig. 3D). In cells treated with siRNA directed against DCAF1, the expression of this protein was decreased while there was no change in the level of DDB1. In cells receiving the control siRNA treatment, DCAF1 and DDB1 both were recovered with UL35 but not with UL35a or LacZ (Fig. 3D). However, the recovery of both DCAF1 and DDB1 with UL35 was reduced in DCAF1-depleted cells, in keeping with the degree of decrease in DCAF1 levels. The decreased association of DDB1 with UL35 after DCAF1 downregulation supports the hypothesis that UL35 associates with DDB1 through DCAF1 as part of a complex, such as the Cul4^DCAF1 complex shown in Fig. 3A.

We also used immunofluorescence microscopy to determine if DCAF1 localization was affected by UL35. U2OS cells were transfected with FLAG-tagged UL35 and stained with anti-FLAG and anti-DCAF1 antibodies (Fig. 4A). In untransfected cells, DCAF1 adopts a diffuse nuclear localization pattern; however, when UL35 is expressed, most of the visible DCAF1 becomes associated with UL35 NBs, indicating the ability of UL35 to relocalize DCAF1. Since DCAF1 forms part of an E3 ubiquitin ligase complex, we wanted to determine if the recruitment of DCAF1 to UL35 NBs was associated with an increase in polyubiquitinated substrates. Since the addition of polyubiquitin chains containing the lysine 48 (K48-Ub) linkage targets proteins for proteasomal degradation, we analyzed UL35-transfected cells for the accumulation of K48-Ub by immunofluorescence microscopy (Fig. 4B). In the presence of UL35, a portion of K48-Ub was found to localize to the UL35 NBs in a pattern similar to that observed for DCAF1, while adjacent untransfected cells contained more diffuse K48-Ub staining. These results indicate that UL35 NBs can act as sites for protein ubiquitination and are consistent with UL35 recruitment of Cul4^DCAF1 E3 ligase complexes to UL35 NBs.

UL35 interacts with the C terminus of DCAF1.

The C-terminal third of DCAF1 (amino acids 1000 to 1501) contains the WD40 region (amino acids 1041 to 1377), which is important for association with both DDB1 and Vpr (48). In addition, the cellular protein Merlin negatively regulates the Cul4-DCAF1 complex (51) through interactions with the C-terminal region between amino acids 1417 and 1501 (33). Thus, we wanted to determine if UL35 functions in a similar manner by associating with the C-terminal region of DCAF1. To this end, U2OS cells were cotransfected with a plasmid expressing untagged UL35 (UL35-wt) and a second plasmid expressing either full-length DCAF1 (FL) or DCAF1 lacking the part of the C terminus (1-1417) (Fig. 5A) that contains the Merlin binding site. Both FL and 1-1417 contain N-terminal FLAG and HA (FH) tags (Fig. 5A). In addition, U2OS cells were cotransfected with FLAG-tagged UL35 and a plasmid encoding the DCAF1 C-terminal region that includes the WD40 motif (important for DDB1 and Vpr interactions) and the Merlin binding region (62). Full-length DCAF1 showed colocalization with UL35 NBs (Fig. 5B) similar to that observed with endogenous DCAF1 (Fig. 4A). However, DCAF1 1-1417 did not colocalize with UL35 in NBs (Fig. 5B), suggesting that the extreme C terminus of DCAF1 is needed for association with UL35. Consistently with this result, the WD40 DCAF1 fragment, which contains the entire C terminus, was strongly associated with UL35 NBs (Fig. 5B), indicating that the C terminus of DCAF1 is required for association with UL35.

We also used Western blot analysis to examine the importance of the DCAF1 C terminus for association with UL35. First, 293A cells were cotransfected with plasmids expressing UL35-wt and FLAG-tagged FL or 1-1417 versions of DCAF1 (Fig. 5C). UL35 was recovered only with full-length DCAF1 and not with 1-1417 or the empty vector negative control (Fig. 5C). Second, we used an antibody directed against the C terminus of DCAF1 (amino acids 1457 to 1507), which spans the Merlin binding site, and attempted to immunoprecipitate endogenous DCAF1 in the presence of UL35 or UL35a (Fig. 5D). Contrary to results obtained with antibody directed against the N terminus of DCAF1 (Fig. 3B), when UL35 was present the recovery of DCAF1 by the C-terminal DCAF1 antibody was substantially reduced relative to that of UL35a-expressing cells. These results suggest that the UL35 interaction with the DCAF1 C terminus prevents the DCAF1 antibody from recognizing its epitope, whereas UL35a, which does not interact with DCAF1, does not affect this interaction. Taken together, these results indicate that residues 1418 to 1501 of DCAF1 are important for interaction with UL35.

UL35 causes cells to accumulate in the G₂ phase of the cell cycle.

The interaction of HIV-1 Vpr with DCAF1 has been shown to induce the accumulation of cells in the G₂ phase of the cell cycle (32, 48, 88, 95). Therefore, we assessed whether UL35 caused an increase in G₂ cells as an indicator of altered DCAF1 activity. To this end, 293A cells were transfected with plasmids expressing FLAG-tagged UL35, UL35a, or β-galactosidase (LacZ), and cell cycle profiles were determined by flow-cytometric analyses for DNA content. UL35-transfected cells consistently showed an accumulation of cells with the DNA content of the G₂ or M phase of the cell cycle (39% on average) compared to the level for the LacZ negative control (19% on average), as shown in the representative experiment in Fig. 6A and the composite quantification of the G₂/M peaks in Fig. 6B. In contrast, UL35a expression did not noticeably affect the cell cycle profile (Fig. 6A and B). To determine whether the increased G₂/M peak seen with UL35 was due to increased levels of G₂ or M cells, Western blotting for phosphorylated histone H3 (pH 3), a marker of M phase, was performed on samples from the same experiment. As shown in Fig. 6C, UL35 did not increase pH 3 levels but rather decreased this cell population, indicating that fewer cells had entered mitosis, as expected if progression through G₂ is slowed or arrested.

We analyzed effects of UL35 and UL35a on cell cycle progression in another cell line, HeLa-Fucci, which is engineered to change colors at different stages of the cell cycle due to the expression of two fluorescently tagged proteins, Cdt1 (mKO2-hCDT1) and geminin (mAG-hGem) (75). mKO2-hCDT1 is expressed in G₁, resulting in red fluorescence, while mAG-hGem is expressed in S, G₂, and M, resulting in green fluorescence. Both proteins are expressed at the G₁/S interphase, resulting in yellow cells. These cells were transfected with the LacZ, UL35, or UL35a plasmid, and FLAG-expressing cells were visualized by microscopy and assessed for green and red fluorescence. As shown in Fig. 6D, UL35 expression resulted in an increase in the proportion of cells with green fluorescence relative to that of LacZ- and UL35a-expressing cells, indicating that UL35 increased the size of the S/G₂/M population. These effects were quantified by FACS for cells with red, green, and yellow fluorescence (Fig. 6E). UL35-expressing cells were consistently seen to increase the S/G₂/M population approximately 2-fold over that seen with the LacZ control (P = 0.01), whereas UL35a had a much smaller (and not statistically significant) effect on this population (P = 0.1). Therefore, these results also are consistent with an increased accumulation of G₂ cells caused by UL35 but not UL35a.

We further examined the cell cycle perturbation of UL35 by determining if UL35 activated a G₂ checkpoint response, as can be detected by the phosphorylation of cdc2 (at Tyr15). To this end, we compared levels of total and Tyr15-phosphorylated cdc2 (pcdc2) in UL35-, UL35a-, and LacZ-transfected samples (Fig. 6F). UL35, but not UL35a, resulted in an increase in pcdc2 without affecting total cdc2 levels, which is consistent with the activation of a G₂ checkpoint. This suggests that the activation of the G₂ checkpoint by UL35 results in the observed accumulation of G₂ cells. We then tested the hypothesis that G₂ checkpoint activation involves the interaction of UL35 with DCAF1 by determining if this effect of UL35 was dependent on DCAF1. Therefore, we treated cells with siRNA against DCAF1 (or negative-control siRNA) and then compared the abilities of UL35, UL35a, and LacZ to induce cdc2 phosphorylation (Fig. 6G). As expected, UL35, but not UL35a, induced cdc2 phosphorylation with control siRNA treatment (Fig. 6G, lanes 1 to 3). However, this induction of pcdc2 was abrogated by the downregulation of DCAF1, such that the level of pcdc2 was identical to that seen in the UL35a or LacZ sample (Fig. 6G, lanes 4 to 6). We conclude that the UL35-mediated G₂ checkpoint activation is dependent on DCAF1.

UL35 induces a DNA damage response.

In addition to inducing G₂ arrest, Vpr has been shown to induce repair foci (containing γ-H2AX and 53BP1), which are indicative of DNA damage, in a DCAF1-dependent manner (3). Therefore, we examined UL35-transfected cells for DNA damage markers. First, we transfected cells with FLAG-tagged UL35, UL35a, or UL35N and immunostained them for FLAG and 53BP1 (Fig. 7). By immunofluorescence microscopy, cells expressing UL35 but not UL35a or UL35N showed an increase in 53BP1 foci in many cells (Fig. 7A). Occasionally, some 53BP1 foci appeared to be associated with UL35 NBs (Fig. 7A, arrows). We quantified the ability of UL35 to induce 53BP1 foci by determining the percentage of cells with five or more 53BP1 foci (Fig. 7B) and determining the overall average number of 53BP1 foci in transfected cells (Fig. 7C). Only about 10% of untransfected control cells had five or more 53BP1 foci with an average of about 1.5 foci per cell, which was similar to the level for LacZ-transfected cells. Compared to untransfected control cells (No Tf), the expression of FLAG-tagged (UL35-F) and untagged UL35 (UL35-wt) caused similar significant increases in 53BP1 foci (P < 0.01 for UL35-wt and UL35-F in Fig. 7B and C), resulting in about 45% of transfected cells with five or more foci and an average of more than six foci per cell. In contrast, neither UL35a (tagged or untagged) nor UL35N induced 53BP1 foci.

We also examined γ-H2AX as a marker for DNA damage by comparing cells expressing FLAG-tagged UL35 or UL35a to untransfected (control) cells (Fig. 8A). Control cells and UL35a-transfected cells looked similar, with very few to no γ-H2AX foci per cell, whereas UL35 induced an obvious increase in γ-H2AX foci in 20 to 30% of the transfected cells (Fig. 8A). There was no obvious correlation between the number of UL35 NBs and the number of γ-H2AX foci, and γ-H2AX foci did not obviously localize to UL35 NBs, suggesting that UL35 NB formation is not the cause of these effects. We also quantified these effects by determining the percentage of cells with four or more γ-H2AX foci per cell (Fig. 8B). This population of cells was found to increase from 10.7% in untransfected cells and 18.0% in LacZ-transfected cells to 40.3% in UL35-transfected cells (P < 0.05 for UL35 versus LacZ; P < 0.01 for UL35 versus No Tf). In contrast, only 4.7% of UL35a-transfected cells had four or more γ-H2AX foci (P > 0.05 for UL35a versus LacZ or No Tf), indicating that only full-length UL35 caused the activation of this DNA damage response. Therefore, the results from both γ-H2AX and 53BP1 foci show that UL35, but not UL35a, induces DNA damage.

Given that UL35 interacts with components of the Cul4^DCAF1 complex and this complex is known to function in DNA damage responses, the results described above are consistent with the hypothesis that UL35 interacts with and disrupts the normal function of DCAF1, resulting in increased DNA damage. We further investigated the involvement of DCAF1 in DNA damage induction by UL35 by examining the induction of γ-H2AX protein levels by UL35 with and without DCAF1 depletion (Fig. 6G). In keeping with the effect on γ-H2AX foci, UL35 but not UL35a expression increased the levels of γ-H2AX (Fig. 6G, lanes 1 to 3) in control siRNA-treated cells. However, the UL35-mediated induction of γ-H2AX was abrogated in cells with decreased levels of DCAF1 (Fig. 6G, lanes, 4 to 6). Therefore, the results as a whole suggest that UL35 induces DNA damage, leading to G₂ checkpoint activation through interactions with DCAF1.