Human Immunodeficiency Virus Type 1 Replication and Regulation of APOBEC3G by Peptidyl Prolyl Isomerase Pin1

Abstract

APOBEC3G (A3G) is a cytidine deaminase that restricts human immunodeficiency virus type 1 (HIV-1) replication. HIV-1 synthesizes a viral infectivity factor (Vif) to counter A3G restriction. Currently, it is poorly understood how A3G expression/activity is regulated by cellular factors. Here, we show that the prolyl isomerase Pin1 protein modulates A3G expression. Pin1 was found to be an A3G-interacting protein that reduces A3G expression and its incorporation into HIV-1 virion, thereby limiting A3G-mediated restriction of HIV-1. Intriguingly, HIV-1 infection modulates the phosphorylation state of Pin1, enhancing its ability to moderate A3G activity. These new findings suggest a potential Vif-independent way for HIV-1 to moderate the cellular action of A3G.

APOBEC3G (A3G) is a cytidine deaminase that edits the C4 position of the cytosine base, converting cytidine to uridine (15, 17). A3G is incorporated into human immunodeficiency virus type 1 (HIV-1) virions and restricts HIV-1 replication in human cells (46). To counter A3G's restriction, HIV-1 encodes an accessory protein, Vif, which blocks A3G incorporation into virions (7, 20, 35, 37, 39, 47, 51). It has been reported that Vif bridges A3G to a cellular ubiquitin E3 ligase complex that contains cullin 5/elongin B and C (64), leading to A3G degradation through a proteasome-mediated mechanism.

While Vif-mediated regulation of A3G has been extensively studied, how A3G is modulated by other cellular factors has largely been unexplored. In analyzing A3G sequence, we noticed the existence of potential Pin1 recognition motifs. Pin1 is a peptidyl prolyl cis-trans isomerase (PPIase) that catalyzes the cis-trans isomerization of X-Pro peptide bonds and promotes protein conformational changes (29, 63). Pin1 has a C-terminal PPIase domain and an N-terminal WW domain that binds protein substrates. In general, Pin1 recognizes target proteins through a phosphorylated serine or threonine residue that is followed by a proline amino acid (e.g., pSer/Thr-Pro). Once bound to a substrate protein, Pin1 can affect protein conformation, phosphorylation status, subcellular localization, and/or polypeptide stability. Many Pin1 substrates have been described, and accordingly, Pin1 has been implicated in tumorigenesis (60), cell cycle control (26, 59), cytokine production (44, 48, 49), and neurodegeneration (27, 30), among other processes. To date, a role for Pin1 in modulating viral infection has, however, not been well characterized.

Here, we report A3G as a new Pin1 substrate. We observed that Pin1 physically associated with A3G. In virus-infected cells, expression of Pin1 destabilized A3G and reduced its incorporation into HIV virions, blunting its virus restriction activity. HIV-1 infection enhanced Pin1 phosphorylation, a modification which seemed to promote Pin1's ability to inactivate A3G function. Taken together, these new findings describe a novel Pin1-mediated mechanism available to HIV-1 for moderating A3G restriction.

MATERIALS AND METHODS

Cell culture and transfection.

293T and HeLa cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal bovine serum and l-glutamine (Invitrogen). MAGI cells (23) were cultured in the same medium supplemented with 150 μg/ml G418 (Invitrogen) and 100 μg/ml hygromycin (Sigma). MT-4 and H9 cells were cultured in RPMI (Invitrogen) supplemented with 10% fetal bovine serum and l-glutamine. Plasmid and small interfering RNA (siRNA) transfection was performed using Lipofectamine 2000 (Invitrogen) as recommended by the manufacturer.

Immunoblot analysis.

Immunoblottings were performed as described previously (25).

Antibodies.

The antibodies used in this experiment are anti-Pin1 (R&D system and Cell Signaling Technology), anti-phospho-Pin1 (ser16) (Cell Signaling Technology), anti-actin (Sigma), anti-Apobec3G (14, 20, 22), anti-FLAG (Sigma), and anti-hemagglutinin (anti-HA; Sigma).

GST pulldown assay.

Glutathione S-transferase (GST), GST-Pin1 fusion (GST-Pin1), and GST-Pin1 mutants {e.g., replacement of tryptophan at position 34 with alanine [GST-Pin1(W34A)] or lysine at position 63 with alanine [GST-Pin1(K63A)]} were expressed from the following vectors: pGEX-6P1, pGEX-Pin1, pGEX-Pin1(W34A), and pGEX-Pin1(K63A). Recombinant proteins were produced in BL21 cells (Invitrogen) following treatment with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside). After proteins were prepared using glutathione-Sepharose resin (Amersham Biosciences), proteins bound to the resin were incubated with cell lysates at 4°C for 2 h in immunoprecipitation (IP) buffer consisting of 50 mM HEPES, 150 mM NaCl, 10% glycerol, 1% NP-40, 1.5 mM MgCl₂, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM Na₃VO₄, 20 mM NaF, and protease inhibitor (complete; Roche). After five washes, resin-bound proteins were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and detected by immunoblotting.

IPs.

The cells were lysed in IP buffer. After centrifugation, the supernatant was incubated with the indicated antibody or with irrelevant mouse control immunoglobulin G (Zymed Laboratories) for 2 to 16 h. Immune complexes were recovered by adsorption to protein G-Sepharose resin (Amersham Biosciences). After being washed in IP buffer four times, the immunoprecipitates were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis followed by immunoblotting.

HIV-1 infection.

To prepare HIV-1 culture supernatant, 293T cells were transfected with the pNL4-3Vif⁻ molecular clone together with the indicated expression plasmid(s). At 2 or 3 days posttransfection, culture medium was recovered and clarified by centrifugation at 1,500 rpm for 10 min and then filtered through a 0.45-μm membrane. After measurement of reverse transcriptase (RT) activity, the levels for HIV-1-containing media were normalized to the equivalent RT activities, and these media were used to infect MAGI cells for 6 h. The infected cells were then washed extensively to remove residual input virus. To harvest virions, HIV-containing culture medium was ultracentrifuged at 35,000 rpm for 2 h at 4°C.

Reverse transcription assay.

RT activity in the medium was measured as described previously (25).

β-Gal staining.

MAGI cells were stained with X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) with a β-galactosidase (β-Gal) staining kit (Invitrogen) according to the manufacturer's protocol.

Plasmid constructions.

The cDNA for Pin1 was cloned by reverse transcription-PCR from a human lymphocyte cDNA library (Clontech) by using primers with the sequences 5′-GTTGAATTCATGGCGGACGAG-3′ and 5′-GTTCTCGAGTCACTCAGTGCG-3′. pCAG-FLAG-Pin1, an expression plasmid for FLAG-tagged Pin1 in mammalian cells, and pGEX-Pin1, an expression plasmid for GST-fused Pin1 in Escherichia coli, were constructed by insertion of the EcoRI-XhoI fragment of a PCR product into the pCAG-FLAG and pGEX-6P1 vectors (Clontech), respectively. The expression plasmids for Pin1 with point mutations changing tryptophan at position 34 or lysine at position 63 to alanine [Pin1(W34A) or Pin1(K63A), respectively] were generated by oligonucleotide-directed mutagenesis. The expression plasmids for A3G, pCAG-HA-A3G, and its mutants were prepared by inserting the EcoRI-XhoI fragment of the PCR product, using pcDNA-APO3G (20) as a template, into the pCAG-HA vector. pNL4-3, a wild-type HIV-1 molecular clone, and pNL4-3Vif⁻, an HIV-1 molecular clone deficient in Vif, have been described previously (1, 14, 20-22). The expression plasmids for A3G, Pin1, and Pin1(W34A) are abbreviated pHA-A3G, pFL-Pin1, and pFL-Pin1(W34A), respectively, in this paper.

RESULTS

Pin1 interacts with A3G.

There are three PPIase families: cyclophilin (CyP), FK506 binding protein, and parvulin polypeptides (57). CyPA, a member of the CyP family, has previously been described to play roles in HIV-1 replication. CyPA can modulate the sensitivity of HIV-1 to a cellular restriction factor, TRIM5α (32, 53). Currently, except for CyPA, how other PPIases affect virus biology remains poorly understood. To address this issue, we investigated the role of Pin1, a parvulin family PPIase, on HIV-1 replication.

We first performed an open-ended screening for the binding between Pin1 and HIV-1 candidate proteins (Gag, Tat, and Vpr) and several cellular proteins described in the literature to influence HIV-1 replication. In the former category, we did not find any interaction between Pin1 and the HIV-1 proteins (data not shown). In the latter group, using recombinant GST-fused Pin1 (GST-Pin1) and 293T cell lysate, we pulled down the human A3G protein (Fig. 1A). This pulldown was confirmed by the co-IP from cells of transfected, HA-tagged A3G with FLAG-tagged Pin1 (Fig. 1B) and by the finding that cell-endogenous Pin1 also coprecipitated with HA-tagged A3G (Fig. 1C). Moreover, in human H9 cells, cell-endogenous A3G and Pin1 were found to associate (Fig. 1D), supporting a physiological intracellular Pin1-A3G interaction. Interestingly, Pin1-A3G association was moderated by treatment with RNase (Fig. 1E). This interaction was weakened by high salt concentration (300 mM and 600 mM) but not by DNase or EDTA treatment (Fig. 1F). We further observed that the introduction of a point mutation at amino acid position 34 of Pin1 drastically decreased its affinity for A3G (Fig. 2C and D). Taken together, the results are compatible with the idea that Pin1-A3G association occurs through direct protein-protein contact assisted through RNA binding. This observation is compatible with the finding of a similar RNA dependence required by other cellular factors for binding A3G (6, 13, 24). We also tested the association of Pin1 with APOBEC3F (A3F). In contrast with the result for A3G, the association of Pin1 with A3F was relatively weak (Fig. 1G and data not shown).

FIG. 1.

Interaction between Pin1 and A3G. (A) GST pulldown assays were performed with 293T cell lysates transfected with A3G expression plasmid by using the recombinant proteins, Pin1 fused with GST (GST-Pin1), or GST alone. The “input” lane shows one-fifth the amount of protein material used in the pulldown assays. Coomassie brilliant blue (CBB) staining patterns of the pulldowns are shown in the lower panel. (B) The lysate from 293T cells ectopically transfected with FLAG-Pin1 (FL-Pin1) with or without cotransfected, HA-tagged A3G (HA-A3G) was immunoprecipitated with anti-HA, anti-FLAG, or irrelevant mouse immunoglobulin G (IgG) as a negative control. Immunoprecipitates were visualized with anti-HA (HA-A3G; upper panel) or anti-FLAG (FL-Pin1; lower panel) by immunoblotting (IB). (C, D) Interaction of endogenous Pin1 with exogenous HA-A3G in 293T cells (C) or endogenous A3G in H9 cells (D) was examined by co-IP analysis as shown in panel B. The lower panel in panel D shows the internal control of cell lysate detected with anti-actin. (E) GST pulldown assays as described for panel A were performed with or without RNase A, using lysate from A3G-transfected HeLa (upper panel) or 293T (lower panel) cells. (F) GST pulldown assays as described for panel A were performed in the absence or presence of DNase, RNase, 100 mM EDTA, 300 mM NaCl, or 600 mM NaCl. (G) GST pulldown assays as described for panel A were performed with 293T cell lysates transfected with the A3F expression plasmid.

FIG. 2.

Mapping the protein regions needed for Pin1 and A3G interaction. (A) GST pulldown assays were performed using recombinant GST-Pin1 or GST alone. Cell lysates were from 293T cells transfected with the HA-A3G plasmid or its indicated point mutants (e.g., S95A or S286A). In one of the pulldown experiments, the lysates were treated first with calf intestine alkaline phosphatase (CIP). Coomassie brilliant blue (CBB) staining of the pulldowns is shown at the bottom. (B) Schematic representation of A3G with amino acid numbering (left). The shaded regions are the two deaminase domains (63). A3G deletion mutants harboring the regions from aa 55 to 384, aa 109 to 384, aa 163 to 384, and aa 1 to 162 were tested for binding to GST alone or GST-Pin1. The abilities of A3G deletion mutants to be pulled down by GST-Pin1 are summarized with +/−. (C) GST-Pin1, GST-Pin1 mutated in residue W34A [GST-Pin1(W34A)], GST-Pin1 mutated in residue K63A [GST-Pin1(K63A)], or GST alone was used to pull down cell lysates from 293T cells transfected with the A3G expression plasmid. CBB staining of the pulldowns is shown at the bottom. (D) 293T cells transfected with FLAG-Pin1(W34A) [FL-Pin1(W34A)], FL-Pin1(K63A), or the empty vector, together with HA-A3G, were immunoprecipitated with anti-FLAG antibody, followed by detection by immunoblot (IB) analysis with anti-HA antibody.

Mapping the interactive regions of Pin1 and A3G.

Pin1 substrates generally conserve an S/T-P (serine/threonine-proline) motif and bind Pin1 in a manner dependent on S/T phosphorylation. To understand how Pin1 and A3G interact, we noted first the existence of two S-P motifs, at residues 95 and 96 and residues 286 and 287, in A3G. The contribution of each S-P motif to Pin1 binding was investigated by individually mutating the serine at position 95 or 286 to alanine. Surprisingly, GST pulldown assays using either the S95A or the S286A mutant showed that both proteins still bound GST-Pin1 (Fig. 2A). Further, alkaline phosphatase-treated A3G did not lose its ability to bind GST-Pin1 (Fig. 2A). Taken together, one interpretation of these results suggests a noncanonical interaction between Pin1 and A3G which does not depend on S/T-P phosphorylation.

Next, we explored protein-protein binding using a series of deletion mutants. We found that A3G deletion mutants encompassing amino acids (aa) 55 to 384, 109 to 384, or 1 to 162, but not 163 to 384, could be captured by GST-Pin1 (Fig. 2B). Thus, the central region of A3G, between the two deaminase domains, likely specifies Pin1 interaction. We also tested two point mutations in Pin1. Mutation of W34 to A or K63 to A abrogated Pin1's WW or PPIase domain, respectively (30, 44). In pulldown assays, GST-Pin1(W34A), but not GST-Pin1(K63A), showed a decreased capture of A3G (Fig. 2C). In agreement with these results, we could coimmunoprecipitate A3G with Pin1(K63A) but not Pin1(W34A) (Fig. 2D).

Pin1-attenuated A3G restriction of HIV-1.

The association of Pin1 with A3G prompted us to ask if such interaction would affect A3G's HIV-1 restriction activity. Vif⁻ HIV-1, unlike Vif⁺ HIV-1, is sensitive to A3G restriction (46). To address if Pin1 expression influences virus restriction, we mock transfected 293T cells or cotransfected pNL4-3Vif⁻, a Vif-deficient HIV-1 molecular clone, into 293T cells with empty vector, with pHA-A3G, with pHA-A3G plus pFL-Pin1, with pHA-A3G plus pFL-Pin1(W34A), or with pFL-Pin1. We recovered virus supernatants from the respective transfectants and used them to individually infect MAGI cells. Infected MAGI cells were washed extensively to remove input virus, and 3 days after infection, MAGI cells were quantified for HIV-1 production by measuring RT activity in the culture media and staining the cells with X-Gal. As shown in Fig. 3A, infection of MAGI cells with virus produced from cotransfection of pHA-A3G with pNL4-3Vif⁻ showed reduced numbers of X-Gal-positive cells compared to infection with virus supernatant produced from transfection with pNL4-3Vif⁻ alone (Fig. 3A, compare panel b to panel a). At the same time, in the MAGI assay, Vif⁻ virus produced when pFL-Pin1 was additionally cotransfected with pHA-A3G and pNL4-3Vif⁻ showed that the virus-restrictive effect of A3G on NL4-3Vif⁻ was attenuated by Pin1 (Fig. 3A, compare panel c to panel b). In contrast, the same experiment conducted with cotransfection using mutant Pin1(W34A), instead of WT Pin1, did not affect A3G's restriction activity (Fig. 3A, compare panels b, c, and d). Moreover, Vif⁻ virus produced from a cotransfection of pNL4-3Vif⁻ with Pin1 alone, in the absence of A3G, infected MAGI cells no differently than control virus produced from transfection of pNL4-3Vif⁻ alone (Fig. 3A, compare panel e to panel a). We did not observe any sign of infection when we used the supernatant from mock-transfected cells as a negative control (Fig. 3A, panel f).

FIG. 3.

Overexpression of Pin1 suppressed the HIV restriction activity of A3G. (A) Infection of MAGI cells with pNL4-3Vif⁻ virus produced using the indicated conditions of A3G and Pin1 overexpression. Individual culture media were separately collected from 293T cells transfected with pNL4-3Vif⁻ plus pCAG (a), pNL4-3Vif⁻ plus pHA-A3G (b), pNL4-3Vif⁻ plus pHA-A3G plus pFL-Pin1 (c), pNL4-3Vif⁻ plus pHA-A3G plus pFL-Pin1(W34A) (d), or pNL4-3Vif⁻ plus pFL-Pin1 (e) or mock-transfected 293T cells (f). Except for the medium from the mock-transfected 293T cells, the level for each virus-containing supernatant was normalized to the equivalent RT activity. MAGI cells were then infected with equivalent RT counts, and the infected cells were stained with X-Gal 3 days later. (B) The number of X-Gal-stained cells in panel A were counted and presented graphically. WT, wild type. (C) Production of supernatant RT activity in the medium of the infected MAGI cells as described for panel A at three days postinfection was measured, and the relative RT activities are shown.

To further verify Pin1's ability to attenuate A3G's restriction of Vif⁻ HIV-1, we generated virus from cotransfection of pNL4-3Vif⁻ plus pHA-A3G with dose escalation of either wild-type pFL-Pin1 (Fig. 3B, lanes 3 to 5) or mutant pFL-Pin1(W34A) (Fig. 3B, lanes 6 to 8). Figure 3B shows X-Gal staining results for MAGI cells infected using Vif⁻ viruses produced from these transfections. We also included in this experiment infection of MAGI cells with virus produced from cotransfection of pNL4-3Vif⁻ with increasing doses of pFL-Pin1 alone (Fig. 3B, lanes 9 to 11). Separately, the X-Gal staining results were compared to direct measurements of virion RT activity released into the supernatant media of the indicated MAGI cell samples (Fig. 3C). Collectively, the findings support that A3G's restrictive effect on pNL4-3Vif⁻ virus is moderated by wild-type Pin1 but not mutant Pin1(W34A).

Pin1 decreased A3G incorporation into HIV-1 virions.

How does Pin1 affect A3G's HIV-restricting function? In the absence of Vif, A3G is incorporated into HIV-1 virions, and this step is critical for mediating restriction activity (17). A simple hypothesis that could explain Pin1's activity would be that it prevents A3G incorporation into HIV-1 virions. Using NL4-3Vif⁻ virus, we checked and found that Pin1 expression, in a dose-dependent manner, indeed decreased virion-associated A3G levels (Fig. 4A, lanes 4 to 6, and B, lanes 2 and 3). The specificity of this finding was verified by employing the Pin1(W34A) mutant, which did not decrease A3G virion incorporation (Fig. 4B, lanes 4 and 5).

FIG. 4.

Overexpressed Pin1 reduced A3G incorporation into HIV-1 virions. (A) Detection of A3G in pNL4-3Vif⁻ virions after overexpression of FLAG-Pin1 (FL-Pin1). Culture supernatants were recovered from 293T cells transfected without (lane 1) or with (lanes 2 to 6) pNL4-3Vif⁻, with an empty plasmid vector (lane 2), or with pHA-A3G (lanes 3 to 6) and increasing amounts of pFL-Pin1 (lanes 4 to 6). The supernatants were ultracentrifuged to pellet virions, which were analyzed by immunoblotting with anti-HA (upper panel) and anti-HIV (lower panel) hyperimmune serum. (B) Detection of A3G in pNL4-3Vif⁻ virions after overexpression of FL-Pin1 or the FL-Pin1(W34A) mutant. The experiments were performed as described for panel A under the indicated transfection of pNL4-3Vif⁻, pHA-A3G, pFL-Pin1, and pFL-Pin1(W34A). The top panel shows A3G in virions as measured by immunoblotting; the bottom panel shows production and isolation of virions as verified by monitoring for p24 capsid protein. (C) Knockdown of cell-endogenous Pin1 increases A3G incorporation into HIV-1 virions. A randomized siRNA (si-control; lane 1) or a Pin1-specific siRNA (si-Pin1; lane 2) was employed. 293T cells were transfected and virions were harvested as described for panel A. The top two panels show monitored A3G amounts in virions and amounts of HIV-1 capsids. The bottom two panels show verified Pin1-specific siRNA knockdown and immunoblotting of actin as internal loading controls.

To establish that cell-endogenous Pin1 has a function similar to that of exogenously transfected Pin1, we performed siRNA knockdown experiments. In the setting of cotransfection of pNL4-3Vif⁻ plus pHA-A3G into cells, our designed siRNA knocked down Pin1 (Fig. 4C, bottom), and this reduction in Pin1, consistent with our above-mentioned findings, increased the amount of HA-A3G in Vif⁻ virions (Fig. 4C, top).

Pin1 downregulated A3G expression.

We next sought to investigate how Pin1 decreases virion-associated A3G. One possible explanation is that Pin1 expression reduced the ambient level of A3G protein in cells. To test this notion, we expressed HA-A3G with either Pin1 or Pin1(W34A) in cells. Using immunoblotting, we found that the steady-state levels of HA-A3G protein were distinctly lower in the Pin1, but not the Pin1(W34A), cells (Fig. 5A). In contrast, Pin1 had only a small effect on A3F expression (Fig. 5B). We next checked the stability of presynthesized A3G protein when in the presence of Pin1. We overexpressed A3G with a small amount of Pin1, followed by treatment of the cells with cycloheximide to block further new protein synthesis; the amount of time maintained in cycloheximide was termed the “chase” period. Hence, we asked whether presynthesized A3G would be more or less stable when maintained over time with cycloheximide in cells that do not overexpress Pin1 (Fig. 5C, top), that overexpress wild-type Pin1 (Fig. 5C, middle), or that overexpress mutant Pin1(W34A) (Fig. 5C, bottom). In this analysis, A3G had a shorter half-life in the presence of transfected Pin1 (Fig. 5C, middle), but not Pin1(W34A) (Fig. 5C, bottom), than the control cells without exogenous Pin1 (Fig. 5C, top). Quantifications of relative half-lives over time are graphed in Fig. 5D. Thus, Pin1 appears to destabilize A3G protein, perhaps through a not-yet-characterized conformational change. Proteasome inhibitor MG132, when employed, showed a partial reversal of Pin1's effect on A3G protein level (Fig. 5E). This result suggests that A3G downregulation is mediated in part by the proteasome pathway, but there may be additional contributions by another pathway(s).

FIG. 5.

Overexpression of FL-Pin1 decreased the level of intracellular A3G. (A) HeLa cells were transfected with pHA-A3G and the indicated expression plasmids for FLAG-Pin1 (FL-Pin1) or FL-Pin1(W34A). An immunoblot analysis were performed with anti-HA (upper), anti-FLAG (middle), and anti-actin as an internal control (lower). The numbers at the top indicate the amounts of transfected plasmid (in micrograms). (B) The expression of A3F was examined in the presence or absence of Pin1 as described for panel A. (C, D) Cycloheximide (CHX) analyses of A3G stability in the presence of transfected Pin1. HeLa cells were transfected with pHA-A3G with an empty plasmid vector (0.05 μg), a expression plasmid for FL-Pin1 (0.05 μg), or a FL-Pin1(W34A) plasmid (0.05 μg). The transfected cells were treated 36 h after transfection with CHX to block de novo protein synthesis; the stability of already synthesized A3G was detected after CHX treatment for 5, 10, and 20 h by immunoblotting. The intensities of bands at the first left lanes were normalized and plotted against the CHX treatment time (in hours) in panel D. (E) A3G expression in the presence or absence of Pin1 was examined as described for panel A with or without 50 μM MG132 for 12 h. The band intensities of HA-A3G are shown at the bottom.

HIV-1 expression cooperated with Pin1 to reduce A3G levels.

Overexpression of A3G is pernicious for HIV-1 replication in cells. While overexpressed Pin1 has a destabilizing effect on A3G protein, we wondered how HIV gene expression might affect Pin1's effect on ambient A3G levels. Interestingly, we observed that while Pin1 transfection alone decreased HA-A3G levels modestly, cotransfection of Pin1 plus NL4-3Vif⁻ was markedly more potent in reducing HA-A3G levels (Fig. 6A).

FIG. 6.

HIV-1 infection enhances Pin1-mediated downregulation of A3G. (A) HIV-1 infection stimulated Pin1 suppression of A3G activity. 293T cells were transfected with pHA-A3G, with or without a small amount of pFL-Pin1 (0.05 μg) or pNL4-3Vif⁻, as indicated. A3G expression (shorter and longer exposures in the upper and middle panels, respectively) and actin as an internal control (lower panel) were detected by immunoblotting. The intensities of the A3G bands were quantified and enumerated at the bottom by setting the respective samples without FLAG-Pin1 (FL-Pin1) as 1. (B) HIV-1 transfection augmented the phosphorylation of Pin1. 293T cells were transfected with FL-Pin1 alone (lane 1), FL-Pin1 plus empty vector (lane 2), FL-Pin1 plus pNL4-3 (lane 3), or FL-Pin1 plus pNL4-3Vif⁻ (lane 4). Phospho-specific anti-Pin1 or total anti-Pin1 antibodies were used to detect phosphorylated and total Pin1 proteins. The upper panel shows total Pin1, the middle panel shows phosphorylated Pin1 at serine 16, and the lower panel shows actin loading controls. (C) The MT-4 T cell-line was infected with HIV-1 (+) or mock infected (−). Total Pin1 (upper), phosphorylated Pin1 (middle), or actin as an internal control (lower) was detected by immunoblotting.

The above-mentioned results raised a query as to whether HIV-1 infection might modulate Pin1 activity in infected cells. To address this question, we first asked whether transfection of the HIV-1 molecular clone into 293T cells would influence the level of Pin1 protein in cells. We found by immunoblotting that neither exogenously cotransfected FLAG-Pin1 (Fig. 6B, top) nor endogenously synthesized Pin1 (data not shown) was affected by transfection into cells of NL4-3 or NL4-3Vif⁻ molecular clones. However, by chance, when a Pin1 phosphor-specific antibody (which recognizes phosphorylation at serine 16) was employed in immunoblotting, a dramatic increase in phosphor-Pin1 was observed in NL4-3- or NL4-3Vif⁻-transfected cells (Fig. 6B, middle). To ask next whether the phosphorylation of cell-endogenous Pin1 is affected by virus infection, we infected MT-4 T cells with HIV-1 and found that while the overall amount of Pin1 was not perturbed, the level of phosphor-Pin1 was significantly increased (Fig. 6C). Currently, we do not understand how phosphorylation of Pin1 might enhance A3G downregulation; the full relevance of Pin1 phosphorylation for A3G downregulation by HIV-1 will require further characterization using Pin1 phosphorylation site mutants. We note that there is a description in the literature of relocalization of Pin1 to the cytoplasm after phosphor-modification at S16 (see Discussion) (31).

DISCUSSION

A3G is a restriction factor that inhibits HIV-1 replication (2, 5, 16, 17, 19, 38). A3G incorporation into HIV-1 virions in producer cells has been reported to be important for its restriction activity when these virions are used to infect new cells (17, 50, 61). HIV-1 Vif, on the other hand, blocks the incorporation of A3G into virions, inhibiting A3G-mediated activity and thus allowing the virus to escape from this restriction (8, 17, 36, 40, 42). Vif can direct A3G to a ubiquitin E3 ligase complex composed of cullin 5, elongin B, elongin C, and Ring box-1, which promotes A3G polyubiquitination, followed by its proteasomal degradation (33, 64). Thus, mechanistic insights into A3G protein quality control as well as its incorporation into virion are important for understanding the molecular basis of A3G-mediated HIV-1 restriction. To date, how cellular factors regulate the steady-state levels of A3G or control its packaging into virions has largely not been investigated, although 7SL RNA has been reported to promote encapsidation of A3G (56). Here, we report for the first time that the prolyl isomerase Pin1 can regulate A3G expression and attenuate its incorporation into HIV-1 virions. In this study, we focused on the effect of Pin1 on A3G in virus producer cells. However, although not explored here, it is also an important issue to further clarify whether Pin1 affects A3G function in resident cells that are infected by HIV-1 virions. This is a topic of interest for a future study.

To date, Pin1 has been reported to influence the degradation of interferon regulatory factor 3 (44), cyclin E (55, 62), and Che-1 (10) through the promotion of protein polyubiquitination followed by proteasomal degradation. In these instances, Pin1 is suggested to induce the conformational change of its substrate proteins and/or recruit a ubiquitin E3 ligase to substrate proteins. For other settings, Pin1 has been described to influence mRNA stability. For example, granulocyte-macrophage colony-stimulating factor (GM-CSF) expression is regulated at the posttranscriptional level through an AU-rich element (ARE) comprising AUUUA repeats in its 3′ untranslated mRNA region (12). Pin1 stabilizes GM-CSF mRNA via the association of the ARE-binding proteins AUF1 and hnRNP C with GM-CSF mRNA and increases GM-CSF expression (48). Pin1 binding with AUF1 also modulates the stability of TGF-β1 mRNA that possesses ARE motifs (49). Although this was not investigated here, we note that A3G has two AUUUA sequences in its 3′ untranslated region. While we have focused on Pin1's activity on A3G protein stability, it would be of interest in the future to check whether Pin1 can affect the half-life of endogenous A3G mRNA.

There are reports that some A3G protein is found in P bodies, and some Pin1 protein is detected in nuclear speckles (31, 58). On the other hand, our immunofluorescence analyses showed that overexpressed A3G and some Pin1 protein are located in the cytoplasm (data not shown). It is possible that Pin1 interacts with cytoplasmic A3G, and it remains currently unclear whether Pin1 interacts with P bodies. Further studies will be needed to clarify the details for how, where, and when Pin1 and A3G colocalize in the cells. Separately, it was also reported that A3G can form high-molecular-mass (HMM) and low-molecular-mass (LMM) complexes (5) and that A3G in an enzymatically active LMM complex functions as a strong restriction factor against HIV. Interestingly, most of the A3G overexpressed in our 293T cells resides in an HMM complex (data not shown), and our fractionation assay showed that Pin1 did not change the localization of A3G between the HMM and the LMM complexes (data not shown). The dynamics of Pin1-A3G interaction and colocalization appear to be complex and require further detailed characterizations in future studies.

The salient observation from our work is that Pin1 has a Vif-independent effect on A3G activity, although Pin1's effect by itself is clearly not sufficient to fully neutralize A3G function. During HIV-1 infection of cells, Vif acts to ameliorate A3G incorporation into virions. However, there may be circumstances when an uninfected cell would desire A3G activity to be moderated. In such a setting, we hypothesize that Pin1 can be an effector that is employed to serve this function. Our current interpretation is that the Vif and Pin1 mechanisms are distinct; nonetheless, there is a suggestion that both could be employed by HIV-1. Thus, we observed that HIV-1 expression can dramatically influence the phosphorylation of Pin1, and this modification in Pin1 appears to enhance a further regulation of A3G expression. In this respect, it has been reported that phosphorylation of Pin1 at serine 16 alters the subcellular localization of Pin1 from the nucleus into the cytoplasm (31). In the case of HIV, it is possible that the relocalization of phosphorylated Pin1 into the cytoplasm increases its opportunity to associate with A3G, which we observe to reside mostly in the cytoplasm. In future studies, we plan to investigate in detail the relative potencies of phosphorylated and unphosphorylated Pin1 for A3G regulation. Of relevance, it has been reported that protein kinase A and protein kinase C can phosphorylate the S16 residue of Pin1 (31). On the other hand, the expression of Tat (3, 4) and gp120 (18) has also been shown to activate protein kinase C. Additionally, since the amount of phosphorylated Pin1 appears to be larger in cells transfected with Vif⁺ NL4-3 than in those transfected with Vif⁻ NL4-3 (Fig. 6A), Vif expression may further contribute to influencing Pin1 phosphorylation. The mechanism(s) by which HIV-1 can affect Pin1 phosphorylation may be complex and multifaceted and will require further investigation.

A3G has the potential to restrict not only HIV-1 but also other viruses, including human T-cell leukemia virus type I, hepatitis B virus (HBV), and foamy virus (9, 28, 34, 43, 45, 54). Since T cells and hepatocytes when stimulated with interferon express A3G (46, 52), viruses, which do not encode a Vif-like protein, logically would need to devise means to neutralize A3G function in the host cells. Primate foamy virus may use its Bet protein to overcome A3G restriction (9, 28, 43). Human T-cell leukemia virus type I may evade A3G restriction via a function in the C terminus of its nucleocapsid protein (11) through an uncharacterized mechanism. However, for many other viruses, including HBV, it is unclear how A3G-mediated restriction could be addressed. On the basis of the findings here, one could reason that viruses without a viral A3G-neutralizing factor, such as Vif, could use Pin1 to moderate A3G's restriction function. It is thus intriguing that Pin1 is upregulated in HBV-positive tissues (41). Further analyses of virus-Pin1-A3G interplay could shed insight into the dynamics of cellular restrictions of viruses and viral responses to silence such attack.