A role for the histone deacetylase HDAC4 in the life-cycle of HIV-1-based vectors

HeLa cells were infected with a pseudotyped HIV-1-based vector (containing a lac Z reporter) at an m.o.i. of 0.1 and harvested at the time points indicated (Fig. 1A). Chromatin immunoprecipitation (ChIP) analysis was used to identify the association of HDAC4 with viral DNA. To do so, DNA isolated from infected cells and the associated proteins were crosslinked, immunoprecipitated with the HDAC4 antibody (see Methods), and associated viral DNA was amplified by real time PCR. Results are expressed as a number of viral DNA amplicons per μl of chromatin immunoprecipitates at each time point. As shown in Fig. 1A, viral DNA was found to be associated with HDAC4 at 4, 6, 8, and 16 hrs post-infection. The amount of HDAC4-associated viral DNA steadily increased from 4 hours, with a peak reached at 8 hours post-infection. The associated viral DNA drastically declined at the 16 hour time point. To determine if vector DNA associates with other HDAC proteins, we have immunoprecipitated lysates from infected cells with the HDAC2 and HDAC6 antibodies. Whereas HDAC2 is a Class I HDAC, we note HDAC6 is a class II HDAC and thus structurally closely related to HDAC4. However, as shown in Fig. 1B, we did not observe any association of these HDACs with viral DNA. We thus conclude that HDAC4 shows a distinct preference for association with vector DNA, when compared to other HDACs.

We note that HDAC4 was reported to associate with DNA of the avian sarcoma virus, but this association was detected only post-integration [24]. To test the hypothesis that integration is required for the association of HDAC4 with the DNA of HIV-1-based vectors, we have infected HeLa cells and treated them with the integrase inhibitor 118-D-24. As shown in Fig. 1C, the inhibitor decreases the association of HDAC4 with vector DNA in a dose-dependent manner. However, we note that the inhibitor effect can be seen only at 8 hours post-infection, when the association of DNA with HDAC4 is at its peak. In contrast, association at 4 hours post-infection is resistant to the inhibitor treatment. These data suggest that while integration does stimulate the association of vector DNA with HDAC4, HDAC4 also associates with vector DNA prior to integration, in an integration-independent manner.

HDAC4 was reported to form foci in irradiated cells. These foci were associated with the formation of double-strand DNA breaks [17]. To determine if infection with HIV-1-based vectors induces the formation of HDAC4 foci, we have infected HeLa cells at a high multiplicity of infection (10), fixed infected cells at predetermined time points and stained with the HDAC4 antibody. We observed that in uninfected cells, HDAC4 is present almost exclusively in the cytoplasm (Fig. 2). Similarly, we have observed that HDAC4 is predominantly cytoplasmic at 4 and 6 hours post-infection. However, we also observed the appearance of HDAC4 foci in infected cells, with the majority of cells containing foci at 8 hrs post-infection (Figs. 2 and 3). Most of the infected cells contained multiple HDAC4 foci. As indicated in Fig. 3, the number of foci correlates well with the multiplicity of infection.

We have observed that integration stimulates the association of HDAC4 with vector DNA and wondered if integration affects the formation of HDAC4 foci. Thus, we have infected HeLa cells in the presence and absence of an integrase inhibitor. We have again detected HDAC4 foci in cells that were infected with the HIV-1-based vector in the absence of the inhibitor (Fig. 4). However, treatment of infected cells with the inhibitor significantly reduced (ca. 3.5 fold) the total number of cells that contained foci (Fig. 4 and Fig. 5). We have also observed a drop in the average number of foci per cell among foci-containing cells, although the difference was within the standard deviation due to a wide range of the numbers of foci (1 to 8 foci per cell among cells infected with the vector and 1 to 6 foci per cell in cells infected with the vector and treated with the inhibitor, Fig. 5B). Treatment with the inhibitor itself had a negligible effect on the intracellular localization of HDAC4 (Figs. 4 and 5). Taken together, our results suggest that although HDAC4 associates with viral DNA even prior to integration, integration stimulates further accumulation of HDAC4 at integration sites, which are then marked by the formation of HDAC4 foci.

We have established a novel interaction between the cellular HDAC4 protein and HIV-1-based vectors. Our results suggested that HDAC4 plays a role in the life-cycle of these vectors. To test this hypothesis, we have knocked down HDAC4 in HeLa cells using siRNA treatment and determined if HDAC4 is required for stable integration of HIV-1- vector DNA. As shown in Fig. 6, HDAC4 had little effect on the efficiency of integration as measured by Alu-PCR. In addition, we have infected siRNA-treated cells with the HIV-1-based vector carrying an EGFP marker and examined EGFP expression using flow cytometry. We have not observed a significant drop in EGFP expression in HDAC4 siRNA-treated cells (data not shown). We conclude that HDAC4 deficiency does not appear to significantly affect the efficiency of integration. Similarly, it appears that HDAC4 is not necessary for the last step of the integration process, termed post-integration repair (PIR), since PIR failure results in a loss of cells in which integrase-mediated joining occurred, and thus again manifests as a decrease in the Alu-PCR signal [25, 26].

HDAC4 is a DSB repair protein, and it had been reported by us and others that these proteins are involved in PIR [23]. However, cellular DSB repair proteins often have overlapping functions and DSB repair systems can partially substitute for each other [27]. It is thus possible that a loss of the HDAC4 protein can be compensated for by other DSB repair systems or proteins. To test this hypothesis, we have induced a DSB repair deficiency in HeLa cells by treatment with an established ATM inhibitor, KU-55933 [28]. The ATM protein is a major player in cellular DSB repair and was reported by us and others to be involved in PIR [26‐28]. At the same time, in these cells, we have overexpressed the HDAC4 protein. Since the normal HDAC4 protein (denoted here as HDAC4-1084) is mainly cytoplasmic, we have also overexpressed a mutant, which lacks a nuclear export signal (HDAC4-1061) and is thus present in the nucleus (Fig. 7, [29]). As expected, the ATM inhibitor reduced the Alu-PCR signal due to the inhibition of PIR (Fig. 7A). We have observed that in ATM-proficient HeLa cells, the overexpressed HDAC4 proteins do not appear to affect the efficiency of integration or PIR, as shown by Alu-PCR (Fig. 7A). However, in HeLa cells that were treated with the ATM inhibitor, the HDAC4-1084 reverses the inhibitor effect and upregulates HIV-1 transduction four fold. The HDAC4-1061 mutant that is constitutively present in the nucleus completely reverses the effect of the ATM inhibitor (Fig. 7A). To investigate the possibility that the differences in the Alu-PCR signals could be due to variations of the exogenous HDAC4 expression levels, we performed a western bloting analysis (Fig. 7B). However HDAC4-1061 and HDAC4-1084 levels appear to be the same in our transfected cells. Taken together our results suggest that HDAC4 is involved in PIR, but its function can be replaced by other DSB protein(s).

Finally, a failure of PIR induces apoptotic death of infected cells. If the effect of nuclear HDAC4 on infection efficiency is due its role in PIR, it should prevent PIR-associated cell death. Thus, cells were treated and infected as above (Fig. 7), except at a high m.o.i. (2), and analyzed by Western blotting for the presence of the 85-kDa PARP fragment, an apoptotic marker generated by caspase-mediated cleavage of the PARP protein [30]. As shown in Fig. 8, ATM inhibition and infection stimulated PARP cleavage. However, the apoptosis was reduced by overexpression of either the full length HDAC4 (HDAC4-1084) or the truncated mutant (HDAC4-1061). This finding is again consistent with an HDAC4 role in PIR.

HeLa cells were maintained in DMEM medium supplemented with 10% fetal bovine serum and antibiotics (Penn/Strep).

All VSV G-pseudotyped HIV-1 based vectors were prepared as described previously [33, 34], and carried either a lac Z or EGFP reporter gene.

Plasmids expressing the full-length HDAC4 (amino acids 1-1084) fused to the EGFP protein (HDAC4-1084) or the HDAC4 C-terminal truncated mutant, lacking the nuclear export signal (amino acids 1-1061) fused to the EGFP protein (HDAC4-1061) have been described, and were a generous gift from Dr. X. J. Yang of McGill University [12, 29]. The PEGFP-C1 control plasmid expressing the EGFP protein under control of the CMV promoter was purchased from Clontech (GenBank Accession # U55763). Plasmids were transfected into HeLa cells using the Lipofectamine™ 2000 transfection reagent (Invitrogen, cat # 11668-027) using company protocols. Cells were infected with the HIV-1-based vector two days post-transfection.

HeLa cells were infected at a multiplicity of infection (m.o.i.) 0.1 for the indicated time intervals. In some cases, an integrase inhibitor was added at the time of infection (118-D-24, NIH AIDS Reagent Program). Cells were then harvested and ChIP was performed as described [35], with the anti-HDAC4 antibody (1 μg/sample, Santa Cruz Biotechnology, cat # sc-11418X) or anti-HDAC2 antibody (1 μg/sample, Abcam, cat # ab16032) or anti-HDAC6 antibody (1 μg/sample, Santa Cruz, cat # sc-11420). Protein-associated vector DNA was detected by real-time PCR, using primers and probes detecting HIV-1 LTR. Forward primer: 5'-TGTGTGCCCGTCTGTTGTGT-3'; Reverse primer: 5'-CCTGCGTCGAGAGAGCTC-3'. To quantitate the viral amplicon, a TaqMan dual 5'-6-carboxyfluorescein-and 3'-6-carboxytetramethylrhodamimine-labeled probe was used: 5'-(FAM)-CAGTGGCGCCCGAACAGGGA-(TAMRA)-3' (Integrated DNA Technologies). Real-time PCR was performed using a LightCycler 1.5 with software 3.5.3 (Roche). Reaction mixtures contained QuantiFast Probe 2× mix (Qiagen), 100 nM probe, and 200 nM primers. The standard cycling conditions were 95°C - 3 min followed by 50 cycles at 95°C - 3 s and 60°C - 30 s. Samples were run in triplicate.

HeLa cells were plated at a density of 2 × 10⁴ and grown on 4-well chamber slides. The following day, the cells were infected with the HIV-based vector at m.o.i. 10 for a time course study at 4, 6, and 8 hours. In another experiment, vector and an integrase inhibitor (118-D-24, final concentration of 100 μM) or the vector only had been incubated for 8 hours prior to fixation. At the indicated time points, cells were washed in PBS and fixed by adding cold methanol-acetone (1:1 volume) at room temperature for 2 min. The slides were incubated with the primary antibody in KB buffer overnight at 4°C. As a control, we used samples incubated in KB buffer with no primary antibody. The primary antibody was the rabbit polyclonal anti-HDAC4 (see above), diluted 1:500 in KB buffer. The secondary antibody, Alexa Fluor 488 donkey anti-rabbit (Invitrogen, cat # A21206) was used at a 1:1000 dilution. Cells were then washed with PBS containing 0.1% Triton. Cells were incubated in the secondary antibodies for 1 hr at room temperature. Cells were then stained directly with 4',6-diamidino-2-phenylindole, dihydrochloride (DAPI) (Invitrogen, cat # D1306) for 5 min at room temperature. The stained cells were washed with KB buffer and mounted with prolong gold anti-fade (Invitrogen, cat # P36930). Images of stained cells were taken using a Nikon Eclipse TE-2000 S with fluorescence optics at an objective magnification of 20×.

Random images of HeLa cells stained as described above were taken using Nikon Eclipse TE-2000 S at a magnification of 20×. All cells were then counted on a randomly selected slide, both to determine the number of cells containing foci and number of foci per cell, if the cell contained foci. This had been performed in duplicate, each time on a different slide.

To detect and quantify fully integrated proviral DNA, a two-step nested Alu-PCR technique was conducted. Cells were infected with the HIV-1-based vector at m.o.i. 0.1. Three days post-infection genomic DNA was extracted (Qiagen, cat # 51306). The first round of Alu-PCR employed a primer targeting the cellular Alu sequence 5' - GCCTCCCAAAGTGCTGGGATTACAG - 3' as well as the primer targeting the HIV-1 LTR/gag region, 5' - TTTTGGCGTACTCACCAGTCG - 3'. This initial amplification step used 100 ng of genomic DNA as template. Samples were subjected to 30 PCR cycles of 95°C - 30 s, 60°C - 45 s, and 72°C - 5 min, and after the final round, samples were kept at 72°C for 10 min. Products of the first round (4 μl of the 50 μl first round reaction) were used in the second, real-time PCR reaction as described above (see ChIP experiments).

A pool of siRNAs targeting HDAC4 (cat # M-003497-03) and a pool of non-targeting, control siRNA (cat # D-001206-14-05) were obtained from Dharmacon. A day after plating 10⁵ HeLa cells per 60 mm dish, cells were transfected with siRNA using Lipofectamine™ RNAiMAX Transfection Reagent (Invitrogen, cat # 13778-075) according to the manufacturer's protocol. The following day, medium was replaced, and cells were transfected again the same way. The next day (three days after cells were plated) cells were infected and assayed for integration, see Alu-PCR methods above. HDAC4 levels were measured three days after plating by western blotting with an anti-HDAC4 antibody (cat # sc-11418, Santa Cruz Biotechnology).

HeLa cells (transfected with either a control C1, HDAC4-1061 or HDAC4-1084 plasmid two days prior to infection, see above) were infected at mo.i. 2. KU-55933 (Calbiochem, cat # 118500-2 MG) was added at the time of infection to a final concentration of 10 μM. One day post-infection, cells were harvested, lysed and cell lysates subjected to western blotting with an anti-PARP antibody (sc-7150, Santa Cruz Biotechnology).