HIV-1 function is mediated by interaction with the damage-specific DNA-binding protein DDB1

Ba¨ rbel Schro¨ felbauer*†, Yoshiyuki Hakata*, and Nathaniel R. Landau*‡

*Infectious Disease Laboratory, The Salk Institute, 10010 North Torrey Pines Road, La Jolla, CA 92037-1099; and †Department of Biotechnology, Institute of Applied Microbiology, University of Natural Resources and Applied Life Sciences, A-1180 Vienna, Austria

Edited by John M. Coffin, Tufts University School of Medicine, Boston, MA, and approved January 11, 2007 (received for review November 21, 2006) The Vpr accessory protein of HIV-1 induces a response similar to A B IP α-HA cells that of DNA damage. In cells expressing Vpr, the DNA damage kDa α-DDB1 sensing kinase, ATR, is activated, resulting in G arrest and apo- 2 188 ptosis. In addition, Vpr causes rapid degradation of the uracil-DNA * α-HA glycosylases UNG2 and SMUG1. Although several cellular proteins 98 DDB1 –+ –+ HA-Vpr HA-Vpr have been reported to bind to Vpr, the mechanism by which Vpr 62 mediates its biological effects is unknown. Using tandem affinity 49 IP α-myc purification and mass spectrometry, we identified a predominant C cells cellular protein that binds to Vpr as the damage-specific DNA- 38 α-DDB1 binding protein 1 (DDB1). In addition to its role in the repair of α-myc damaged DNA, DDB1 is a component of an E3 ubiquitin ligase that 38 –+ –+ degrades numerous cellular substrates. Interestingly, DDB1 is tar- myc-Vpr myc-Vpr geted by specific regulatory proteins of other viruses, including 17 14 Vpr simian virus 5 and . We show that the interaction with IP α-HA cells DDB1 mediates Vpr-induced and UNG2/SMUG1 degra- D 6 α dation and impairs the repair of UV-damaged DNA, which could -DDB1 account for G2 arrest and apoptosis. The interaction with DDB1 may α-HA explain several of the diverse biological functions of Vpr and - + - + NTAP R HA-R R HA-R suggests potential roles for Vpr in HIV-1 replication. AP vpr NT NL4-3 NL4-3

proteasomal degradation ͉ UNG2 ͉ DNA repair ͉ G2 arrest Fig. 1. Identification of DDB1 as Vpr interacting protein. (A) Protein com- plexes containing TAP-tagged Vpr from lysates of transfected 293 cells were purified over streptavidin and calmodulin and a portion was visualized on a pr is a small nuclear accessory protein of HIV-1 and other silver-stained SDS/PAGE. DDB1 and Vpr are indicated with arrows. The three Vlentiviruses that, when expressed in cells, causes G2 cell- bands were excised and subject to MALDI-TOF mass spectrometry. Two of the cycle arrest and apoptosis. Although Vpr is not required for virus bands were identified as DDB1 and Vpr. This identification was reproduced in replication in vitro, it is well conserved in primary virus isolates three independent analyses with separate preparations on different instru- and contributes to pathogenesis (1). Whereas ⌬vpr SIV was ments. The third (indicated by the asterisk) was not identified. (B) HA-Vpr was found to be nearly as pathogenic as wild-type, a vpr/ double expressed in transfected 293T cells and immunoprecipitated with anti-HA mutant was highly attenuated, suggesting an important and mAb. Coimmunoprecipitated DDB1 was detected on an immunoblot probed partially redundant role in pathogenesis (2). Vpr is the only with anti-DDB1. (C) Myc-tagged SIVmac Vpr was expressed in transfected cells, immunoprecipitated with anti-Myc mAb and the coimmunoprecipitated HIV-1 accessory protein to be specifically packaged into virions DDB1 was detected on an immunoblot. (D) CEMss cells were infected with a and localizes to the nucleus of infected cells. Several roles for replication-competent NL4-3 [NL4-3(HA-Vpr)] engineered to express HA- Vpr in virus replication and pathogenesis have been proposed tagged Vpr or with control ⌬vpr NL4-3. After 4 days, HA-Vpr was immuno- (reviewed in refs. 3 and 4). It causes a modest increase in virion precipitated with anti-HA mAb. Coimmunoprecipitated DDB1 was detected production (5) and may facilitate nuclear import of the prein- on an immunoblot probed with anti-DDB1. (B, C, and D Right) Immunoblot tegration complex (6). Vpr has been found both to cause analysis of the cell lysates to confirm expression of the relevant proteins. packaging of UNG2 (7) and to cause its proteasomal degrada- tion (8). Vpr activates the ataxia telangiectasia-mutated (ATM) and Rad3-related protein (ATR) (9–13) causing the phosphor- interaction with DDB1 mediates the proteasomal degradation of ylation of the ATR substrates Chk1 (10) and the histone 2A UNG2 and SMUG1, is required for Vpr-induced apoptosis, and variant H2AX (11), resulting in G2 arrest. Vpr also induces prevents the cells from repairing UV damage. Interestingly, BRCA-1 and GADD45a (14) and generates nuclear foci con- taining H2AX and BRCA-1 (11). These phenotypes resemble that of DNA damage although Vpr itself does not seem to Author contributions: B.S. and N.R.L. designed research; B.S. and Y.H. performed research; damage DNA. B.S. contributed new reagents/analytical tools; B.S. analyzed data; and B.S. and N.R.L. wrote Activation of ATR is the furthest activity upstream in the the paper. cell-cycle regulatory pathway known to be affected by Vpr, but The authors declare no conflict of interest. this activity is not thought to be caused by a direct interaction. This article is a PNAS direct submission. Several cellular binding partners for Vpr have been reported Abbreviations: DDB1, DNA-binding protein 1; ATM, ataxia telangiectasia-mutated; ATR, ATM and Rad3-related protein; TAP, tandem affinity purification; CPD, cyclobutane py- including UNG2, HSP70, HHR23A, p300, and VprBP (reviewed rimidine dimer. in ref. 15), but these interactions have not been shown to induce ‡To whom correspondence should be sent at the present address: New York University G2 arrest or cause apoptosis. The identification of the immediate School of Medicine, Department of Microbiology, 550 First Avenue, New York, NY 10016. target of Vpr would provide valuable insight into its function. E-mail: [email protected]. Using tandem affinity purification (TAP) coupled with mass This article contains supporting information online at www.pnas.org/cgi/content/full/ spectrometry, we identified the damaged DNA-binding protein 0610167104/DC1. (DDB1) as a specific cellular binding partner for Vpr. The © 2007 by The National Academy of Sciences of the USA

4130–4135 ͉ PNAS ͉ March 6, 2007 ͉ vol. 104 ͉ no. 10 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0610167104 Downloaded by guest on September 24, 2021 A IP α-HA cells B IP α-HA cells α-DDB1 α-DDB1 α-DDB1

α-myc α-myc α-myc-Cul4 α-Roc1 α-Roc1 α-Roc1 α-HA α α-HA -HA –+HA-Vpr ++myc-Cul4 –+ –+ HA-Vpr –+ –+ HA-Vpr –+ –– ++ siDDB1 –– ++ myc-Cul4 ++ ++ ++ myc-Cul4 ++ ++

C IP α-HA cells IP α-HA cells α-DDB1 α-DDB1

α-HA α-HA

P A – A – – – WT WT WT WT 65AI70 30P I70A A30 L64PQ H71GR80AR85A A L64PQ65 H71GR80AR85A W54R W54R HA-Vpr HA-Vpr HA-Vpr HA-Vpr

Fig. 2. Vpr interacts with DDB1-Cul4A E3 ubiquitin ligase complexes. (A Left) HA-Vpr and Myc-Cul4 were expressed in cotransfected 293T cells, and HA-Vpr was immunoprecipitated with anti-HA mAb. Coimmunoprecipitated DDB1, myc-Cul4, and Roc1 were detected on an immunoblot. (Right) Expression of the proteins was confirmed by immunoblot analysis of the cell lysates. (B Left) DDB1 was knocked down by siRNA transfection. To detect the association of Vpr with DDB1 Cul4A and Roc1, Vpr was immunoprecipitated, and coimmunoprecipitated proteins were detected on an immunoblot. (Right) The efficiency of DDB1 knockdown was determined by immunoblot analysis of the cell lysates. (C) HA-Vpr point mutants were expressed in transfected 293T cells. HA-Vpr was immunoprecipitated, and the coimmunoprecipitated DDB1 was detected on an immunoblot (Left). Stability of the mutant Vpr was determined by immunoblot analysis of the cell lysates (second panel from the Left). The W54R Vpr mutant was tested in a separate experiment (Right two panels).

viruses such as simian virus 5 (SV5) and hepatitis B also target nucleus where it binds with high affinity to UV-damaged DNA DDB1 (16–18). to mediate repair. In addition to its role in NER, DDB1 forms part of a ubiquitin ligase containing Cul4A and Roc1 that targets Results cell-cycle regulatory proteins such as Cdt1 and p27kip for pro- Vpr Interacts with DDB1. To identify interacting cellular proteins, teasomal degradation (20, 21). a TAP-tagged Vpr expression vector was generated in which Vpr induces the degradation of UNG2 and SMUG1 in asso- tandem streptavidin and calmodulin-binding sites were linked to ciation with Cul4A (8). To determine whether DDB1 might play the N terminus of Vpr. The TAP-tagged Vpr was expressed in a role in this degradation, we tested for the presence of a complex transiently transfected 293T cells and shown to be functional for of Vpr with DDB1, Cul4A, and Roc1. 293T cells were cotrans- G2 arrest by DNA content profiling [supporting information (SI) fected with HA-Vpr and myc-Cul4A expression vector. HA-Vpr Fig. 6]. Complexes of the tagged Vpr were purified from lysates was immunoprecipitated, and the coimmunoprecipitated DDB1, of the transfected cells by TAP over streptavidin and calmodulin Cul4A, and Roc1 were detected on an immunoblot (Fig. 2A). columns and the purified protein complexes were visualized by The analysis confirmed the presence of a complex containing Coomassie staining after SDS/PAGE. Protein bands specific to Vpr, DDB1, Cul4A, and Roc1. the TAP-tagged Vpr (Fig. 1A) were excised and subjected to To determine whether the interaction with Cul4A and Roc1 mass spectrometry. Two proteins were identified in the 100-kDa was mediated by the interaction of Vpr with DDB1, siRNA band: DDB1 and VprBP, a previously reported binding partner transfection was used to knockdown DDB1. DDB1 was knocked for Vpr that was affinity purified on a bacterially synthesized down with specific or control siRNA, and the next day the cells recombinant Vpr column (19). In repeated analyses, we could were transfected with myc-Cul4A with or without HA-Vpr. Vpr not identify the 150-kDa protein. complexes were immunoprecipitated, and the coimmunoprecipi- To confirm the association of Vpr with DDB1 in cells, the tated DDB1, Cul4A, and Roc1 were detected on an immunoblot. complex was coimmunoprecipitated with DDB1 in transfected siRNA knocked down DDB1 by 80–90% (Fig. 2B). DDB1 cell lysates. An HA-tagged Vpr (HA-Vpr) was expressed in knockdown prevented the association with Cul4A and Roc1. HeLa cells and immunoprecipitated with anti-HA mAb. The Complexes of Vpr with Cul4A and Roc1 were detected in coimmunoprecipitated DDB1 was detected on an immunoblot. control-transfected cells but not in those knocked down for DDB1 was found to specifically coimmunoprecipitate with Vpr DDB1. These findings demonstrated the dependence of DDB1 (Fig. 1B). DDB1 also coimmunoprecipitated with a myc-tagged on formation of the complex with the Cul4A-Roc1 E3 ubiquitin SIV Vpr (Fig. 1C). In addition, the complex was coimmunopre- ligase. cipitated from HIV-1-infected T cells. For this analysis, CEMss cells were infected with replication-competent NL4-3 HIV-1 Identification of a Stable Vpr Point Mutant that Fails to Bind DDB1. To that encodes a HA-tagged Vpr or with ⌬vpr NL4-3 (Fig. 1D). probe the specificity of the interaction, a panel of mutated Vpr proteins was tested for DDB1 binding. HA-Vpr point mutants Vpr Interacts with the DDB1-Cul4-Roc1 E3 Ubiquitin Ligase. DDB1 is were expressed in transfected 293T cells and immunoprecipi- a 128-kDa UV-damaged DNA-binding protein involved in nu- tated. The coimmunoprecipitated DDB1 was detected on an cleotide excision repair (NER) that forms a heterodimer with the immunoblot. The majority of the mutants, including A30P, I70A,

48-kDa protein DDB2 (17, 18). Upon damage to DNA by UV R80A, Q65A, H71G, and R85A were poorly expressed (Fig. 2C, MICROBIOLOGY light, the DDB1/DDB2 complex (UV-DDB) translocates to the second panel), a property that is typical of Vpr point mutants

Schro¨felbauer et al. PNAS ͉ March 6, 2007 ͉ vol. 104 ͉ no. 10 ͉ 4131 Downloaded by guest on September 24, 2021 A B with specific siRNA, but not control siRNA prevented Vpr- induced UNG2 degradation (Fig. 3C). We further tested the UNG2 UNG2 DDB1 mutant W561A that does not bind Cul4A (24) and found that it was impaired for Vpr-induced UNG2 degradation (Fig. Vpr Vpr 3D). The effect was modest, probably because of the endogenous DDB1. Taken together, the results suggested that Vpr interacts Tubulin DDB1 with the DDB1-Cul4A-Roc1 E3 ubiquitin ligase to target UNG2 for proteasomal degradation. –WTL64PVpr Tubulin Vpr Interferes with UV Damage-Induced Nuclear Localization of DDB1. –+–+Vpr UV damage causes the DDB1/DDB2 complex to translocate to ––++siDDB the nucleus (25, 26). To determine whether Vpr might affect the D translocation of DDB1 to the nucleus, HA-Vpr was expressed in C transfected HeLa cells. The cells were exposed to UV light and UNG2 UNG2 subsequently Vpr and DDB1 were visualized by immunofluo- rescence. In unirradiated cells, Vpr localized to the nucleus (Fig. Vpr Vpr 4A) whereas DDB1 was present throughout the cell. In low dose 2 Cul4A UV irradiated cells (UV-B 30 J/m ) without Vpr, DDB1 accu- mulated in the nucleus (cells lacking Vpr are shown in SI Fig. 7). DDB1 Tubulin Expression of Vpr prevented the nuclear accumulation of DDB1 –+–+Vpr –+–+ Vpr in cells exposed to low and high doses of UV light. Biochemical analysis of DDB1 localization confirmed these results. Cells were ––++siCul4 WT W561A DDB1 transfected with Vpr expression vector and then fractionated Fig. 3. Vpr recruits the DDB1-Cul4A E3 ubiquitin ligase to target UNG2 for into nucleus and cytoplasm. Effective fractionation was con- proteasomal degradation. (A) 293T cells were cotransfected with HA-UNG2 ex- firmed by immunoblot quantitation of the nuclear protein, pression vector and empty vector, Vpr, or L64P Vpr expression vector. UNG2, Vpr, heterogeneous nuclear ribonucleoprotein C, and cytoplasmic and tubulin in cell lysates were detected on an immunoblot. (B) DDB1 was protein, GAPDH (SI Fig. 8). In unirradiated cells, Vpr was knocked down with siRNA. The cells were then transfected with Vpr and UNG2 present in the nuclear fraction and DDB1 was in the cytoplasm expression vector. UNG2, Vpr, DDB1, and tubulin were detected on an immuno- blot. (C) Cul4 was knocked down in 293T cells. The cells were then transfected (Fig. 4B). UV irradiation caused DDB1 to accumulate in the with Vpr and UNG2 expression vector. UNG2, Vpr, Cul4A, and tubulin were nucleus but in the presence of Vpr, DDB1 remained in the cy- detected on an immunoblot. (D) 293T cells were cotransfected with UNG2, Vpr, toplasm. L64P Vpr did not affect DDB1 translocation to the or empty expression vector and myc-DDB1 or W561A myc-DDB1 expression nucleus. These data suggest that Vpr blocks the UV damage vector. UNG2, Vpr, and DDB1 in the cell lysates were detected on an immunoblot. induced nuclear accumulation of DDB1.

Vpr Impairs DNA Repair. UV light causes the formation of (our unpublished observations). Only L64P and W54R were cyclobutane pyrimidine dimers (CPD) and pyrimidine- stably expressed. Importantly, L64P was stably expressed but pyrimidone (6-4) photoproducts (6-4PP) (17, 18). To test failed to interact efficiently with DDB1. The L64P mutant has whether the binding of Vpr to DDB1 might interfere with been described by Jian and Zhao (22), who previously showed DNA repair, HeLa cells were transfected with Vpr expression that this mutant was particularly proapoptotic. Two mutants, vector and two days later UV irradiated. Resulting CPDs in the H71G and R85A, were expressed at low levels but maintained genomic DNA was then quantitated by ELISA. In control UV their interaction with DDB1. W54R, a mutant that has been irradiated cells, UV light induced CPDs which were rapidly shown previously not to bind UNG2 (23), was stably expressed repaired (Fig. 4C). In cells that expressed Vpr, CPDs remained but maintained its interaction with DDB1. Two mutants, H71G relatively stable over the time course. To confirm that the and R85A, maintained their interaction with DDB1 despite their repair was DDB1-mediated, the experiment was repeated in reduced expression. No conclusion could be drawn for A30P, cells in which DDB1 was knocked down. In control cells, 80% Q65A, I70A, and R80A, which were poorly expressed. of CPDs generated by UV irradiation were repaired within 24 h whereas, in cells expressing Vpr, repair was prevented. In Vpr Induces the DDB1-Cul4A-Roc1 Ubiquitin Ligase to Target UNG2 for cells in which DDB1 had been knocked down, the CPDs failed Degradation. The association of Vpr with the DDB1-containing to be repaired (Fig. 4C Right). These results suggested that Vpr E3 ligase suggested that this complex might mediate the degra- interferes with DDB1-mediated repaired of UV-damaged dation of UNG2 and SMUG1. To determine whether Vpr DNA. binding to DDB1 was required for UNG2 degradation, we tested whether the DDB1-binding-defective Vpr mutant, L64P, could Vpr Prevents the Binding of DDB to UV-Damaged DNA. The impaired degrade UNG2. HA-UNG2 was expressed in cotransfected 293T ability to repair UV-damaged DNA in the presence of Vpr could cells with wild-type or L64P Vpr and detected on an immuno- be caused by a failure of UV-DDB to bind to damaged DNA. To blot. The results showed that Vpr, but not L64P Vpr, caused a determine whether Vpr interfered with the ability of DDB1 to bind significant reduction in UNG2 expression (Fig. 3A). To test UV-damaged DNA, nuclear extracts were prepared from cells that whether Vpr-induced degradation of UNG2 depends on DDB1, expressed a transfected Vpr expression vector and incubated with we knocked down DDB1 and tested for Vpr-induced degrada- a biotinylated UV irradiated DNA oligonucleotide. The DD- tion of UNG2. Knockdown of DDB1 was efficient and specific B:DNA complexes were detected by an electrophoretic mobility (Fig. 3B) and prevented the Vpr-induced degradation of UNG2. shift assay (EMSA). A shifted band was observed with the UV Similar results were obtained with SMUG1 (data not shown). irradiated but not the unirradiated control oligonucleotide probe These findings suggested that the interaction with DDB1 is (Fig. 4D). Nuclear extract from cells that expressed L64P Vpr or required for Vpr-induced degradation of UNG2 and SMUG1. empty vector resulted in two bands on the EMSA. These bands We further knocked down Cul4 to determine whether it is were much weaker in cells that expressed Vpr. Addition of a 100 M involved in Vpr-induced UNG2 degradation. Cul4 knockdown excess of unlabeled CPD prevented the shift. We concluded that

4132 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0610167104 Schro¨felbauer et al. Downloaded by guest on September 24, 2021 A Vpr DDB1 DAPI Merge B α-DDB1 control control α-HA cn cn cn – WT L64P Vpr

α 2 -DDB1 UVB 30 J/m UV irratiated α-HA cn cn cn – WT L64P Vpr

UVB 60 J/m2

no vpr siControl C vpr siControl 120 160 no vpr siDDB no vpr vpr siDDB 140 100 vpr 120 80 100 60 80 40 60 40 20 remaining CPDs (%) remaining CPDs (%) 20 0 0 0 4 8 12162024 04812162024 time post UV (h) time post UV (h) D E IP α-Flag cells

α-DDB1 α-Flag

α-HA – + – + Flag-DDB2 – + – + UV –++–++–++ – – + + HA-Vpr – – + + competitor ––+––+––+ No Vpr WT L64P Vpr

Fig. 4. Interaction of Vpr with DDB1 disrupts the UV-DDB complex. (A) Cells were UV irradiated with the indicated dose, and 6 h later Vpr (red) and DDB1 (green) were visualized by immunofluorescence with specific antibody and labeled second antibody. The nuclei (blue) were stained with DAPI. (B) 293T cells were transfected with wild-type or L64P HA-Vpr expression vector or empty vector control. After 2 days, the cells were UV irradiated, and, after an additional6h, nuclear and cytoplasmic fractions were prepared. DDB1 and Vpr in the fractions were detected on an immunoblot. (C) HeLa cells transfected with Vpr expression vector or empty vector (Left). The cells were irradiated with 10 J/m2 UV-C and harvested at the indicated time points. The signal from the control cells was subtracted, and counts at t0 were set to 100%. At the indicated time points, genomic DNA was prepared, and the CPD content was measured by ELISA. (Right) The cells were transfected with control or DDB1-specific siRNA before irradiation. The results are representative of at least two repetitions. (D) HeLa cells were transfected with wild-type, L64P Vpr, or empty vector, and, after 2 days, nuclear extracts were prepared. The extracts were incubated with irradiatedor unirradiated 5Ј-biotinylated double stranded oligonucleotide in the presence or absence of a 100-fold molar excess of unlabeled competitor oligonucleotide. Interaction of the oligonucleotide with the UV-DDB complex was detected by electrophoresis mobility shift assay (EMSA). (E) 293T cells were transfected with Flag-DDB2 and HA-Vpr or empty vector. Flag-DDB2 was immunoprecipitated, and coimmunoprecipitated DDB1 and Vpr were detected by immunoblot analysis (Left). An immunoblot on the cell lysates is shown (Right).

Vpr reduces the affinity of the UV-DDB complex for CPD The Interaction of Vpr with DDB1 May Cause G2 Arrest and Apoptosis. containing DNA. To determine whether Vpr induces apoptosis via DDB1, we tested the proapoptotic activity of L64P Vpr. HeLa cells were Vpr Disrupts the UV–DDB Complex. The DNA repair function of transfected with expression vector encoding wild-type or L64P DDB1 requires the interaction with DDB2. To determine whether HA-Vpr. Three days later, the extent of apoptosis was deter- Vpr might block this association, we transfected 293T cells with mined by TUNEL and G arrest was measured by DNA content. Flag-DDB2 expression vector with or without HA-Vpr. Two days 2 Wild-type Vpr induced G arrest and induced apoptosis in post transfection, DDB2 was immunoprecipitated and DDB1 that 2 coimmunoprecipitated was detected on an immunoblot. In the 40–50% of cells. In contrast, L64P Vpr did not induce apoptosis absence of Vpr, DDB1 and DDB2 coimmunoprecipitated. The (Fig. 5A)orG2 arrest (Fig. 5B). To further support these expression of Vpr prevented the coimmunoprecipitation (Fig. 4E). findings, we determined the effect of DDB1 knockdown on

These data suggest that Vpr interferes with the interaction of DDB1 cell-cycle progression and apoptosis. DDB1 was knocked down MICROBIOLOGY with DDB2. in HeLa cells and the cells were retransfected with Vpr expres-

Schro¨felbauer et al. PNAS ͉ March 6, 2007 ͉ vol. 104 ͉ no. 10 ͉ 4133 Downloaded by guest on September 24, 2021 A pcDNA WT Vpr L64P 8.6% 46.1% 4.5% counts

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 100 101 102 103 104 TUNEL B pcDNA WT Vpr L64P counts 0 50 100 150 200 2500 50 100 150 200 250 0 50 100 150 200 250

DNA content

C siControl siDDB1 counts 00 50 100 150 200 250 00 50 100 150 200 250

DNA content

D pcDNA WT Vpr 50 40 18.3% 43.9% siControl 30

10 0 10 1 10 2 10 3 10 4 10 0 10 1 10 2 10 3 10 4 20 FL1-H FL1-H apoptosis (%)

counts 10 15.3% 16.4% siDDB1 l 1 ro

10 0 10 1 10 2 10 3 10 4 0 1 2 3 4 10 10 10 10 10 siDDB siCont DB1+Vpr siD TUNEL siControl+Vpr

Fig. 5. Knockdown of DDB1 relieves Vpr-induced apoptosis. (A) HeLa cells were transfected with wild-type or L64P Vpr expression vector or empty vector. Apoptosis was detected 3 days later by TUNEL. (B) Cells were fixed and stained with propidium iodide and analyzed by FACS. (C) HeLa cells were transfected with control siRNA or siRNA against DDB1. Three days posttransfection, the cells were fixed and stained with propidium iodide. The cell-cycle profiles were analyzed by FACS. (D) HeLa cells were transfected with control siRNA or DDB1-specific siRNA and, 24 h later, with wild-type Vpr. Three days later, apoptosis was measured by TUNEL.

sion vector. In the absence of Vpr, knockdown of DDB1 itself and the subsequent downstream effects. The failure of DDB1 to caused the cells to arrest in G2/M (Fig. 5C). In addition, bind DNA and mediate the repair of DNA lesions generated knockdown of Vpr largely prevented Vpr-induced apoptosis during S phase might be sensed by the cell as DNA damage and (Fig. 5D). Microscopic inspection of the culture supported the would activate ATR, resulting in G2 arrest followed by apoptosis. TUNEL results. Cells that expressed Vpr were apoptotic in Interestingly, a viral connection to DDB1 has been described. appearance. Apoptosis was almost completely prevented by Paramyxoviruses such as SV5, and human parainfluenza knockdown of DDB1 (data not shown). virus type 2 encode regulatory proteins that interact with DDB1 (16–18). The paramyxovirus V protein hijacks DDB1-Cul4A to Discussion target STAT2 for proteasomal degradation, interfering with the Several lines of evidence suggest that the interaction with DDB1 induction of IFN response genes (27–30). Vpr is also similar to is physiologically significant. First, a Vpr point mutant that failed the X protein of (HBX) and V, both of which to bind DDB1 failed to induce G2 arrest and apoptosis. Second, slow or arrest the , induce a DNA damage response and knockdown of DDB1 prevented Vpr-induced UNG2 degrada- cause global alterations in cellular (31–33). HBX, tion. Third, Vpr binding to DDB1 interfered with the repair of like Vpr, induces apoptosis through its interaction with DDB1 UV-damaged DNA. Fourth, the interaction with DDB1 was (32, 34–36). Like Vpr, HBX and SV5-V disrupt the UV-DDB conserved in SIV Vpr. Fifth, knockdown of DDB1 results in G2 complex (31, 32). Whether the interaction of Vpr with DDB1 arrest, a phenotype that is also induced by Vpr. Our findings causes the degradation of STAT proteins remains to be deter- suggest a model in which Vpr activates the degradation of mined, although in a preliminary experiment, we did not detect cellular substrates by hijacking an E3 ubiquitin ligase containing decreased steady-state expression of STATs upon Vpr expres- DDB1-Cul4A-Roc1 (SI Fig. 9). sion in HeLa cells (data not shown). Our results present a potential explanation for several of the Our findings also present a second potential explanation for known cellular functions of Vpr. In the cell, Vpr induces a G2 cell-cycle arrest. The DDB1-Cul4A ubiquitin ligase targets response that mimics the response to DNA damage in which several proteins involved in cell-cycle regulation for proteasomal ATR is activated to phosphorylate Chk1 and nuclear ␥H2AX degradation including Cdt1 and p27kip (20, 21). Vpr could cause and 53BP1 foci form (9–12). The interaction of Vpr with DDB1 G2 arrest by interfering with the degradation of proteins that could cause the accumulation of damaged DNA by preventing regulate cell-cycle progression. This possibility would be consis- DNA repair. The damaged DNA could trigger activation of ATR tent with our finding that DDB1 knockdown caused G2 arrest. To

4134 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0610167104 Schro¨felbauer et al. Downloaded by guest on September 24, 2021 test this model, we overexpressed DDB1 in cells that expressed vided by P. R. Raychaudhuri. The T7 tag was replaced with a limiting amounts of Vpr. However, Vpr overexpression did not myc-tag by PCR, and the W561A mutation was introduced by abrogate Vpr-induced G2 arrest (data not shown). site-directed mutagenesis. NTAP-Vpr was generated by subclon- In addition to DDB1, the TAP analysis identified VprBP (19). ing Vpr from pcHA-Vpr into the pNTAP (Stratagene) at the Interestingly, VprBP was recently identified as a binding partner Bam-HI and HindIII sites. UNG2 and SMUG1 expression for DDB1. VprBP was proposed to be a member of a novel class vectors have been reported (8). pNL43(HA-Vpr) was con- of DDB1-Cul4A associated WD40 domain proteins (DCAF) structed by inserting a HindIII site at the start codon of vpr of that may be substrates for the DDB1-Cul4A E3 ubiquitin ligase pNL43. pNL43 was than cleaved with HindIII and at the unique (21, 37). Our findings do not distinguish between direct binding EcoRI site in vpr, and the resulting fragment was replaced with of DDB1 to Vpr or bridging through VprBP. a HindIII/EcoRI fragment from pcHA-Vpr. The modification The findings presented here provide an important clue toward does not interfere with virus replication in T cell lines. understanding the role of Vpr in HIV-1 replication. They suggest Additional experimental methods are in SI Materials and that Vpr mediates some, if not all of its biological activities Methods. through its interaction with DDB1. How these functions pro- mote HIV-1 replication remains to be determined. These find- We thank Hitoshi Endo (Jichi Medical School, Tochigi, Japan) and ings raise interesting possibilities regarding the role of Vpr in Pradip Raychaudhuri (University of Illinois at Chicago, Chicago, IL) for HIV-1 replication that can be experimentally addressed. reagents; Wolfgang Fischer for mass spectrometry; Esther Francisco, Materials and Methods Jody Chou, and Dimas Espinola for technical assistance; and Hui Chen and Erica Dhuey for critical reading of the manuscript. This work was pcHA-Vpr expresses an N-terminal HA-tagged NL4-3 Vpr in supported by National Institutes of Health Grants AI058864-03 and the cytomegalovirus promoter driven vector pcDNAI/amp (In- DA014494-05, the University of California at San Diego Center for vitrogen). The L64P mutation was introduced by site-directed AIDS Research, and the American Foundation for AIDS Research mutagenesis. 3X-Flag-DDB2 expression vector was a gift from (AmfAR). N.R.L. is an Elizabeth Glaser Fellow of the Pediatric AIDS Hitoshi Endo (38). The T7-DDB1 expression vector was pro- Foundation.

1. Lang SM, Weeger M, Stahl-Hennig C, Coulibaly C, Hunsmann G, Muller J, 17. Tang J, Chu G (2002) DNA Repair (Amsterdam) 1:601–616. Muller-Hermelink H, Fuchs D, Wachter H, Daniel MM, et al. (1993) J Virol 18. Wittschieben BB, Wood RD (2003) DNA Repair (Amsterdam) 2:1065–1069. 67:902–912. 19. Zhang S, Feng Y, Narayan O, Zhao LJ (2001) Gene 263:131–140. 2. Gibbs JS, Lackner AA, Lang SM, Simon MA, Sehgal PK, Daniel MD, 20. Bondar T, Kalinina A, Khair L, Kopanja D, Nag A, Bagchi S, Raychaudhuri Desrosiers RC (1995) J Virol 69:2378–2383. P (2006) Mol Cell Biol 26:2531–2539. 3. Sherman MP, De Noronha CM, Williams SA, Greene WC (2002) DNA Cell 21. Hu J, McCall CM, Ohta T, Xiong Y (2004) Nat Cell Biol 6:1003–1009. Biol 21:679–688. 22. Jian H, Zhao LJ (2003) J Biol Chem 278:44326–44330. 4. Muthumani K, Desai BM, Hwang DS, Choo AY, Laddy DJ, Thieu KP, Rao 23. Selig L, Benichou S, Rogel ME, Wu LI, Vodicka MA, Sire J, Benarous R, RG, Weiner DB (2004) DNA Cell Biol 23:239–247. Emerman M (1997) J Virol 71:4842–4846. 5. Goh WC, Rogel ME, Kinsey CM, Michael SF, Fultz PN, Nowak MA, Hahn BH, 24. Li T, Chen X, Garbutt KC, Zhou P, Zheng N (2006) Cell 124:105–117. Emerman M (1998) Nat Med 4:65–71. 25. Liu W, Nichols AF, Graham JA, Dualan R, Abbas A, Linn S (2000) J Biol Chem 6. Heinzinger NK, Bukinsky MI, Haggerty SA, Ragland AM, Kewalramani V, 275:21429–21434. Lee MA, Gendelman HE, Ratner L, Stevenson M, Emerman M (1994) Proc 26. Wakasugi M, Kawashima A, Morioka H, Linn S, Sancar A, Mori T, Nikaido Natl Acad Sci USA 91:7311–7315. O, Matsunaga T (2002) J Biol Chem 277:1637–1640. 7. Mansky LM, Preveral S, Selig L, Benarous R, Benichou S (2000) J Virol 27. Ulane CM, Rodriguez JJ, Parisien JP, Horvath CM (2003) J Virol 77:6385– 74:7039–7047. 6393. 8. Schrofelbauer B, Yu Q, Zeitlin SG, Landau NR (2005) J Virol 79:10978–10987. 28. Lin GY, Paterson RG, Richardson CD, Lamb RA (1998) Virology 249:189– 9. Lai M, Zimmerman ES, Planelles V, Chen J (2005) J Virol 79:15443–15451. 200. 10. Roshal M, Kim B, Zhu Y, Nghiem P, Planelles V (2003) J Biol Chem 29. Ulane CM, Kentsis A, Cruz CD, Parisien JP, Schneider KL, Horvath CM 278:25879–25886. 11. Zimmerman ES, Chen J, Andersen JL, Ardon O, Dehart JL, Blackett J, (2005) J Virol 79:10180–10189. Choudhary SK, Camerini D, Nghiem P, Planelles V (2004) Mol Cell Biol 30. Palosaari H, Parisien JP, Rodriguez JJ, Ulane CM, Horvath CM (2003) J Virol 24:9286–9294. 77:7635–7644. 12. Zimmerman ES, Sherman MP, Blackett JL, Neidleman JA, Kreis C, Mundt P, 31. Leupin O, Bontron S, Strubin M (2003) J Virol 77:6274–6283. Williams SA, Warmerdam M, Kahn J, Hecht FM, et al. (2006) J Virol 32. Bontron S, Lin-Marq N, Strubin M (2002) J Biol Chem 277:38847–38854. 80:10407–10418. 33. Lin-Marq N, Bontron S, Leupin O, Strubin M (2001) Virology 287:266–274. 13. Nakai-Murakami C, Shimura M, Kinomoto M, Takizawa Y, Tokunaga K, 34. Becker SA, Lee TH, Butel JS, Slagle BL (1998) J Virol 72:266–272. Taguchi T, Hoshino S, Miyagawa K, Sata T, Kurumizaka H, et al. (2006) 35. Lin GY, Lamb RA (2000) J Virol 74:9152–9166. 26:477–486. 36. Sitterlin D, Bergametti F, Transy C (2000) Oncogene 19:4417–4426. 14. Andersen JL, Zimmerman ES, DeHart JL, Murala S, Ardon O, Blackett J, 37. Angers S, Li T, Yi X, Maccoss MJ, Moon RT, Zheng N (2006) Nature Chen J, Planelles V (2005) Cell Death Differ 12:326–334. 493:590–593. 15. Kino T, Pavlakis GN (2004) DNA Cell Biol 23:193–205. 38. Inoki T, Yamagami S, Inoki Y, Tsuru T, Hamamoto T, Kagawa Y, Mori T, 16. Barry M, Fruh K (2006) Sci STKE, pe21. Endo H (2004) Biochem Biophys Res Commun 314:1036–1043. MICROBIOLOGY

Schro¨felbauer et al. PNAS ͉ March 6, 2007 ͉ vol. 104 ͉ no. 10 ͉ 4135 Downloaded by guest on September 24, 2021