HIV-1 Vpr Induces DNA Double-Strand Breaks

Recent observations imply that HIV-1 infection induces chromosomal DNA damage responses. However, the precise molecular mechanism and biological relevance are not fully understood. Here, we report that HIV-1 infection causes double-strand breaks in chromosomal DNA. We further found that Vpr, an accessory gene product of HIV-1, is a major factor responsible for HIV-1–induced double-strand breaks. The purified Vpr protein promotes double-strand breaks when incubated with isolated nuclei, although it does not exhibit endonuclease activity in vitro. A carboxyl-terminally truncated Vpr mutant that is defective in DNA-binding activity is less capable of Vpr-dependent double-strand break formation in isolated nuclei. The data suggest that double-strand breaks induced by Vpr depend on its DNA-binding activity and that Vpr may recruit unknown nuclear factor(s) with positive endonuclease activity to chromosomal DNA. This is the first direct evidence that Vpr induces double-strand breaks in HIV1–infected cells. We discuss the possible roles of Vpr-induced DNA damage in HIV-1 infection and the involvement of Vpr in further acquired immunodeficiency syndrome–related tumor development. (Cancer Res 2006; 66(2): 627-31)


Introduction
A high incidence of malignant tumors, such as non-Hodgkin's lymphoma, Kaposi's sarcoma, and invasive cervical cancer [acquired immunodeficiency syndrome (AIDS)-defining cancers], is epidemiologically associated with HIV-1 infection (1,2).These neoplasms are attributable mainly to diseases that accompany immunodeficiency, including coinfection with EBV, human herpes virus 8, and human papillomavirus (1,2).In addition to these AIDS-defining cancers, several non-AIDS-defining cancers also occur with a higher incidence in HIV-infected individuals (3,4).These reports lead to the assumption that HIV-1 has the potential to induce neoplasms before AIDS develops.Recently, DNA damage responses have been observed in precancerous lesion before inactivation of p53 (5,6).Interestingly, it has been reported that HIV-1 infection induces DNA damage responses by activating Rad3-related or ataxia-telangiectasia mutated proteins and pro-moting phosphorylation of their downstream substrates (7,8).The elucidation of the factor triggering the DNA damage responses to HIV-1 infection is essential to determine the as yet unknown mechanism causing AIDS-related neoplasms.In the present study, we found that HIV-1 infection induces double-strand breaks of chromosomal DNA, as detected using pulsed-field gel electrophoresis (PFGE).We further showed that vpr, an accessory gene of HIV-1 encoding a virion-associated nuclear protein, which induces cell cycle accumulation at G 2 -M phase and increases ploidy (9), was a factor responsible for double-strand breaks.We discuss the potential ability of Vpr-induced double-strand breaks to develop into neoplasms in HIV-1 infection.

Materials and Methods
Cell culture.MIT-23 and DVpr, a mock transfectant, were established from HT1080 (JCRB9113; the Health Science Research Resources Bank) as previously described (9).In MIT-23, Vpr expression is controlled by the rtet promoter on incubation with 3 Ag/mL doxycycline (Sigma, St. Louis, MO) for 48 hours.
Virus infection.Vesicular stomatitis virus G protein (VSV-G)-pseudotyped HIV-1 was produced by cotransfection with a plasmid encoding VSV-G (pHIT/G) and the pNL-Luc-E À R + or pNL-Luc-E À R À proviral clone (10).(10).The preparation and titration of viruses are described elsewhere (11).Briefly, the concentration of p24 antigen in the culture supernatant was measured using a p24 Gag antigen capture ELISA kit (ZeptoMetrix, Buffalo, NY).The infectivity of the prepared viral stock was examined using MAGIC5 cells.HT1080 cells were infected for 48 hours with viruses that had 200 ng/mL of p24 Gag antigen, giving a multiplicity of infection (MOI) of 0.7.
Immunostaining.Immunostaining was carried out as described (9).A rabbit polyclonal Rad51 antibody raised against the bacterially expressed protein and a mouse monoclonal antibody raised against synthesized peptides of full-length of Vpr (mAb8D1) were used as the primary antibody.Goat anti-rabbit IgG conjugated with Alexa Fluor 488 (Molecular Probes, Inc., Eugene, OR) and goat anti-mouse IgG conjugated with Cy3 (Zymed Laboratories, Inc., San Francisco, CA) were used as the secondary antibodies.Images were captured on a phase contrast microscope, BX50 (Olympus Corp., Tokyo Japan), or a Radiance 2100 laser scanning confocal microscope (Carl Zeiss, Oberkochen, Germany).
Overexpression and purification of Vpr and its mutant.The HIV-1 vpr gene was ligated into the NdeI and BamHI sites of the pET15b vector (Novagen, Madison, WI).The Vpr protein and VprDC12 mutant were produced in the Escherichia coli BL21 (DE3) Codon(+)RIL strain (Novagen) by induction with isopropyl-h-D-thiogalactopyranoside (IPTG; Nacalai Tesque, Inc., Kyoto, Japan) and were purified as described in Supplementary Method.The concentration of the purified Vpr protein was determined with a Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA) using bovine serum albumin (BSA) as the standard.
Isolation of nuclei.Cells scraped from culture dishes were washed once with ice-cold PBS and resuspended in 3 mL of ice-cold 20 mmol/L Tris-HCl buffer (pH 7.6) containing 60 mmol/L KCl, 15 mmol/L NaCl, 5 mmol/L MgCl 2 , 1 mmol/L DTT, 250 mmol/L sucrose, 0.6% NP40, and protease inhibitor mixture (Sigma).The cell suspension was incubated for 10 minutes on ice and the sucrose concentration was adjusted to 1.6 mol/L.Then, the sample was loaded onto a sucrose cushion of 2.3 mol/L sucrose solution and centrifuged at 35,000 Â g for 30 minutes.The isolated nuclei were obtained in the 2.3 mol/L sucrose fraction.For immunostaining, isolated nuclei were cytocentrifuged to the MAS-coated slide glass (Matsunami Glass IND., LTD., Tokyo, Japan) for 6 minutes at 800 rpm (Thermo Shandon, Chadwick Road, United Kingdom).
PFGE assay.Isolated nuclei were incubated with 10 Amol/L of purified Vpr or VprDC12 for 15 hours at 30jC.The cells (isolated nuclei) were embedded in agarose plugs at a density of 3 Â 10 5 cells/100 AL.The plugs were treated with proteinase K solution [0.5 mol/L EDTA (pH 8.0), 1% sarcosyl, and 0.5 mg/mL proteinase K] for 38 hours at 50jC.After PFGE was done in a CHEFF Mapper (Bio-Rad Laboratories), the gels were stained with Vistra Green (Amersham Bioscience, Piscataway, NJ).
The DNA-binding assay.The Vpr protein was incubated with $X174 single-stranded DNA (ssDNA; 20 Amol/L) or $X174 superhelical dsDNA (10 Amol/L) in 10 AL of 8 mmol/L Tris-HCl buffer (pH 8.5) containing 1 mmol/L DTT and 100 Ag/mL BSA.The reaction mixtures were incubated for 1 hour at 37jC and were analyzed by electrophoresis on a 0.8% agarose gel in 1Â TAE buffer (40 mmol/L Tris acetate and 1 mmol/L EDTA) at 3.3 V/cm for 2 hours.The bands were visualized using ethidium bromide staining.
Nuclease activity.The Vpr protein (18.8 Amol/L) or DNaseI (Invitrogen Corporation, Carlsbad, CA; 0.02 unit/AL) were incubated with $X174 superhelical double-stranded DNA (dsDNA; 2.5 Amol/L) in 40 AL of 15 mmol/L Tris-HCl buffer (pH 8.5) containing 1 mmol/L DTT and 100 Ag/mL BSA, in the presence of 5 mmol/L MgCl 2 , MnCl 2 , ZnSO 4 , or CaCl 2 .The reaction mixtures were incubated at 37jC for 30 minutes.After incubation, the samples were treated with proteinase K (0.3 mg/mL) in the presence of 0.1% SDS and the DNA was extracted using phenol-chloroform.The DNA was precipitated by ethanol and was analyzed by electrophoresis on a 0.8% agarose gel in 1 Â TAE buffer at 6.6 V/cm for 30 minutes.The bands were visualized with ethidium bromide staining.

Vpr expression induces chromosomal double-strand breaks.
To test whether HIV-1 infection causes double-strand breaks, we used PFGE, which was able to clearly detect the double-strand breaks induced by X-ray irradiation (Fig. 1A, lane 2; ref. 12).HT1080 cells were infected with HIV-1 that had 200 ng/mL of p24 Gag antigen, giving a MOI of 0.7, and the cellular DNA was fractionated using PFGE.Figure 1A (lane 6) shows that HIV-1 infection induced double-strand breaks.Interestingly, the amount of HIV-1-dependent double-strand breaks was reduced significantly (Fig. 1A, lane 5) when the vpr gene was deleted from the HIV-1 viral genome (HIV-1DVpr).To show that HIV-1-dependent double-strand breaks are attributable to Vpr expression, we examined double-strand break formation in Vpr stable transfectant, MIT-23 (9), in which Vpr expression is controlled by the rtet promoter by doxycycline, and, in DVpr, a mock transfectant.As shown in Fig. 1B, double-strand breaks were observed in the Vpr-expressing cells (lane 5, arrow) but not in the mock transfectants (lane 4).Furthermore, Rad51 foci, which are formed at double-strand break sites (13), were observed with Vpr expression (Fig. 1C).These results indicate that Vpr is responsible for double-strand break formation.The double-strand breaks shown in Fig. 1B were not the result of an apoptotic process as the DNA ladder typically observed in apoptotic cells (14) was not detected (data not shown).
Vpr has no endonuclease activity.Next, we studied whether Vpr directly induces double-strand breaks.The recombinant Vpr protein was purified to near homogeneity (Fig. 2A) and the DNAbinding activity of Vpr was examined.As shown in Fig. 2B, purified Vpr bound both ssDNA (lanes 2-6) and dsDNA (lanes 8-12) in an ATP-and Mg 2+ -independent manner (15).Then, we examined whether Vpr has nuclease activity.Superhelical dsDNA containing small amounts of nicked circular dsDNA was incubated with Vpr in the presence of various divalent cations.After the incubation, the proteins were removed and the DNA was examined by electrophoresis.If Vpr induces a double-strand break or nick, the superhelical dsDNA would give rise to linear or nicked circular forms, producing a different electrophoretic pattern.However, the DNA incubated with Vpr in the absence (lane 2) or presence of any divalent cation examined (lanes 4, 6, 8, and 10) showed the same migration pattern with control (lane 1), indicating that Vpr does not cleave DNA (Fig. 2C).Positive control experiments showed that the DNA was digested by DNaseI with MgCl 2 , MnCl 2 , or CaCl 2 (lanes 5, 7, and 11) but not with ZnSO 4 (lane 9; Fig. 2C).Therefore, these results indicate that Vpr lacks endonuclease or nicking activity.
Vpr induces double-strand breaks in vitro.In a second approach, we tested whether purified Vpr induces doublestrand breaks in nuclei isolated from HT1080 cells (Fig. 3A).First, we confirmed by a laser confocal microscopy that Vpr localizes in nuclei after incubation in vitro (Fig. 3B).The nuclear DNA was then analyzed for double-strand breaks by using PFGE (Fig. 3C).Interestingly, purified Vpr induced doublestrand breaks in the DNA of the isolated nuclei (Fig. 3C, lane 5, arrow).By contrast, few double-strand breaks were detected without Vpr (Fig. 3C, lane 4).Because Vpr alone did not show endonuclease activity (Fig. 2C), these results suggest that Vpr interacts with intrinsic nuclear protein(s), which required for double-strand break formation.To identify candidates for the Vprinteracting nuclear proteins, we did the Ni-NTA pull-down assay.In this assay, recombinant His 6 -tagged Vpr was incubated with the extract from isolated nuclei and Ni-NTA beads precipitated proteins bound to His 6 -tagged Vpr (Fig. 3D).As shown in Fig. 3D, His 6 -tagged Vpr associated with numerous proteins that were not detected in the control precipitates (lane 2, asterisks).
The DNA-binding activity of Vpr is correlated with doublestrand break formation.The COOH-terminal region of Vpr is arginine rich and is thought to be an important site for DNA binding to Vpr (15).Nuclear magnetic resonance analysis shows that Vpr has three a-helixes (amino acids 17-33, 38-50, and 56-77) in solution, whereas the COOH-terminal region from amino acid residues 84 to 96 is disordered (16).This suggests that the deletion of the COOH-terminal 12 amino acid residues does not affect the tertiary structure of Vpr.We purified a Vpr mutant protein lacking the COOH-terminal 12-amino-acid residues (VprDC12; Fig. 4A), and examined its DNA-binding activity.Purified VprDC12 was significantly defective in both ssDNA-and dsDNA-binding activity compared with wild-type Vpr (Fig. 4B).Interestingly, VprDC12 induced double-strand breaks in isolated nuclei but its efficiency was reduced significantly (Fig. 4C, lane 6).These results indicate that the DNA-binding ability of Vpr is important for the induction of double-strand breaks by Vpr.

Discussion
Here, we present evidence that HIV-1 Vpr induces double-strand breaks.Our data are consistent with previous observations in Vpr-expressing cells: the up-regulation of gene amplification events that are believed to be introduced by broken DNA strands (17) and the activation of activating Rad3-related/ataxia-telangiectasia mutated, followed by the phosphorylation of their downstream substrate, a histone H2A variant, H2AX, and g-H2AX and BRCA1 focus formation (8). Biochemical analyses using purified Vpr indicated that Vpr alone has no endonuclease activity (Fig. 2C), suggesting that a cellular factor(s), possibly with endonuclease activity, is required for Vpr-dependent double-strand breaks.The factor(s) required for double-strand breaks must preexist in nuclei because double-strand breaks were observed upon incubating a mixture of isolated nuclei and purified Vpr in vitro (Fig. 3C).As one possible mechanism, Vpr may recruit a nuclease factor to chromosomal DNA, given that the Vpr-dependent double-strand breaks were correlated with the DNA-binding activity (Figs.4B  and C).Alternatively, Vpr itself may acquire endonuclease activity after modification in the nucleus.Further analyses are necessary to clarify this point.In the HIV-1 life cycle, DNA breakage and repair are thought to be essential steps for integrating the double-stranded viral cDNA into the host genome.In this study, we found that Vpr is one molecule responsible for the double-strand breaks that occur upon HIV-1 infection.However, it is also noteworthy that some double-strand breaks were induced in the cells with HIV-1DVpr (Fig. 1A, lane 5), suggesting that other viral factors are also involved.It has been shown that integrase activates the ataxiatelangiectasia mutated-dependent pathway (7) and, thus, the double-strand breaks observed with HIV-1DVpr infection are probably owing to integrase.For viral integration to occur, the amount of double-strand breaks induced by HIV-1DVpr (Fig. 1A, lane 5) may be sufficient, because viral production in peripheral blood mononuclear cells was not alleviated by infection with Vpr-deleted HIV-1 (18). 4 Vpr-induced double-strand breaks may be surplus to those required for viral integration (Fig. 1A, lane 6).The resultant DNA damage may reduce the integrity of the host genome.
Recently, DNA damage signaling was observed at an early stage of tumor development, suggesting that the DNA damage response is a mechanism to prevent the progression of preneoplasic lesions (5).If DNA repair is not accomplished correctly or is skipped because of unregulated checkpoint controls, the genomic structure would be altered severely (19).The progression of malignant tumors in AIDS-defining cancers is well documented in oncovirus infections (1,2).If DNA damage increases the probability of neoplasia, Vpr-induced double-strand breaks with oncovirus infection may accelerate tumor progression during the clinical course of AIDS.In addition to AIDS-defining cancers, non-AIDS-defining cancers also occur at a higher incidence and the factor responsible for such oncogenesis is now a critical issue (3,4).Vpr-induced DNA damage may result in  these AIDS-related malignancies.It is essential to explore the molecular mechanism of Vpr-induced double-strand breaks to clarify their role in HIV-1 infection and their effect on the stability of the host cell genome.

Figure 1 .
Figure 1.Vpr induces double-strand breaks in vivo.A, PFGE analysis of double-strand breaks after HIV-1 infection.HT1080 cells were infected with the same amount of HIV-1 or HIV-1DVpr (MOI = 0.7) and subjected to PFGE.As a positive control, uninfected cells were analyzed immediately after 7.5 Gy of X-ray irradiation.Molecular mass markers (lane 1), control cells (lanes 3 and 4 ), cells subjected to X-ray irradiation (lane 2 ), and cells infected with HIV-1DVpr (lane 5 ) or HIV-1 (lane 6) are shown.Arrow, position corresponding to the double-strand breaks.B, PFGE analysis in Vpr-expressing cells.Molecular mass markers (lane 1 ), cells irradiated with 7.5 Gy (lane 2), control cells (lane 3 ), mock transfectants (lane 4), and cells with Vpr expression (lane 5) are shown.Arrow, double-strand breaks.C, Rad51 focus formation with Vpr expression.An immunohistochemical analysis was used to detect Rad51 in cells with (right) or without (left ) Vpr expression.Bar, 10 Am.

Figure 3 .
Figure 3. Purified Vpr induces double-strand breaks in vitro.A, a scheme of the protocol used to detect Vpr-induced double-strand breaks in isolated nuclei.B, Vpr localization in isolated nuclei.Isolated nuclei from HT1080 after incubation with Vpr were immunostained by a-Vpr (mAb8D1) and the images were captured by a laser confocal microscopy.The Z-series of optical sections collected at 1 Am steps of the cells were presented.Vpr (red; middle ), DNA staining by Hoechst (blue; top) and their merged images (bottom ) are shown.Without Vpr incubation, any signals by a-Vpr immunostaining were not detected in isolated nuclei (data not shown).Bar, 10 Am.C, PFGE analysis of double-strand breaks in isolated nuclei treated with Vpr.Molecular mass markers (lane 1), control cells (lane 2), cells subjected to X-ray irradiation (lane 3 ), and isolated nuclei without (lane 4 ) or with 10 Amol/L Vpr (lane 5).Arrow, double-strand breaks.D, Ni-NTA pull-down assay with His 6 -tagged Vpr on isolated nuclei.Precipitated proteins bound to His 6 -tagged Vpr (lane 2 ) and the control precipitates (lane 1 ) are indicated.*, His 6 -Vpr-specific bands. 4M. Shimura, unpublished data.