HIV-1 Vpr Enhances PPARβ/δ-Mediated Transcription, Increases PDK4 Expression, and Reduces PDC Activity

Abstract

HIV infection and its therapy are associated with disorders of lipid metabolism and bioenergetics. Previous work has suggested that viral protein R (Vpr) may contribute to the development of lipodystrophy and insulin resistance observed in HIV-1–infected patients. In adipocytes, Vpr suppresses mRNA expression of peroxisomal proliferator-activating receptor-γ (PPARγ)-responsive genes and inhibits differentiation. We investigated whether Vpr might interact with PPARβ/δ and influence its transcriptional activity. In the presence of PPARβ/δ, Vpr induced a 3.3-fold increase in PPAR response element-driven transcriptional activity, a 1.9-fold increase in pyruvate dehydrogenase kinase 4 (PDK4) protein expression, and a 1.6-fold increase in the phosphorylated pyruvate dehydrogenase subunit E1α leading to a 47% decrease in the activity of the pyruvate dehydrogenase complex in HepG2 cells. PPARβ/δ knockdown attenuated Vpr-induced enhancement of endogenous PPARβ/δ-responsive PDK4 mRNA expression. Vpr induced a 1.3-fold increase in mRNA expression of both carnitine palmitoyltransferase I (CPT1) and acetyl-coenzyme A acyltransferase 2 (ACAA2) and doubled the activity of β-hydroxylacyl coenzyme A dehydrogenase (HADH). Vpr physically interacted with the ligand-binding domain of PPARβ/δ in vitro and in vivo. Consistent with a role in energy expenditure, Vpr increased state-3 respiration in isolated mitochondria (1.16-fold) and basal oxygen consumption rate in intact HepG2 cells (1.2-fold) in an etomoxir-sensitive manner, indicating that the oxygen consumption rate increase is β-oxidation–dependent. The effects of Vpr on PPAR response element activation, pyruvate dehydrogenase complex activity, and β-oxidation were reversed by specific PPARβ/δ antagonists. These results support the hypothesis that Vpr contributes to impaired energy metabolism and increased energy expenditure in HIV patients.

The HIV-1 accessory protein viral protein R (Vpr) is a 96–amino-acid peptide that contributes to nuclear import of the viral preintegration complex, including retroviral DNA. Vpr also affects other cellular functions, such as the host response to infection, including modulation of cytokine production (1). Vpr-induced apoptosis has been linked to G2/M cell cycle arrest (2) and permeabilization of the mitochondrial membrane via a specific interaction with the permeability transition pore complex, which comprises the voltage-dependent anion channel in the outer membrane and the adenine nucleotide translocator in the inner membrane (3). Interestingly, Vpr circulates in plasma independently of intact HIV-1 and can transduce a variety of cells that cannot be infected directly by HIV-1 (4). Vpr also acts as a transcriptional regulator of nuclear hormone receptors. For instance, Vpr enhances glucocorticoid receptor-mediated transcriptional activity by directly interacting with this steroid receptor and its coregulator p300/cAMP response element-binding protein-binding protein (5). Conversely, acting as a corepressor, Vpr inhibits peroxisomal proliferator-activating receptor-γ (PPARγ)-mediated transactivation of genes whose enhancers include PPAR response elements (PPREs) (6).

PPARs, comprising PPARα, PPARβ/δ, and PPARγ, are a family of nuclear receptors that function as transcription factors and gene modulators to regulate a wide range of biological processes including differentiation, immunity, and the cellular metabolism of lipid and glucose (7). As heterodimers with the retinoic acid X receptor (RXR), the 3 PPARs bind to PPREs and regulate gene transcription (8, 9). The PPAR/RXR heterodimers interact with both coactivators (eg, PPARγ coactivator 1) (10) and corepressors (eg, receptor interacting protein 140) (11). Although the 3 PPAR isoforms share similar structures, they possess distinct metabolic functions in different tissues (12). PPARα activates mitochondrial fatty acid oxidation, and PPARγ regulates lipid anabolism or lipogenesis (12), whereas the function of PPARβ/δ is less well understood.

PPARβ/δ is expressed in many tissues and participates in the regulation of fatty acid oxidation as well as other functions (13). PPARβ/δ agonists may have a role in treating diabetes, inflammation, and oxidative stress (14). Endogenous ligands for PPARβ/δ include fatty acids, triglycerides, and prostacyclin (15, 16). Activation of PPARβ/δ stimulates the expression of pyruvate dehydrogenase (PDH) kinase (PDK) isoforms, particularly PDK4 (16), through multiple PPREs located in their promoter regions, and thus modulates the activity of the PDH complex (PDC). PDC is the rate-limiting enzyme for conversion of pyruvate to acetyl-coenzyme A (CoA), providing acetyl-group carbon atoms for the mitochondrial tricarboxylic acid cycle (16).

In the present work, we addressed the hypothesis that Vpr acts as a coactivator of PPARβ/δ-regulated transcription and thereby has distinct effects on the metabolism of carbohydrates and lipids and on mitochondrial function.

Materials and Methods

Plasmids

Preparation of pCDNA3-Vpr, -VprL64A, and -VprL64, 67,68A, which express wild-type Vpr or the transcriptional coactivator activity-defective Vpr mutants harboring leucine to alanine replacements at amino acid positions 64, 67, and/or 68, have been described previously (5). pCMV-PPARβ/δ, which expresses mouse PPARβ/δ under the control of the cytomegalovirus promoter, was a gift from Dr. R. M. Evans (Salk Institute, San Diego, California). pCDNA3-hPXR, which expresses the human PXR, was a gift from Dr. F. J. Gonzalez (National Institutes of Health, Bethesda, Maryland). pCDNA3-RXRγ was constructed by subcloning the coding sequence of the human RXRγ obtained from CMV27103 (provided by Dr W. Lamph, Ligand Pharmaceutical, San Diego, California) into EcoRI and HindIII sites of pCDNA3 (Invitrogen, Carlsbad, California). The full-length pGEX4T3-PPARβ/δ and the ligand-binding domain (LBD) were constructed by subcloning corresponding cDNA fragment of the mouse PPARβ/δ (full-length, amino acids 2-440; LBD, amino acids 137-440) into BamHI and NotI sites of pGEX-4T3 (GE Healthcare Bio-Sciences Corp, Piscataway, New Jersey). pPPRE-Luc, which contains the luciferase gene under the control of a synthetic PPRE linked to the proximal portion of the HSV-TK promoter, was a gift of Dr. A. D. Miller (Fred Hutchinson Cancer Center, Seattle, Washington). pCMV-β-Gal, which expresses β-galactosidase under the control of the CMV promoter, was purchased from Stratagene (La Jolla, California).

Compounds and antibodies

The 15d-prostaglandin J2 (15dPJ2), WY14643, and GW1929 were purchased from Enzo (Farmingdale, New York). GW501516 was purchased from AXXORA (San Diego, California). GSK6060, etomoxir sodium salt, and 2,4-dinitrophenol (DNP) were purchased from Sigma-Aldrich (St Louis, Missouri). GSK3787 was purchased from Tocris Bioscience (Bristol, United Kingdom). Anti-PDHE1α, anti–phospho-PDHE1α (Ser293), and anti–phospho-PDHE1α (Ser300) were gifts from Dr. H. Pilegaard (Copenhagen, Denmark). Anti-PDK2, -PDK4, and -PDH phosphatase 2 (PDP2) antibodies were gifts from Dr. R. A. Harris (Indianapolis, Indiana). PPARβ/δ and β-actin antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, California). XF assay medium (DMEM) was purchased from Seahorse Bioscience (Billerica, Massachusetts). PPARβ/δ-specific small interfering RNA (siRNA) (PPARβ/δ Stealth Select RNAi PPARDHSS108292), and a negative control siRNA (Stealth RNAi Negative Control Duplexes) were purchased from Life Technologies (Grand Island, New York).

Cell culture and transient transfection

Human hepatocellular carcinoma HepG2, human rhabdomyosarcoma A204, and human proximal renal tubular epithelial HK-2 cells were grown in a 6-well (surface area 4.67 cm²) plate at 5 × 10⁵ cells per well in 2 mL DMEM (Life Technologies, Gaithersburg, Maryland), supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. On the following day, approximately 60% confluent cells were used for transfection assay. In some experiments, cells were treated with 100 ng/mL soluble Vpr (sVpr), the synthetic full-length Vpr peptide previously reported (6) for 24 hours before addition of PPARβ/δ ligands.

Standard transfections were performed as described previously (6). Lipofectamine 2000 reagent and serum-free Opti-MEM were used according to the manufacture's instructions (Invitrogen). Transfected wells for experiments described in this study received 0.2 μg each of pCMV-PPARβ/δ, pCDNA3-RXRγ, mutant or wild-type pCDNA3-Vpr along with 0.1 μg/well pPPRE-Luc and pCMV-β-Gal each. pCDNA3 vector without Vpr (0.2 μg/well) was used as control and as plasmid DNA to equalize the amount of transfected DNA in Figure 1D. For the PPARβ/δ knockdown study, HepG2 cells were transfected for 24 hours with 100 nM PPARβ/δ-specific siRNA or a negative control siRNA. The cells were subsequently exposed to sVpr (1 μg/mL) in the presence or absence of PPARβ/δ ligand GW501516 (1 μM) and were harvested to measure PDK4 mRNA and PPARβ/δ protein levels in real-time PCR and in Western blots, respectively.

Figure 1.

Vpr enhances PPARβ/δ-induced transcriptional activity in HepG2, HK2, and A204 cells. HepG2 (A, D, and E), HK2 (B), and A204 (C) cells were transfected with PPARβ/δ- and RXRγ-expressing plasmids together with PPRE-luciferase and pCMV-β-Gal in the presence or absence of the Vpr-expressing plasmid. Cells in A–D were incubated in the presence or absence of 1 μM GW501516. Increasing amounts of Vpr-expressing plasmids were included in transfection in D, and to keep the total amount of DNA constant, plasmid without Vpr was used. Increasing concentrations of GW501516 were employed in E, whereas the total amount of transfected plasmid DNA for Vpr was same. RLU, relative luciferase unit. Bars and circles represent mean ± SD of the luciferase activity normalized by β-galactosidase activity. **, P < .01, as indicated. For all experiments, n = 3.

Reporter assays

Luciferase and β-galactosidase activities were determined, as previously described (6). Twelve hours after transfection, HepG2 cells were treated with fresh culture medium supplemented with PPARα, PPARγ, and/or PPARβ/δ agonists. In a few experiments, HepG2 cells were pretreated for 2 hours with 16 μM PPARβ/δ antagonist GSK0660,10 μM PPARβ/δ antagonist GSK3787, or 0.3 μM etomoxir. After the antagonist treatment, the PPARβ/δ agonist GW501516 was added to the culture medium for 24 hours. Treated cells were then lysed using a lysis buffer (Promega, Madison, Wisconsin). Luminescence analyses were performed with a luminometer (EG&G Berthold, Bad Wildbad, Germany).

Measurement of PDC activity

PDC activity was measured using a radiometric assay, with modification, as described by Constantin-Teodosiu et al (17). For measurement of the active PDC activity (dephosphorylated form), HepG2 cell pellets were homogenized in a buffer containing 200 mM sucrose, 50 mM KCl, 5 mM EGTA, 50 mM Tris-HCl (pH 7.8), 50 mM sodium fluoride (NaF, a nonspecific serine/threonine phosphatase inhibitor of PDP1–2), 50 mM sodium pyrophosphate, 5 mM dichloroacetate (a PDK1–4 inhibitor), and 0.1% Triton X-100. Total PDH activity (combination of phosphorylated and dephosphorylated form) was measured, and HepG2 cell pellets were homogenized in the buffer described above but lacking NaF, sodium pyrophosphate, and Triton X-100. The enzymatic reaction was initiated by addition of pyruvate (1 mM), whereas the reaction was terminated 10 minutes thereafter by adding perchloric acid (1 M). An acetyl-CoA standard curve was generated in parallel with each assay. For the measurement of acetyl-CoA content, acetyl-CoA present in the samples was first converted to [¹⁴C]citrate, which was further separated from unreacted radiolabeled substrate by using the Dowex resin (50WX8, 100–200 mesh). Radioactivity of produced [¹⁴C]citrate was measured by using a scintillation counter, and the amount of acetyl-CoA was determined. All measurements were performed in duplicate from a set of 6 samples in each group. Enzymatic activities of active and total PDCs were calculated as micromoles per gram protein per minute. Figure 4 shows normalized data as compared with enzymatic activity in vector without any treatment.

Figure 4.

Vpr suppresses the active but not total PDC activity in HepG2 cells. HepG2 cells were transfected with Vpr-, PPARβ/δ-, and RXRγ-expressing plasmids. The transfected HepG2 cells were treated for 2 hours with PPARβ/δ inhibitor GSK0660 (16 μM) before addition of 1 μM PPARβ/δ agonist GW501516. A, PDC activity in the presence or absence of GW501516 and /or GSK0660. B, Total PDC activity (both dephosphorylated and phosphorylated form). Bars represent mean ± SD values. ***, P < .001, as indicated; n.s., not significant. For all experiments, n = 6.

Western blot analysis

HepG2 cell pellets were lysed in 1× RIPA lysis buffer (Millipore, Billerica, Massachusetts) containing one Complete Mini tablet of protease inhibitor cocktail (Roche Applied Bioscience, Indianapolis, Indiana) per 10 mL 1× RIPA buffer. Protein concentrations were determined using the bicinchoninic acid protein assay kit (Pierce Biotechnology, Rockford, Illinois). Identical amounts (20 μg) of cell lysates were run on 4% to 12% Bis-Tris gels (Invitrogen), and proteins were then transferred to the nitrocellulose membrane. Blotting efficiency and position of protein standards were assessed by Ponceau staining. Membranes were blocked by overnight incubation at 4°C with 2% dry milk dissolved in Tris-buffered saline (pH 7.4) containing 0.05% Tween 20 (TBS-T). The membranes were then incubated for 2 hours at room temperature with the primary antibody (1 μg/mL), which recognized either total PDH or PDH phosphorylated at serine 293 (site 1) or serine 300 (site 2). After washing with TBS-T 3 times, the membranes were further incubated for 1 hour at room temperature with the secondary antibody DAKO rabbit antisheep-horseradish peroxidase (1:5000) in TBS-T containing 2% dry milk. Membranes containing primary antibodies (diluted 1:500 in 5% bovine serum albumin [BSA]) against PDP2, PDK2, PDK3, and PDK4 were incubated overnight at 4°C by gentle shaking. After washing with TBS-T 3 times, the membranes were further incubated with the secondary antibody donkey antirabbit-horseradish peroxidase P (diluted at 1:50 000) in Western blots. After incubation with the secondary antibody, the membranes were washed for 1 hour at room temperature with TBS-T containing 2% dry milk. Blotted proteins were finally visualized with the chemiluminescence reaction using the Super Signal West Femto maximum sensitivity substrate (Pierce Biotechnology). They were then exposed to x-ray films, and signal intensity of each protein band was measured with the Image J software (National Institutes of Health, Bethesda, Maryland).

Real-time quantitative PCR

Total RNA was purified from HepG2 cells using the RNeasy Mini kit (QIAGEN, Valencia, California), treated with deoxyribonuclease (Promega), and reverse transcribed to cDNA with the TaqMan reverse transcription reagents (Applied Biosystems, Foster City, California).

For reverse transcription-linked quantitative PCR, the following primers were used: PDK2, forward primer 5′-CCGTCTCCAACCAGAACATCC-3′ and reverse primer 5′-CGTAGGCATCTTTGACCACCTC-3′; PDK3, forward primer 5′-TGG GATTGGTTCAGAGTTGG-3′ and reverse primer 5′-CGGAA AGAGATGCGGTTG-3′; and PDK4, forward primer 5′-AGAGCCTGATGGATTTGGTG-3′ and reverse primer 5′-GCTTGGGTTTCCTGTCTGTG-3′. Primers for ACAA1 (PPH13327A-200), ACAA2 (PPH13528A-200), CPT1A (PPH15298E-200), and GAPDH (PPH00150E-200) were purchased from SA Biosciences/QIAGEN. Real-time PCRs were performed in triplicate using the SYBR Green PCR Master Mix in the 7500 real-time PCR system (Applied Biosystems), as described previously (6).

Glutathione S-transferase pull-down assay

Glutathione S-transferase (GST) pull-down assay was performed as reported previously (6). Briefly, ³⁵S-labeled wild-type Vpr and a Vpr mutant defective in leucines 64, 67, and 68 were generated by in vitro transcription and translation reaction with wheat germ extract (Promega) and pCDNA3-Vpr and pCDNA3-VprL64,67,68A as templates, respectively. They were tested for interaction with the GST-fused full-length or LBD of mouse PPARβ/δ immobilized on glutathione-Sepharose beads in a buffer containing 50 mM Tris-HCl (pH 8.0), 50 mM NaCl, 1 mM EDTA, 0.1% Nonidet P-40, 10% glycerol, and 0.1 mg/mL BSA at 4°C for 1.5 hours. After vigorous washing, proteins were eluted and separated on 4% to 20% SDS-PAGE gels. Approximately 3% total input of labeled Vpr was loaded as a control.

Chromatin immunoprecipitation assay

HepG2 cells were transfected with pCDNA3-Vpr together with pCDNA3-PPARβ/δ and pCDNA3-RXRγ using Lipofectamine 2000 reagent. Twenty-four hours after transfection, cells were treated with 1 μM GW501516 or vehicle (dimethylsulfoxide). After an additional 24 hours, cells were treated with 1% formaldehyde for 10 minutes at 37°C and were processed for chromatin immunoprecipitation (ChIP) assay using a ChIP kit (EMD Millipore, Billerica, Massachusetts). Samples containing DNA/protein complexes were incubated overnight with polyclonal rabbit anti-Vpr antibody (6), anti-PPARβ/δ antibody, or rabbit control IgG (Santa Cruz Biotechnology). Immune complexes were collected by adding protein A-agarose/salmon sperm DNA (EMD Millipore), and cross-linked DNA and bound proteins were uncoupled by heating at 65°C for 4 hours. The promoter region (−2876 to −2617) of the human PDK4 gene, which contains a PPRE (located at −2720 to −2708), and its region containing the transcription start site (TSS) (−207 to +91) were amplified quantitatively in SYBR Green real-time PCR (Applied Biosystems) by using corresponding primer pairs (PPRE, forward 5′-GTATGTGTACTGGGGGGAC-3′ and reverse 5′-CAGATGGCTCTTTTCGTTCC-3′; TSS, forward 5′-CCGCCTTCATCTTGACGCCC-3′ and reverse 5′-CCAAGTTCCAGTGACTCCTCC-3′) (18), SYBR Green PCR Master Mix (Applied Biosystems) and a StepOnePlus real-time PCR system (Applied Biosystems). Obtained threshold cycle values of ChIP samples were normalized for those of corresponding inputs, and their relative precipitation was demonstrated as fold precipitation against the baseline (samples obtained from cells in the absence of Vpr transfection and GW501516 treatment).

Measurements of oxygen consumption rates in isolated mitochondria

Mitochondria from HepG2 cells were isolated by differential centrifugation (19). Measurement of mitochondrial oxygen consumption rates was performed at 25°C in a chamber (600 μL) connected to a Clark-type O₂ electrode (Instech, Plymouth Meeting, Pennsylvania) and the O₂ monitor (model 5300; YSI, Yellow Springs, Ohio). In this chamber, isolated mitochondria were incubated in the respiration buffer containing 120 mmol/L KCl, 5 mmol/L MOPS, 1 mmol/L EGTA, 5 mmol/L KH₂PO₄, and 0.2% (vol/wt) BSA. After addition of glutamate/malate (10/2 mmol/L), the state-3 oxygen consumption was measured by addition of ADP (0.5 mmol/L). The uncoupled oxygen consumption was finally measured by adding 50 μmol/L of the mitochondrial uncoupler DNP.

Measurements of oxygen consumption rates in live intact cells

The XF24 extracellular flux analyzer (Seahorse Bioscience) was used to evaluate cellular oxygen consumption rates (OCRs) according to the company's directions. Briefly, HepG2 cells were seeded in XF24 well plates purchased from Seahorse Bioscience, at the density of 4.0 × 10⁴ cells per well (surface area 0.33 cm²) in 100 μL regular culture medium and were incubated overnight at 37°C. On the following day, the cells were either treated with sVpr or transfected with the proportionally reduced amounts of the indicated control or Vpr-expressing plasmids based on lower surface area of the XF24 plates as compared with the 6-well plate used in transfection experiments. After the transfection, the cells were incubated for 2 hours with the PPARβ/δ antagonist GSK0660 (16 μM), the PPARβ/δ antagonist GSK3787 (10 μM) or the carnitine palmitoyltransferase-1 (CPT1) inhibitor etomoxir (0.3 μM) followed by 24 hours treatment with the PPARβ/δ agonist GW501516 (1 μM). Before OCR measurements, the experimental XF24 plate containing the cells was washed with bicarbonate-free DMEM assay medium (Seahorse Bioscience) containing 25 mM glucose and 1 mM sodium pyruvate, and the cells were preincubated for 1 hour at 37°C without a CO₂ supply in 625 μL assay medium.

Before OCR measurements, the XF24 was calibrated using a calibration cartridge according to the company's directions. After the calibration, baseline OCR measurements were conducted in the test plate for 4 min. The averages of four baseline rates were calculated, and the results were demonstrated as percentages of control or relative OCR obtained by comparing experimental OCR to the OCR in wells containing vector without any treatment.

Measurements of β-hydroxylacyl CoA dehydrogenase activity

β-Hydroxylacyl CoA dehydrogenase (HADH) activity was measured as described previously (20). Briefly, HepG2 cell pellets were homogenized in 1:10 wt/vol AMP-activated protein kinase homogenization buffer (50 mM Tris-HCl, 1 mM EDTA, 10% glycerol [wt/vol], 0.2% Brij-35). After homogenization, the cell lysates were microfuged at 8000 rpm for 30 minutes at 4°C. To a 96-well UV plate, 260 μL imidazole (50 mM, pH 7.4), 20 μL reduced nicotinamide adenine dinucleotide (0.15 mM), and 10 μL cell lysate were added per well. The absorbance was followed at 340 nM for 5 minutes with 40-second intervals both before (baseline) and after adding 10 μL of 3 mM acetoacetyl CoA.

Statistical analysis

Data are presented as the mean ± SD. All experiments were performed using a minimum of triplicate wells and were repeated at least 3 times; a representative experiment is shown. PDC activity was measured in duplicate from 6 different samples in each group. Data were analyzed with ANOVA, and posttesting involved comparison of selected groups using the Bonferroni test (Prism; GraphPad, San Diego, California). P values <.05, <.01, and <.001 were considered as statistically significant.

Results

We first tested the effect of GW501516, a PPARβ/δ agonist, on the transcriptional activity of a PPRE-driven reporter in human hepatoma HepG2, human renal proximal tubular HK2, and human rhabdomyosarcoma A204 cells. In the absence of exogenous PPARβ/δ and RXRγ expression, GW501516 resulted in some stimulation of transcriptional activity of this promoter only in the HepG2 and HK2 cells. However, after expression of both receptors, a greatly enhanced GW501516 stimulation of the PPRE-driven promoter activity was seen in all 3 cell lines (Supplemental Figure 1, published on The Endocrine Society's Journals Online web site at http://mend.endojournals.org), consistent with our previous report (6).

We next examined the effect of Vpr on PPARβ/δ-induced transcriptional activity in the presence of PPARβ/δ and RXRγ overexpression. In all 3 cell types, Vpr significantly increased GW501516-stimulated transcriptional activity of the PPRE-driven promoter, indicating that Vpr acts as an enhancer of PPARβ/δ-induced transcriptional activity (Figure 1, A–C). Furthermore, this effect exhibited a dose-response relationship for both Vpr (Figure 1D) and GW510516 (Figure 1E) in HepG2 cells. Therefore, in subsequent experiments, HepG2 cells were transfected with 0.2 μg/well of Vpr-expressing plasmid and treated with 1 μM/well of the PPARβ/δ agonist GW501516.

To evaluate the role of the Vpr nuclear receptor coactivator domain (L64XXL67L68) in enhancing PPARβ/δ-mediated transcriptional activity, we transfected HepG2 cells with plasmids expressing wild-type or coactivator motif-defective mutant Vpr L64A or triple-mutant Vpr L64,67,68A, in the presence of PPARβ/δ- and RXRγ-expressing-plasmids and treated them with 1 μM GW501516 for 24 hours (Figure 2). Wild-type Vpr transfection was again associated with a significant increase in PPARβ/δ-mediated transcriptional activity, which was lacking after transfection with mutant VprL64A or VprL64,67,68A. These results suggest, that an intact LXXLL coactivator motif is necessary for Vpr to enhance PPARβ/δ-induced transcriptional activity.

Figure 2.

Coactivator activity-defective Vpr mutants do not enhance PPARβ/δ-induced transcriptional activity in HepG2 cells. HepG2 cells were transfected with PPARβ/δ- and RXRγ-expressing plasmids together with PPRE-luciferase and pCMV-β-Gal in the presence or absence of the wild-type (WT) Vpr-expressing plasmid or mutant plasmids for VprL64A or VprL64,67,68A and were incubated in the presence or absence of 1 μM GW501516. RLU, relative luciferase unit. Bars represent mean ± SD values of the luciferase activity normalized by β-galactosidase activity. **, P < .01, as compared with vector treated with GW501516; n.s., not significant. For all experiments n = 3.

PPARα, PPARγ, and PPARβ/δ bind to the same PPRE sequence, yet they elicit distinct effects in response to specific or common ligands (21). WY14643 has >1000-fold selectivity for PPARα over PPARβ/δ and PPARγ (22), GW501516 has >1000 selectivity for PPARβ/δ over PPARα and PPARγ (23), and GW1929 has >1000-fold selectivity for PPARγ over PPARα and PPARβ/δ (24) To address the issue of response specificity, we transfected HepG2 cells with PPARβ/δ and RXRγ expression plasmids and a PPRE promoter-driven luciferase reporter and treated the cells with 15dPJ2 or WY14643 (PPARα agonists), GW1929 (a PPARγ agonist), or GW501516 (a PPARβ/δ agonist). As shown in Figure 3A, Vpr in the presence of the PPARβ/δ receptor agonist GW501516 stimulated promoter activity, indicating its specificity for the PPARβ/δ receptor. In the same experiment, GW1929, a PPARγ-specific ligand, showed a weaker transcriptional response to the PPARβ/δ receptor, whereas 15dPJ2 or WY1463 did not stimulate PPARβ/δ promoter activity. In experiments to further define PPARβ/δ receptor specificity, we found that GSK0660, a PPARβ/δ-specific antagonist, completely abolished the ability of Vpr to enhance PPARβ/δ transcriptional activity (Figure 3B).

Figure 3.

Vpr enhances PPARβ/δ-agonist–induced transcriptional activity of the PPAR-driven promoter, whereas PPARβ/δ antagonism, either by compounds or by knocking down this receptor, abolishes such enhancement in HepG2 cells. A, HepG2 cells were transfected with PPARβ/δ- and RXRγ-expressing plasmids together with PPRE-luciferase and pCMV-β-Gal in the presence or absence of the Vpr-expressing plasmid and were incubated with 1 μM of indicated PPAR agonists. B, Transfected HepG2 cells were treated with 1 μM GW501516 in the presence or absence of 16 μM of the PPARβ/δ antagonist GSK0660. RLU, relative luciferase unit C, HepG2 cells were transfected with PPARβ/δ-specific siRNA or a negative control siRNA in the presence or absence of (1 μg/mL) sVpr and/or 1 μM PPARβ/δ agonist GW501516. D, HepG2 cells were transfected with PPARβ/δ-specific siRNA or a negative control siRNA, and PPARβ/δ protein levels along with those of β-actin were examined in Western blots by using their specific antibodies. Bars represent mean ± SD values of the luciferase activity normalized by β-galactosidase activity (A and B) and of PDK4 mRNA and protein levels normalized to those of β-actin in C and D. A representative gel image is shown in the top of D. **, P < .01; ***, P < .001, compared with the condition indicated; n.s., not significant. For all experiments, n = 3.

Activity of the mitochondrial PDC, composed of PDHE1 (E1α and E1β), -E2, and -E3 enzymes, is regulated by 4 PDKs (PDK1–4) and 2 PDPs (PDP1–2) (16). PDC acts as an energetic switch, which when activated, directs cells to use glucose and glucose precursors in preference to lipids. PDK phosphorylates PDHE1α at serine 293 and thereby suppresses its enzymatic activity. PPARβ/δ is known to stimulate expression of PDK4 (25). We therefore examined whether PPARβ/δ is required for Vpr-mediated induction of endogenous PPARβ/δ-responsive PDK4 mRNA by knocking down this receptor in HepG2 cells. PPARβ/δ agonist GW501516 strongly stimulated PDK4 mRNA expression, and addition of sVpr into media further enhanced its mRNA expression in these cells, in agreement with our previous report (6) (Figure 3C). After transfection with PPARβ/δ-specific siRNA, there was a significant reduction in PDK4 mRNA expression in the presence of GW501516 and attenuated sVpr-induced enhancement of PDK4 mRNA expression (Figure 3C). Transfection of these cells with PPARβ/δ-specific siRNA strongly reduced PPARβ/δ protein levels (Figure 3D). Taken together, these results indicate that Vpr enhanced the transcriptional activity of PPARβ/δ.

These results prompted us to examine the influence of Vpr on the activity of the PDC in HepG2 cells. Active PDC (i.e., PDC containing the dephosphorylated form of PDH) was 47% suppressed in the Vpr-transfected cells as compared with vector-transfected cells (Figure 4A). Of interest, GSK0660, a PPARβ/δ antagonist, abolished the negative effect of Vpr on PDC activity, suggesting that Vpr inhibits PDC activity through the PPAR pathway. To confirm this, vector- or Vpr-transfected HepG2 cells were treated with the PPARβ/δ agonist GW501516. In these experiments, PDC activity was further reduced by GW50516 treatment of Vpr-transfected cells when compared with control, and PDC activity was fully restored after treatment with the PPARβ/δ antagonist GSK 0660 (Figure 4A). In contrast to their effects on active PDC, neither GW50516 nor Vpr or their combination altered the activity of total PDC (PDC containing both dephosphorylated and phosphorylated forms of PDH) (Figure 4B).

PPARβ/δ is known to suppress PDC activity by stimulating the expression of PDK4 (26), which phosphorylates serine 293 of PDHE1α (27). Hence, we next examined the effect of transfected Vpr on the expression of PDK4 protein and phosphorylation of PDHE1α in HepG2 cells. Representative immunoblots are shown in Figure 5. GW501516 stimulated PDK4 protein expression, and Vpr further increased GW501516-stimulated PDK4 protein expression (Figure 5A). Consistent with the decrease in PDC activity shown in Figure 4A, GW501516 increased phosphorylated PDHE1α, and Vpr further enhanced the GW501516-stimulated increase of PDHE1α phosphorylation as shown by immunoblotting with a specific phosphoserine 293 antibody (Figure 5C). This was accompanied by no change in total PDH E1α protein expression (Figure 5B). Vpr also did not influence phosphorylation of PDH at serine 300 (site 2) (data not shown). Furthermore, neither GW501516 nor Vpr affected the levels of PDP2, a major isoform of PDP that dephosphorylates PDH in HepG2 cells (data not shown).

Figure 5.

Vpr enhances PPARβ/δ agonist GW501516-induced PDK4 protein expression and phosphorylation of PDHE1α at serine 293 in HepG2 cells. HepG2 cells were transfected with Vpr-expressing plasmid and were incubated in the presence or absence of 1 μM GW501516 as indicated. Cell lysates were subjected to SDS-PAGE and transferred to nitrocellulose membrane, followed by incubation with anti-PDK4, anti-PDHE1α, anti–phospho-PDHE1α (Ser293) and anti-voltage-dependent anion channel (VDAC) (loading control) antibody as indicated. The relative protein expression, normalized with VDAC, was quantified by densitometry and are presented: A, PDK4; B, PDHE1α; C, phosphorylated PDHE1α (Ser293). Bars represent mean ± SD values of band density of indicated proteins corrected with that of control. *, P < .05; **, P < .01; ***, P < .001, compared with the two conditions indicated; n.s., not significant. For all experiments, n = 3.

We also examined the effects of Vpr on mRNA expression of PDK2, PDK3, and PDK4 in HepG2 cells incubated in the presence or absence of GW501516. In agreement with the results at the protein level, Vpr (either plasmid-expressed Vpr or sVpr) enhanced GW501516-induced PDK4 mRNA expression (Figure 6, A and B), but neither GW501516 nor Vpr affected the mRNA levels of PDK2 and PDK3 in these cells (data not shown). Collectively, the observations that Vpr enhances PPARβ/δ-stimulated induction of PDK4 mRNA (Figure 6, A and B) and protein (Figure 5A) expression as well as PDK4-mediated phosphorylation (Figure 5C) and inactivation of PDH (Figure 4A), provide a potential explanation for hyperlactatemia in HIV-1-infected patients (28).

Figure 6.

Transfected Vpr and sVpr enhanced PPARβ/δ-induced PDK4, CPT1, and ACAA2 mRNA expression in HepG2 cells. HepG2 cells were transfected with Vpr-expressing plasmid (A, C, and D) or treated with sVpr protein (B), in the presence or absence of 1 μM GW501516 and mRNA expression of PDK4 (A and B), CPT1 (C), and ACAA2 (D) was examined in quantitative RT-PCR. Bars represent mean ± SD values of PDK4 mRNA expression normalized by β-actin mRNA, and CPT1 and ACAAA2 were normalized by GAPDH mRNA. *, P < .05; **, P < .01; ***, P < .001, compared with the conditions indicated. For all experiments, n = 3.

Because Vpr strongly regulated PPARβ/δ transcriptional activity on the PDK4 gene and we previously reported its physical interaction with PPARγ (6), we examined physical interaction of this viral protein on PPARβ/δ. As expected, wild-type Vpr, but not a Vpr mutant lacking in leucines located at amino acid positions 64, 67, and 68 (hence defective in the LXXLL motif), directly interacted with the LBD of human PPARβ/δ in a ligand-dependent fashion (Figure 7A). In ChIP assays employing anti-Vpr antibody along with anti-PPARβ/δ and control antibodies, Vpr was attracted to the reported PPRE of the PDK4 gene also in a ligand-dependent fashion in HepG2 cells (Figure 7B). These results indicated that Vpr physically interacts with PPARβ/δ in a ligand-dependent fashion both in vitro and in vivo, further supporting the hypothesis that Vpr potentiated PPARβ/δ-induced transcriptional activity by functioning as its coactivator.

Figure 7.

Vpr interacted with PPARβ/δ in vitro and in vivo. A, Vpr physically interacted with the LBD of PPARβ/δ in a GST pull-down assay. In vitro translated and labeled Vpr was incubated with bacterially produced full-length (FL) or the LBD of the human PPARβ/δ in the presence or absence of 1 μM GW501516. Samples were run on 4% to 20% SDS-PAGE gels. Top and bottom panels indicate results for the wild-type Vpr and a Vpr mutant containing L64, 67, and 68A replacement, respectively. B, Vpr was attracted to the PPRE of the PDK4 gene in an agonist-dependent fashion in HepG2 cells. HepG2 cells were transfected with Vpr-expression plasmids and treated with 1 μM GW501516. Cells were then treated with 1% formaldehyde for 10 minutes to cross-link the protein-DNA complex, and ChIP assay was performed with normal rabbit IgG (control antibody), anti-Vpr, or anti-PPARβ/δ antibody. DNA fragment containing the reported PPRE or the transcription start site (TSS) was amplified by using their specific primers in the SYBR Green-based real-time PCR, and fold precipitation was calculated against the baseline (samples obtained from cells in the absence of Vpr transfection and GW501516 treatment) after correcting each threshold cycle value with that of corresponding input. **, P < .01, compared with the conditions indicated. For all experiments, n = 3.

Energy expenditure is increased with HIV infection, even among asymptomatic patients on antiretroviral therapy (29). This could be due to a variety of mechanisms, including altered mitochondrial substrate oxidation. Because inhibitory effects of Vpr on PDH activity could diminish the rate of carbon flux from glycolysis into the tricarboxylic acid (TCA) cycle, and because PPARβ/δ agonism could increase fatty acid oxidation (30) and possibly mitochondrial uncoupling (31), we determined whether Vpr could mediate changes in cellular respiration. We examined the effect of Vpr on cellular respiration and oxygen consumption in isolated mitochondria from HepG2 cells. Vpr slightly increased state-3 (ADP-stimulated) respiration (OCR) by 16% and increased GW501516-stimulated state-3 OCR by 71% (Figure 8A). Similar effects were found for maximum OCR measured in the presence of the uncoupler DNP as shown in Figure 8B. These observations in isolated mitochondria were extended to intact HepG2 cells using the Seahorse Bioscience device. Consistently, basal OCR was higher in Vpr-transfected cells and increased further with GW501516 treatment. These effects were blocked with GSK3787, a specific potent PPARβ/δ antagonist, indicating that HIV-1 Vpr regulates mitochondrial respiration through PPARβ/δ (Figure 8, C and D).

Figure 8.

Vpr increases OCR in both the isolated mitochondria and in the cultured HepG2 cells. Increase in oxygen consumption rate can be reduced by CPT1 inhibitor etomoxir and PPARβ/δ antagonist GSK3787. HepG2 cells were transfected with Vpr-, PPARβ/δ-, and RXRγ-expressing plasmids and were incubated in the presence or absence of 1 μM GW501516. In the matched control wells, cells were treated with either the CPT1 inhibitor etomoxir or the PPARβ/δ antagonist GSK3787. A and B, Vpr effect on state-3 mitochondria respiration (A) and the maximum O₂ consumption (B) in mitochondria isolated from HepG2 cells. C–E, The effect of Vpr on the enhancement of PPARβ/δ-induced OCR and the action of GSK3787 (C and D) and etomoxir (E). Bars represent mean ± SD values of percent increase compared with baseline *, P < .05; **P < .01; ***P < .001, as indicated; n.s., not significant. For all experiments, n = 3.

To further understand the mechanism whereby Vpr increases resting energy expenditure and oxygen consumption, we measured β-oxidation–dependent oxygen consumption in transfected HepG2 cells by treatment with etomoxir, an inhibitor of CPT1, the first and rate-limiting enzyme involved in the entry of long-chain fatty acids into mitochondria for β-oxidation (32). As shown in Figure 8E, etomoxir inhibited the GW501516-stimulated increase in OCR, supporting the proposed crucial role for the PPARβ/δ in β-oxidation of fatty acids. Importantly, etomoxir abrogated the Vpr-induced GW501516-stimulated OCR increase (Figure 8E), suggesting that this effect is β-oxidation–dependent. Consistent with this concept, Vpr also increased mRNA expression of CPT1 (Figure 6C) as well as another β-oxidation enzyme, acetyl CoA acyltransferase-2 (ACAA2) (Figure 6D), whereas it did not affect the expression of ACAA1 (data not shown).

We next measured what effect Vpr had on the activity of HADH, the third reaction in fatty acid β-oxidation that is specific for the l-stereoisomer of hydroxylacyl-CoA. As shown in Figure 9, the activity of HADH increased significantly when cells were either transfected with Vpr or treated with sVpr and GW501516 (Figure 9, A and B). Remarkably, the Vpr- and GW501516-stimulated increase in HADH activity was abrogated in cells pretreated with the PPARβ/δ antagonist GSK0660 (Figure 9C), suggesting that the effect of Vpr on β-oxidation is PPARβ/δ-dependent.

Figure 9.

Vpr enhances HADH enzymatic activity in PPARβ/δ-transfected HepG2 cells. A–C, HepG2 cells were transfected with Vpr-expressing plasmid (A and C) or treated with sVpr protein (B) in the presence or absence of 1 μM PPARβ/δ agonist GW501516 or 16 μM PPARβ/δ antagonist GSK0660 (C). Both transfected Vpr and sVpr protein enhanced PPARβ/δ agonist GW501516-induced HADH activity (A and B), which was abrogated in the presence of PPARβ/δ antagonist GSK0660 (B). Bars represent mean ± SD. *, P < .05; **, P < .01; ***, P < .001, as indicated; n.s., not significant. For all experiments, n = 3.

Discussion

The principal findings of this study are as follows: in HepG2 cells, 1) Vpr enhances PPARβ/δ-induced transcriptional activity, leading to increased expression of PDK4 mRNA and protein; 2) Vpr physically interacts with the PPARβ/δ LBD through its LXXLL motif in an agonist-dependent fashion; 3) Vpr increases PDH phosphorylation, resulting in its inactivation, which would reduce pyruvate carbon utilization by mitochondria for oxidation; 4) Vpr stimulates transcription of the fatty acid transport genes CPT1 and ACAA2 and increases the activity of HADH, which is associated with a fatty acid-dependent (etomoxir-inhibited) increase in cellular oxygen consumption. Taken together, these findings suggest that Vpr, acting through PPARβ/δ coactivation, shifts cellular fuel preference away from carbohydrates and toward fatty acids and stimulates mitochondrial function and oxygen consumption. The enhanced efficiency of fatty acid oxidation and oxygen consumption via this mechanism could contribute to the increased energy expenditure phenotype of HIV-1-infected patients (29).

One consequence of the cascade of Vpr effects is that pyruvate derived from anaerobic glycolysis is less efficiently used as an oxidative substrate and, hence, is likely shunted toward lactate production. This might contribute to the propensity for lactic acidosis observed in HIV-1-infected patients, although mitochondrial toxicity of nucleoside/nucleotide reverse transcriptase inhibitors is a major contributor to this phenomenon (33). Because Vpr is a circulating molecule and can transduce noninfected cells directly (34), these findings may have relevance for tissues and organs that are either not susceptible or less susceptible to HIV-1 infection, such as liver, muscle, and adipose tissue. In transfected cell experiments outlined here, Vpr levels in the cell layers were estimated at 0.2 μg/mg total protein (Supplemental Figure 2), whereas plasma Vpr level in HIV-1-infected patients are estimated at 10 ng/mL, approximately equivalent to 0.1 ng/mg of plasma protein (16). The higher levels in cell culture may model tissues in vivo, where infected cells release Vpr, which enters adjacent uninfected cells, or may represent the limitations of transient transfection experiments to quantitatively model in vivo conditions.

We found that Vpr increases mitochondrial oxygen consumption in vitro (isolated mitochondria) and elevates cellular O₂ uptake in vivo (cultured cells) in part by enhancing the biological action of PPARβ/δ. It has been shown that Vpr directly acts on mitochondria and induces apoptosis (2), and in skeletal muscle, GW501516 (at higher doses) can have a direct PPARβ/δ-independent effect on mitochondrial uncoupling (35). Recently, Wu et al (36) reported in CD4⁺ and/or CD8⁺ cells that RNA transcripts relating to enzymes of oxidative phosphorylation and the TCA cycle are up-regulated in viremic patients on antiviral therapy compared with either nonviremic patients or long-term nonprogressors. These data suggest that our finding in HepG2 cells, including the role of Vpr and PPARβ/δ to stimulate mitochondrial function, may have broader relevance to other tissues.

Taken together, our results shed light on the role of viral gene products in regulating energy balance in HIV-1-associated disease. For example, a recent study linked increased metabolism of branched-chain amino acids to activation of PPARβ/δ, resulting in muscle atrophy (37).

We found that, in the presence of a PPARβ/δ agonist, Vpr enhances the transcriptional activity of a PPRE-driven promoter in 3 human cell lines (liver, kidney, and muscle), and this requires a coactivator domain of Vpr. These results suggest that Vpr acts as a coactivator of PPARβ/δ in a fashion similar to its effect on the glucocorticoid receptor (5). We previously reported that Vpr physically interacts with PPARγ at its LBD through its LXXLL domain and repressed PPARγ transcriptional activity in 3T3-L1 cells transfected with PPARγ and tested against its specific agonist ciglitizone (6). Data presented in Figure 3A also show some transcriptional activity of PPARγ agonist GW1929 in PPARβ/δ-transfected HepG2 cells. It is likely that Vpr interacts with the LBD of PPARβ/δ, similar to its interaction with this subdomain of PPARγ. We do not know the molecular mechanisms by which Vpr differentially regulates the transcriptional activity of these 2 PPARs, but cofactors attracted differently to these receptors in the transcriptional complex could be part of the explanation. Future work might examine whether Vpr interacts with these cofactors and, if so, how these cofactors contribute to Vpr-mediated differential regulation of the transcriptional activities of PPAR family members.

Despite sharing a common DNA target sequence (PPRE), PPARα, PPARγ, and PPARβ/δ have distinct effects on cellular energetics and fuel substrate preferences. Tissue specificity and cell specificity of PPAR activity may depend upon several factors, including differential PPAR expression, availability of endogenous ligands specific to each PPAR, and differential integration of PPARs into particular transcriptional complexes (38) This may explain why in mouse 3T3-L1 adipocytes, PPARγ increases fat synthesis and storage, whereas PPARβ/δ increases fatty acid β-oxidation, flux through the TCA cycle, and oxidation of branched-chain fatty acids (37). Disorders of glucose and lipid metabolism have been noted in patients with HIV-1 infection before the advent of antiviral therapy, suggesting a role for the virus itself or its sequelae, and the prevalence and severity of several aspects of these metabolic disorders have been increased by antiviral therapies (39, 40). Our findings suggest that Vpr via interactions with PPARγ and PPARβ/δ in distinct tissues may contribute to the pathogenesis of these metabolic disorders.

In summary, HIV Vpr is associated with multiple disorders of bioenergetics and energy metabolism via PPARβ/δ (Figure 10). Vpr acts as an enhancer of PPARβ/δ-induced transcriptional activity, inducing PDK4 expression, thereby leading to PDHE1α phosphorylation and inactivation of PDC. In addition, Vpr increases CPT1 and ACAA2 mRNA expression and HADH enzyme activity. These heterogeneous effects may contribute to the development of some features of a complex metabolic phenotype in HIV patients, specifically, increased energy expenditure, hyperlactatemia, and muscle atrophy.

Figure 10.

Scheme of carbohydrate and fatty acids metabolism. Arrows represent the flow of the metabolic pathways, and dashed lines indicate the stimulation or inhibition of Vpr/PPARβ/δ on the metabolites in the pathway. The inhibitory effect of Vpr/PPARβ/δ on pyruvate catabolism is mediated by reduced PDC activity. Abbreviation: NADH, reduced nicotinamide adenine dinucleotide.