HIV-1 Accessory Protein Vpr Inhibits the Effect of Insulin on the Foxo Subfamily of Forkhead Transcription Factors by Interfering With Their Binding to 14-3-3 Proteins: Potential Clinical Implications Regarding the Insulin Resistance of HIV-1–Infected Patients

pCDNA3-Vpr and -VprR80A, which express wild-type and R80A mutant Vpr, respectively, were described previously (12). pCMV-FLAG-Vpr and -VprR80A, which respectively express FLAG epitope-tagged wild-type Vpr and R80A mutant Vpr, were also described previously (12). HA-FKHR-L1 (HA-Foxo3a), HA-FKHR-L1-T32, S253, S315A, and HA-AKT-ΔPH, which express the hemagglutinin epitope-tagged wild-type Foxo3a (FKHR-L1) and Foxo3a (FKHR-L1) mutant that harbors mutations replacing a threonine and serines located at amino acids 32, 253, and 312 to alanines, and constitutively active Akt were gifts from Dr. M. Greenberg (Harvard Medical School, Boston, MA) (40). pEGFP-C1-Foxo3a, which expresses the enhanced green fluorescent protein (EGFP)-fused Foxo3a, was constructed by subcloning the full coding sequence of Foxo3a (FKHR-L1) from HA-FKHR-L1 into pEGFP-C1 (Clontech, Palo Alto, CA). 3xIRS-Luc, which has the luciferase gene under the control of three FOXO-responsive elements, was a gift from Dr. K.L. Guan (University of Michigan Medical School, Ann Arbor, MI) (44). pGEXT-4T3-Foxo3a and -Cdc25C were constructed by subcloning Foxo3a and Cdc25C cDNAs into pGEX-4T3 in an in-flame manner, respectively. pHook-1, pSV40-β-Gal, and Bluescript SK (+) were purchased from Invitrogen (Carlsbad, CA), Promega (Madison, WI), and Stratagene (La Jolla, CA), respectively. A probe for the Northern blot analysis of the G6Pase gene was a gift from Dr. J. Chou (National Institutes of Health, Bethesda, MD).

Human cervical carcinoma HeLa cells and human hepatoma HepG2 cells were purchased from American Type Culture Collection (Rockville, MD). Cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% of fetal bovine serum, 100 units/ml penicillin, 1 μg/ml streptomycin sulfate, and 25 mmol/l HEPES. Cells were transfected with either the Lipofectin or CaPO₄ method, as described previously (12,13). For the experiments to detect EGFP, HeLa cells were seeded in 20-mm-diameter dishes 1 day before transfection and 0.3 μg per well of pEGFP-C1-Foxo3a and Vpr-expressing plasmids were used for the transfection. Cells were treated with 100 nmol/l insulin after 24 h of transfection or were transfected with 0.2 μg per well of HA-AKT-ΔPH to phosphorylate expressed Foxo3a. The EGFP-fused Foxo3a was detected under a Leica DM IRB inverted microscope (Leica Microsystems, Wetzlar, Germany). Images were captured with a charge-coupled device camera (Shimazu, Kyoto, Japan) and analyzed with the Open Lab software (Improvision, Coventry, U.K.). For reporter assays, HeLa cells were transfected with 1 μg per well of HA-FKHR-L1 (HA-Foxo3a) or HA-FKHR-L1-T32, S253, or S312A, and 1 μg per well of Vpr-expressing plasmids, together with 1.5 μg per well of 3xIRS-Luc and 0.5 μg per well of pSV40-β-Gal. For keeping the same amounts of DNA, empty vectors or Bluescript SK (+) were used. Twenty-four hours after the transfection, cells were treated with 100 nmol/l insulin, and cell lysates were collected after an additional 24-h incubation. Luciferase and β-galactosidase activities were determined as described previously (12). Luciferase activity was divided by β-galactosidase activity to account for transfection efficiency.

HeLa cells were grown in 175-cm² flasks and transfected with 15 μg of each indicated plasmid. Cells were treated with 100 nmol/l insulin 24 h after transfection. After an additional 24 h, cells were lysed with buffer that contained 50 mmol/l Tris-HCl (pH 8.0), 120 mmol/l NaCl, 0.5% NP-40, and 1 Tab/50 ml Complete tablets (Roche Molecular Biochemicals, Indianapolis, IN), and extracts were centrifuged at 400g at 4°C for 10 min. Supernatants were collected to obtain the whole homogenates (20). They were incubated with anti–14-3-3β (K-19) antibody (Santa Cruz Biotechnology, Santa Cruz, CA), which cross-reacts with multiple 14-3-3 proteins, at 4°C for 3 h. Protein-antibody complexes were subsequently harvested by adding protein A/G agarose PLUS (Santa Cruz Biotechnology) and were washed three times with the buffer that contained 20 mmol/l Tris-HCl (pH 8.0), 100 mmol/l NaCl, 0.5% NP-40, and 1 Tab/50 ml Complete tablets. Samples were separated on 12 or 16% SDS-PAGE gels and blotted to a nitrocellulose membrane, which was subsequently treated with anti-hemagglutinin antibody (Santa Cruz Biotechnology). Expressed Vpr was detected with anti-FLAG (M2) antibody (Sigma, St. Louis, MO) in whole homogenates. Inputs of hemagglutinin (HA)-Foxo3a and 14-3-3 were detected in 10% of whole homogenates. Foxo1 was detected in 10% whole homogenates by using anti-Foxo1 antibody (Santa Cruz Biotechnology) in Western blots. Phosphorylated Foxo1, Foxo3a, and Akt were also detected in Western blot by using anti-Foxo1, -Foxo3a, and -Akt, which specifically recognize these molecules that are phosphorylated at specific amino acids (Santa Cruz Biotechnology). The membranes were treated with secondary antibodies, and blotted proteins were visualized by the ECL Plus Western Blotting Detection System (Amersham Pharmacia Biotech, Piscataway, NJ) and exposed to high-performance chemiluminescence films (Hyperfilm ECL; Amersham Pharmacia Biotech).

³⁵S-labeled Vpr was generated by in vitro translation and tested for interaction with bacterially produced and purified GST-Foxo3a or -Cdc25C immobilized on glutathione-sepharose beads in buffer that contained 50 mmol/l Tris-HCl (pH 8.0), 50 mmol/l NaCl, 1 mmol/l EDTA, 0.1% NP-40, 10% glycerol, and 0.1 mg/ml BSA at 4°C for 1.5 h. After vigorous washing with the buffer, proteins were eluted and separated on a 14% SDS-PAGE gel. Gels were fixed and exposed on film. Expression and purification of GST, and GST-Foxo3a and -Cdc25C were confirmed by running these proteins on a 4–20% SDS-PAGE gel and by visualizing with SimplyBlue SafeStain (Invitrogen).

HepG2 cells were plated on 150-mm-diameter dishes 1 day before the experiments. They were transfected with pCDNA3-Vpr and pHook-1. Cells were subsequently treated with 100 nmol/l insulin or vehicle for 24 h. Transfection-positive cells were then collected through the epitope expressed from pHook-1 following the company’s instruction (13), and total RNA was isolated with TRIZOL (Invitrogen). Purified total RNA was run on a formaldehyde-denaturing gel and was blotted to a nylon membrane, which was subsequently treated with ³²P-labeled probe directed to G6Pase or glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The membrane was washed in SSPE buffers and was exposed to an X-ray film. Band density of G6Pase or GAPDH was measured by using the NIH Image software (National Institutes of Health, Bethesda, MD). The expression of G6Pase was corrected for the expression of GAPDH. Results represent the mean and SE of three independent experiments. Akt inhibitor 1L-6-hydroxymethyl-chio-inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate was purchased from Calbiochem-Novabiochem (La Jolla, CA).

HepG2 cells (1 × 10⁶) were cultured in serum-free medium for 48 h, plated on a 100-mm dish, and transfected with pCDNA3-Vpr wild-type or R80A, and pHook-1. The cells were subsequently treated with 100 nmol/l insulin or vehicle for 24 h. Transfection-positive cells were then collected through the epitope expressed from pHook-1 as per company instructions (13), and total RNA was isolated with TRIZOL (Invitrogen). The reverse transcription reaction was carried out using random hexamers with TaqMan Reverse Transcription Reagents (Applied Biosystems, Foster City, CA), as reported previously (45). For detecting mRNA levels of G6Pase, manganese superoxide dismutase (MnSOD) and SCP2 and control human acidic ribosomal phosphoprotein P0 (RPLP0), primer pairs (G6Pase: forward primer 5′-CTACTCTTTCCATCTTTCAG-3′, reverse primer 5′-GAGTTAGGAAGCACTGATC-3′; MnSOD: forward primer 5′-CAGATAGCTCTTCAGCCTG-3′, reverse primer 5′-CACGCTGCGCTTGAAGAAG-3′; SCP2: forward primer 5′-GAAGCTGTTCCAACCAG-3′, reverse primer 5′-GAGTTAGGAAGCACTGATC-3′; RPLP0: forward primer 5′-CGCGACCTGGAAGTCCAACT-3′; reverse primer 5′-CCATCAGCACCACAGCCTTC-3′) were used. The real-time PCR, consisting of the heat activation of the Taq polymerase (at 95°C for 10 min) and the subsequent 62 PCR cycles (denaturing at 95°C for 15 s, annealing/extension at 62°C for 1 min) was performed in triplicate using the SYBR Green PCR Master Mix (Applied Biosystems) in an ABI PRIZM 7700 SDS lightcycler (Applied Biosystems). Obtained threshold cycle values of G6Pase, MnSOD, and SCP2 were normalized for those of RPLP0, and their relative mRNA expressions were shown as fold induction over the baseline. The dissociation curves of used primer pairs showed a single peak, and samples after PCRs had a single expected DNA band in an agarose gel analysis (data not shown).

To test our hypothesis, we first examined the effects of Vpr on the nucleus to cytoplasm translocation of EGFP-Foxo3a induced by insulin in HeLa cells (Fig. 1A). In the absence of insulin, the majority of cells had EGFP-Foxo3a in the nucleus with minimal distribution in the cytoplasm (Fig. 1A, top left). Coexpression of Vpr did not affect the subcellular localization of EGFP-Foxo3a (Fig. 1A, bottom left). Addition of 100 nmol/l insulin, however, strongly shifted the intracellular localization of EGFP-Foxo3a from the nucleus to the cytoplasm (Fig. 1A, top right), whereas addition of Vpr moderately retained EGFP-Foxo3a in the nucleus (Fig. 1A, bottom right). These results indicate that Vpr antagonized insulin’s effect on the nucleus to cytoplasm translocation of Foxo3a.

We next used a constitutively active form of an Akt mutant instead of insulin, which directly phosphorylates Foxo3a, bypassing several intermediate phosphorylation-dependent reactions induced by insulin (Fig. 1B). Expression of this mutant Akt induced cytoplasmic redistribution of EGFP-Foxo3a in a manner similar to insulin (Fig. 1B, top right). Vpr also antagonized this effect of Akt on EGFP-Foxo3a localization, partially retaining it in the nucleus (Fig. 1B, bottom right).

We examined >100 cells that were cotransfected with EGFP-Foxo3a- and Vpr-expressing plasmids and treated with 100 nmol/l insulin and categorized them into five groups that have different distribution patterns from complete nuclear localization to complete cytoplasmic localization. We demonstrated cell numbers in each group as percentage of the total numbers of examined cells (Fig. 1C). In this analysis, the spectrum of cells that demonstrated different EGFP-Foxo3a distribution was from the cells that showed complete nucleus versus cytoplasm localization in the absence of insulin, consistent with the previous experiment. Addition of 100 nmol/l insulin strongly shifted EGFP-Foxo3a from the nucleus into the cyto-plasm. Expression of wild-type Vpr weakly shifted localization of EGFP-Foxo3a in the nucleus in the absence of insulin. In the presence of insulin, however, Vpr expression partially retained EGFP-Foxo3a in the nucleus, suggesting that Vpr moderately antagonized insulin’s effect on Foxo3a subcellular distribution. In contrast to wild-type Vpr, VprR80A mutant, which is defective in binding to 14-3-3 (20), did not affect the nucleus-to-cytoplasm translocation of EGFP-Foxo3a induced by insulin.

We previously demonstrated that Vpr through its COOH-terminal portion binds 14-3-3 outside the phosphopeptide-binding pocket of the latter located COOH-terminal (20). Because Vpr inhibited insulin-induced nucleus-to-cytoplasm translocation of EGFP-Foxo3a, we hypothesized that binding of Vpr to 14-3-3 might interfere with association of 14-3-3 and FOXO proteins induced by insulin. We thus examined this possibility in a coimmunoprecipitation assay in HeLa cells (Fig. 2A). In the absence of insulin, Foxo3a was coprecipitated with 14-3-3 (Fig. 2A, top gel, lane 1). Addition of 100 nmol/l insulin increased the association of 14-3-3 and Foxo3a (Fig. 2A, top gel, lane 2). However, once Vpr was coexpressed, this viral protein inhibited the coprecipitation of 14-3-3 and Foxo3a that was induced by insulin (Fig. 2A, top gel, lane 4). In control studies, insulin stimuli phosphorylated Foxo3a (Fig. 2A, third top gel) and upstream Akt (Fig. 2A, bottom three gels), indicating that Vpr did not inhibit the Foxo3a/14-3-3 association by suppressing phosphorylation reactions on these molecules induced by insulin. In addition, VprR80A, which cannot bind 14-3-3, did not affect the association of these two molecules in contrast to the wild-type Vpr (Fig. 2B). Finally, Vpr did not bind Foxo3a in vitro (Fig. 2C). These results further strengthen the possibility that Vpr interfered with the association of 14-3-3 and Foxo3a via direct binding to 14-3-3.

We next examined the effects of Vpr on Foxo3a-induced transactivation of a responsive promoter to address the functional significance of the observed inhibitory effect of Vpr on the insulin-induced association of Foxo3a and 14-3-3 (Fig. 3). We used 3xIRS-Luc, which contains three FOXO-responsive elements that drive the luciferase reporter gene, as a reporter plasmid. Overexpression of Foxo3a stimulated this promoter activity by fourfold, and addition of 100 nmol/l insulin suppressed Foxo3a-stimulated transactivation by 60%. Wild-type Vpr reduced insulin-induced inhibition of the Foxo3a-stimulated transcription, whereas it did not have a discernible effect on the basal activity of the promoter. VprR80A, however, failed to antagonize insulin-induced suppression of the Foxo3a activity. A Foxo3a mutant (Mut), which has amino acid substitutions at threonine 32 and serines 253 and 312 to alanines and, thus, is unable to bind 14-3-3, stimulated the promoter activity in an insulin-independent manner. Vpr did not affect the transcriptional activity of this mutant Foxo3a. These results are consistent with the observation obtained in the subcellular localization study and the coimmunoprecipitation experiments, in which Vpr respectively suppressed cytoplasmic redistribution of EGFP-Foxo3a and inhibited the association of Foxo3a and 14-3-3 induced by insulin. However, Vpr might also affect FOXO-induced transcriptional activity by a mechanism(s) independent of 14-3-3. Indeed, insulin and Akt may regulate FOXOs in multifaceted ways not all fully understood (42,43).

We examined the effect of Vpr on the expression of endogenous G6Pase mRNA, a gene whose transcription is stimulated by FOXO proteins and, thus, is negatively regulated by insulin (29). We transfected HepG2 cells with the Vpr-expressing plasmids, harvested transfection-positive cells, and examined G6Pase mRNA expression in Northern blots. Expression of Vpr did not have an obvious effect on the G6Pase mRNA abundance, whereas 100 nmol/l insulin suppressed it by 65%. Addition of Vpr expression with the same concentration of insulin almost reversed the suppressive effect of insulin on the G6Pase mRNA expression (Fig. 4A, top two gels). In the same experiment, Foxo1, which is a major FOXO protein in HepG2 cells, was phosphorylated at threonine 24 in Western blots (Fig. 4A, bottom two gels). Mean ± SE values of the G6Pase expression corrected for the expression of GAPDH from three independent experiments are shown in Fig. 4B. In this cell line, an Akt kinase inhibitor 1L-6-hydroxymethyl-chio-inositol 2-(R)-2-O-methyl-3-O-octadecylcarbonate reversed insulin-induced suppression of G6Pase mRNA expression in Northern blots and suppressed phosphorylation of Foxo1 in Western blots (Fig. 4C), indicating that insulin, at least in part, impaired G6Pase mRNA expression by inhibiting Foxo activity in this cell line. Taken together, these findings indicate that Vpr disrupts insulin-induced regulation of G6Pase gene expression by altering the interaction of Foxo with 14-3-3. However, it is possible that Vpr inhibits some of insulin’s effect on G6P gene expression via non–14-3-3–mediated mechanisms.

Because there are several conflicting reports indicating limited effect of FOXO proteins on G6Pase expression (31,46), we also tested Vpr on two other Foxo-responsive genes, MnSOD and SCP2, using real-time PCR in the same HepG2 cells (Fig. 4D). Expression of wild-type Vpr or the R80A mutant did not affect mRNA expression of G6Pase, MnSOD, or SCP2, whereas addition of insulin significantly suppressed it. Wild-type Vpr strongly attenuated the negative effect of insulin on their mRNA expression of these genes, whereas the 14-3-3 binding–defective Vpr R80A mutant exerted no such effects. Taken together, our results suggest that Vpr suppresses insulin’s effect on FOXO-induced expression of endogenous genes, such as those of G6Pase, MnSOD, and SCP2, possibly by inhibiting insulin-induced association of 14-3-3 and FOXO proteins in HepG2 cells.

In the present study, we investigated the effect of HIV-1 Vpr on the insulin-induced modulation of Foxo3a activity. We demonstrated that Vpr antagonized both the insulin- and Akt-induced translocation of Foxo3a from the nucleus to the cytoplasm. It also inhibited the association of Foxo3a with 14-3-3 induced by insulin in a coimmunoprecipitation assay. Furthermore, Vpr antagonized the negative effect of insulin on Foxo3a-induced transactivation of a responsive promoter and also inhibited the suppressive effect of insulin on the mRNA expression of the endogenous G6Pase, MnSOD, and SCP2 genes in HepG2 cells. Because the VprR80A mutant, which is defective in binding to 14-3-3, lost all of these activities and because wild-type Vpr inhibited insulin-induced coprecipitation of 14-3-3 and Foxo3a, our results strongly suggest that Vpr inhibits binding of 14-3-3 to phosphorylated FOXO proteins and, thus, antagonizes insulin’s negative regulation of their activity. It is still possible, however, that Vpr regulates insulin/FOXO-responsive genes by 14-3-3–bypassing mechanisms (42,43,46).

The above effects of Vpr may be relevant to several biologic changes observed in HIV-1–infected lymphocytes and monocytes or nonimmune tissues, such as adipose tissue, that are primary targets of HIV-1 (47). However, the majority of insulin-responsive tissues, such as muscle and liver, have not been reported to be infected directly with HIV-1. The latter may not be necessary for the metabolic actions of Vpr, which is expressed and secreted by host cells infected by HIV-1 and is detected in extracellular fluids, such as plasma and cerebrospinal fluid, of HIV-1–infected patients (14). This is corroborated by extracellular administration of synthetic Vpr, which readily penetrates the cell membrane and causes cell cycle arrest or exerts other effects in HIV-1–uninfected cells (15,16), indicating that Vpr may also act as a paracrine or endocrine “hormone,” acting on neighboring or distant cells that are not infected with HIV-1.

AIDS patients frequently develop a pathologic state known as AIDS-associated insulin resistance and lipodystrophy syndrome, especially when given long-term therapy with anti-HIV drugs, such as protease inhibitors and nucleotide/nonnucleotide reverse transcriptase inhibitors (26). AIDS patients who have this syndrome have carbohydrate intolerance and dyslipidemia and may also develop overt diabetes, along with their cardiovascular sequelae, particularly atherosclerotic coronary artery disease (24,25,48–50). The cause of this syndrome is not clear but seems to be multifactorial. The HIV-1 infection itself seems to contribute to the development of these pathologic changes or increases the vulnerability of patients to develop the syndrome upon uptake of anti–HIV-1 drugs; some AIDS patients develop the characteristic features and metabolic disturbances of the syndrome before treatment (26,51,52). Our results suggest that Vpr might contribute to the insulin resistance of AIDS patients through both its FOXO-related insulin action blockade and its previously described glucocorticoid receptor coactivator activity (12,13). A recent report indicated that FOXO plays an important role in the adipocyte differentiation (53); thus, it is also possible that Vpr influences development of lipodystrophic phenotype seen in this syndrome by modulating FOXO activity.

Vpr inhibited insulin-induced association of 14-3-3 and Foxo3a in a coimmunoprecipitation assay. This indicates that binding of Vpr to 14-3-3 interferes with the interaction of phosphorylated serine and threonine residues of Foxo3a to a phosphopeptide-binding pocket located in 14-3-3. Vpr binds 14-3-3 at the COOH-terminal portion of the latter, outside the phosphopeptide-binding site, including the eighth α-helix of 14-3-3 (20). Because this portion of 14-3-3 also plays a role in the creation of the phosphopeptide-binding cleft, possibly by stabilizing the structure of 14-3-3 (54,55), it seems that Vpr reduces the binding activity of 14-3-3 to the phosphorylated serine and threonine residues of FOXO proteins. In contrast, we previously reported that Vpr facilitated binding of 14-3-3 and Cdc25C independent of the phosphorylation state of Cdc25C (20). These results indicate that Vpr influences the binding activity of 14-3-3 to its partners in different directions, depending on the partner molecule with which it associates.

In addition to the forkhead transcription factors, 14-3-3 proteins influence the function of numerous proteins involved in growth factor, hormone, and cytokine signaling, as well as apoptosis, and cell cycle regulation, such as IRS-1, Raf-1, MEKK1, Bcr, calmodulin/calmodulin kinase, protein kinase C, c-Cbl, BAD, and NF-AT, through direct binding to their phosphorylated serine/threonine residues (21,56). Modulation of the activities of these proteins by Vpr might contribute to the development of other pathologic manifestations observed in HIV-1–infected patients.

We conclude that the HIV-1 accessory protein Vpr counteracts the negative regulation of insulin on the FOXO subfamily of the forkhead proteins possibly by inhibiting association between these proteins and 14-3-3. Our results may be relevant to pathogenetic mechanisms that account for the development of insulin resistance and related metabolic disturbances observed in HIV-1–infected individuals.

We thank Drs. G.N. Pavlakis, A. Gragerov, J. Nakae, and D. Accili for insightful comments; Drs. M. Greenberg, K.L. Guan, and J. Chou for kindly providing plasmids; and K. Zachman and L. Chheng for excellent technical assistance.