The cellular antiviral protein APOBEC3G interacts with HIV-1 reverse transcriptase and inhibits its function during viral replication

Abstract

The cytidine deaminase APOBEC3G (A3G) exerts a multifaceted antiviral effect against HIV-1 infection. First, A3G was shown to be able to terminate HIV infection by deaminating the cytosine residues to uracil in the minus strand of the viral DNA during reverse transcription. Also, a number of studies have indicated that A3G inhibits HIV-1 reverse transcription by a non-editing-mediated mechanism. However, the mechanism by which A3G directly disrupts HIV-1 reverse transcription is not fully understood. In the present study, by using a cell-based coimmunoprecipitation (Co-IP) assay, we detected the direct interaction between A3G and HIV-1 reverse transcriptase (RT) in produced viruses and in the cotransfected cells. The data also suggested that their interaction did not require viral genomic RNA bridging or other viral proteins. Additionally, a deletion analysis showed that the RT-binding region in A3G was located between amino acids 65 and 132. Overexpression of the RT-binding polypeptide A3G(65-132) was able to disrupt the interaction between wild-type A3G and RT, which consequently attenuated the anti-HIV effect of A3G on reverse transcription. Overall, this paper provides evidence for the physical and functional interaction between A3G and HIV-1 RT and demonstrates that this interaction plays an important role in the action of A3G against HIV-1 reverse transcription.

Fig 1

Inhibitory effect of A3G on HIV-1 reverse transcription. (A) Vif⁻ HIV-1 virions were produced from 293T cells in the presence or absence of A3G. Virus-associated proteins, including HIV-1 RT, Gag p24, and incorporated HA-A3G were analyzed by WB with the corresponding antibodies (left panels). Also, equal amounts of viruses (normalized by amounts of HIV-1 Gag p24) were used to infect CD4⁺ C8166 T cells. Twelve hours postinfection, total DNA was extracted from the infected cells. HIV-1 late reverse transcription (LRT) products were analyzed by real-time PCR (right panels). The level of LRT in HIV-1 Vif⁻ virus-infected cells was set to 100%, with errors bars representing the standard deviations from three independent experiments. (B) Schematic structure of the RT/IN-deleted HIV-1 provirus HxBruΔRI-Vif⁻ and the Vpr-RT-IN and Vpr-RT expression plasmids, which have been previously described (5, 7, 65). PROM, promoter; PCS, protease cleavage site. (C) WB analysis of the RT/IN trans-complemented HxBruΔRI-Vif⁻ viruses, which were produced from cotransfected 293T cells with HxBruΔRI-Vif⁻ and Vpr-RT-IN or Vpr-RT in the presence or absence of A3G. Briefly, viruses were collected, lysed, and loaded into SDS-PAGE gels. The presence of viral proteins and HA-A3G was analyzed by WB with corresponding antibodies. (D) Real-time PCR analysis of the level of LRT in C8166 T cells infected with equal amounts of the trans-complemented HxBruΔRI viruses in the presence or absence of A3G. The level of LRT was measured at 12 h postinfection by real-time PCR as in panel A. The values presented are the means and standard deviations from three independent experiments. Panels A and D show a representative WB image from three independent experiments.

Fig 2

A3G interacts with HIV-1 reverse transcriptase in the virus. (A) To test the A3G-RT interaction in the virus, the HIV-1 Vif⁻ provirus was cotransfected with the pAS1B-HA or pAS1B-HA-A3G expression plasmid in 293T cells. Seventy-two hours posttransfection, viruses were purified from the culture supernatants. The viral lysates were immunoprecipitated with an anti-HA antibody followed by WB with anti-RT or anti-HA antibodies (lanes 1 and 2). The expression of virus-associated RT and HA-A3G was also detected by directly loading viral lysates into SDS-PAGE gels followed by WB with the corresponding antibodies (lanes 3 and 4). A nonspecific band below RT p51 was found in each testing sample. (B) The A3G-RT interaction was not mediated by the RNA bridge. HIV-1 Vif⁻ viruses produced from 293T cells in the presence or absence of HA-A3G were either untreated (lanes 1 and 3) or treated with RNase (lane 2) and then lysed. After lysis, 50% of viral samples were subjected to immunoprecipitation with anti-HA antibody. The precipitated samples (left upper panels) and other 25% of viral samples (left lower panels) were then loaded into SDS-PAGE gels followed by WB with anti-RT, anti-IN, and anti-HA antibodies, respectively. Meanwhile, the remaining 25% of viral lysates were subjected to viral RNA isolation followed by reverse transcription. The reverse-transcribed DNA was detected by PCR and analyzed by agarose gel electrophoresis (right panel). NC stands for negative control. (C) To test the effect of A3G on HIV-1 RT processing, the Vif⁻ viruses produced in 293T cells at the various expression levels of HA-A3G were collected, lysed, and loaded into SDS-PAGE gels. The RT, Gag p24, and HA-A3G expression levels were analyzed by WB using anti-RT (top panel), anti-p24 (middle panel), and anti-HA (lower panel) antibodies. The total amounts of RT p66 and p51 from two independent experiments were measured by imaging analysis. The p66/p51 ratios in the viral lysates are shown at the bottom of the RT panel. (D) The Vif⁻ viruses produced in A3G⁻ C8166 cells (lanes 1 and 3) or A3G-expressing C8166 T cells (lanes 2 and 4) after 2 to 3 days of infection were collected, lysed, and loaded into SDS-PAGE gels. The levels of RT and Gag p24 were detected by anti-RT antibody (top panel) or anti-p24 antibody (lower panel). The p66/p51 ratios in lysates from two independent experiments are shown at the bottom of the RT panel. Panels A and B show representative WB results from three independent experiments.

Fig 3

A3G-RT interaction in the absence of other viral proteins. (A) Schematic structure of the T7-tagged RT, HA-tagged, and ProLabel-tagged A3G expression plasmids. ProLabel is a fragment of the split β-galactosidase, which has no enzymatic activity. However, the ProLabel-tagged fusion protein can recombine with a Ω fragment of β-galactosidase to reconstitute an active enzyme which is able to cleave chemiluminescence substrate. (B) To test RT/A3G interaction in the absence of other viral proteins, the pAS1B-HA-A3G plasmid was cotransfected with SVCMVin-T7 or SVCMVin-T7-RT expression plasmid into 293T cells. After 48 h, cells were lysed, and 10% of them (input) were analyzed by WB using anti-HA or anti-T7 antibodies to detect HA-A3G and T7-RT (lower two panels). Meanwhile, the remaining cell lysates were immunoprecipitated by the anti-T7 antibody to pull down T7-RT protein, and the coprecipitated HA-A3G was detected by WB with anti-HA antibody (upper panel). (C) The A3G-RT interaction was also detected by a chemiluminescence-based enzyme complementation (ProLabel [PL]) assay. Briefly, T7-RT and ProLabel-A3G (PL-A3G) were coexpressed in 293T cells. After 48 h, cells were lysed and immunoprecipitated by an anti-T7 antibody. The Co-IP samples (left panel) and cell lysates (right panel) were analyzed by the ProLabel activity assay. Values are averages from three independent experiments, and error bars indicate standard deviations from the means. Statistical significance was calculated using Student's t test, and P values are shown above the bars. RLU, relative light units.

Fig 4

Mapping the binding domains in both RT and A3G required for their interaction. (A) Schematic representation of the different T7-tagged RT wt/mutant constructs used in the domain-mapping experiments. Full-length T7-RT is shown at the top. The numbers indicate the amino acid positions in RT. The domains are marked in the full-length RT. Different RT truncated mutants are shown at the bottom. (B) Interaction of HA-A3G with T7-RT wt/mut. HA-A3G was cotransfected with T7-RTwt/mut in 293T cells for 48 h. The association of A3G with truncated RT was analyzed by Co-IP with an anti-T7 antibody followed by WB with an anti-HA antibody (upper panel). The immunoprecipitated T7-RTwt/mut was detected by an anti-T7 antibody (middle panel). The HA-A3G expression levels in the cell lysates were evaluated by WB with an anti-HA antibody (lower panel). (C) Schematic representation of the different HA-A3G wt/mutant constructs used in the domain-mapping experiments. Full-length HA-A3G is shown at the top. The numbers indicate the amino acid positions in A3G. The domains marked are the cytidine deaminase domain (CDD), the linker (LINK), and the pseudoactive site (PAS). Different A3G truncated mutants are shown at the bottom. (D) Interaction of T7-RT with HA-A3G wt/mut was detected in 293T cells cotransfected with SVCMVin-T7-RT plasmid and each of the A3G mutants. After 48 h of transfection, the interaction of each A3G truncated mutant with RT was analyzed by Co-IP with anti-T7 antibody followed by WB with an anti-HA antibody (upper panel). The immunoprecipitated T7-RT was detected by anti-T7 antibody (middle panel). The expression levels of HA-A3Gwt/mutant in the cell lysates were detected with an anti-HA antibody (lower panel). The images represent the results of three independent experiments.

Fig 5

A3G_65-132 both disrupts the A3G-RT interaction and attenuates A3G's inhibitory effect on HIV-1 reverse transcription. (A) A3G_65-132 blocks the binding of A3Gwt with HIV-1 RT, HA-A3Gwt, SVCMVin-T7-RT, and each HA-A3G truncated mutant plasmid was cotransfected into 293T cells for 48 h. The interaction between HA-A3Gwt and T7-RT was analyzed by Co-IP with anti-T7 antibody followed by WB with an anti-HA antibody (upper panel). The immunoprecipitated T7-RT was visualized by WB with anti-T7 antibody (second panel). The expression of HA-A3Gwt in the cell lysates was detected with anti-A3G antibody (third panel). HA-A3G mutants were detected with an anti-HA antibody (lower panel). (B) To evaluate the virus incorporation of A3G in the presence of HA-A3G_65-132, the HxBru-Vif⁻ provirus was cotransfected with HA-A3Gwt or HA-A3G_65-132 alone or with the HA-A3Gwt and HA-A3G_65-132 expression plasmids. After 48 h of cotransfection, virions were collected by ultracentrifugation, lysed, and loaded on SDS-PAGE gels. The presence of A3G and viral protein Gag p24 in each viral sample was analyzed by WB with anti-A3G (upper panel) and anti-p24 (lower panel) antibodies. (C and D) To test the effect of A3G_65-132 on HIV-1 reverse transcription and viral replication, equal amounts of viruses (normalized by p24 ELISA) were used to infect CD4⁺ C8166 T cells. Twelve hours postinfection, total DNA was extracted from the infected cells. The level of HIV-1 LRT was analyzed by real-time PCR. Meanwhile, at different time points after infection, the supernatants from the infected cell cultures were collected (C), and viral replication was monitored by measuring the level of the HIV-1 Gag p24 antigen in the supernatant (D). Panels A and B show representative WB results from three independent experiments. The values presented in panel C are the means and standard deviations from three independent experiments. P values determined by the t test are shown above the bars of compared groups. The HIV p24 values presented in panel D are the means and ranges of values from two independent experiments.