The HIV1 Protein Vpr Acts to Promote G2 Cell Cycle Arrest by Engaging a DDB1 and Cullin4A-containing Ubiquitin Ligase Complex Using VprBP/DCAF1 as an Adaptor*

The roles of the HIV1 protein Vpr in virus replication and pathogenesis remain unclear. Expression of Vpr in dividing cells causes cell cycle arrest in G2. Vpr also facilitates low titer infection of terminally differentiated macrophages, enhances transcription, promotes apoptosis, and targets cellular uracil N-glycosylase for degradation. Using co-immunoprecipitation and tandem mass spectroscopy, we found that HIV1 Vpr engages a DDB1- and cullin4A-containing ubiquitin-ligase complex through VprBP/DCAF1. HIV2 Vpr has two Vpr-like proteins, Vpr and Vpx, which cause G2 arrest and facilitate macrophage infection, respectively. HIV2 Vpr, but not Vpx, engages the same set of proteins. We further demonstrate that the interaction between Vpr and the ubiquitin-ligase components as well as further assembly of the ubiquitin-ligase are necessary for Vpr-mediated G2 arrest. Our data support a model in which Vpr engages the ubiquitin ligase to deplete a cellular factor that is required for cell cycle progression into mitosis. Vpr, thus, functions like the HIV1 proteins Vif and Vpu to usurp cellular ubiquitin ligases for viral functions.

HIV1 2 Vpr is a 14-kDa, virion-associated protein that is expressed late in virus infection. The specific role of Vpr in HIV biology remains unclear. Expression of Vpr in dividing cells causes arrest in the G 2 phase of the cell cycle by an as yet unidentified mechanism (1,2). The phosphorylation status of proteins that regulate the cell cycle indicates that Vpr triggers a cellular state similar to that resulting from DNA damage. ATR and Chk1 become activated, and this, like Vpr-mediated arrest, can be blocked by caffeine (3)(4)(5).
Cells arrested in G 2 by Vpr provide an improved environment for HIV replication (6). Although the increase in viral replication itself is relatively small, the effect over several rounds of replication is significant (6). Furthermore, Vpr mod-estly enhances virus transcription through interactions with cellular transcription factors (7)(8)(9)(10)(11).
Vpr also promotes HIV infection of terminally differentiated macrophages (12)(13)(14) by aiding transport of viral preintegration complexes into the cell nucleus (12). Although it is a relatively small protein, Vpr has at least two nuclear import signals and one nuclear export signal (15)(16)(17)(18)(19)(20)(21)(22), and it remains associated with the preintegration complex until after it enters the nucleus. The preintegration complex, however, bears a number of other nuclear import signals (23)(24)(25)(26), and recent work has suggested that all of the import signals may be dispensable for HIV infection of non-dividing cells (27).
In vitro, the importance of Vpr for HIV replication is most apparent during the establishment of low-titer infections of terminally differentiated macrophages (12)(13)(14). The significance of Vpr for HIV biology in vivo is reflected by its highly conserved sequence and its expression by virtually all primary virus isolates. Furthermore, HIV2 and several SIV strains have two Vpr-like open reading frames that resulted from apparent duplication of a single parental gene (28). In these viruses one gene product, Vpr, causes G 2 arrest, and the other, Vpx, boosts infection of terminally differentiated macrophages (29). The duplication of Vpr supports the findings that HIV1 Vpr has multiple functions and the supposition that their importance for virus function outweighs the cost of maintaining extra coding sequences. The duplication may also have enabled further refinement of Vpr functions by promoting independent evolution.
Schröfelbauer et al. (30) recently demonstrated that Vpr promotes proteasomal degradation of the cellular enzyme uracil-N-glycosylase (UNG). This function was linked with cullins 1 and 4 (cul1 and cul4) and shown to enhance virus replication in the presence of the cellular cytidine deaminase APOBEC3G. The virus likely benefits from reduced UNG levels in the presence of APOBEC3G because uracil residues formed in the proviruses as a result of cytidine deamination will not be targeted for excision.
The goal of our work is to determine how HIV1 Vpr acts to cause cell cycle arrest by identifying its cellular protein partners. Our work is among the first to identify endogenous cellular partners that co-immunoprecipitate with Vpr rather than relying on yeast-based interaction screens. Our results show that Vpr mediates G 2 arrest by engaging an ubiquitin ligase complex containing DDB1 and cul4A through the previously identified cellular Vpr partner VprBP. Our results strongly suggest that Vpr promotes the degradation of an as yet unidentified factor that is required for cell cycle progression from G 2 to mitosis.
Plasmids and Cell Transfection-The plasmid pcDNA3. 1huVpr expresses Vpr using human preferred (humanized) codons. The Vpr reading frame was cloned into pcDNA3.1(Ϫ) (Invitrogen) between the XbaI and HindIII sites. The plasmids pcDNA3.1FLAG-huVpr and pcDNA3.1HIV2FLAG-huVpr express HIV1 or HIV2 Vpr, respectively, with a single FLAG epitope tag at the amino terminus. The LaminC-enhanced green fluorescent protein expression construct was a gift from Dr. Ronald Goldman. The T7-DDB1 expression vector was a gift from Dr. Pradip Raychaudhuri. The pSPORT6-based VprBP expression vector was purchased from Open Biosystems (Image clone ID 5547856). pSport6 FLAG-VprBP was constructed by excising the SalI to BamHI fragment from pSPORT6 VprBP and replacing it with one containing the FLAG epitope tag preceded by a start codon and followed by six consecutive glycine residues. The plasmid for expression of the non-degrading T187A mutant of p27 Kip was a gift from Dr. Masaki Matsumoto. The plasmid for expression of dominant negative cul1 (DN cul1) was a gift from Dr. Zhen-Qiang Pan. The dominant negative cul4A (DN cul4A) expression vector was constructed by amplifying cul4A coding sequences with oligonucleotides HACul4A 5Ј (5Ј-GCACTCTAGAACCATG-GCCTACCCCTACGACGTGCCCGACTACGCCATGGCG-GACGAGGCCCCG-3Ј) and Cul4AStop 3Ј H3 (5Ј-GTACAA-GCTTCAGTGCTGCAGCAGCGCC-3Ј) and cloning the DNA into the XbaI and HindIII sites of pcDNA3.1(Ϫ). This vector directs expression of the first 339 amino acids of cul4A with an amino-terminal HA epitope tag. The cul4A expression vector used for the amplification was a gift of Dr. Peter Jackson. Plasmid DNA was introduced into cells using conventional calcium phosphate transfection in the presence of chloroquine. A total of 4 g of DNA was added to each 6-well plate well (200 l volume of calcium phosphate mix into 2 ml of media), and 20 g of DNA was added to each 10-cm plate (1-ml volume of calcium phosphate mix into 10 ml of media).
Immunoprecipitation and Tandem Mass Spectroscopy Analysis-3 ϫ 10 6 293T cells were seeded into 10-cm plates and transfected with 20 g of pcDNA3.1, pcDNA3.1HIV1huVpr, pcDNA3.1HIV1FLAG-huVpr, or pcDNA3.1HIV2FLAG-huVpr. Twenty-four hours after transfection the cells were lysed with 1.0 ml of cold ELB buffer (50 mM HEPES, pH 7.3, 400 mM NaCl, 0.2% Nonidet P-40, 5 mM EDTA, 0.5 mM dithiothreitol, and Complete TM protease inhibitor mixture (Roche Applied Science, as per instructions)). The lysates were cleared twice by centrifugation at 10,000 ϫ g for 10 min at 4°C. The supernatants were incubated with 50 l of anti-FLAG M2 agarose resin (Sigma-Aldrich) for 2 h at 4°C with constant rotation. The beads were then washed three times with ELB. Bound proteins were eluted with 50 l of 200 mg/ml FLAG peptide (Sigma-Aldrich) at room temperature for 10 min. After boiling in Laemmli buffer, the proteins were resolved by polyacrylamide gel electrophoresis and subjected to silver-staining or Western blotting. Protein bands were excised, dried, and identified using tandem mass spectroscopy (Stanford Protein and Nucleic Acid Facility).
To determine whether Vpr forms complexes with DDB1 through VprBP, 293T cells were transfected with HIV1 Vpr or HIV1 FLAG-Vpr together with empty control vector, VprBPdirected short hairpin RNA (shRNA), or DDB1-directed shRNA. Cells were lysed with ELB buffer 1 day after transfection and subjected to immunoprecipitation with anti-FLAGagarose beads as described above.
Experiments to determine whether cul4A co-immunoprecipitates with Vpr were performed as follows. 293T HEK cells were transfected with HIV1 FLAG-Vpr, HIV2 FLAG-Vpr, or HIV2 FLAG-Vpx expression vector and lysed 24 h later in buffer consisting of 5 M CHAPS, 50 M NaCl, and 20 M Tris, pH7.5. The lysates were cleared by centrifugation before the addition of beads bearing FLAG-specific antibody. The samples were placed on a rotator for 2 h at 4°C. The beads were then washed three times with the lysis buffer. Western blots of the pre-and post-immunoprecipitation samples were probed for DDB1, cul4A, and for HIV1 or HIV2 FLAG-Vpr or HIV2 FLAG-Vpx.
Immunoblotting-Proteins immunoprecipitated as described above were resolved by polyacrylamide gel electrophoresis and transferred to Immobilon-PSQ membranes (Millipore). The membranes were blocked with 5% nonfat milk in PBS with 0.1% Tween 20. The membranes were washed with PBS, 0.1% Tween20 and then incubated in the same buffer overnight at 4°C in the presence of primary antibody diluted 1:2000. The primary antibodies used were VprBP-specific polyclonal antiserum (a gift from Dr. Ling-Jun Zhao), anti-FLAG M2 (Sigma-Aldrich), anti-HA monoclonal 12CA5 (Roche Applied Science), DDB1-specific polyclonal antiserum (Invitrogen), anti-cul4A (a gift from Dr. Yue Xiong), and anti-rpS3 (a gift from Dr. Andrew Deutsch). The membranes were washed 3 times with PBS, 0.1% Tween 20 and then incubated with secondary, horseradish peroxidase-conjugated antibody diluted 1:20,000 with 5% nonfat milk in PBS, 0.1% Tween20 (ZyMax Grade, Invitrogen). The immobilized proteins were detected with Immobilon Western horseradish peroxidase substrate (Millipore).
Short Hairpin RNA-The DDB1-directed shRNA vector and VprBP-directed shRNA vector were purchased from Open Biosystems (Clone ID V2HS_151134 and V2SH_74153, respectively). These expression vectors were transfected with calcium phosphate as described above.
Cell Cycle Analysis-293T cells were seeded at a density of 3 ϫ 10 5 cells per well in 6-well plates. Cells were transfected with expression vectors for DDB1 or VprBP-or DDB1-directed shRNA or DN cul1 or DN cul4A together with a constant amount of empty, HIV1 Vpr, or HIV2 Vpr expression vector. Each transfection contained 4 g of DNA. Empty expression vector was added to transfections if less than 4 g of expression vector was used. LaminC-GFP (0.17 g/culture) was co-transfected into experimental samples to allow identification of transfected cells by flow cytometry. Cells were harvested 48 h after transfection. After 2 washes with PBS the cells were suspended in 200 l of PBS and 200 l of propidium iodide staining solution (10 mM PIPES, 0.1 M NaCl, 2 mM MgCl 2 , 0.1% Triton X-100, 0.2 mg/ml RNase A, and 0.02 mg/ml propidium iodide adjusted to pH 6.8) for 20 min before flow cytometry. This procedure removes the cytoplasm and allows analysis of nuclear DNA content. Data were collected on a BD FACSCanto flow cytometer and analyzed with FlowJo analysis software.
DNA Damage Experiments-293T cells were seeded at a density of 3 ϫ 10 5 cells per well in 6-well plates. Each transfection contained 4 g of DNA. Cells were transfected with 1 g of empty expression vector and 3 g of expression vectors for DDB1 or VprBP-or DDB1-directed shRNA, DN cul1, or DN cul4A. LaminC-GFP (0.17 g/culture) was co-transfected to allow identification of transfected cells by flow cytometry. 20 h after transfection the cell culture medium was replaced with fresh media containing either Me 2 SO, doxorubicin (0.25 nM) or etoposide (1 M) dissolved in Me 2 SO. The cells were then cultured for an additional 18 h before the cell cycle profile was analyzed as described above.

Cellular Proteins DDB1 and VprBP Co-immunoprecipitate
with Both HIV1 and HIV2 Vpr-The goal of our work was to learn how the HIV1 protein Vpr promotes G 2 cell cycle arrest by identifying its cellular protein partners. We used co-immunoprecipitation to isolate native cellular components that engage amino-terminal FLAG epitope-tagged Vpr. The single FLAG tag at the amino terminus does not interfere with the capacity of Vpr to cause cell cycle arrest. To assure that both our experimental and control immunoprecipitations were from closely matched cells, we expressed untagged Vpr in our control cultures. Furthermore, because a large number of cellular interaction partners have previously been identified for Vpr, we chose to compare the identity of the proteins that co-isolate with HIV1 Vpr with those that co-isolate with HIV2 Vpr as a specificity control. Expression of both Vpr types causes G 2 arrest, whereas the two share only about 50% amino acid identity. We lysed the transfected cells and immunoprecipitated HIV1 FLAG-Vpr or HIV2 FLAG-Vpr using anti-FLAG M2-agarose affinity beads. Lysates from cells expressing untagged HIV1 Vpr served as our immunoprecipitation background controls. FLAG-Vpr and associated proteins were eluted from the beads using FLAG peptide. The eluates were resolved on SDS-PAGE gels, and proteins were visualized by silver-staining. Protein bands present in the HIV1 or HIV2 FLAG-Vpr samples and not in those from the cultures transfected with untagged Vpr were excised and identified using tandem mass spectroscopy.
Four proteins co-immunoprecipitated with HIV1 FLAG-Vpr, and two of these also co-isolated with HIV2 FLAG-Vpr (Fig. 1A). The 32-kDa protein, rpS3, was isolated only with HIV1 FLAG-Vpr and not with HIV2 FLAG-Vpr. It was first identified as a ribosomal component but later also found to excise abasic residues from DNA (31). Similarly, only HIV1 FLAG-Vpr engaged the 125-kDa protein matrin 3. This protein is a component of the nuclear matrix (32) and has been implicated in nuclear retention of hyper-edited RNA (33). Both HIV1 and HIV2 FLAG-Vpr engaged DDB1 (127-kDa) and VprBP (180-kDa). DDB1 was identified as a DNA damagebinding protein (34), but later work showed that it acts as a platform for the assembly of cul4A-and B-containing ubiquitin ligase complexes (35). Interestingly, Paramyxovirus V proteins act as adaptors to DDB1 and thereby cause the degradation of Stat proteins to hinder interferon-␣ signaling via the Jak/Stat (Jak/signal transducers and activators of transcription) pathway (36). VprBP had only been characterized as a binding partner for Vpr, and little was known about its normal cellular function (37). A number of recent publications, however, have shown that VprBP is a protein partner for DDB1, possibly acting as an adaptor between the ubiquitin ligase complex and a specific target protein or proteins.
Both DDB1 and VprBP co-purified with HIV1 and HIV2 Vpr but not with HIV2 Vpx (data not shown). These proteins, thus, met our criteria for putative cellular partners for Vpr that could be involved in Vpr-mediated G 2 cell cycle arrest. We hypothesize that Matrin 3 and rpS3, which engage only HIV1 Vpr but not HIV2 Vpr, are either involved in HIV1-specific Vpr functions, possibly overlapping with those of HIV2 Vpx, or represent spurious interactions.
VprBP Acts to Bridge Vpr with DDB1-We consistently isolated both VprBP and DDB1 with Vpr, leading us to hypothesize that the proteins are in a complex together. This premise was supported by the recent observation that VprBP is a cellular partner for DDB1 in the absence of Vpr (38,39) and the previous finding that Vpr engages VprBP (37). We further hypothesized that Vpr engages DDB1 by using VprBP as an adaptor.
If VprBP functions as an adaptor between Vpr and DDB1, we reasoned that reduction of steady-state VprBP levels should hinder isolation of DDB1 through Vpr. We, thus, depleted VprBP in HIV1 Vpr or HIV1 FLAG-Vpr-expressing cells with shRNA directed against VprBP and immunoprecipitated HIV1 FLAG-Vpr with anti-FLAG antibody-coated beads. HIV1 Vpr expression vector-transfected cells served as our negative con-FIGURE 1. VprBP and DDB1 both co-immunoprecipitate with HIV1 and HIV2 Vpr. Cultures of 293T cells were transfected with expression vectors for HIV1 Vpr, HIV1 FLAG-Vpr, or HIV2 FLAG-Vpr as indicated. 24 h after transfection the cells were lysed in ELB buffer, the lysates were cleared by centrifugation, and beads coated with FLAG-specific monoclonal antibody (mAb) were added. After incubation at 4°C for 2 h, the beads were washed with cell lysis buffer, and the bound proteins were eluted with FLAG peptide. The eluates were resolved on either a 7.5% (A) or a 12% SDS-PAGE gel (B). The proteins were visualized using standard silver-staining procedures. The bands indicated were excised and identified as indicated by tandem mass spectroscopy. mw, molecular weight standards.
trols. The quantity of DDB1 that was co-isolated with FLAG-Vpr fell to 16% that co-isolated from the empty vector control. This corresponded with the reduction of VprBP co-isolation, which fell to 19% that of the empty vector control ( Fig. 2A). To assure that the reduction of VprBP affected only the amount of DDB1 that was isolated through Vpr, we re-probed the Western blots for rpS3, one of the other proteins that co-immuno-precipitates with HIV1 FLAG-Vpr (Fig. 1). As we expected, the amount of rpS3 that was co-isolated with HIV1 FLAG-Vpr remained relatively unchanged ( Fig. 2A).
We further reasoned that if Vpr engages VprBP and not DDB1 directly, decreasing cellular DDB1 levels should not impact the quantity of VprBP co-immunoprecipitated with Vpr. We depleted DDB1 in HIV1 Vpr-or HIV1 FLAG-Vprexpressing cells with shRNA directed against DDB1 and immunoprecipitated HIV1 FLAG-Vpr with anti-FLAG antibody-coated beads. The cells transfected with HIV1 Vpr expression vector served as our negative controls. As expected, the quantity of DDB1 recovered decreased (52% of the control), but the quantity of VprBP did not (Fig. 2B). The relative level of rpS3, our internal co-immunoprecipitation control, similarly did not decline (Fig. 2B). This observation further supports our model that Vpr engages VprBP directly and likely uses it as an adaptor to DDB1. Of note, we consistently observed higher levels of Vpr in both VprBPand DDB1-depleted cells than in non-depleted control cells (Fig. 2, A  and B). This suggests that Vpr may be getting degraded as a component of an active ubiquitin ligase complex. Consistent with this observation is the finding by Schröfelbauer et al. (30) that cells express more Vpr when cultured in the presence of the proteasome inhibitor MG132.
Cul4A is a part of the DDB1-containing ubiquitin ligase complex. We, therefore, examined whether it is also a part of the HIV1 Vpr⅐VprBP⅐DDB1 complex. Cells were transfected with HIV1 FLAG-Vpr, HIV2 FLAG-Vpr, or HIV2 FLAG-Vpx expression vectors and lysed with a CHAPS-containing buffer as used by Schröfelbauer et al. (30). FLAG-specific antibodies were used to isolate HIV1 or 2 FLAG-Vpr or HIV2 FLAG-Vpx from each of the cleared culture lysates. Cul4A was co-isolated with HIV1 and HIV2 FLAG-Vpr but not with HIV2 FLAG-Vpx (Fig. 2C).
Our data together with a recently published structure of DDB1, based on x-ray crystallography and mapping of the Vpr/VprBP interaction, support a model in which Vpr engages a DDB1-con-FIGURE 2. Vpr engages the DDB1⅐cul4A ubiquitin ligase complex using VprBP as an adaptor. 293T HEK cells were transfected with expression constructs as indicated. 24 h after transfection the cells were lysed in ELB buffer, the lysates were cleared by centrifugation, and beads coated with FLAG-specific monoclonal antibody were added. After incubation at 4°C for 2 h, the beads were washed with cell lysis buffer, and the bound proteins were eluted with FLAG peptide. The eluted proteins were resolved on SDS-PAGE gels and blotted, and the specific proteins were identified using either anti-FLAG M2 monoclonal antibody for FLAG-Vpr or VprBP, DDB1, or rpS3-specific polyclonal antisera for the corresponding proteins. Either VprBP (A) or DDB1 (B) was specifically depleted using shRNA. FLAG-Vpr was isolated to determine whether depletion of VprBP or of DDB1 would diminish the quantity of the other that could be co-immunoprecipitated with FLAG-Vpr. Endogenous rpS3, another protein that was found to associate with HIV1 Vpr, was used as an immunoprecipitation (IP) control. C, 293T HEK cells were transfected as indicated with HIV1 or HIV2 FLAG-Vpr or with HIV2 FLAG-Vpx. 24 h after transfection the cells were lysed with CHAPS-containing buffer and cleared by centrifugation. The FLAGtagged proteins were purified with anti-FLAG M2 resin and eluted with FLAG peptide. The proteins were resolved on SDS-PAGE gels and blotted. FLAG-tagged Vpr and Vpx were detected with anti-FLAG M2 monoclonal antibody, and DDB1 and cul4A were detected with specific antisera. D, our model for interaction between Vpr and the DDB1⅐cul4A complex. Please note that elements for which we have only indirect evidence are shown are gray.
taining ubiquitin ligase complex using VprBP as an adaptor molecule (Fig. 2D). The structural data show that DDB1 consists of three interconnected propeller motifs, one of which likely provides an interaction surface for the WD40 domain of VprBP (38).
Vpr-mediated G 2 Cell Cycle Arrest Is DDB1-dependent-Expression of HIV1 or HIV2 Vpr causes dividing cells to arrest in the G 2 phase of the cell cycle. Because both of these Vpr species engage VprBP and DDB1, we hypothesized that both act to halt cell cycle progression through a complex containing these proteins. Furthermore, the SV5 V protein, which apparently functions through DDB1 directly, also causes cells to arrest in G 2 (40 -42).
The observation that SV5 V-mediated arrest can be blocked by overexpression of DDB1 provided the starting point for our functional analysis of the role that Vpr plays in the ubiquitin ligase complex (42). We hypothesized that SV5-mediated arrest observed by Precious et al. (42) is due to ubiquitylation and degradation of a factor necessary for cell cycle progression. Overexpression of DDB1 in the context of the model shown in Fig. 2D could interfere with ubiquitylation by providing a molecular sink separately for SV5 V and for cul4A complexed with ROC and E2. This blocks ubiquitylation by separating ubiquitin ligase components from the target proteins.
We tested whether overexpression of DDB1 interferes with Vpr-mediated cell cycle arrest as it interfered with SV5 V-mediated arrest. 293T cells were transfected with either FLAG-Vpr alone or together with increasing quantities of DDB1 expression vector. Total amounts of DNA were held constant by adding empty expression vector. We measured cellular DNA content to determine the cell cycle profile of the transfected cells. Our results showed that DDB1 could block not only HIV1 but also HIV2 Vpr-mediated G 2 cell cycle arrest in a dose-dependent manner (Fig. 3A, a-f and g-l, respectively). Of note, HIV2 Vpr does not cause arrest as efficiently as HIV1 Vpr. Other studies have, however, shown that HIV2 Vpr has a shorter half-life than HIV1 Vpr and that this could account for the difference in the number of cells arrested in G 2 (43). The decreased half-life may be because HIV2 Vpr is more specialized for interaction with ubiquitin ligases and, thus, more susceptible to ubiquitylation and degradation as a bystander protein. Alternatively, HIV2 Vpr may have evolved to be more labile to optimize the cell cycle arrest function.

DDB1 Depletion Modestly Reverses Vpr-mediated G 2 Arrest-
Our model predicts that depletion of DDB1 should also break the connection between Vpr complexed with any putative ubiquitylation target protein(s), and the ubiquitin ligase components. To test this, we transfected cells with Vpr expression vector together with increasing quantities of DDB1-directed shRNA expression vector and measured the cellular DNA content using flow cytometry. As expected, we observed a reproducible, dose-dependent decrease in Vpr-mediated cell cycle arrest, consistent with our model (Fig. 4A, a-f). The amount of arrest inhibition, however, was modest and may reflect low turnover of DDB1 in the Vpr-containing complexes or possibly an increase in cells with DNA-repair defects caused by interference with other DDB1 functions. Fig. 4Ai shows the cell cycle profile of cells transfected with DDB1-directed shRNA expression vector alone. This culture showed modest accumulation in G 2 over the negative control culture transfected with empty expression vector (Fig. 4Ag).
VprBP Is Necessary for Vpr-mediated G 2 Arrest-The SV5 V protein apparently engages DDB1 directly (40,42). Our protein interaction data, however, show that VprBP provides a physical link between Vpr and DDB1 (Fig. 2). We, therefore, hypothesized that if the VprBP-mediated interaction between Vpr and DDB1 is necessary for Vpr-mediated G 2 arrest, then depleting VprBP should reduce or eliminate engagement of the DDB1⅐cul4A complex by Vpr and thereby block Vpr-mediated G 2 arrest. We transfected 293T cells with either HIV1 FLAG-Vpr expression vector alone or together with increasing quantities of expression vector for VprBP-directed shRNA and measured the cellular DNA content using flow cytometry. We found that depletion of VprBP decreased the number of cells arrested in G 2 in a dose-dependent manner (Fig. 5A, a-g).
Additional Components of the Ubiquitin Ligase Complex Are Necessary for Vpr-mediated G 2 Arrest-We showed above that VprBP connects Vpr with DDB1. Little, however, is known about the cellular function of VprBP, especially in the context of the DDB1⅐cul4A-containing ubiquitin ligase complex. We, therefore, examined whether other parts of the ubiquitin ligase complex are required for Vpr-mediated G 2 arrest to determine whether Vpr is usurping the functional ubiquitin ligase complex or only its parts.
We used DN cul4A to determine whether ROC1 and E2 are required for Vpr-mediated G 2 arrest. Cul4A links DDB1 with the regulator of cullins (ROC1) protein, which in turn engages the ubiquitin-conjugating enzyme E2 (Fig. 2D). E2 is crucial for attaching ubiquitin onto target proteins. Previously characterized DN cul1 engages Skp1, the DDB1 analog in the Skp1⅐Cdc53⅐F-box protein ubiquitin ligase complex, but it lacks the carboxyl-terminal amino acids required for ROC1 binding and for a complex-activating neddylation (44,45). Using DN cul1 and the structural data available for cul4A as guides, we inserted a stop codon after amino acid 339 of 759 to create a structurally similar DN cul4A. This is the first published example of a mammalian DN cul4A, so we tested its function on a substrate that we know is targeted by the DDB1⅐cul4A complex. We found that UNG HA-tagged at the amino terminus is constitutively targeted for degradation in a DDB1⅐cul4A-dependent manner. Expression of DDB1-directed shRNA greatly increased levels of HA-UNG (Fig. 6, lanes 2 and 3). Expression of HA-UNG together with HA-tagged, DN cul4A also increased levels of HA-UNG albeit not as efficiently (Fig. 6, lanes 4 and 5). DN cul1 has no apparent effect on HA-UNG levels (Fig. 6, lane 6) but does, as expected, increase endogenous levels of the cyclin-dependent kinase inhibitor p27 (Fig. 6, lane 6). Interestingly, DDB1-directed shRNA and DN cul4A both caused decreases in p27 levels (Fig.  6, lanes 3-5).
We next tested whether expression of DN cul1 and cul4A blocks Vpr-mediated G 2 arrest. Both did in a dose-dependent manner (Fig. 7A, a-f and g-l). Neither vector expressed alone, however, caused significant redistribution of cells in the cell cycle (Fig. 7A, m, q, and r). This provided indirect evidence that engagement of ROC1 and E2 is required by both the Skp1⅐Cdc53⅐ F-box protein-and DDB1⅐cul4A-containing ubiquitin ligase complexes for cells to accumulate in G 2 in response to Vpr expression.
Expression of DN cul1 causes accumulation of endogenous p27 (Fig. 6, lane 6), and expression of a non-degrading p27 T187A mutant (ND p27) caused marked G 1 cell cycle arrest (Fig. 7Ao). These results leave open the possibility that the apparent block of Vpr-mediated arrest by DN cul1 is due to cul1-mediated G 1 arrest. Alternatively, cells may fail to accumulate in G 2 in response to Vpr expression because either DN cullin or both interfere with execution of the G 2 checkpoint.
Modification of the DDB1⅐cul4A Complex Does Not Interfere with DNA Damage-triggered G 2 Cell Cycle Arrest-The DDB1⅐cul4A complex is involved in degradation of the DNA replication licensing factor CDT1 in response to UV radiation (46 -50), the inhibitor of ribonucleotide reductase assembly Spd1 (51,52), p27, and the G 1 /S phase transition-promoting protein cyclin E (53). We, therefore, tested whether our modifications of the DDB1⅐cul4A complex were giving the appearance of blocking Vpr-mediated arrest by either trapping cells in G 1 before Vpr could exert its effects or by blocking G 2 checkpoint function.
We examined the cell cycle profiles of cultures transfected with empty expression vector or with expression vectors for non-degrading p27 mutant T187A, DDB1, DDB1-directed shRNA, DN cul1 or cul4A, or VprBP-directed shRNA (Fig. 8A, a-g and o-u, respectively). Expression of non-degrading p27 caused almost complete G 1 arrest, as expected, and served as our G 1 arrest control. Expression of the other proteins or shRNAs caused only very minor alterations in the cell cycle profile. Replicate cultures in the same experiments were exposed to one of two DNA-damaging drugs to determine whether the cells retained the capacity to arrest in G 2 in response to DNA damage. We used doxorubicin or etoposide, which both inhibit topoisomerase II function and thereby prevent the DNA double helix from being resealed after DNA replication. Doxorubicin-treated cultures, with the exception of those expressing the non-degrading p27, showed a robust accumulation of cells in G 2 (Fig. 8A, h-n). Of these, DN cul1-transfected cells exhibited the lowest fraction of cells in G 2 followed by those transfected with DDB1 expression vector (Fig. 8Al). Cells expressing DN cul4A and VprBP-or DDB1-directed shRNA exhibited cell cycle profiles much like the cells transfected with empty vector (Fig. 8A, j, m, n, and k versus h). This indicates that although expression of each of these products blocks Vpr-mediated G 2 arrest (Figs. 3-5 and 7), the resulting alterations do not interfere with G 2 arrest triggered by drug-mediated DNA damage.
The cell cycle profiles of etoposide-treated cultures were distinct from those treated with doxorubicin in that these accumulated cells in more sharply defined G 2 peaks (Fig. 8A, h-n  versus v-z and aЈ and bЈ). The pattern of cell cycle arrest was nevertheless similar to that observed in the doxorubicintreated cells. The DN cul1-expressing cultures showed only slightly more cells in G 2 than those left untreated (Fig. 8A, z versus s) indicating that expression of this protein not only blocks Vpr-mediated arrest but also arrest occurring in response to DNA damage. All of the cultures in which we modulated components of the DDB1⅐cul4A complex, with the exception of that expressing DDB1-directed shRNA, showed only slightly less arrest than the empty vector transfected control (Fig. 8A, x, y, aЈ, and bЈ versus v). These observations support our findings that Vpr acts to cause G 2 cell cycle arrest through the DDB1⅐cul4A ubiquitin ligase complex. None of the modifications that we made to the composition of the DDB1⅐cul4A ubiquitin ligase complex, unlike expression of ND p27, caused G 1 arrest to prevent cells from reaching G 2 . Fur-thermore, the G 2 checkpoint remained intact as the cells accumulated in G 2 in response to drug-induced DNA damage. We cannot, however, be sure from our results whether Vpr also acts through the Skp1⅐Cdc53⅐F-box protein complex as dominant negative cul1 reversed both Vpr and DNA damage-induced G 2 arrest.

DISCUSSION
We used co-immunoprecipitation and tandem mass spectroscopy to discover that exogenously expressed HIV1 and HIV2 Vpr physically engage endogenous cellular VprBP and DDB1. We further demonstrated that these two cellular interaction partners are in the same complex, that Vpr requires VprBP as an adaptor to engage DDB1, and that cul4A also co-immunoprecipitates with HIV1 and HIV2 Vpr but not with HIV2 Vpx.
We confirmed the functional relevance of the physical interactions between Vpr and the ubiquitin ligase complex components. Vpr-mediated G 2 arrest was blocked by manipulation of the complex composition by several means, including VprBP depletion, DDB1 overexpression or depletion, and expression  , panels a, h, o, and v) or with 3 g of expression construct for the indicated proteins (A, ND p27 in panels b, i, p, and w; DDB1 in panels c, j, q, and x; DDB1-directed shRNA in panels d, k, r, and y; DN cul1 in panels e, l, s, and z; DN cul4A in panels f, m, t, and aЈ; VprBP-directed shRNA in panels g, n, u, and bЈ) together with 1 g of empty vector. 20 h after transfection the cell media was replaced with media supplemented with either Me 2 SO (A, panels a-g) or an equivalent quantity of Me 2 SO containing doxorubicin (A, panels h-n). In a second experiment that was carried out similarly, cells were treated with media supplemented with either Me 2 SO (A, panels o-u) or an equivalent quantity of Me 2 SO-containing etoposide (A, panels v-z, aЈ and bЈ). 18 h later cell nuclei were isolated and treated with RNase A. The DNA content was measured by flow cytometry after staining with propidium iodide. Panel B shows the (G 2 ϩ M)/G 1 ratios as a bar graph to facilitate comparison. of DN cul4A. Finally, we determined that interference with the composition of the VprBP⅐DDB1⅐cul4A complex did not itself change the distribution of cells in the cell cycle in a manner that could provide an alternative explanation for our functional observations. Specifically none of the changes that blocked Vpr-mediated G 2 arrest did so by retaining cells in G 1 or by interfering with the capacity of the cell to arrest in G 2 .
During the course of our studies, two groups working to identify adaptor proteins that recruit specific ubiquitylation targets to the DDB1⅐cul4A complex determined that VprBP (alternatively designated DCAF1) is a cellular partner for DDB1 (38,39). This provided further confirmation of our observations.
Our work complements and extends a recent publication by Le Rouzic et al. (54), showing an association between exogenously expressed, epitope-tagged Vpr, DDB1, and the VprBP/ DCAF1 WD40 domain. Interestingly, when the authors of this work isolated FLAG-tagged DDB1 from cells transfected with HA-tagged Vpr, they did not co-isolate Vpr in the absence of Myc-tagged VprBP/DCAF1 WD40 domain. Our findings, relying on endogenous cellular proteins, suggest that endogenous levels of VprBP/DCAF1 should have been sufficient to bridge Vpr and DDB1. Importantly, this work also shows that Vpr mutant Q65R fails to cause G 2 arrest or engage VprBP. Le Rouzic et al. (54), thus, reconfirmed the association between Vpr and VprBP/DCAF1 that Zhao and co-workers (37) previously identified, physically linked VprBP/DCAF1 with DDB1 and further showed that VprBP/DCAF is important for Vprmediated arrest. Our work extends their findings by showing that these interactions also take place among endogenous, intact cellular proteins, demonstrating that DDB1 and cul4A are also important for Vpr-mediated G 2 arrest and showing that neither these changes nor depletion of VprBP/DCAF1 causes significant G 1 arrest or interferes with the capacity of cells to arrest in G 2 . Furthermore, we also demonstrate that DN cul4A can block Vpr-mediated arrest, which suggests that Vpr is not just usurping parts of the DDB1⅐cul4A complex but actually requires its full function as an ubiquitin ligase. Our findings differ in that we easily co-immunoprecipitated VprBP/DCAF1 and DDB1 together with HIV1 and HIV2 Vpr and not with HIV2 Vpx. Le Rouzic et al. (54) on the other hand found that Vpx but not Vpr from SIV mac engages VprBP/DCAF1 in a yeast two-hybrid screen. Vpx from SIV mac and HIV2 are closely related, as are the corresponding Vpr proteins. It is possible that small differences between HIV2 and SIV mac account for the different results. Our finding that the VprBP⅐DDB1⅐cul4A complex is important for Vpr-mediated cell cycle arrest, however, corresponds better with the observation that both HIV2 and SIV mac Vpr cause arrest, whereas HIV2 and SIV mac Vpx do not.
Schröfelbauer et al. (55) also recently found that Vpr engages a complex containing DDB1. This study extended their previous work, which showed that expression of Vpr causes proteasomal degradation of UNG (30). Here they showed that UNG degradation, which is not linked to cell cycle arrest (56), is mediated through a DDB1⅐cul4A ubiquitin ligase complex. This work did not find a clear role for VprBP in UNG degradation. Of note, we found that VprBP depletion does not alter levels of UNG in the presence or absence of Vpr (data not shown). Vpr engages UNG directly and may, thus, promote its degradation by linking it with DDB1 or by first engaging another as yet unidentified adaptor protein. Schröfelbauer et al. (55) further showed that Vpr interferes with DNA repair in response to UV irradiation, likely by retaining DDB1 in the cytoplasm. This scenario is very plausible because Vpr, in commandeering DDB1 for other functions, may restrict the pool available for UV damage repair. Cytoplasmic retention of DDB1 was only observed in response to UV irradiation, and it remains to be seen what role it plays in Vpr-mediated arrest.
Schröfelbauer et al. (55) also show that depletion of DDB1, in the absence of Vpr, causes accumulation of cells in G 2 . We also observed a redistribution of cells to G 2 upon depletion of DDB1, albeit not as pronounced (Fig. 4A, g versus i or Fig. 8A, a versus d) or consistent (Fig. 8A, o versus r). This disparity may be attributable to experimental differences such as our use of a shRNA expression vector instead of siRNA or our cell analysis 2 days after transfection rather than 3. The accumulation of cells in G 2 in response to DDB1 depletion may provide an explanation for the inefficient block of Vpr-mediated G 2 arrest that we observed (Fig. 4).
Our experiments suggest a more direct role for Vpr in G 2 arrest than cytoplasmic sequestration of DDB1. We propose that Vpr blocks cells in G 2 by causing degradation of an as yet unidentified protein that is required for progression of cells from G 2 to mitosis (Fig. 2D). Some of our experiments do not distinguish between our model and that proposed by Schröfelbauer et al. (55). Overexpression of DDB1 could either, as we propose, separate elements of the ubiquitin ligase complex and thereby render it inactive, or it could overcome the proposed cytoplasmic sequestration of DDB1 by Vpr by supplying more DDB1 than Vpr can sequester. Depletion of VprBP could similarly separate Vpr and a putative ubiquitylation target from the ubiquitin ligase complex, or it could free DDB1 for DNA repair or other degradation functions. Our observation that DDB1 depletion reverses Vpr-mediated arrest, however, supports only our model. We were able to reverse Vpr-mediated arrest consistently, albeit modestly by depleting DDB1. If Vpr causes arrest by sequestering DDB1, we would expect a drastic exacerbation of Vpr-mediated arrest when less DDB1 is available in the cell. Furthermore, our model predicts that expression of DN cul4A separates Vpr and a putative ubiquitylation target from ROC1 and the E2 ligase and, thus, prevents ubiquitylation and Vpr-mediated arrest. DN cul4A should also block other DDB1 functions, so we would again expect increased G 2 arrest both in the presence and absence of Vpr if G 2 arrest were caused by DDB1 sequestration.
Finally, we did not propose that Vpr acts to block degradation of a protein that hinders the progression from G 2 to M because we would expect more Vpr-mediated G 2 arrest after depletion of VprBP or G 2 arrest upon VprBP depletion in the absence of Vpr. Similarly, overexpression of DDB1, depletion of DDB1, and expression of dominant negative cul4A all interfere with the normal function of the ubiquitin ligase complex but block, rather than exacerbate, Vpr-mediated G 2 arrest. While this manuscript was under review, DeHart et al. (57) published work with findings similar to ours. Furthermore, in this work they demonstrated that both the proteasome inhibitor epoxomicin and overexpression of non-branching ubiquitin (Ub(K48R)) inhibit accumulation of cells in G 2 in response to Vpr expression. This appears to support the concept that Vprmediated G 2 arrest depends upon ubiquitylation and proteasome-mediated degradation rather than on blocking these processes. However both Ub(K48R) and epoxomicin block proteasomal degradation non-specifically, and thus, the possibility remains that the Vpr-mediated G 2 arrest in these experiments was masked by G 1 arrest or inhibition of the G 2 checkpoint.
In summary, our work and that recently published by Le Rouzic et al. (54), Schröfelbauer et al. (55), and DeHart et al. (57) are shedding new light on the functions of HIV1 Vpr. Although this work reveals aspects of the interaction between Vpr and the DDB1⅐cul4A ubiquitin ligase complex, the relative effects of Vpr-mediated UNG degradation, cell cycle arrest, and DDB1 sequestration on HIV biology remain to be determined.