HIV-1 Vpr Loads Uracil DNA Glycosylase-2 onto DCAF1, a Substrate Recognition Subunit of a Cullin 4A-RING E3 Ubiquitin Ligase for Proteasome-dependent Degradation*

The human immunodeficiency virus type 1 (HIV-1) accessory protein, Vpr, interacts with several host cellular proteins including uracil DNA glycosylase-2 (UNG2) and a cullin-RING E3 ubiquitin ligase assembly (CRL4DCAF1). The ligase is composed of cullin 4A (CUL4A), RING H2 finger protein (RBX1), DNA damage-binding protein 1 (DDB1), and a substrate recognition subunit, DDB1- and CUL4-associated factor 1 (DCAF1). Here we show that recombinant UNG2 specifically interacts with Vpr, but not with Vpx of simian immunodeficiency virus, forming a heterotrimeric complex with DCAF1 and Vpr in vitro as well as in vivo. Using reconstituted CRL4DCAF1 and CRL4DCAF1-Vpr E3 ubiquitin ligases in vitro reveals that UNG2 ubiquitination (ubiquitylation) is facilitated by Vpr. Co-expression of DCAF1 and Vpr causes down-regulation of UNG2 in a proteasome-dependent manner, with Vpr mutants that are defective in UNG2 or DCAF1 binding abrogating this effect. Taken together, our results show that the CRL4DCAF1 E3 ubiquitin ligase can be subverted by Vpr to target UNG2 for degradation.

One potential target of Vpr is uracil-DNA glycosylase-2 (UNG2), which removes uracil lesions from single-stranded and double-stranded DNA in the base excision repair pathway. Initially identified in a yeast two-hybrid screen (38), UNG2 has been implicated as a Vpr-dependent substrate of the CRL4 E3 ubiquitin ligase (39,40).
However, a number of studies investigating the roles of UNG2 in HIV replication resulted in opposing views regarding HIV biology. For example, some studies found that virion-associated UNG2 modulates the innate mutation rate of HIV-1 in primary monocyte-derived macrophages as well as actively dividing cells (8,41) and that UNG2 is required for replication in macrophages (41). Furthermore, control of dUTP misincorporation into viral DNA by host-derived UNG2, in concert with reverse transcriptase, was shown to be essential to HIV replication (42).
Other reports suggested that UNG2 exerts a negative effect on HIV-1 replication by influencing HIV-1 long terminal repeat-directed transcription (43). In addition, UNG2 was also shown to trigger the degradation of nascent viral DNA, previously modified by APOBEC3G (44). As a counteraction of HIV-1 against the antiviral activity of UNG2, Vpr is proposed to down-regulate UNG2. For example, Vpr-mediated proteasomal degradation of UNG2 was shown to be associated with viral replication when APOBEC3G levels are low (39). Further, the level of UNG2 has also been shown to be negatively influenced by transcriptional effects of Vpr (45).
Another group of studies, however, suggested that UNG2 does not contribute at all to the degradation of retroviral DNA, mediated by APOBEC3G (46 -48), and plays no role in HIV-1 biology. In particular, using a natural human UNG Ϫ/Ϫ cell line as producer or target cells, viral infectivity was not impaired (46), and cell lines that stably express the UNG inhibitor, Ugi, also exhibited no effect. In support of the above, APOBEC3Gmediated retroviral degradation was shown to be independent of uracil DNA glycosylase activity in chicken cells devoid of SMUG1, another abundant uracil DNA glycosylase (49).
To gain further insight into the interplay between Vpr, the proteasomal pathway, and the cellular host protein UNG2, we set out to reconstitute the CRL4 E3 ubiquitin ligase in complexes with DCAF1 and Vpr. We show that Vpr facilitates ubiquitination of UNG2 by recruiting the protein to DCAF1, the substrate recognition subunit of CRL4 DCAF1 . Our results suggest that Vpr hijacks CRL4 DCAF1 and targets UNG2 for degradation, using the ubiquitin-proteasome pathway, similar to how Vif causes APOBEC3G degradation by hijacking the CRL5 E3 ubiquitin ligase (50 -54).

EXPERIMENTAL PROCEDURES
Cloning and Plasmid Construction-The cDNAs encoding residues 1-79 (Vpr-⌬C) of HIV-1 Vpr (provided by V. Ayyavoo at the University of Pittsburgh, Pittsburgh, PA) and residues 1-102 (Vpx-⌬C) of SIVmac Vpx (provided by J. Skowronski at Case Western Reserve University, Cleveland, OH) were cloned into pET43 vectors (EMD Chemicals) modified to include a tobacco etch virus protease site at the C terminus of a NusA fusion protein. N-terminally HA-or Myc-tagged full-length Vpr WT (Vpr-FL), HA-tagged Vpr-⌬C WT, W54R, and H71R (mutants were provided by J. Skowronski) were cloned into the pCDNA3.1 vector (Invitrogen). The cDNAs for UNG2, DCAF1, UBA1, and UbcH5b were purchased from Open Biosystems. The cDNA coding for full-length UNG2 and ⌬N-UNG2 (residues of 99 -313) were cloned into the pE-SUMO vector (LifeSensors) and the pCDNA3.1 vector, encoding proteins with a Myc tag at the N terminus. DCAF1 constructs coding for residues, 817-1507 (DCAF1A), 987-1396 (DCAF1B), and 1057-1396 (DCAF1C) were amplified and cloned into the pENT-TOPO vector (Invitrogen). The HAtagged full-length DCAF1 (DCAF1-FL) and FLAG-tagged DCAF1B were cloned into the pCDNA3.1 vector. The C-terminally His 6 -tagged UBA1 and UbcH5b were cloned into the pET21, and the N-terminally His 6 -and FLAG-tagged ubiquitin (provided by C. Prives, Columbia University, New York, NY) was cloned into the pET28 expression vector (EMD Chemicals). The His 6 -tagged ubiquitin was cloned into pCDNA3.1 vector. The N-terminally His 10 -tagged RBX1 (Open Biosystems) was amplified and cloned into the pBlueBac4.5 plasmid. Cloning of CUL4A and DDB1 into the pBlueBac4.5 plasmid was performed as described previously (55,56). The HA-tagged DDB1 was cloned into pCDNA3.1, and the HA-tagged CUL4A and RBX1 were cloned into the pIRES vector (Clontech).
Protein Expression and Purification-Proteins including His 6 -tagged NusA-Vpr-⌬C, NusA-Vpx-⌬C, SUMO-UNG2, UBA1, UbcH5b, and ubiquitin were expressed in Escherichia coli Rosetta 2 (DE3), cultured in Luria-Bertani medium, using 0.4 mM isopropyl-1-thio-␤-D-galactopyranoside for induction at 18°C for 16 h. Soluble forms of His-tagged proteins were first purified using a 5-ml Ni-NTA column (GE Healthcare). Aggregated material was removed by gel filtration column chromatography using a HiLoad Superdex 200 16/60 column (GE Healthcare), equilibrated with a buffer containing 25 mM sodium phosphate, pH 7.5, 150 mM NaCl, 1 mM DTT, 10% glycerol, and 0.02% sodium azide. NusA-Vpr-⌬C and NusA-Vpx-⌬C were further purified over a 5-ml Hi-Trap QP column (GE Healthcare) at pH 7.5 using a 0 -1 M NaCl gradient. UNG2 was purified over a 5-ml Hi-Trap SP column (GE Healthcare) at pH 7.5 using a 0 -1 M NaCl gradient after digestion with ubiquitin-like-protein specific protease (ULP1) (LifeSensors). Baculoviruses expressing C-terminally His 6 -tagged DCAF1A, -B, and -C were prepared using BaculoDirect C-term (Invitrogen), according to the manufacturer's protocol. Recombinant baculovirus encoding N-terminally His 10 -and FLAG-tagged DDB1 was prepared as described previously (57). Baculoviruses expressing N-terminally His 10 -tagged RBX1 and CUL4A were prepared utilizing the Bac-N-Blue transfection kit (Invitrogen) following the manufacturer's protocols. Proteins were expressed in SF21 cells (Invitrogen) after infection with a multiplicity of infection of 2 for 40 h. DCAF1A and DCAF1B were co-expressed with DDB1, whereas DCAF1C was expressed alone in SF21 cells. CUL4A and RBX1 were also co-expressed. Cells were harvested, and proteins were purified over a 5-ml Ni-NTA column followed by a Superdex 200 16/60 gel filtration column, equilibrated with a buffer containing 25 mM sodium phosphate, pH 7.5, 150 mM NaCl, 1 mM DTT, 10% glycerol, and 0.02% sodium azide. Proteins were further purified over a 5-ml Hi-Trap QP column at pH 7.5 using a 0 -1 M NaCl gradient. For preparation of the multiprotein complexes, each component was mixed at a molar ratio of 1:1, and the protein complexes were purified over an 8-ml Mono Q column (GE Healthcare) at pH 7.5 using a 0 -1 M NaCl gradient.
Mammalian Cell Lines and Transient Transfection-Human embryonic kidney cell lines (HEK293 from ATCC) were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. Cells were plated in 6-well or 10-cm plates 24 h prior to transient transfection and grown to about 95% confluency. The cells were transfected with either 4 g or 24 g of the indicated pCDNA3.1 plasmids encoding specific cDNAs using Lipofectamine 2000 (Invitrogen), according to the manufacturer's protocol. The cells were treated with 10 -15 M MG132 after 42 h and at 6 h prior to harvesting when indicated.
Immunoprecipitation and Immunoblotting-Transiently transfected cells were harvested and treated with 240 l of a lysis buffer containing phosphate-buffered saline (PBS), 1% Tween 20, 0.3% Nonidet P-40, and 0.2 mM phenylmethylsulfonyl fluoride. The lysate was incubated with 40 l of anti-FLAG agarose affinity gel for 4 h, and the beads were washed four times with 500 l of the lysis buffer. Bound proteins were eluted with 50 l of FLAG peptides at a concentration of 100 g/ml in the lysis buffer and analyzed by immunoblotting. When purified recombinant proteins were used for immunoprecipitation, bovine serum albumin (Sigma) was added to the reaction mixture at a concentration of 1 mg/ml in the lysis buffer. Protein complexes captured by anti-FLAG agarose affinity gel (Sigma) were directly treated with 50 l of SDS-PAGE sample loading buffer. Proteins were separated by 4 -20% gradient SDS-PAGE and subsequently identified either by Coomassie Brilliant Blue staining or by immunoblotting. For detection of proteins, anti-HA (Covance), anti-Myc (Sigma), anti-FLAG (Sigma), anti-actin (Sigma), and anti-UNG (Abnova) antibodies were used.
In Vivo Ubiquitination Assays-Transiently transfected cells were harvested and lysed as described as above. The lysate was incubated with 50 l of Ni-NTA agarose beads (GE Healthcare) for 12 h, and the beads were washed four times with 500 l of the lysis buffer. The bound proteins were eluted with 50 l of lysis buffer containing 500 mM imidazole, separated by SDS-PAGE, and analyzed by immunoblotting.
Analytical Size Exclusion Column Chromatography-Proteins or protein mixtures (50 l) at a concentration of 12 or 24 M in PBS buffer with 5% glycerol and 0.02% sodium azide were incubated at 4°C for 30 min and injected into a 24-ml analytical Superdex 200 column at a flow rate of 0.8 ml/min. Peak fractions were collected at 0.5-ml intervals and concentrated 10-fold with Amicon concentrators (Millipore). Proteins in each fraction were separated by 4 -20% gradient SDS-PAGE and stained with Coomassie Brilliant Blue for visualization.
Multi-angle Light Scattering-Data were obtained using an analytical Superdex 200 column (1 ϫ 30 cm, GE Healthcare) with in-line multi-angle light scattering (HELEOS, Wyatt Technology), variable wavelength UV (Agilent 1100 Series, Agilent Technology), and refractive index detectors (Optilab rEX, Wyatt Technology). About 100 l of 10 M protein solutions were injected into the pre-equilibrated column in 25 mM sodium phosphate buffer (pH 7.5), 150 mM NaCl, 10% glycerol, and 0.02% (w/v) sodium azide at a flow rate of 0.5 ml/min at room temperature and then eluted with the same buffer. The molecular masses of eluted species were determined using the ASTRA program (Wyatt Technology).

Vpr Mediates the Interaction between UNG2 and DCAF1-
Analytical gel filtration column chromatography was carried out to assess any protein-protein interactions between Vpr, UNG2, and DCAF1. The individual proteins, ⌬N-UNG2 (resi-dues 99 -313), NusA-Vpr-⌬C (residues 1-79), and DCAF1C (residues 1057-1396), were injected into a Superdex 200 column, and their retention volumes were determined (Fig. 1, A-C). Each protein was found to elute at a distinct volume. Protein peaks were collected and analyzed by SDS-PAGE (Fig.  1M, lanes 1-3). Different protein pairs were then analyzed for co-elution. Consistent with the previous in vivo result that showed an interaction between Vpr and the C-terminal WD40 domain of DCAF1 (residues 1041-1377) (24), NusA-Vpr-⌬C co-eluted with DCAF1C at an elution volume of 12.5 ml (Fig.  1D). The DCAF1C elution peak at 16.0 ml was absent in the DCAF1C-NusA-Vpr-⌬C mixture, and the peak at 12.5 ml contained both NusA-Vpr-⌬C and DCAF1C (Fig. 1M, lane 4). ⌬N-UNG2 and DCAF1C did not interact, eluting at the same volumes as when analyzed individually (Fig. 1, E, and M, lanes 5  and 6). Instead, ⌬N-UNG2 was found to co-elute with NusA-Vpr-⌬C (Fig. 1, F, and M, lanes 7 and 8), confirming the previous yeast-two hybrid results (38). To test whether DCAF1 binds UNG2 in the presence of Vpr, all three proteins were mixed and injected into the column. A peak eluting at 12.3 ml (Fig. 1G) was found to contain all three proteins (Fig. 1M, lane 10), indicating that a triple complex of DCAF1C-NusA-Vpr-⌬C-⌬N-UNG2 was formed. The NusA-Vpr-⌬C-⌬N-UNG2 peak eluting at 11 ml (Fig. 1F) was not observed in this experiment, suggesting that DCAF1C formed protein complexes with all of NusA-Vpr-⌬C-⌬N-UNG2. In these two experiments (Fig. 1, F and G), free ⌬N-UNG2 was observed at 17.9 ml because a 2-fold excess molar concentration of ⌬N-UNG2 was used over NusA-Vpr-⌬C and DCAF1C. After tobacco etch virus (TEV) protease cleavage, the triple complex eluted at 14.8 ml, and the cleaved NusA eluted at 13.4 ml (Fig. 1, H, and N, lanes 12 and 13). These data suggest that the NusA protein, fused to the N terminus of Vpr, does not influence the observed DCAF1-Vpr-UNG2 complex formation. Full-length UNG2 exhibited a similar binding behavior toward NusA-Vpr-⌬C and NusA-Vpr-⌬C-DCAF1C when analyzed by analytical gel filtration column (data not shown). Neither NusA nor NusA-Vpx-⌬C interacted with UNG2 ( Fig. 1, I-L, and N, lanes 14 -19). Taken together, the analytical gel filtration data suggest that DCAF1 binds to the Vpr-UNG2 complex, but not UNG2 alone.
Because DCAF1 is a substrate recognition subunit protein, interacting with the DDB1 adaptor protein associated with CUL4A-RBX1, we tested whether Vpr can load UNG2 onto the DDB1-DCAF1 complex. Each individual recombinant protein was purified and analyzed by SDS-PAGE (supplemental Fig.  1A). In immunoprecipitation experiments with anti-FLAG antibodies, UNG2 was pulled down to a similar extent with either DDB1-DCAF1A (residues 817-1507) or DDB1-DCAF1B (residues 987-1396) only in the presence of NusA-Vpr-⌬C, but not NusA (supplemental Fig. 1B, compare lanes 9 and 14 with  lanes 10 and 15). UNG2 was not detected when only DDB1 was used as pulldown bait (supplemental Fig. 1B, lanes 19 and 20). The above data suggest that DCAF1 interacts with DDB1 as well as the Vpr-UNG2 complex.
CRL4 DCAF1 E3 Ligase Mediates Ubiquitination of UNG2 via Vpr-To determine whether the CRL4 DCAF1 E3 ligase targets UNG2 for ubiquitination in a Vpr-dependent manner, we carried out ubiquitination assays in vitro. For these experiments, the CRL4 E3 ligase was first purified by mixing CUL4A-RBX1 with DDB1-DCAF1B and then subjected to ion exchange chromatography (Fig. 2, A and B). Analytical gel filtration of the purified complex showed that the E3 ligase complex was homogeneous and eluted as a single peak at 11.10 ml (data not shown). Further, CUL4A-RBX1, DDB1-DCAF1B, and the mixture of these two protein complexes were analyzed by in-line multi-angle light scattering after elution from the analytical gel filtration column (Fig. 2C). The molecular mass of the CRL4 DCAF1 E3 ubiquitin ligase was estimated to be 300 kDa, which is in good agreement with the theoretical value of 281 kDa. The estimated molecular masses of CUL4A-RBX1 and DDB1-DCAF1B were 120 and 180 kDa, respectively. The theoretical molecular masses of each complex are 101 and 180 kDa, respectively.
Using these protein complexes, we tested whether full-length UNG2 can be ubiquitinated via the CRL4 DCAF1 E3 ligase in a Vpr-dependent manner (Fig. 3A). CUL4A was autoubiquitinated only in the presence of both E1 and E2 (Fig. 3A, lanes  1-3). Neither NusA nor NusA-Vpr-⌬C mediated ubiquitination of full-length UNG2 when only E3 was present in the reaction mixture (Fig. 3A, lanes 4 -6). Ubiquitinated UNG2 was readily observed in the presence of E1, E2, and E3 with NusA- and CUL4A-RBX1 (absorbance at 280 nm in pink inverted triangles, and molecular mass in light pink filled diamonds) were estimated to be 180 and 120 kDa, respectively. The theoretical molecular mass of each complex is 180 and 101 kDa, respectively. 1.5-fold molar excess of DDB1-DCAF1B was incubated with CUL4-RBX1. The mixture was injected into the analytical gel filtration column. The molecular mass of CRL4 DCAF1B (absorbance at 280 nm in black filled diamonds, and molecular mass in gray filled circles) was estimated to be 300 kDa. Abs, absorbance; arb. Unit, arbitrary unit.
Because NusA, fused to the Nterminal part of Vpr, may sterically interfere with poly-Ub chain formation on UNG2 by the E3 ligase, we purified the CRL4 DCAF1-Vpr E3 ligase (Fig. 3B) and tested its activity toward ubiquitination of UNG2. The E3 ligase complexed with Vpr-⌬C was confirmed to be homogeneous by analytical gel filtration column chromatography and light scattering (data not shown). Polyubiquitination of UNG2 was apparent with the E3 ligase complexed with Vpr-⌬C (Fig. 3C, both panels, compare lane 1 with lane 2, and supplemental Fig. 2). Therefore, our results indicate that Vpr targets UNG2 to the CRL4 DCAF1 E3 ligase for polyubiquitination.
Vpr Down-regulates UNG2 in a DCAF1-dependent Manner in Vivo-To confirm our in vitro results in vivo, we tested whether UNG2 is down-regulated in the presence of both Vpr and DCAF1 in HEK293 cells. Previously, the level of UNG2 protein in cells was shown to be reduced in the presence of CUL4A and CUL4B, DDB1, or Vpr (39,40). However, a role for DCAF1 in UNG2 regulation by Vpr had not been demonstrated. Therefore, protein levels in transiently transfected HEK293 cells expressing DCAF1B, Vpr, and UNG2, individually and in combination, were probed with appropriate antibodies (Fig. 4A). For transient expression of proteins, a 4-fold excess of pCDNA3.1-DCAF1B-DNA over Vpr-DNA was used. However, we observed that DCAF1 protein levels in transfected cells were still lower than Vpr levels, using the same epitope tag for detection (data not shown). The level of UNG2 was not affected by co-expression with Vpr alone but decreased greatly by co-expression with Vpr and DCAF1B (Fig. 4A,  compare lanes 6 and 8). When the cells were treated with 15 M MG132, a proteasome inhibitor, the level of UNG2 stayed constant, irrespective of co-expression with Vpr or Vpr and DCAF1B (Fig.  4A, compare lanes 1-4). Taken together, these results confirm that UNG2 is down-regulated by Vpr and CRL4 DCAF1 in a proteasome-dependent manner.   2 g), and FLAG-tagged DCAF1B (residues 987-1396, 2.2 g) in 6-well plates as indicated. The parent plasmid pCDNA3.1 was used to balance the total amount of DNA (4 g) in each transfection when needed. After 42 h of transient transfection, cells were treated with either MG132 (15 M, lanes 1-4) or vehicle only (dimethyl sulfoxide (DMSO), lanes [5][6][7][8]. After 6 h of treatment, cells were harvested, and the cell lysates were analyzed by immunoblotting with anti-Myc, anti-FLAG, and anti-actin antibodies as indicated. The average of relative intensity of Myc-UNG2 from three independent experiments was determined using a densitometer and is indicated in each lane. B, HEK293 cells, plated in 10-cm plates, were transfected with Myc-tagged Vpr-FL (3.6 g), Myc-tagged UNG2 (3.6 g), and FLAG-tagged DCAF1B (16.8 g) and an appropriate amount of the parent pCDNA3.1 plasmid. After 42 h of transfection, cells were treated with 10 M MG132 for 6 h. The cell lysates were immunoprecipitated (IP) with anti-FLAG antibodies. The pulled down mixtures were separated by SDS-PAGE and analyzed by immunoblotting with anti-Myc and anti-FLAG antibodies as indicated. C, HEK293 cells were transfected with HA-tagged Vpr-⌬C (3.6 g), Myc-tagged UNG2 (3.6 g), FLAG-tagged DCAF1B (16.8 g), and the appropriate amount of the parent plasmids in 10-cm plates. N-terminally HA-tagged WT, W54R (defective in UNG2 binding), or H71R (defective in DCAF1 binding) of Vpr-⌬Cs (residues 1-79) was used for transfection. Cells were harvested 48 h after transfection. Cell lysates were immunoprecipitated with anti-FLAG antibodies, and the proteins bound to antibodies were separated by SDS-PAGE and analyzed by immunoblotting with anti-Myc, anti-HA, anti-actin, and anti-FLAG antibodies as indicated. D, HEK293 cells were co-transfected with constructs expressing HA-tagged, full-length DCAF1 (DCAF1-FL), DDB1, CUL4A-RBX1, and Vpr (5,5,4, and 2 g, respectively) as indicated. The cells were also co-transfected with Myc-tagged UNG2 and His 6 -tagged ubiquitin (4 g each). At 40 h after transfection, cells were harvested and lysed. His 6 -Ub-conjugated proteins were pulled down with Ni-NTA agarose beads. The proteins that bound to Ni-NTA beads were eluted and separated by SDS-PAGE and analyzed by immunoblotting with anti-Myc antibody.
We also investigated whether triple complexes of DCAF1-Vpr-UNG2 can be detected in HEK293 cells (Fig. 4B). Cells were treated with MG132 to preserve the level of UNG2, and FLAG-tagged DCAF1B was immunoprecipitated with anti-FLAG antibody. UNG2 was detected only when Vpr was coexpressed with DCAF1B (Fig. 4B, lane 4), confirming our in vitro data that DCAF1 interacts with the Vpr-UNG2 complex, but not with UNG2 alone.
Further evidence supporting Vpr-DCAF1-dependent downregulation of UNG2 was provided using co-expression of Vpr W54R and H71R mutants that are defective in UNG2 and DCAF1 binding, respectively (28,58). As expected, co-expression of these mutants with DCAF1 and UNG2 did not affect the level of UNG2 (Fig. 4C, top panel, compare lane 4 with lanes 7  and 9). Surprisingly, the WT Vpr with a C-terminal deletion (Vpr-⌬C) was more potent in down-regulating UNG2 than the full-length Vpr (compare Fig. 4A, lanes 7 and 8, with Fig. 4C, top  panel, lanes 4 and 5). Cell lysates were subjected to immunoprecipitation by anti-FLAG antibodies and probed with anti-HA, anti-Myc, and anti-FLAG antibodies (Fig. 4C, lower panels). As expected, the WT and W54R Vpr-⌬C were pulled down by DCAF1, whereas the H71R mutant was not. Myc-UNG2 was not detected in any lane following the pulldown because neither mutant is competent for forming a stable heterotrimeric complex comprising DCAF1-Vpr-UNG2. In the case of WT-Vpr, degradation of UNG2 is induced.
We then performed in vivo ubiquitination assays to further demonstrate Vpr-DCAF1-dependent proteasomal degradation of UNG2. Co-expression of Vpr and full-length DCAF1 (DCAF1-FL) resulted in lower cellular levels of Myc-UNG2 (compare Fig. 4D, lanes 1-3). Moreover, co-expression of the CRL4 complex, which consists of DDB1, CUL4A, and RBX1, resulted in extensive down-regulation of UNG2 (Fig. 4D, lane  4). Pulldowns of ubiquitinated proteins with Ni-NTA beads, detected by immunoblotting with anti-Myc antibody, clearly showed the presence of polyubiquitinated Myc-UNG2 proteins when both CRL4 DCAF1 and Vpr were co-expressed. Taken together, both in vivo and in vitro results, reported here, demonstrate that UNG2 loading onto the DCAF1 receptor of CRL4 E3 ubiquitin ligase is mediated by Vpr followed by UNG2 degradation via the ubiquitin-proteasome pathway.

DISCUSSION
Although certain aspects of the interplay between HIV-1 Vpr and CRL4 DCAF1 E3 ubiquitin ligase in cell cycle arrest and viral infectivity have been described recently (23)(24)(25)(26)(27)(28)(29)(30)(31), many gaps are still present in our understanding of Vpr activity in proteasomal-dependent degradation of host cell proteins. For example, previous reports on Vpr-mediated down-regulation of UNG2 (39,40) did not reveal the substrate receptor subunit that is utilized by Vpr to down-regulate UNG2. Here, we demonstrate for the first time that HIV-1 Vpr mediates UNG2 binding to the substrate recognition subunit, DCAF1 (Fig. 1), and that a stable heterotrimeric complex is formed. The heterotrimeric complex then binds to the DDB1 subunit of the CRL4 E3 ubiquitin ligase (supplemental Fig. 1B), facilitating ubiquitination of UNG2 ( Fig. 3C and supplemental Fig. 2). In this regard, it is worth pointing out that the activity of Vpr differs from that of other viral proteins known to subvert cullin 4A-RING E3 ligases (Fig.  5). For example, paramyxovirus SV5-V, hepatitis B virus X, and woodchuck hepatitis virus X proteins bind directly to DDB1, the adaptor protein of CRL4, inducing the degradation of cell host proteins (59 -62).
DCAF1 is the largest substrate receptor subunit (165 kDa) among all members of the DCAF family that associate with CUL4A-DDB1 E3 ubiquitin ligase. The WD40 domain (ϳ35 kDa) is responsible for binding to DDB1 and is located in the C-terminal region (residues 1057-1377) of DCAF1. The protein constructs used in our studies (residues 987-1396 and residues 817-1396, supplemental Fig. 1B) bind DDB1 well. In addition, binding to Vpr and Vpr-UNG2 by the C-terminal domain (residues 987-1396) was shown to mediate ubiquitination of UNG2 by CRL4 E3 ligase in vitro. However, it cannot be ruled out that other unidentified regions of DCAF1 could alter E3 ubiquitin ligase activity.
There still is no consensus on the biological role(s) of UNG2 in HIV-1 replication. As presented in the Introduction, there are diverging views regarding the relevance of the UNG2-Vpr interaction for HIV-1 biology as follows. 1) UNG2 is a positive regulator of HIV-1, modulating viral mutation rate; 2) UNG2 is an antiviral factor, which Vpr down-regulates; and 3) UNG2 does not contribute to degradation of viral DNA. The biochemical and in vivo data presented here clearly support the view that Vpr down-regulates UNG2 in a proteasome-dependent manner. We therefore propose an analogous role for Vpr as has been established for Vif, namely targeting antiviral host cell factors for degradation in a proteasome-dependent pathway. It has been well established that Vif hijacks CRL5 E3 (CUL5-RBX1, Elongin B and C) to induce degradation of APOBEC3G by polyubiquitination (50 -54). Our results presented here demonstrate that Vpr uses the DCAF1 substrate receptor protein for loading UNG2 onto the CRL4 E3 ubiquitin E3 ligase for degradation in an analogous fashion. The stable triple complex of DCAF1-Vpr-UNG2 with Vpr as a bridging factor is essential for proteasomal degradation of UNG2 (Fig. 4C). Further, when DCAF1 was co-expressed, the levels of UNG2 dropped significantly, and this effect was clearly reversed by treatment with the proteasome inhibitor MG132. These data link UNG2 levels to proteasome-regulated protein homeostasis (Fig. 4A), supporting the hypothesis that indeed Vpr is responsible for the downregulation of UNG2 protein levels in vivo (39,40). It has also been reported that Vpr negatively influences the level of endogenous UNG2 via a transcriptional effect involving the UNG2 promoter (45). Our transient expression data do not rule out the possibility that UNG2 is down-regulated at the transcriptional level by Vpr because both Vpr and UNG2 were ectopically expressed. In fact, all previously reported and current data show that Vpr acts at multiple levels, transcriptional and posttranslational, further lending credence to a possible biological role of UNG2 in antiviral replication.
Given the unresolved questions concerning the biological roles of UNG2 in HIV-1 replication, there is no doubt that more studies are needed to dissect the roles of Vpr and UNG2 in the delicate balancing act between viral and cellular survival. The reconstitution of the CRL4 DCAF1-Vpr E3 ubiquitin ligase complex detailed here provides a means to further discover and validate potential cell cycle-related targets of Vpr.