HIV-1 Vpr Protein Induces Proteasomal Degradation of Chromatin-associated Class I HDACs to Overcome Latent Infection of Macrophages*

Mechanisms underlying HIV-1 latency remain among the most crucial questions that need to be answered to adopt strategies for purging the latent viral reservoirs. Here we show that HIV-1 accessory protein Vpr induces depletion of class I HDACs, including HDAC1, 2, 3, and 8, to overcome latency in macrophages. We found that Vpr binds and depletes chromatin-associated class I HDACs through a VprBP-dependent mechanism, with HDAC3 as the most affected class I HDAC. De novo expression of Vpr in infected macrophages induced depletion of HDAC1 and 3 on the HIV-1 LTR that was associated with hyperacetylation of histones on the HIV-1 LTR. As a result of hyperacetylation of histones on HIV-1 promotor, the virus established an active promotor and this contributed to the acute infection of macrophages. Collectively, HIV-1 Vpr down-regulates class I HDACs on chromatin to counteract latent infections of macrophages.

HIV-1 genome encodes 4 accessory proteins, including Nef, Vif, Vpu, and Vpr. These proteins are known to down-regulate cellular proteins through different mechanisms to facilitate various stages of viral infection. For instance, Nef down-regulates surface expression of CD4 and MHC-I (1); Vif down-regulates APOBEC3G and APOBEC3F (2), and Vpu down-regulates surface CD4 and tetherin (3). Recent studies also show a growing list of cellular proteins whose degradation is induced by Vpr while the outcomes of their degradation remain to be elucidated.
HIV-1 Vpr is well-documented to interact with the Cul4-DDB1[VprBP] E3 ubiquitin ligase to direct proteasomal degradation of a number of cellular proteins and induce G2/M cell cycle arrest (4 -7). Vpr enhances proteasomal degradation of the natural substrates of the Cul4-DDB1[VprBP] E3 ubiquitin ligase (8 -10). For instance, Vpr has been shown to enhance proteasomal degradation of UNG2 and dicer (11,12). These proteins are recruited by Vpr-binding protein (VprBP) 3 to the E3 ubiquitin ligase complexes as their natural substrates. By binding to VprBP, Vpr enhances degradation of the natural substrates recruited by VprBP (10,13). Furthermore, minichromosome maintenance complex component 10 (MCM10) was recently shown to be the natural substrate of the Cul4-DDB1[VprBP] E3 ubiquitin ligase (14) and our laboratory showed that proteasomal degradation of MCM10, is also enhanced by Vpr (10).
It is believed that Vpr induces most of its functions by interacting with VprBP. To investigate the biological roles of Vpr, we investigated proteins that interact with VprBP. It was recently shown that VprBP directly interacts with nucleosomes by recognizing the unacetylated N-terminal tail of histone H3 (15). In this complex, VprBP was also shown to bind histone deacetylase 1 (HDAC1). HDACs are a class of enzymes that remove acetyl groups from ⑀-N-acetyl lysine amino-acids of histones and other proteins and also regulate a number of key cellular processes, including gene expression, protein complex formation, and protein stability and localization (16 -19). In mammals, 18 HDACs have been described and they are classified into 4 classes based on their phylogeny and function (20,21). There is a high level of homology among the members of class I HDACs, which comprises HDAC1, 2, 3, and 8. Class I HDACs localize to the nucleus, except for HDAC3, which can be found both in the cytoplasm and the nucleus (22,23). In this study, we found that Vpr binds and depletes chromatin-associated class I HDACs, enabling the virus to overcome latent infection in primary macrophages.

Experimental Procedures
Cell Lines and Primary Cells-HEK293T and HeLa cells were obtained from ATCC. TZM-bl cells were obtained from the NIH AIDS Research and Reference Reagent Program. Doxycycline-inducible HeLa cell lines (HeLa-iFlag-Vpr, HeLa-iFlag-Q65R) and the control cell line (HeLa-iMock) were described previously (10). Expression of Flag-Vpr was induced by adding 1 g/ml doxycycline. HeLa and the inducible HeLa cell lines were maintained in DMEM supplemented with 10% FBS, 100 units/ml penicillin, and 100 g/ml of streptomycin. To isolate monocyte-derived macrophages (MDMs), hepatitis B virus and HIV-1-seronegative adult donors that had signed a consent form, provided peripheral blood samples. PBMCs were isolated from whole blood using Ficoll-density gradient centrifugation.
Monocytes were primarily isolated by plastic adherence. For differentiation into mature macrophages, monocytes were cultured in RPMI supplemented with 50 ng/ml macrophage-colony stimulating factor M-CSF, 100 units/ml penicillin, 100 g/ml streptomycin, and 10% FBS. After 7 days, cells were washed with PBS to rinse off the nonadherent cells. Purity of MDMs was verified by flow cytometry using antibodies against CD14 and CD16.
Lentiviral Transduction and Viral Infection-HeLa cells were transduced with lentiviral vectors at an MOI of 2.0. Primary MDMs were infected at an MOI of 1.0 with fully replicative viruses or with 1 ng of p24 per 10 5 cells when infecting with single cycle VSV-G pseudotyped viruses.
siRNA Transfection-VprBP, HDAC1, HDAC3 and nontargeting siRNAs were transfected into inducible HeLa cell lines and/or primary MDMs using Lipofectamine RNAiMax Reagent according to the manufacturer's instructions.
Cell Fractionation-In order to fractionate HeLa and inducible HeLa cell lines, cells were lysed in 1 ml of 0.5% Triton lysis buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 0.5% Triton, and protease inhibitor mixture). Cells were incubated for 10 min with mild agitation at 4°C and then centrifuged at 6000 rpm for 10 min at 4°C to pellet chromatin and other large insoluble debris. Supernatant was collected as the soluble fraction, and the pellet was resuspended in 2 ml of Benzonase buffer (50 mM Tris pH 8, 150 mM NaCl, 1.5 mM MgCl 2 , 0.1 mg/ml BSA, and protease inhibitor mixture). The resuspended pellet was centrifuged and supernatant was discarded. 1 l of Benzonase enzyme (25 units/l) was added to 1 ml of Benzonase buffer and this was used to resuspend the pellet and incubate it on ice for 60 min. The Benzonase-treated pellet was centrifuged at 13000 rpm for 10 min at 4°C. Supernatant was collected as the fraction containing chromatin-bound proteins. Purity of the fractions was confirmed by the presence of GAPDH in soluble protein fraction and histone H3 in chromatin-bound protein fraction.
Western Blotting and Immunoprecipitation-Anti-Flag and anti-HA immunoprecipitations were performed using 40 l of commercial antibodies conjugated to agarose beads. For anti-VprBP, HDAC1, HDAC2, HDAC3, and HDAC8 immunoprecipitations, antibodies (2 g per immunoprecipitation) were incubated overnight at 4°C with 50 l of protein A-Sepharose beads in 1 ml of PBS supplemented with 5% FBS. All the immunoprecipitations were performed in the presence of 150 mM NaCl and 0.5% Triton X-100 for 2 h at 4°C. For VprBP pulldown, 25 g/ml ethidium bromide was also added to the solution to avoid the nonspecific pull-down of chromatin-bound proteins through DNA bridges. After thorough washes in 0.5% Triton lysis buffer, the anti-Flag immunoprecipitated proteins were eluted by adding 100 g/ml Flag peptides. Immunoprecipitations using protein A-Sepharose beads and monoclonal anti-HA-agarose beads were released by treating the beads with 0.1 M glycine, pH 2.0, for 10 min on ice. Immunoprecipitated proteins and cellular fractions (30 g) were resuspended in Laemmli buffer, heat-denatured for 5 min, and separated Vpr Down-regulates Class I HDACs to Overcome Latency in MDMs FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6 JOURNAL OF BIOLOGICAL CHEMISTRY 2697 on 12% SDS-PAGE gels. Western blots were performed as described previously (10).
ELISA-Supernatants of the infected cells were collected, inactivated with Triton X-100 (1% final concentration), and stored at Ϫ80°C until analyzed. Levels of HIV-1 p24 were determined using HIV-1 p24 Antigen Capture assay (ABL Inc.) according to the manufacturer's instructions.
Luciferase Assay-Primary MDMs were infected at an MOI of 1.0 with single cycle VSV-G pseudotyped pNL4.3.Luc.Env-(WT/⌬Vpr) that expresses luciferase in the infected cells. Seven days after infection, cells were washed with PBS and lysed directly in the tissue culture plates using Triton X-100 (1% final concentration). Luciferase was measured in the cell lysates using luciferase assay system (Promega).
Chromatin Immunoprecipitation (ChIP)-To analyze enrichment of class I HDACs and acetylated histones on the HIV-1 LTR, 5 ϫ 10 5 MDMs were infected with pNL4.3.AD8. IRES_GFP_Nef-(WT and ⌬Vpr) at an MOI of 1.0. After 7 days, cells were looked under a fluorescent microscope for GFP-positive infected cells. If percentage of the infected cells was higher than 5%, we proceeded with the experiment. Cells were then fixed with 0.5% formaldehyde. After 5 min, fixation was quenched by adding glycine to a final concentration of 125 mM for 5 min and cells were then washed with PBS and harvested. Cells were lysed in TpA (0.25% Triton, 10 mM Tris pH8, 10 mM EDTA, 0.5 mM EGTA, protease inhibitor) for 5 min on ice and centrifuged. This would disrupt the cellular membrane and release the non-integrated proviruses. Cell nuclei were then washed with TpB (200 mM NaCl, 10 mM Tris, pH 8, 1 mM EDTA, 0.5 mM EGTA, protease inhibitor) and lysed in TpS (0.5% SDS, 0.5% Triton, 10 mM Tris pH8, 140 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, protease inhibitor). Chromatin was sheared using Bioruptor XL sonicator (Diagenode) to obtain DNA fragments ranging from 200 to 800 bps.
For each immunoprecipitation, 2 g of antibody was conjugated to 10 l of protein A and protein G Dynabeads (1:1 ratio) by overnight incubation of the beads/antibody in 500 l of ChIP dilution buffer (1% Triton, 10 mM Tris, pH 8, 150 mM NaCl, 2 mM EDTA) at 4°C. ChIPs were performed by adding the sonicated DNA to the antibody-conjugated Dynabeads and incubation at 4°C for 4 h. Following extensive washes, the DNA of ChIP samples and nonimmunoprecipitated chromatin (input) were decross-linked in TpE (0.3% SDS, 50 mM Tris, pH 8, 10 mM EDTA, 0.4 M NaCl) by overnight incubation at 65°C and extracted using the standard phenol-chloroform method. The DNA was then subjected to quantitative PCR.
To assess whether Vpr affects class I HDACs at transcription level, HeLa-iFlag-Vpr and HeLa-iMock were induced by adding 1 g/ml doxycycline. After 30 h, cells were harvested, and total RNA was extracted from 10 6 cells using RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Using SuperScript II Reverse Transcriptase (Invitrogen) and oligo(dT) [12][13][14][15][16][17][18] Primers (Invitrogen), RNA was reverse transcribed into cDNA. The cDNA was diluted 100-fold with DNase-free water. Quantitative PCR was performed on the diluted cDNA using SYBR Select Master Mix. Primers used in qPCR include: Data were normalized to GAPDH and calculated using the 2 Ϫ(⌬⌬CT) method.
Quantitative Alu-PCR-To compare integration of HIV-1 provirus in the presence or absence of Vpr, genome of infected MDMs was analyzed using Alu-PCR. Briefly, primary MDMs were infected with the single cycle VSV-G pseudotyped pNL4.3Env-.IRES.GFP WT and ⌬Vpr viruses. Four days postinfection, GFP-positive and -negative cells were sorted and the genomic DNA was extracted from 20,000 sorted and unsorted cells using DNeasy Blood & Tissue Kit (QIAGen). DNA was amplified using the forward primer Alu-F 5Ј-GCCTCCCAAA-GTGCTGGGATTACAG-3Ј that binds human Alu gene and the reverse primer HIV-LTR-outer-R 5Ј-TGCTAGAGATTT-TCCACACTGA-3Ј that binds HIV-1 LTR. One microliter of the amplified DNA was used for qPCR using HIV-LTR-F 5Ј-CTGGCTAACTAGGGAACCCACT-3Ј and HIV-LTR-inner-R 5Ј-CTCAATAAAGCTTGCCTTGAGTGCTC-3Ј primers. To amplify the total HIV-1 provirus (integrated and non-integrated), DNA was amplified using HIV-LTR-F and HIV-LTR-inner-R. Amplification of a specific region on HIV-1 LTR was detected using the probe LTR-Probe 5Ј-(FAM)-CTC-AATAAAGCTTGCCTTGAGTGCTC-(TAMRA)-3Ј. Results of Alu-PCR were normalized to the total HIV-1 provirus.
Flow Cytometry and Sorting-To analyze the cell cycle profile of inducible HeLa cell lines, they were induced by adding doxycycline and after 30 h their DNA content was labeled by pro-pidium iodide staining and cells were analyzed as previously described (10). To examine purity and differentiation of MDMs, their surface markers were labeled using CD14 Mouse Anti-Human mAb, PE-Texas Red conjugate and APC/Cy7 anti-human CD16 for 30 min on ice. Cells were washed three times and analyzed using flow cytometry. To analyze infectivity of the single cycle HIV-1 in MDMs, they were infected with VSV-G-pseudotyped NL4 -3-deltaE-EGFP (WT, ⌬Vpr, Q65R, and R80A). Four days post-infection, MDMs were analyzed for expression of GFP. All the analyses were performed using CyAn ADP Analyzer.
To analyze acute versus latent infections, primary MDMs were infected with single cycle VSV-G pseudotyped GFP reporter WT/⌬Vpr viruses for 4 days. The GFP-positive and -negative cells were sorted in PBS using an Influx cell sorter (BD Biosciences) and analyzed using Alu-PCR or Western blot. To analyze viral expression in macrophages, primary MDMs were infected with GFP reporter WT/⌬Vpr viruses for 7 days. The GFP-positive cells were then sorted using an Influx cell sorter (BD Biosciences) and analyzed using Western blot.
Statistical Analysis-Student's t test or Analysis of variance (ANOVA) with Bonferroni's multiple comparison test was used for statistical analysis using GraphPad Prism 6.0. A value of p Ͻ 0.05 was considered statistically significant. Results were expressed as mean Ϯ S.E.M. or mean Ϯ S.D., and represent data from a minimum of three independent experiments unless otherwise stated.

HIV-1 Vpr Induces Depletion of Class I HDACs on
Chromatin-Since VprBP was reported to interact with HDAC1 on chromatin, we developed a fractionation method to examine the impact of Vpr on this interaction in a localized manner. Our fractions included soluble and chromatin-bound protein fractions, in which the soluble fraction contained both cytoplasmic proteins and soluble nuclear proteins, while the chromatin-bound protein fraction contained nuclear proteins released by treatment of the chromatin pellet with Benzonase DNase.
A doxycycline-inducible cell line, HeLa-iFlag-Vpr cells, was used as previously described (10). Expression of Flag-tagged HIV-1 Vpr was induced by adding doxycycline to HeLa-iFlag-Vpr cells and the quantities of class I HDACs were monitored in a time course experiment (Fig. 1A). A significant depletion of HDAC1 was found after 10 h in chromatin fraction followed by its depletion in the soluble fraction only after 30 h. Depletion of HDAC2 was only found significant in chromatin fraction 30 h post-induction with no significant depletion in the soluble fraction. Depletion of HDAC3 was initiated as early as 5 h postinduction in the chromatin fraction and then followed by its depletion in the soluble fraction 20 h post-induction. Depletion of HDAC8, which was only found in chromatin fraction, appeared significant after 20 h. Our results indicated that depletion of class I HDACs was more drastic on chromatin as the chromatin associated class I HDACs were depleted before the soluble class I HDACs. Chromatin depletion of HDAC3 was found to occur before other HDACs, suggesting more sensitiv-ity of HDAC3 to expression of Vpr while HDAC2 was found the least sensitive member of class I HDACs.
Expression of Vpr did not seem to affect the mRNA levels of class I HDACs (Fig. 1B) suggesting the down-regulation of HDACs occurs at protein level. Furthermore, we tested the effect of Vpr on other classes of HDACs (Fig. 1C). Expression of Vpr did not affect the protein levels of other classes of HDACs.
Vpr-induced Depletion of Class I HDACs Is Not Resulted from G 2 /M Arrest-Vpr is well-documented to induce a cell cycle arrest at G 2 /M phase through activation of ATR pathway (24). The cell cycle arrest could potentially change protein contents of the cells. To examine whether the depletion of class I HDACs is a by-product of G 2 /M cell cycle arrest we inhibited Vpr-induced G2/M cell cycle arrest by caffeine, an ATR inhibitor. As shown in Fig. 2A, treatment of induced HeLa-iFlag-Vpr cells with caffeine efficiently inhibited Vpr-induced G 2 /M arrest. However, inhibition of G 2 /M arrest did not affect the ability of Vpr for depletion of class I HDACs.
We also induced G 2 /M cell cycle arrest by treatment of HeLa cells with nocodazole. Despite the potent G 2 /M arrest induced by Nocodazole, the cell cycle arrest did not induce depletion of HDACs (Fig. 2B).
Vpr-induced Depletion of Class I HDACs Is Proteasomal Dependent-To determine whether Vpr usurps a proteasomal pathway for degradation of class I HDACs, we used transduction to avoid the interference between proteasomal inhibitor and doxycycline induction. Normal HeLa cells were transduced with lentiviral vectors for expression of the Flag-tagged Vpr or Q65R and R80A mutants that have been previously reported defective for depletion of proteasomal targets of Vpr (10,12,13). HeLa cells were also transduced with lentiviral vectors for expression of simian immunodeficiency virus Vpx or empty lentiviral vectors. The transduced cells were treated with the proteasomal inhibitor, MG132, or DMSO 10 h before harvesting (Fig. 3A). Western blot analysis of the fractionated cells indicated that the wild-type Vpr was able to efficiently deplete all class I HDACs in the chromatin fraction of the cells treated with DMSO. A milder effect of Vpr on HDAC1 and HDAC3 was also observed in the soluble fraction that was only found significant for HDAC3. However, treatment of the cells with MG132 restored HDACs in both fractions. Q65R and R80A mutants as well as Vpx were not able to deplete HDACs in the presence or absence of MG132. Deficiency of the mutants for depletion of HDACs, especially Q65R that is defective for interaction with VprBP, provided further support for proteasomal dependence of depletion of HDACs.
To provide direct evidence demonstrating that the Vpr-induced depletion of HDACs is in fact proteasomal dependent, we cotransfected HEK293T cells with expression vectors for HA-ubiquitin, Vpr and C-terminally Flag-tagged HDAC1 or HDAC3. Cells were fractionated and pulled down using anti-HA antibodies (Fig. 3, B and C). Western blot analysis of the eluates indicated that ubiquitination of HDAC1 and 3 only occurred on chromatin. Interestingly, expression of Vpr increased the basal ubiquitination of HDAC1 and HDAC3 to 2.2-and 3.4-fold, respectively.

Vpr-induced Depletion of Class I HDACs Is
VprBPdependent-To characterize composition of the complex involved in the Vpr-induced depletion of HDACs, we decided to test involvement of VprBP. We transfected inducible HeLa cell lines with non-targeting siRNA or siRNA against VprBP. Expression of Flag-Vpr or Flag-Q65R was induced by doxycycline treatment. Analysis of the fractionated cell lysates demonstrated that the wild-type Vpr was able to deplete class I HDACs on chromatin only in the presence of VprBP since depletion of VprBP abolished the Vpr ability to induce depletion of class I HDACs (Fig. 4A). The Q65R mutant, which is defective for binding to VprBP, was not able to induce depletion of HDACs in the presence or absence of VprBP, providing further support for the key role of VprBP in Vpr-induced depletion of HDACs. Immunoprecipitation of Vpr indicated that the wild-type Vpr was able to interact with class I HDACs only in the presence of VprBP and only in the chromatin fraction. Depletion of VprBP abolished the Vpr ability to bind HDACs, suggesting Vpr indirectly binds HDACs through VprBP on the chromatin. The Q65R mutant was not able to interact with HDACs and VprBP, however, Q65R was also found defective for efficient localization on chromatin (Fig. 4A, compare lanes 9  and 10 with lanes 11 and 12). We previously found that MG132 treatment stabilized the Q65R mutant on chromatin (Fig. 3A).
We took advantage of MG132 treatment to stabilize the Q65R mutant on DNA to provide a bona fide control that can be pulled down in chromatin fraction (Fig. 4B). Pull down of Q65R at quantities comparable with those of the wild-type Vpr indicated that Q65R was not able to efficiently interact with VprBP, and therefore it was not able to pull down class I HDACs. Of note, expression of Vpr did not seem to affect the localization or expression levels of VprBP in any of the fractions (Fig. 4, A and  B).
We then tested interaction of VprBP with class I HDACs by immunoprecipitation of the endogenous VprBP. Ethidium bromide was also used in the immunoprecipitations to dissociate chromatin-bound proteins from bridging DNA fragments. VprBP was able to pull down class I HDACs only in the chromatin fraction (Fig. 4C). This interaction was not due to nonspecific pull-down through DNA bridges since it also occurred in the presence of ethidium bromide. Individual pull-down of class I HDACs showed that HDACs bind VprBP and DDB1 only in the chromatin fraction and these bindings occur independently of Vpr (Fig. 4D).
Vpr Enhances HIV-1 Infection in Macrophages through a VprBP-dependent Mechanism-Vpr is known to enhance HIV-1 replication in macrophages. To examine whether this effect of Vpr is related to depletion of HDACs, we depleted  FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6

JOURNAL OF BIOLOGICAL CHEMISTRY 2701
VprBP in MDMs and infected them with the wild-type or Vprdefective (⌬Vpr) viruses. Replication kinetics of the viruses was measured by quantification of p24 in the supernatant of the infected cells (Fig. 5A) and it was shown that in the presence of VprBP, the WT virus infected MDMs more efficiently than the ⌬Vpr virus. Interestingly, depletion of VprBP abolished the effect of Vpr such that there was no significant difference between the WT and ⌬Vpr viruses. We also showed that the Q65R and R80A mutants did not increase infection of MDMs at levels comparable to the WT virus (Fig. 5B).
Since class I HDACs are known to suppress gene expression, we tested whether Vpr-induced depletion of HDACs increases the HIV-1 gene expression. Primary MDMs were infected with the WT and ⌬Vpr viruses that had a luciferase gene as an indicator of active transcription of the viral genome. Measuring the luciferase reporter indicated that transcription of the WT virus was 3.58-fold higher than that of the ⌬Vpr virus (Fig. 5C). We also compared infectivity of the single cycle VSV-G pseudotyped GFP-reporter WT HIV-1 or the corresponding ⌬Vpr, Q65R, and R80A mutants. As shown in Fig. 5, D and E, the ⌬Vpr, Q65R, and R80A viruses infected MDMs at comparable levels. But the WT virus showed ϳ3-fold increase in the efficiency of the infection, suggesting the efficient infection of MDMs by the WT virus is dependent on Vpr and its interaction with VprBP. Conversely, when we compared the expression of p24 and GFP in equal numbers of sorted cells, we did not observe a significant difference between the WT and ⌬Vprinfected MDMs (Fig. 5F). The mRNA level of HIV-1 was also compared in total population of WT and ⌬Vpr-infected MDMs (Fig. 5G). Additionally, GFP-positive infected MDMs were sorted and their HIV-1 mRNA was compared (Fig. 5H). Although the viral mRNA had increased in the total population, when equal numbers of individual infected cells were compared, no significant difference was observed in the presence or absence of Vpr. This suggested that Vpr increases the total number of infected MDMs but not the viral expression in individual infected cells and opened new venues for new models and further investigations as discussed in the following sections.
Vpr Depletes Class I HDACs on the HIV-1 LTR of Infected Macrophages-Fractionation requires a large number of cells that is not feasible in the case of primary MDMs. Therefore, we performed chromatin immunoprecipitations (ChIPs) to examine whether Vpr induces depletion of HDACs on the HIV-1 LTR of infected MDMs. HIV-1 has two conserved nucleosomes on its LTR, known as nuc-0 and nuc-1 (25,26). We designed primers to test depletion of class I HDACs and hyperacetylation of histones on nuc-0 and nuc-1 as well as nuc-2 that is located downstream of LTR (Fig. 6A). Primary MDMs were infected with the WT and ⌬Vpr viruses. After 7 days, cells were treated with MG132 or DMSO for 20 h and then we performed ChIPs against individual class I HDACs. As shown in Fig. 6B, the WT virus depleted 50 -60% of HDAC1 on the HIV-1 LTR of the DMSO-treated cells. Interestingly, treatment with MG132 restored the levels of HDAC1 on HIV-1 genome. The levels of HDAC2 were 10 -20% lower in the WT HIV-1 LTR but this was not statistically significant. HDAC3, however, showed about 80% depletion on the HIV-1 LTR and, similarly to HDAC1, it was restored after treatment with MG132. We could not find an enrichment of HDAC8 on the HIV-1 LTR.
We also examined the effect of Vpr on the gapdh coding region. Unlike LTR that is only found in infected cells, gapdh is found in both infected and uninfected macrophages. To ensure that our ChIP assay only analyzes gapdh in the infected cells, GFP-positive infected cells were sorted before ChIP (Fig. 6C). Unlike the HIV-1 LTR, Vpr did not affect enrichment of HDAC1 and HDAC3 on the gapdh gene, suggesting that Vpr degrades HDACs in a localized manner that does not affect all cellular genes.
Vpr Induces Hyperacetylation of Core Histones on the HIV-1 LTR-Class I HDACs remove acetyl groups from the lysine residues of histones and therefore depletion of HDACs is expected to induce hyperacetylation of histones. To examine whether the Vpr-induced depletion of HDACs is followed by hyperacetylation of histones, we infected MDMs with the WT and ⌬Vpr viruses. After 7 days, cells were ChIPed against acetylated histone H3 (Ace H3K9) and acetylated histone H4 (Ace H4K5) (Fig. 7, A and B). Comparison of MDMs infected with the WT and ⌬Vpr viruses indicated that Ace H3K9 was enriched on nuc-0, nuc-1, and nuc-2 of the WT viruses, 3.26-, 3.82-, and 3.45-fold, respectively. Similarly, Ace H4K5 was enriched on nuc-0, nuc-1, and nuc-2 of the WT viruses, 2.98-, 3.07-, and 2.34-fold, respectively. Interestingly, there was no significant difference between the gapdh gene of the WT and ⌬Vpr-infected MDMs. As a negative control, we examined enrichment of Ace H3K9 and H4K5 on an intergenic region, where no enrichment of hyperacetylated histones is expected, and our results showed that there was no significant enrichment on this region. To test whether hyperacetylation of histones follows the same pattern as depletion of HDACs we examined if the hyperacetylation was proteasomal dependent. We infected MDMs with the WT and ⌬Vpr viruses and after 7 days we treated cells with MG132 or DMSO for 20 h. Using ChIP assay, cells were then analyzed for the enrichment of acetylated histones (Ace H3K9 and H4K5) on nuc-0 (Fig. 7, C and D). Comparison of the WT versus ⌬Vpr infected MDMs, indicated that expression of Vpr by the WT virus induced about 3-fold enrichment of both Ace H3K9 and Ace H4K5 on nuc-0, while MG132 treatment significantly abolished the ability of the Vpr such that there was no significant difference between the WT and ⌬Vpr-infected MDMs.
We included an additional control against total histone H3 to ensure that there is no significant difference in localization of histones between the WT and ⌬Vpr-infected viruses (Fig. 7E).  FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6 Although there was a good enrichment of histones on nuc-0, nuc-1, and nuc-2, there was no significant difference between MDMs infected with the WT and ⌬Vpr viruses.

Vpr Down-regulates Class I HDACs to Overcome Latency in MDMs
Vpr Overcomes Latent Infections of Macrophages-We found that Vpr induces depletion of HDAC1 and HDAC3 on the HIV-1 LTR of infected MDMs and this depletion was associated with hyperacetylation of histones in this region. Hyperacetylation is normally associated with active transcription which is consistent with our finding in total population of infected MDMs (Fig. 5, A-C). However, we did not observe an increase in the expression of p24 or GFP when we compared equal numbers of infected cells (Fig. 5F). We hypothesized that by hyperacetylation of histones on the HIV-1 LTR, Vpr may overcome latent infections in macrophages to increase the number of acutely infected cells, as consistent with our findings (Fig. 5, D and E). To test our hypothesis, we infected primary MDMs with the single cycle VSV-G pseudotyped GFP-reporter WT or ⌬Vpr HIV-1. After 4 days, cells were harvested and Alu-PCR was performed on the total population (Fig. 8A) or cells were sorted and Alu-PCR was performed on equal numbers of the GFP-positive and -negative cells, separately (Fig. 8B). Alu-PCR showed no significant difference in the integration of HIV-1 in the total population of unsorted cells, in the presence or absence of Vpr (Fig. 8A). Alu-PCR did not also show a significant difference when integrated DNA was compared in equal numbers of GFP-positive cells. Interestingly, when GFPnegative cells were compared, MDMs infected with ⌬Vpr virus had a significantly higher occurrence of integrated HIV-1 genome, suggesting the presence of more latent infections in the absence of Vpr (Fig. 8B).
Since Vpr was found to inhibit HDACs in macrophages, we decided to test whether the lack of Vpr could be complemented  FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6 in ⌬Vpr viruses using an HDAC inhibitor. In order to test our hypothesis, we infected primary MDMs with the single cycle VSV-G pseudotyped GFP-reporter WT or ⌬Vpr virus. After 3 days, cells were treated with SAHA or DMSO for 24 h. Four days after infection, the GFP-positive and -negative cells were sorted and analyzed using Alu-PCR (Fig. 8C). Consistent with our previous results (Fig. 8B), when cells treated with DMSO, we found more HIV-1 DNA integrated in the genome of MDMs infected with the ⌬Vpr virus. However, when cells treated with SAHA, relatively equal copies of integrated HIV-1 genome were found in the MDMs infected with the ⌬Vpr or WT viruses, suggesting SAHA has reactivated the latent proviruses in MDMs infected with the ⌬Vpr virus. Interestingly, SAHA significantly increased infectivity of the ⌬Vpr virus from 1.2 to 3.9% but had a very slight effect on the WT virus (3.5% in DMSO-treated cells and 4.0% in SAHA-treated cells). Taken together, de novo expression of Vpr or treat-ment with SAHA helped the virus counteract latent infections in macrophages.

Vpr Down-regulates Class I HDACs to Overcome Latency in MDMs
To investigate the effects of Vpr-induced depletion of HDACs, we also examined the impact of depletion of class I HDACs on infection of MDMs. Primary MDMs were infected with the single cycle VSV-G pseudotyped GFP-reporter WT HIV-1 or the corresponding ⌬Vpr virus. After 2 days, HDAC1 and HDAC3 were depleted in the infected cells. Four days postinfection, cells were analyzed for GFP expression (Fig. 8 D). While in the cells transfected with the non-targeting siRNA, the ⌬Vpr virus infected 1.06% of the cells, depletion of HDAC1 and HDAC3 increased infectivity of the ⌬Vpr virus to 2.98 and 3.03%, respectively. Interestingly, siRNA transfection of MDMs with non-targeting siRNA resulted in 3.42% infection with the WT virus but siRNA depletion of HDAC1 and HDAC3 did not significantly change infectivity of the WT virus (3.58 and 3.38%, respectively).

Discussion
HIV-1 Vpr has been well documented to enhance proteasomal degradation of a number of proteins by interacting with the Cul4-DDB1[VprBP] E3 ubiquitin ligase (4,27). Nonetheless, degradation of none of these proteins fully explains the effect of Vpr in increasing viral infectivity in macrophages. Here we report novel protein targets for Vpr that explain the previously described functions of this protein. We showed that overexpression of Vpr resulted in depletion of all the four members of class I HDACs with no significant effect on other classes of HDACs. The depletion was more drastic on chromatin and when tested in time course and dose-response experiments, depletion of chromatin-associated HDACs preceded that of soluble HDACs. VprBP and DDB1 were found to bind class I HDACs and we showed that VprBP plays a crucial role in Vprinduced depletion of HDACs. Our results suggest targeted proteasomal depletion of class I HDACs by Vpr through an E3 ubiquitin ligase on chromatin that contains both VprBP and DDB1 (Fig. 9, upper panels). However, even in an overexpressed system, the chromatin depletion of class I HDACs was never complete, suggesting a specific mechanism that does not deplete the entire pool of the chromatin-associated class I HDACs. This suggests that Vpr may target a fraction of chromatin-associated class I HDACs on specific regions of the genome.
In overexpressed systems, HDAC3 showed the highest sensitivity to expression of Vpr by being depleted at lower concentrations of Vpr while HDAC2 showed the lowest sensitivity by being depleted at higher concentrations of Vpr. De novo expression of Vpr in infected MDMs, however, resulted in depletion of HDAC1 and 3 on the HIV-1 LTR. We did not observe a significant depletion of HDAC2 on the HIV-1 LTR of infected MDMs. HDAC8 was not found to localize on HIV-1 LTR and therefore we did not observe its depletion in MDMs. Nonetheless, we cannot rule out the possibility of depletion of class I HDACs in other genomic regions, such as cellular genes, which were not addressed in this study.
It has previously been reported that depletion or inhibition of HDACs, especially HDAC3, reactivates latent HIV-1 infection (28 -32). Interestingly, Vpr has been reported to activate HIV-1 in latently infected cells (33)(34)(35), as well as endogenous retroviruses, LINEs and SINEs (36,37). In this study, we showed that FIGURE 7. Vpr induces hyperacetylation of histones on the HIV-1 LTR in a proteasomal-dependent manner. A and B, MDMs from 4 donors were infected with the GFP reporter WT or ⌬Vpr viruses. After 7 days, chromatin immunoprecipitations were performed against acetylated histone H3 (Ace H3K9) and acetylated histone H4 (Ace H4K5). Quantitative PCR was performed against nuc-0, nuc-1, and nuc-2, gapdh and an intergenic region (INT). C and D, MDMs were infected with the WT and ⌬Vpr viruses. After 7 days, cells were treated with 5 M MG132 or DMSO for 20 h. Using ChIP, cells were then analyzed for occupancy of acetylated histones (Ace H3K9 and Ace H4K5) on nuc-0. E, MDMs were infected with the WT and ⌬Vpr viruses. After 7 days, cells were ChIPed using antibody against total histone H3. The pulled-down chromatin was analyzed using quantitative PCR against nuc-0, nuc-1, nuc-2, and gapdh. FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6

JOURNAL OF BIOLOGICAL CHEMISTRY 2707
Vpr depletes HDAC1 and HDAC3 on the HIV-1 LTR of infected MDMs which was found in line with hyperacetylation of histones on this region. Hyperacetylation of histones is normally associated with active transcription and we showed that expression of Vpr was associated with active transcription and expression of the virus in a VprBP-dependent manner. However, Vpr did not enhance viral expression when equal numbers of WT and ⌬Vpr infected cells were compared. As depicted in Fig. 9 (lower panels), Vpr does not enhance integration of more proviral DNAs in the infected cells. Instead, Vpr counteracts latent infection of macrophages. Expression of Vpr maintains the HIV-1 LTR active, resulting in active expression of the viral genome. This function of Vpr increases the proportion of acutely infected macrophages. In the absence of Vpr, there are more latently infected macrophages and we showed that these cells carry a latent proviral genome integrated in their chromatin (Fig. 8, A and B). As shown in Fig. 8C, the latent infections established in the absence of Vpr, can be reactivated by SAHA to restore a higher level of infectivity that is comparable with the infectivity of the wild-type virus. Interestingly, SAHA did not significantly reactivate viral infection in the wild-type infected macrophages, indicating that there is a lower prevalence of latently infected cells in this population.
We also showed that depletion of HDAC1 and HDAC3 using siRNA induces a similar effect as that of Vpr by activating the virus in the ⌬Vpr-infected MDMs. Taken together, our results suggest that Vpr depletes members of class I HDACs in a dosedependent manner, with HDAC3 as the most affected HDAC, on chromatin of infected cells to maintain an active HIV-1 promoter.
A previous study has shown that components of Sin3a-HDAC1 complex are incorporated into HIV-1 virions. The study also showed that the virion incorporation of HDAC1 plays a role in efficient early post-entry events (38). Since Vpr has already been shown to incorporate one of its target proteins, UNG2, into viral particles (39), it is worth examining whether Vpr plays a role in incorporation of HDACs into the  . Vpr induces ubiquitination of class I HDACs on chromatin that results in a higher proportion of acutely infected macrophages versus latently infected macrophages. Vpr binds chromatin-associated class I HDACs through a VprBP-dependent mechanism and enhances ubiquitination and degradation of the targeted HDACs. This function of Vpr results in hyperacetylation of histones on certain regions including HIV-1 LTR. As a result, the wild-type virus counteracts latent infections and results in more acutely infected cells, despite the fact that the wild-type and ⌬Vpr HIV-1 virions infect equal numbers of macrophages in the first round of infection. There is a higher ratio of acute versus latent infections in the presence of Vpr but the total number of infected macrophages is not affected by Vpr in the first round of infection. FEBRUARY 5, 2016 • VOLUME 291 • NUMBER 6 virions or if there is a link between degradation and virion incorporation of HDACs.

Vpr Down-regulates Class I HDACs to Overcome Latency in MDMs
In addition to the viral genome, Vpr-induced depletion of HDACs has the potential to activate or enhance expression of cellular genes that need to be further investigated. Maintaining an active LTR may be only one of the advantages of downregulation of HDACs. There may be other consequences in down-regulation of HDACs on other genomic regions of the infected cells and even non-infected bystander cells that needs to be elucidated.