A Dominant Negative Human Peroxisome Proliferator-Activated Receptor (PPAR) Is a Constitutive Transcriptional Corepressor and Inhibits Signa
http://www.100md.com
内分泌学杂志 2005年第4期
Departments of Clinical Biochemistry (R.K.S., A.J.V.-P., S.O.) and Medicine (M.G., V.K.K.C., S.O.), University of Cambridge, Addenbrooke’s Hospital, Cambridge CB2 2QQ, United Kingdom; Institut National de la Santé et de la Recherche Médicale U.508 (A.M.), Institut Pasteur de Lille, 59019 Lille Cedex, France; Metabolic Research Laboratory (D.W., G.F.G.), Oxford Centre for Diabetes, Endocrinology and Metabolism, Nuffield Department of Clinical Medicine, University of Oxford, Churchill Hospital, Oxford OX3 7LJ, United Kingdom; MRC-Laboratory of Molecular Biology (J.W.R.S.), Cambridge CB2 2QH, United Kingdom
Address all correspondence and requests for reprints to: Robert Semple, Department of Clinical Biochemistry, Addenbrooke’s Hospital, Cambridge CB2 2QR, United Kingdom. E-mail: rks16@cam.ac.uk.
Abstract
Several missense mutations in the ligand-binding domain of human peroxisome proliferator-activated receptor (PPAR) have been described in subjects with dominantly inherited severe insulin resistance associated with partial lipodystrophy, hypertension, and dyslipidemia. These mutant receptors behave as dominant-negative inhibitors of PPAR signaling when studied in transfected cells. The extent to which such dominant-negative effects extend to signaling through other coexpressed PPAR isoforms has not been evaluated. To examine these issues further, we have created a PPAR mutant harboring twin substitutions, Leu459Ala and Glu462Ala, within the ligand binding domain (PPARmut), examined its signaling properties, and compared the effects of dominant-negative PPAR and PPAR mutants on basal and ligand-induced gene transcription in adipocytes and hepatocytes. PPARmut was transcriptionally inactive, repressed basal activity from a PPAR response element-containing promoter, inhibited the coactivator function of cotransfected PPAR- coactivator 1, and strongly inhibited the transcriptional response to cotransfected wild-type receptor. In contrast to PPAR, wild-type PPAR failed to recruit the transcriptional corepressors NCoR and SMRT. However, PPARmut avidly recruited these corepressors in a ligand-dissociable manner. In hepatocytes and adipocytes, both PPARmut and the corresponding PPAR mutant were capable of inhibiting the expression of genes primarily regulated by PPAR, -, or - ligands, albeit with some differences in potency. Thus, dominant-negative forms of PPAR and PPAR are capable of interfering with PPAR signaling in a manner that is not wholly restricted to their cognate target genes. These findings may have implications for the pathogenesis of human syndromes resulting from mutations in this family of transcription factors.
Introduction
NATURALLY OCCURRING MUTATIONS in the nuclear hormone receptor peroxisome proliferator- activated receptor (PPAR) are now well established as the cause of a stereotyped human syndrome of severe insulin resistance, partial lipodystrophy, hypertension, and dyslipidemia, often manifesting in women as polycystic ovary syndrome and preeclampsia (1, 2). The majority of mutations (mut) reported to date lie within the ligand-binding domain (LBD) of the receptor and, when studied in vitro, exhibit dominant-negative activity when coexpressed with their wild-type (wt) counterpart (3, 4). The molecular basis of the dominant-negative activity of mutations in the PPAR LBD has been most closely examined in relation to an artificial double-mutant receptor (Leu468Ala/Glu471Ala) (5). This receptor was found to bind DNA and heterodimerize with retinoid X receptor (RXR) normally. Its marked dominant-negative activity reflected aberrant interaction with nuclear receptor coregulators: release of corepressor molecules such as nuclear receptor corepressor (NCoR) and silencing mediator of retinoid and thyroid receptors (SMRT) was impaired, whereas the ligand-induced recruitment of coactivators such as cAMP response element-binding protein was also greatly diminished. Similar findings have more recently been reported for two of the naturally occurring PPAR mutants (3). It is notable that no naturally occurring mutations within other PPAR isoforms ( and ) have yet been reported.
We have now elected to investigate whether equivalent mutations in another PPAR isoform would also result in a dominant-negative receptor and, if so, determine whether they would do so through entirely analogous molecular mechanisms. Thus, we first created the corresponding mutations in PPAR to those described previously for PPAR and examined the signaling properties of the mutant receptor in detail. In transient transfection assays using an artificial PPAR response element reporter gene construct, we demonstrated that the dominant-negative mutant PPAR failed to mediate ligand-induced transcriptional activation and moreover interfered with the action of not only PPAR-specific ligands acting through cotransfected wild-type PPAR but also a PPAR-specific ligand and a PPAR-specific ligand acting through their cognate receptors.
The availability of dominant-negative mutants of both PPAR and PPAR, which could be adenovirally transduced into differentiated cells, allowed us to extend our investigation of the specificity of dominant-negative activity within this receptor family. By adenovirally transducing adipocyte and hepatocyte cell lines with either the PPAR or PPAR dominant-negative receptors and examining their effects on basal and ligand-induced transcription of canonical target genes for each PPAR isoform, we demonstrated that cross-inhibition of PPAR signaling is also seen for transcriptional responses mediated by endogenous PPARs.
The physiological relevance of these observations is unclear, but if such cross-inhibition occurs in vivo, this suggests that the metabolic syndrome seen in humans with PPAR dominant-negative mutations might represent the composite effects of the receptor mutants on signaling through more than one coexpressed isoform of PPAR.
Materials and Methods
Plasmid constructs
The full-length human PPAR cDNA was cloned by RT-PCR from total HepG2 cell RNA using the following oligonucleotides: sense, 5'-CAG GGT ACC ACC ATG GAC TAC AAA GAC GAT GAC GAC AAG ATG GTG GAC ACG GAA AGC CCA CTC T-3', and antisense, 5'-TGG CGG CCG CTC AGT ACA TGT CCC TGT AGA TCT CC-3'. The sense oligonucleotide contained a Kozak translation signal (ACCATG) and a FLAG epitope (MDYKDDDDK residues). An amplicon of 1450 bp was obtained after 30 cycles (94 C for 1 min, 63 C for 1 min, and 72 C for 1 min 30 sec) with a 1:1 ratio of pfu turbo (Stratagene, La Jolla, CA)/AmpliTaq Gold (PerkinElmer, Norwalk, CT). The PCR product was purified and cloned into pcDNA3.1 (+) (Invitrogen, Paisley, UK) using the KpnI and NotI restriction sites, making pcDNA3-PPARwt. The insert was sequenced to confirm its integrity.
The leucine and glutamate at positions 459 and 462, respectively, were mutated into alanine by PCR site-directed mutagenesis using the Quickchange kit (Stratagene) according to the supplier’s instructions with the following oligonucleotides: sense, 5'-GCT GCG CTG CAC CCG GCA CTG CAG GCG ATC TAC AGG GAC ATG-3', and antisense, 5'-CAT GTC CCT GTA GAT CGC CTG CAG TGC CGG GTG CAG CGC AGC-3' (mutations underlined). After confirmation by direct sequencing, the insert was double digested with KpnI and NotI and ligated into empty pcDNA3.1 (+) vector between KpnI and NotI sites.
The hinge region and LBD (amino acids 167–468) of PPARwt and PPARmut were amplified using the following oligonucleotides: sense, 5'-CAT GAA TTC TCA CAC AAC GCG ATT CGT TT3–3' and antisense 5'-CAC GAA TTC TCA GTA CAT GTC CCT GTA GAT-3' from pcDNA3-PPARwt and pcDNA3-PPARmut, and the PCR product (909 bp) was cloned into the EcoRI site of AASV-VP16 vector to give VP16-PPARwt LBD and VP16-PPARmut LBD. The L468A/E471A PPAR double mutant, human PPAR2, Gal4-NCoR, and Gal4-SMRT (5); peroxisome proliferator response element (PPRE)3-TKLUC (6); UASTKLUC (7); CRBPIITKLUC (8); MALTKLUC, RSV-TR?1, and RSV-RXR (9); retinoic acid receptor (RAR)?2TKLUC and RSV-RAR1 (10) constructs have been described previously. The PcDNA3-PPAR has been described previously (11) and was a gift from Dr. R. Vogel (Merck, Whitehouse Station, NJ). The hemagglutinin-tagged human PPAR- coactivator 1 (PGC1) plasmid was a gift from Dr. A. Kralli (University of Basel, Basel, Switzerland).
Transfection assays
HepG2 cells were maintained in DMEM containing 4.5 g glucose/liter (Sigma, Poole, UK) supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. Cells were split into 24- or 96-well plates and transfected 8 h later in the same medium using Fugene (Roche, Stockholm, Sweden). A ratio of 3 μl Fugene to 1 μg DNA was used for all transfections. The transfection efficiency was estimated by cotransfecting the vector pRL-cytomegalovirus (CMV) (Promega, Southampton, UK) encoding the Renilla luciferase gene. Fifteen hours after transfection, the medium was removed, the cells were washed once with PBS, and fresh medium containing the relevant ligand added for 30 h. The cells were then harvested and both Firefly and Renilla luciferase activities were measured in 20 μl of lysate according to the manufacturer’s instructions (dual luciferase assay, Promega). The following specific ligands were used: PPAR-GW7647 (12) (Sigma), RXR-LG100268 (13) (a gift from Mark Leibowitz, Ligand Pharmaceuticals, San Diego, CA), PPAR-BRL49653 (14) (Alexis, San Diego, CA), and PPAR-L165041 (15) (a gift from Dr. D. Moller, Merck).
Recombinant adenovirus construction
The PPARmut recombinant adenovirus was created using the AdEAsy kit according to the supplier’s instructions (QBiogene, France). Briefly, the human PPARmut cDNA was cloned into pShuttle-CMV between KpnI and NotI restriction sites. The vector was linearized with PmeI and cotransformed with the pAdEasy-1 vector into BJ5183 competent cells by electroporation to effect recombination between the two vectors. Recombinants were then digested with BstXI. DH5 cells were transformed with one positive recombinant clone, and DNA was extracted using the large-construct maxiprep kit (Qiagen, West Sussex, UK). Direct sequencing was used to verify the integrity of the cDNA and the presence of the mutations. The DNA was linearized with PacI and QBI-293A cells plated in 6-well plates with agarose-DMEM overlay were infected with the viral DNA to produce viral particles. Plaques appeared after 14 d and were picked and transferred onto fresh QBI-293A cells. Cytopathic effect (CPE) started to appear after 8 d. A second round of amplification was performed and CPE was complete after 4 d. The viral particles were released by three cycles of freezing/thawing and large-scale amplification performed by QBiogene (Canada).
Use of recombinant adenovirus in cultured cells
Male Wistar rats, fed and housed as described previously (16), were used for the preparation of hepatocytes when they weighed between 200 and 300 g. Hepatocytes were prepared under sterile conditions (17) and were suspended [(0.75–1.0) x 106 cells/ml)] in Waymouth’s medium MB752/1 supplemented with amino acids, antibiotics, and 10% (vol/vol) heat-inactivated FBS. Then 3.0 ml of this suspension was added to culture dishes coated with rat-tail collagen, and after 3 h, the cells became attached as a monolayer. The medium was removed and the cells were washed twice with 3.0ml PBS before addition of 3.0 ml serum-free Waymouth’s medium supplemented with amino acids, antibiotics, 1 μM dexamethasone, 1 mM pyruvate, and 10 mM lactate (supplemented medium). To each plate was added adenoviral storage buffer, green fluorescent protein (GFP)-expressing adenovirus, or PPARmut-expressing adenovirus. Plaque-forming units (PFUs) of each virus (1.26 x 108) were used per plate, giving a multiplicity of infection (MOI) of approximately 60 PFU/cell, in line with previous publications. After 3 h of incubation in virus-containing medium, the cells were washed twice with PBS before addition of supplemented medium containing the relevant concentration of PPAR agonist GW7647 or dimethylsulfoxide. After 16 h RNA was extracted from the cells using the RNEasy kit (Qiagen).
HepG2 experiments were carried out in DMEM containing 4.5 g/liter glucose (Sigma) supplemented with 10% FBS and antibiotics in 12-well plates. Adenoviral vector or vehicle (5 x 106 PFU) was added per well (MOI around 20 PFU/cell), and cells were incubated for 12 h before washing with PBS and replacement with medium containing GW7647 or dimethylsulfoxide. At this stage, infection rates of greater than 95% for the GFP and GFP/PPARmut viruses were verified by fluorescence microscopy of the living cells. After 48 h RNA was extracted using the RNEasy kit.
3T3-L1 preadipocytes (American Type Culture Collection, Manassas, Va) were maintained at less than 70% confluence in DMEM containing 4.5 g/liter glucose supplemented with 10% newborn calf serum, penicillin/streptomycin, and 2 mM glutamine (all Sigma). For each experiment, cells were seeded into 12-well plates and grown until 2 d post confluence. Differentiation medium consisted of the FBS-containing medium supplemented with insulin (5 μg/ml), dexamethasone (0.1 μg/ml), and 3-isobutyl-1-methylxanthine (110 μg/ml). Where cells were to be fully differentiated, this was replaced after 3 d by the FBS-containing medium supplemented with insulin (5 μg/ml) alone and after a further 3 d by the FBS-containing medium alone. A pilot study to assess the validity of adipocyte fatty acid binding protein 4 (aP2) as a specific transcriptional target of PPAR in 3T3-L1 preadipocytes was carried out. The cells were induced to differentiate in the presence of BRL49653or vehicle and RT-PCR used to quantify mRNA levels aP2. In comparison with differentiation medium (DM) alone, BRL49653plus DM resulted in a more than 20-fold increase in aP2 mRNA levels at 48 h. Thus, BRL49653stimulation concomitant with the first 2 d of differentiation was chosen to test the effect of the mutant PPARs. For adenovirus experiments, cells at 2 d post confluence were exposed to maintenance medium containing either adenoviral storage buffer or 2 x 109 PFU of adenovirus per well (MOI around 5000 PFU/cell unless otherwise indicated) for 12 h with agitation. Adenovirus-expressing GFP alone (QBiogene) was used to assess infection rates in conjunction with fluorescence microscopy. The monolayers were then washed with PBS, and cells were incubated in DM containing different concentrations of BRL49653for 48 h before RNA extraction using the RNEasy kit (Qiagen). Oil Red O staining employing a standard protocol was used to assess differentiation in those experiments in which cells were differentiated for 12 d.
Real-time quantitative PCR
On the basis of published work, we identified peroxisomal fatty acyl-coenzyme A oxidase (pFAO), pyruvate dehydrogenase kinase (PDK) 4 and cytochrome IVA1 as robustly responsive targets of PPAR in rat liver and PDK4 and liver type carnitine palmitoyl transferase 1 (CPT1) as targets in human liver and derived cell lines. Primer Express software (version 1.0, PerkinElmer Applied Biosystems, Foster City, CA) was used to design the probes and primers for real-time quantitative PCR shown in Table 1. Five hundred nanograms of total RNA was reverse transcribed for 1 h at 37 C in a 20-μl reaction including 1 x reverse transcription buffer and 100 U Moloney murine leukemia virus reverse transcriptase, 250 ng random hexamers, 1.25 mmol/liter deoxynucleotide triphosphates (all Promega), and the solution was made up to 100 μl. Two microliters of the resulting cDNA were used in each 25 μl PCR, in which 300 nmol/liter of forward and reverse primers and 150 nmol/liter of fluorogenic probe were used. Reactions were carried out at least in duplicate for each sample on an ABI 7700 sequence detection system (PerkinElmer Biosystems) according to the manufacturer’s instructions, and target values were normalized to either glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or 18S rRNA as indicated (rodent and human reagents from PerkinElmer).
TABLE 1. Sequences of primers and probes used for quantitative real-time PCR
Results
Figure 1 shows the amino acid sequence comparison of the distal LBD of human, rat, and Xenopus PPARs. The leucine and glutamate residues that confer on human PPAR dominant-negative activity when mutated into alanine (5) are conserved in all homologs shown. Therefore, we decided to apply this mutational strategy to create a dominant-negative mutant of PPAR. Thus, wild-type PPAR cDNA was amplified and cloned from HepG2 cell cDNA, and Leu459 and Glu462 were replaced by alanine using site-directed mutagenesis to create PPARmut.
FIG. 1. C-terminal 141 amino acids of PPAR, -, and - from human (h), rat (r), and Xenopus (x). Human PPAR residues L459 and E462 (mutated to alanine) and their equivalents in the other PPARs are shaded, and residues conserved in all species are underlined.
The transcriptional activity of PPARmut was then investigated by cotransfecting HepG2 cells with expression vectors encoding RXR, PPARwt, or PPARmut and a reporter construct consisting of three copies of the PPRE of the acyl-coenzyme A oxidase gene coupled to the Firefly luciferase gene (PPRE3-TKLUC), with the Renilla luciferase vector as a transfection control. In the presence of 100 nM GW7647 and LG100268, PPARwt induced luciferase activity by 4-fold (Fig. 2A). The PPARmut, in contrast, exhibited intense suppression of both basal (Fig. 2A, inset) and ligand-stimulated transcription (Fig. 2A). Ligand dose response testing (Fig. 2B) confirmed the complete inability of PPARmut to mediate a transcriptional response to ligand. Thus, the introduction of the two mutations in the ligand-binding domain of PPAR was sufficient to abolish its transactivation capacity.
FIG. 2. Transcriptional activity of PPARwt and PPARmut. HepG2cells were transfected with 300 ng pcDNA3.1, pcDNA3-PPARwt, or pcDNA3-PPARmut, together with 300 ng RSV-hRXR, 60 ng PPRE3-TKLUC reporter, and 20 ng pRL-CMV. The activity is expressed relative to the wild-type maximum (100%): with 100 nM of both GW7647 and L165041 (A) or vehicle (open bars). Inset, Detail of baseline levels.*, P < 0.05; **, P < 0.01 (two-tailed Student’s t test). B, Dose response of reporter activity to equimolar GW7647 and L165041. Results are the mean ± SD of five experiments in duplicate.
Recruitment of transcriptional coregulators
To elucidate the mechanism of this loss of function, we went on to examine the interactions of PPARmut with a panel of transcriptional coregulator molecules. First, corepressor recruitment by PPARwt and PPARmut was assessed in a mammalian two-hybrid assay using the receptor-interacting domains of SMRT and NCoR fused to the Gal4 DNA binding domain (DBD) and the LBD of PPARwt or PPARmut fused to the activation domain of VP16 (Fig. 3). HepG2 cells were cotransfected with the Gal4-corepressor chimera, together with either the wild-type or mutant fusion construct and the UAS-TKLUC reporter. In the case of the wild-type receptor, recruitment of neither NCoR (Fig. 3A) nor SMRT (Fig. 3B) could be detected. However, in marked contrast, the PPARmut LBD was able to recruit either NCoR or SMRT in the unliganded state and release them in a dose-dependent manner after addition of GW7647.
FIG. 3. Mammalian two-hybrid assay with nuclear receptor interaction domains of NCoR (A) and SMRT (B). HepG2 cells were transfected with 100 ng of vectors encoding VP16 only, VP16 fused to the wild-type PPAR LBD (VP16-LBDwt), or VP16 fused to the double-mutant PPAR LBD (VP16-LBDmut) together with 100 ng of expression vector encoding the Gal4 DBD fused to the nuclear receptor interaction domains of NCoR (Gal4-NCoR) (A) or SMRT (Gal4-SMRT) (B) and 500 ng of the reporter construct UASTKLUC, with results normalized to Renilla luciferase activity as previously described. The dose response of reporter activity to GW7647 is shown. The activity is expressed relative to the maximum obtained for PPARmut in the absence of ligand (100%). Results are the mean ± SD of four experiments in duplicate.
Second, we explored the capacity of a known nuclear receptor coactivator to potentiate transcriptional activation by PPARwt and PPARmut. HepG2 cells were cotransfected with PPARwt or PPARmut plasmids, the PPRE3-TKLUC reporter plasmid, and incremental amounts of hemagglutinin-tagged hPGC1 expression vector and stimulated with ligand, with results normalized to Renilla luciferase activity. A clear dose response of reporter activity to increasing amounts of hPGC1 was observed both in the presence of the reporter plasmid alone (in which case endogenous PPAR expression presumably accounts for the transcriptional activity) and with overexpressed PPARwt, although it should be noted that basal activity was much higher in the presence of exogenous receptor (Fig. 4). Similar but enhanced transactivation profiles were recorded after the addition of GW7647. Strikingly, the presence of PPARmut virtually abolished transcriptional activity at all concentrations of PGC1 and ligand (Fig. 4).
FIG. 4. Suppression of PGC1-mediated transcriptional activation by PPARmut. HepG2 cells in 96-well plates were cotransfected with 8 ng PPRE3-TKLUC, 3 ng of either pcDNA3-PPARwt or pcDNA3-PPARmut, increasing amounts of plasmid encoding hemagglutinin-tagged human PGC1 as shown, and empty pcDNA3.1 (vector) to balance the total amounts of transfected DNA. Results were normalized to Renilla luciferase activity, and results are the mean ± SD of three experiments, expressed as percentages of the maximal recorded transcriptional activity. DMSO, Dimethylsulfoxide.
Dominant-negative activity of PPARmut
To evaluate the possible dominant-negative activity of the mutant, HepG2 cells were cotransfected with either PPARwt or PPARwt and PPARmut (Fig. 5). The addition of PPARmut reduced the ability of PPARwt to transactivate the luciferase gene by 50%, demonstrating that the mutant possesses dominant-negative activity (Fig. 5A). This effect shows dose responsiveness, with a 75% reduction of activity seen with a 10-fold excess of mutant.
FIG. 5. Dominant-negative activity of PPARmut over PPARwt (A), PPAR (B), and PPAR (C). A, HepG2 cells were transfected with 600 ng pcDNA3-PPARwt or 300 ng pcDNA3-PPARwt plus 300 ng or 3000 ng pcDNA3-PPARmut as indicated. In all cases 60 ng PPRE3-TKLUC and 20 ng pRL-CMV were cotransfected with pcDNA3.1 to equalize the total amounts of transfected DNA. Three hundred nanograms pcDNA3-PPARwt had previously been established to give the same transcriptional activity as 600 ng. Dose responses of reporter activity to GW7647 are shown, the activity being expressed relative to the wild-type maximum obtained with 100 nM GW7647 (100%). B is the same as for A but with pcDNA3-PPARwt, and BRL49653as ligand. C is the same as for A but with pcDNA3-PPAR, and L165041 as ligand. Results were normalized to Renilla luciferase activity, and results are the mean ± SD of five experiments (A and B) or three experiments (C), expressed as percentages of the transcriptional activity in response to 100 nM ligand.
We also tested the effect of the PPARmut on PPAR-mediated transcription. The presence of the thiazolidinedione ligand BRL49653at 100 nM resulted in an 8-fold activation of luciferase activity (Fig. 5B). However, addition of equimolar PPARmut was associated with a 75% decrease in PPAR-mediated transcriptional activity at 100 nM BRL49653 Further suppression was again seen with a 10-fold excess of mutant. Using the same reporter construct, we also demonstrated potent dominant-negative activity of the mutant PPAR over PPAR (Fig. 5C). Thus, in this cotransfection system, there appears to be no selectivity of the inhibitory effect of the PPARmut construct for PPAR-mediated over PPAR- and PPAR-mediated transcriptional activation, albeit in the context of an artificial, minimal PPRE-containing promoter.
To assess whether the dominant-negative activity of the mutant construct is selective or specific for PPAR-mediated responses, we also investigated the effect of coexpressing PPARmut with other nuclear hormone receptors in the presence and absence of their specific ligand. Expression of PPARmut either at the same level or in 5-fold excess had no significant effect on robust transcriptional responses mediated by RAR, thyroid hormone receptor ?1, or RXR in the presence of their cognate ligands, suggesting that its powerful dominant-negative activity is selective for PPAR-mediated responses (Fig. 6).
FIG. 6. Selectivity of dominant-negative activity of PPARmut in vitro. HepG2 cells in 96-well plates were transfected with reporter plasmid and nuclear receptor-encoding plasmid as shown with or without an equal amount or 5-fold excess of cotransfected pcDNA3-PPARmut. In all cases 3 ng pRL-CMV were cotransfected, and empty pcDNA3.1 was added to equalize amounts of transfected DNA. Reporter responses to specific ligand (as shown) were assayed, and results are the mean ± SD of three experiments, normalized to the baseline activity seen in the absence of overexpressed nuclear receptor in eachcase. A, RXR; B, thyroid hormone receptor (TR)?1; C, RAR1. ATRA, All-trans-retinoic acid.
Effect of PPARmut adenoviral expression on native PPAR-responsive genes
Because the PPAR-dependent promoter used hitherto was artificial and possibly unrepresentative of isoform-specific PPAR-driven responses in living cells, we went on to look at the effect of PPARmut expression on well-characterized canonical PPAR-mediated responses in primary rat hepatocytes, HepG2 cells, and murine 3T3-L1 preadipocytes. To this end, we created an adenovirus-based expression vector encoding the PPARmut cDNA and compared its ability to inhibit PPAR-mediated responses with the inhibitory activity of a previously described PPARmut/GFP-coexpressing adenovirus (5). Cells in culture were infected with vehicle, a GFP-expressing adenovirus, PPARmut-expressing adenovirus, or GFP/PPARmut-expressing adenovirus under conditions that have been optimized to permit high levels of infection. The infected cells were subsequently exposed to a range of ligand concentrations (PPAR-GW7647, PPAR-BRL49653, PPAR-L165041), and PPAR target gene mRNA expression was assessed by real-time quantitative PCR.
In rat hepatocytes, PDK4 showed a steep dose-dependent response to PPAR agonist (GW7647) in both uninfected cells and cells infected with GFP-expressing virus, with a 15- to 20-fold induction of mRNA levels (Fig. 7A). In the presence of the PPARmut, not only was induction of expression ablated, but there was also baseline suppression of PDK4 mRNA levels, and indeed they did not approach the uninfected baseline until very high concentrations of ligand were added (Fig. 7A). The PPARmut adenovirus also significantly attenuated the response to the PPAR agonist but to a lesser degree: no suppression of baseline activity was seen, and the peak stimulation of mRNA expression was around 3-fold. Similar patterns were seen for pFAO (fig 7B) and CYPIVA1 (data not shown), although the degree of induction of mRNA was less (8- to 10-fold).
FIG. 7. Effect of PPARmut or PPARmut on endogenous PPAR target genes in primary rat hepatocytes. Primary rat hepatocytes were infected with adenovirus expressing GFP (Ad-GFP), PPARmut (Ad-mut), or PPARmut and GFP (Ad-mut) before stimulation by 16 h of exposure to different concentrations of GW7647, after which mRNA levels of PDK4 (A) or pFAO (B) were measured by RT-PCR and normalized to GAPDH mRNA levels. Representative experiments from three are shown, with error bars representing the SD of duplicates of RT-PCR.
In HepG2 cells, addition of PPAR agonist GW7647 resulted in modest induction of CPT1 (Fig. 8A) and strong induction of PDK4 (Fig. 8B). In both cases, PPARmut and PPARmut attenuated this, most markedly in the case of PDK4 with, once again, a more robust effect seen with the PPARmut-expressing virus.
FIG. 8. Effect of PPARmut or PPARmut on endogenous PPAR target genes in HepG2 cells. HepG2 cells were infected with Ad-mut, Ad-GFP, or Ad-mut and then exposed to GW7647 for 48 h before measurement of PDK4 (A) and CPT1 (B) mRNA levels by quantitative RT-PCR, normalized to GAPDH. The means of three experiments are shown, with error bars representing the SD.
Next, to determine whether the adenoviral constructs were equally capable of attenuating the responsiveness of selective PPAR target genes, we turned to the murine 3T3-L1 preadipocyte cell line. Preliminary experiments established that addition of 100 nM BRL49653to standard proadipogenic DM for 48 h resulted in a 20-fold greater increase in mRNA levels of aP2 than with DM alone, which we thus regarded as a robust PPAR-specific response against which to test the effects of the mutant constructs. This response was dramatically attenuated in the presence of either the PPARmut or PPARmut, with the PPARmut once again exhibiting a greater effect (Fig. 9A). A correlate of this was seen in the failure of 3T3-L1 cells infected with either construct to accumulate lipid in response to standard differentiation conditions (Fig. 9C) (5).
FIG. 9. Effect of PPARmut or PPARmut on endogenous PPAR and PPAR target genes. A, Two-day postconfluent 3T3-L1 preadipocytes were infected with Ad-mut, Ad-GFP, or Ad-mut for 12 h before 48 h of exposure to adipogenic DM and different concentrations of BRL49653 aP2 mRNA levels were then determined by quantitative real-time PCR and normalized to 18S rRNA levels. B, HepG2 cells were infected with Ad-mut, Ad-GFP, or Ad-mut and then exposed to GW7647 for 48 h before measurement of PDK4 mRNA levels by quantitative RT-PCR, normalized to GAPDH. The means of three experiments are shown, with error bars representing the SD. C, Two-day postconfluent 3T3-L1 preadipocytes were infected with no adenovirus, Ad-GFP, or Ad-mut at the range of MOIs shown before adipogenic differentiation as described using a standard protocol. Lipid accumulation was assessed by staining with oil Red O at d 10. Representative images are shown. Maximal infection rates of 90% were achieved at an MOI of 109 Ad-GFP PFU per well (not shown).
Because PDK4 has been shown previously to be a robustly responsive gene to PPAR stimulation as well as PPAR stimulation (18), we finally tested the effect of adenoviral overexpression of either PPARmut or PPARmut in HepG2 cells on this response. Both mutant receptors were found to inhibit the response of PDK4 to stimulation with the specific PPAR agonist L165041 (Fig. 9B), suggesting that the dominant-negative activity of these agents extends also to the action of PPAR.
Discussion
The PPAR family consists of three structurally related nuclear receptors, , , and , which are intimately involved in the sensing of, and metabolic response to, nutritional status. Whereas PPAR is best established as a pivotal mediator of adipocyte differentiation and the cellular trapping and accumulation of lipid, PPAR is principally involved in the control of fatty acid oxidation, lipoprotein assembly, and amino acid catabolism and plays a central role in the metabolic response to fasting. The physiological roles of PPAR, which is widely expressed, are, as yet, less well defined. Although all three isoforms are capable of binding and responding to a wide array of lipid mediators, the question of whether each isoform has a specific high-affinity endogenous ligand is still unresolved.
Murine genetic knockout models have contributed substantially to our understanding of the biology of the PPARs (19, 20). Human genetics has also made important contributions: most notably, four different germline mutations of PPAR resulting in amino acid substitutions in the LBD of the receptor have been described in association with a stereotyped clinical syndrome of severe insulin resistance, dyslipidemia, hypertension, hepatic steatosis, and partial lipodystrophy. Two of these (Pro467Leu and Val290Met) have been shown to exert potent dominant-negative activity over their wild-type counterpart (3, 4). In contrast, Phe388Leu reportedly failed to exhibit dominant-negative activity (21). However, subsequent studies in our own laboratory with both the Phe388Leu and previously uncharacterized Arg425Cys mutants have revealed comparable ability to interfere with wild-type signaling (Gurnell, M., and V. K. K. Chatterjee, unpublished data). The importance of dominant negativity, as opposed to haploinsufficiency, is supported by the absence of a marked metabolic phenotype in human subjects carrying a PPAR frameshift/premature stop mutation, which produces a transcriptionally inactive truncated receptor that is unable to act as a dominant negative (22). It is further attested to by the discovery that a proportion of thyroid follicular carcinomas express a fusion protein consisting of the DBD of the transcription factor paired box transcription factor 8 and the LBD of PPAR. This chimeric species not only fails to transactivate on PPAR-responsive promoters but also suppresses the transactivation function of wild-type PPAR on such promoters, an activity that has been implicated in the pathogenesis of these neoplasia (23).
In contrast, no pathogenic mutations in PPAR or PPAR have been described to date in human subjects. A human PPAR splice variant (PPARtr) lacking exon 6 (resulting in deletion of part of the hinge region and the entire LBD) has been reported, comprising 20–50% of native PPAR in some human cells. Although this splice variant was unable to bind DNA, it was able to exert dominant-negative activity after forced nuclear localization (24).
To explore PPAR-mediated dominant-negative activity further, we exploited a strategy previously successfully employed for PPAR (5) to create a powerful artificial dominant-negative PPAR (PPARmut) by mutagenesis to alanine of two residues in the activation function 2 domain that critically regulate coactivator recruitment. Analysis of the structure of liganded PPAR bound to coactivator peptide (PDB 1K7L) (25) (Fig. 10) provides support for this strategy as applied to PPAR. Thus, glutamate 462 makes two hydrogen bonds to backbone amides in the coactivator helix and complements the N-terminal-positive charge of the helix dipole. This charge clamp in the wild-type molecule anchors the amphipathic, helical LXXLL motifs of many coactivators, thereby orienting the helices and apposing hydrophobic interaction surfaces on coregulatory molecules and nuclear receptor (26). Second, leucine 459 forms part of closely packed nonpolar interaction with helices 3 and 12 and the helical LXXLL coactivator motif. Interrupting both the charge clamp and hydrophobic coactivator-binding pocket would be predicted to have a marked destabilizing influence on the PPAR-coactivator interaction.
FIG. 10. Location of L459 and E462 in the three-dimensional crystal structure of PPAR complexed to a chemical agonist and coactivator peptide (PDB 1K7L).
Both the previously described double-mutant PPAR (PPARmut) and PPARmut act as constitutive repressors in the basal state, exhibit strong dominant-negative inhibition of transcriptional activation by their wild-type counterpart in response to a specific ligand, and recruit NCoR and SMRT avidly. However, the corepressor association was significantly more attenuated by ligand in the case of PPARmut than PPARmut. In the same system, wild-type PPAR interacted with both SMRT and NCoR (5), whereas no association was seen in this study for wild-type PPAR. This suggests that, whereas the double mutation in each case enhances corepressor interaction, the baseline levels of corepressor recruitment of the two receptors differ. Whether this is due to intrinsic differences in corepressor binding of the unliganded receptors or rather due to different degrees of receptor occupancy by endogenous ligand remains undetermined (5). The failure of PGC1 to augment transcriptional activation by cotransfected PPARmut is consistent with structural predictions (Fig. 10) and provides evidence that, as for PPARmut, the dominant-negative activity of PPARmut is due to impaired coactivator recruitment in the liganded state as well as enhanced corepressor association.
NCoR and SMRT possess a bipartite nuclear receptor interaction domain, which permits binding to specific residues in the hydrophobic pocket of nuclear receptors. Ligand binding, by establishing the charge clamp, introduces steric constraints on the length of helix that can be accommodated in the hydrophobic pocket, resulting in dissociation of corepressors with their longer helical interaction domain and binding of the smaller LXXLL-containing domain of coactivators (27, 28, 29). Not only has the crystal structure of a PPAR-agonist-coactivator complex been described but, uniquely to date among nuclear hormone receptors, the PPAR structure in complex with a chemical antagonist and a corepressor peptide from SMRT (30). However, how these well-delineated interactions relate to transcriptional regulation by PPARs in a cellular context is less clear. Although unliganded nuclear receptors such as thyroid hormone receptor and RAR are known to recruit corepressors in the absence of exogenous ligand, whether PPARs do so in vivo is contentious. Ligand-dissociable interaction between PPAR and NCoR has been shown in a yeast two-hybrid assay, and cotransfection studies in human embryonic kidney 293 cells have shown NCoR to suppress the constitutive reporter activity seen in the presence of PPAR and RXR, this repression being alleviated by specific ligand dose dependently (31). However, consistent with our results, PPAR-Gal4 fusion proteins have been shown to attenuate transcriptional activation by Gal4 in yeast but not mammalian cells (32), and it has not proved possible to detect either DNA-bound PPAR-RXR-NCoR or PPAR-RXR-SMRT complexes (32). Our failure to detect an interaction between NCoR or SMRT and the wild-type PPAR LBD in a mammalian two-hybrid system adds to these conflicting data.
It is possible that these discrepant results are explained by variable levels of endogenous PPAR ligand in different cell types. It may be that in some cells the generation of low levels of intracellular ligand means that native PPAR is never truly in the unliganded state, so that helix 12 is always in an active conformation, precluding corepressor interaction. Leu459Ala and Glu462Ala double mutation may mimic the binding of a chemical antagonist, destabilizing helix 12, and strongly promoting corepressor interaction, even in the presence of endogenous ligand. However, as shown in Fig. 3, potent exogenous ligand retains the ability to promote dissociation of PPARmut from corepressors in the two-hybrid system.
The dominant-negative activity of PPARmut was shown on an artificial PPRE and native robustly PPAR-responsive promoters in both HepG2 cells and rat primary hepatocytes, establishing that PPARmut is a potent dominant-negative inhibitor of PPARwt on endogenous promoters, even in primary culture in which PPAR expression levels are high. Similarly, concerns that cross-inhibition of PPAR and PPAR by PPARmut was due only to the use of a minimum artificial PPRE and that, in a more natural promoter context, selectivity would be exhibited were addressed by examining the effect of PPARmut expression on lipid accumulation and the induction of aP2 expression during differentiation of 3T3-L1 preadipocytes in the presence of a potent PPAR ligand (Fig. 9). Despite reports that the aP2 enhancer PPRE, which mediates activation by PPAR, is the most PPAR-selective of a raft of PPREs examined in vitro (33), expression of PPARmut was sufficient to block both aP2 transcriptional up-regulation and lipid accumulation. Together with our demonstration that PPARmut also severely attenuates the transcriptional response to a PPAR ligand in HepG2 cells, this suggests that the lack of selectivity of the mutant receptor’s inhibitory activity may be generalizable to specific responses mediated by all three PPARs. Moreover reexamination of the properties of PPARmut showed that it, too, is not PPAR selective with regard to suppression of PPAR-mediated transcriptional responses from endogenous promoters, albeit to a lesser extent than PPARmut. Although the lack of specificity using adenoviral delivery of the mutant PPARs could be attributed to high levels of expression from the constitutive CMV promoter in the viral vectors, a similar lack of specificity was seen in cotransfection experiments with stoichiometric levels of expression of wild-type and mutant receptors.
The lack of selectivity of the artificial PPAR mutants described here may also have pathophysiological relevance for human subjects harboring dominant-negative PPAR mutations, which are believed to exhibit mechanistically similar, albeit less potent, dominant-negative effects (3, 4). Indeed, before loss of PPAR signaling alone can be deemed to be the cause of the severe metabolic syndrome seen in subjects harboring dominant-negative PPAR mutations, it must be established whether dominant-negative inhibition by these naturally occurring mutants can spill over onto other PPARs. Evidence for such inhibitory cross-talk is also provided by a recent study of wild-type PPAR, which, in the unliganded state, was shown to inhibit PPAR and PPAR action, probably due to occupancy of the relevant PPAR response elements allied to its greater affinity for SMRT and perhaps other corepressors (32). Thus, it is conceivable that in some or all of the tissues in which the receptor isoforms are coexpressed, cross-inhibition of PPAR and/or PPAR signaling by a mutant PPAR with an aberrant affinity for corepressor may occur.
In summary, we have generated and characterized a double-mutant PPAR with potent dominant-negative activity toward its wild-type counterpart due to enhanced corepressor recruitment and impaired interaction with coactivators. The mutant receptor exhibits dominant-negative activity both in cotransfection reporter assays and when tested against endogenous PPAR responses in cell culture, and this inhibition is manifest against all three PPAR isoforms. This is also true of the previously reported analogous dominant-negative PPAR mutant, which may have implications for the pathogenesis of the metabolic syndrome seen in human subjects harboring dominant-negative PPAR mutations.
References
Savage DB, Tan GD, Acerini CL, Jebb SA, Agostini M, Gurnell M, Williams RL, Umpleby AM, Thomas EL, Bell JD, Dixon AK, Dunne F, Boiani R, Cinti S, Vidal-Puig A, Karpe F, Chatterjee VK, O’Rahilly S 2003 Human metabolic syndrome resulting from dominant-negative mutations in the nuclear receptor peroxisome proliferator-activated receptor-. Diabetes 52:910–917
Gurnell M, Savage DB, Chatterjee VK, O’Rahilly S 2003 The metabolic syndrome: peroxisome proliferator-activated receptor and its therapeutic modulation. J Clin Endocrinol Metab 88:2412–2421
Agostini M, Gurnell M, Savage DB, Wood EM, Smith AG, Rajanayagam O, Garnes KT, Levinson SH, Xu HE, Schwabe JW, Willson TM, O’Rahilly S, Chatterjee VK 2004 Tyrosine agonists reverse the molecular defects associated with dominant-negative mutations in human peroxisome proliferator-activated receptor . Endocrinology 145:1527–1538
Barroso I, Gurnell M, Crowley VE, Agostini M, Schwabe JW, Soos MA, Maslen GL, Williams TD, Lewis H, Schafer AJ, Chatterjee VK, O’Rahilly S 1999 Dominant-negative mutations in human PPAR associated with severe insulin resistance, diabetes mellitus and hypertension. Nature 402:880–883
Gurnell M, Wentworth JM, Agostini M, Adams M, Collingwood TN, Provenzano C, Browne PO, Rajanayagam O, Burris TP, Schwabe JW, Lazar MA, Chatterjee VK 2000 A dominant-negative peroxisome proliferator-activated receptor (PPAR) mutant is a constitutive repressor and inhibits PPAR-mediated adipogenesis. J Biol Chem 275:5754–5759
Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM 1995 15-Deoxy- 12, 14-prostaglandin J2 is a ligand for the adipocyte determination factor PPAR . Cell 83:803–812
Tone Y, Collingwood TN, Adams M, Chatterjee VK 1994 Functional analysis of a transactivation domain in the thyroid hormone ? receptor. J Biol Chem 269:31157–31161
Heyman RA, Mangelsdorf DJ, Dyck JA, Stein RB, Eichele G, Evans RM, Thaller C 1992 9-Cis retinoic acid is a high affinity ligand for the retinoid X receptor. Cell 68:397–406
Collingwood TN, Adams M, Tone Y, Chatterjee VK 1994 Spectrum of transcriptional, dimerization, and dominant-negative properties of twenty different mutant thyroid hormone ?-receptors in thyroid hormone resistance syndrome. Mol Endocrinol 8:1262–1277
Wentworth JM, Schoenfeld V, Meek S, Elgar G, Brenner S, Chatterjee VK 1999 Isolation and characterisation of the retinoic acid receptor- gene in the Japanese pufferfish, F. rubripes. Gene 236:315–323
Schmidt A, Endo N, Rutledge SJ, Vogel R, Shinar D, Rodan GA 1992 Identification of a new member of the steroid hormone receptor superfamily that is activated by a peroxisome proliferator and fatty acids. Mol Endocrinol 6:1634–1641
Brown PJ, Stuart LW, Hurley KP, Lewis MC, Winegar DA, Wilson JG, Wilkison WO, Ittoop OR, Willson TM 2001 Identification of a subtype selective human PPAR agonist through parallel-array synthesis. Bioorg Med Chem Lett 11:1225–1227
Boehm MF, Zhang L, Zhi L, McClurg MR, Berger E, Wagoner M, Mais DE, Suto CM, Davies JA, Heyman RA 1995 Design and synthesis of potent retinoid X receptor selective ligands that induce apoptosis in leukemia cells. J Med Chem 38:3146–3155
Lehmann JM, Moore LB, Smith-Oliver TA, Wilkison WO, Willson TM, Kliewer SA 1995 An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor (PPAR). J Biol Chem 270:12953–12956
Berger J, Leibowitz MD, Doebber TW, Elbrecht A, Zhang B, Zhou G, Biswas C, Cullinan CA, Hayes NS, Li Y, Tanen M, Ventre J, Wu MS, Berger GD, Mosley R, Marquis R, Santini C, Sahoo SP, Tolman RL, Smith RG, Moller DE 1999 Novel peroxisome proliferator-activated receptor (PPAR) and PPAR ligands produce distinct biological effects. J Biol Chem 274:6718–6725
Duerden JM, Bartlett SM, Gibbons GF 1989 Long-term maintenance of high rates of very-low-density-lipoprotein secretion in hepatocyte cultures. A model for studying the direct effects of insulin and insulin deficiency in vitro. Biochem J 263:937–943
Hebbachi AM, Brown AM, Gibbons GF 1999 Suppression of cytosolic triacylglycerol recruitment for very low density lipoprotein assembly by inactivation of microsomal triglyceride transfer protein results in a delayed removal of apoB-48 and apoB-100 from microsomal and Golgi membranes of primary rat hepatocytes. J Lipid Res 40:1758–1768
Muoio DM, MacLean PS, Lang DB, Li S, Houmard JA, Way JM, Winegar DA, Corton JC, Dohm GL, Kraus WE 2002 Fatty acid homeostasis and induction of lipid regulatory genes in skeletal muscles of peroxisome proliferator-activated receptor (PPAR) knock-out mice. Evidence for compensatory regulation by PPAR. J Biol Chem 277:26089–26097
Miles PD, Barak Y, Evans RM, Olefsky JM 2003 Effect of heterozygous PPAR deficiency and TZD treatment on insulin resistance associated with age and high-fat feeding. Am J Physiol Endocrinol Metab 284:E618–E626
Wang YX, Lee CH, Tiep S, Yu RT, Ham J, Kang H, Evans RM 2003 Peroxisome-proliferator-activated receptor activates fat metabolism to prevent obesity. Cell 113:159–170
Hegele RA, Cao H, Frankowski C, Mathews ST, Leff T 2002 PPARG F388L, a transactivation-deficient mutant, in familial partial lipodystrophy. Diabetes 51:3586–3590
Savage DB, Agostini M, Barroso I, Gurnell M, Luan J, Meirhaeghe A, Harding AH, Ihrke G, Rajanayagam O, Soos MA, George S, Berger D, Thomas EL, Bell JD, Meeran K, Ross RJ, Vidal-Puig A, Wareham NJ, O’Rahilly S, Chatterjee VK, Schafer AJ 2002 Digenic inheritance of severe insulin resistance in a human pedigree. Nat Genet 31:379–384
Kroll TG, Sarraf P, Pecciarini L, Chen CJ, Mueller E, Spiegelman BM, Fletcher JA 2000 PAX8-PPAR1 fusion oncogene in human thyroid carcinoma. Science 289:1357–1360
Gervois P, Torra IP, Chinetti G, Grotzinger T, Dubois G, Fruchart JC, Fruchart-Najib J, Leitersdorf E, Staels B 1999 A truncated human peroxisome proliferator-activated receptor splice variant with dominant-negative activity. Mol Endocrinol 13:1535–1549
Xu HE, Lambert MH, Montana VG, Plunket KD, Moore LB, Collins JL, Oplinger JA, Kliewer SA, Gampe Jr RT, McKee DD, Moore JT, Willson TM 2001 Structural determinants of ligand binding selectivity between the peroxisome proliferator-activated receptors. Proc Natl Acad Sci USA 98:13919–13924
Heery DM, Kalkhoven E, Hoare S, Parker MG 1997 A signature motif in transcriptional coactivators mediates binding to nuclear receptors. Nature 387:733–736
Perissi V, Staszewski LM, McInerney EM, Kurokawa R, Krones A, Rose DW, Lambert MH, Milburn MV, Glass CK, Rosenfeld MG 1999 Molecular determinants of nuclear receptor-corepressor interaction. Genes Dev 13:3198–3208
Hu X, Lazar MA 1999 The CoRNR motif controls the recruitment of corepressors by nuclear hormone receptors. Nature 402:93–96
Nagy L, Kao HY, Love JD, Li C, Banayo E, Gooch JT, Krishna V, Chatterjee K, Evans RM, Schwabe JW 1999 Mechanism of corepressor binding and release from nuclear hormone receptors. Genes Dev 13:3209–3216
Xu HE, Stanley TB, Montana VG, Lambert MH, Shearer BG, Cobb JE, McKee DD, Galardi CM, Plunket KD, Nolte RT, Parks DJ, Moore JT, Kliewer SA, Willson TM, Stimmel JB 2002 Structural basis for antagonist-mediated recruitment of nuclear co-repressors by PPAR. Nature 415:813–817
Dowell P, Ishmael JE, Avram D, Peterson VJ, Nevrivy DJ, Leid M 1999 Identification of nuclear receptor corepressor as a peroxisome proliferator-activated receptor interacting protein. J Biol Chem 274:15901–15907
Shi Y, Hon M, Evans RM 2002 The peroxisome proliferator-activated receptor, an integrator of transcriptional repression and nuclear receptor signaling. Proc Natl Acad Sci USA 99:2613–2618
Juge-Aubry C, Pernin A, Favez T, Burger AG, Wahli W, Meier CA, Desvergne B 1997 DNA binding properties of peroxisome proliferator-activated receptor subtypes on various natural peroxisome proliferator response elements. Importance of the 5'-flanking region. J Biol Chem 272:25252–25259
Address all correspondence and requests for reprints to: Robert Semple, Department of Clinical Biochemistry, Addenbrooke’s Hospital, Cambridge CB2 2QR, United Kingdom. E-mail: rks16@cam.ac.uk.
Abstract
Several missense mutations in the ligand-binding domain of human peroxisome proliferator-activated receptor (PPAR) have been described in subjects with dominantly inherited severe insulin resistance associated with partial lipodystrophy, hypertension, and dyslipidemia. These mutant receptors behave as dominant-negative inhibitors of PPAR signaling when studied in transfected cells. The extent to which such dominant-negative effects extend to signaling through other coexpressed PPAR isoforms has not been evaluated. To examine these issues further, we have created a PPAR mutant harboring twin substitutions, Leu459Ala and Glu462Ala, within the ligand binding domain (PPARmut), examined its signaling properties, and compared the effects of dominant-negative PPAR and PPAR mutants on basal and ligand-induced gene transcription in adipocytes and hepatocytes. PPARmut was transcriptionally inactive, repressed basal activity from a PPAR response element-containing promoter, inhibited the coactivator function of cotransfected PPAR- coactivator 1, and strongly inhibited the transcriptional response to cotransfected wild-type receptor. In contrast to PPAR, wild-type PPAR failed to recruit the transcriptional corepressors NCoR and SMRT. However, PPARmut avidly recruited these corepressors in a ligand-dissociable manner. In hepatocytes and adipocytes, both PPARmut and the corresponding PPAR mutant were capable of inhibiting the expression of genes primarily regulated by PPAR, -, or - ligands, albeit with some differences in potency. Thus, dominant-negative forms of PPAR and PPAR are capable of interfering with PPAR signaling in a manner that is not wholly restricted to their cognate target genes. These findings may have implications for the pathogenesis of human syndromes resulting from mutations in this family of transcription factors.
Introduction
NATURALLY OCCURRING MUTATIONS in the nuclear hormone receptor peroxisome proliferator- activated receptor (PPAR) are now well established as the cause of a stereotyped human syndrome of severe insulin resistance, partial lipodystrophy, hypertension, and dyslipidemia, often manifesting in women as polycystic ovary syndrome and preeclampsia (1, 2). The majority of mutations (mut) reported to date lie within the ligand-binding domain (LBD) of the receptor and, when studied in vitro, exhibit dominant-negative activity when coexpressed with their wild-type (wt) counterpart (3, 4). The molecular basis of the dominant-negative activity of mutations in the PPAR LBD has been most closely examined in relation to an artificial double-mutant receptor (Leu468Ala/Glu471Ala) (5). This receptor was found to bind DNA and heterodimerize with retinoid X receptor (RXR) normally. Its marked dominant-negative activity reflected aberrant interaction with nuclear receptor coregulators: release of corepressor molecules such as nuclear receptor corepressor (NCoR) and silencing mediator of retinoid and thyroid receptors (SMRT) was impaired, whereas the ligand-induced recruitment of coactivators such as cAMP response element-binding protein was also greatly diminished. Similar findings have more recently been reported for two of the naturally occurring PPAR mutants (3). It is notable that no naturally occurring mutations within other PPAR isoforms ( and ) have yet been reported.
We have now elected to investigate whether equivalent mutations in another PPAR isoform would also result in a dominant-negative receptor and, if so, determine whether they would do so through entirely analogous molecular mechanisms. Thus, we first created the corresponding mutations in PPAR to those described previously for PPAR and examined the signaling properties of the mutant receptor in detail. In transient transfection assays using an artificial PPAR response element reporter gene construct, we demonstrated that the dominant-negative mutant PPAR failed to mediate ligand-induced transcriptional activation and moreover interfered with the action of not only PPAR-specific ligands acting through cotransfected wild-type PPAR but also a PPAR-specific ligand and a PPAR-specific ligand acting through their cognate receptors.
The availability of dominant-negative mutants of both PPAR and PPAR, which could be adenovirally transduced into differentiated cells, allowed us to extend our investigation of the specificity of dominant-negative activity within this receptor family. By adenovirally transducing adipocyte and hepatocyte cell lines with either the PPAR or PPAR dominant-negative receptors and examining their effects on basal and ligand-induced transcription of canonical target genes for each PPAR isoform, we demonstrated that cross-inhibition of PPAR signaling is also seen for transcriptional responses mediated by endogenous PPARs.
The physiological relevance of these observations is unclear, but if such cross-inhibition occurs in vivo, this suggests that the metabolic syndrome seen in humans with PPAR dominant-negative mutations might represent the composite effects of the receptor mutants on signaling through more than one coexpressed isoform of PPAR.
Materials and Methods
Plasmid constructs
The full-length human PPAR cDNA was cloned by RT-PCR from total HepG2 cell RNA using the following oligonucleotides: sense, 5'-CAG GGT ACC ACC ATG GAC TAC AAA GAC GAT GAC GAC AAG ATG GTG GAC ACG GAA AGC CCA CTC T-3', and antisense, 5'-TGG CGG CCG CTC AGT ACA TGT CCC TGT AGA TCT CC-3'. The sense oligonucleotide contained a Kozak translation signal (ACCATG) and a FLAG epitope (MDYKDDDDK residues). An amplicon of 1450 bp was obtained after 30 cycles (94 C for 1 min, 63 C for 1 min, and 72 C for 1 min 30 sec) with a 1:1 ratio of pfu turbo (Stratagene, La Jolla, CA)/AmpliTaq Gold (PerkinElmer, Norwalk, CT). The PCR product was purified and cloned into pcDNA3.1 (+) (Invitrogen, Paisley, UK) using the KpnI and NotI restriction sites, making pcDNA3-PPARwt. The insert was sequenced to confirm its integrity.
The leucine and glutamate at positions 459 and 462, respectively, were mutated into alanine by PCR site-directed mutagenesis using the Quickchange kit (Stratagene) according to the supplier’s instructions with the following oligonucleotides: sense, 5'-GCT GCG CTG CAC CCG GCA CTG CAG GCG ATC TAC AGG GAC ATG-3', and antisense, 5'-CAT GTC CCT GTA GAT CGC CTG CAG TGC CGG GTG CAG CGC AGC-3' (mutations underlined). After confirmation by direct sequencing, the insert was double digested with KpnI and NotI and ligated into empty pcDNA3.1 (+) vector between KpnI and NotI sites.
The hinge region and LBD (amino acids 167–468) of PPARwt and PPARmut were amplified using the following oligonucleotides: sense, 5'-CAT GAA TTC TCA CAC AAC GCG ATT CGT TT3–3' and antisense 5'-CAC GAA TTC TCA GTA CAT GTC CCT GTA GAT-3' from pcDNA3-PPARwt and pcDNA3-PPARmut, and the PCR product (909 bp) was cloned into the EcoRI site of AASV-VP16 vector to give VP16-PPARwt LBD and VP16-PPARmut LBD. The L468A/E471A PPAR double mutant, human PPAR2, Gal4-NCoR, and Gal4-SMRT (5); peroxisome proliferator response element (PPRE)3-TKLUC (6); UASTKLUC (7); CRBPIITKLUC (8); MALTKLUC, RSV-TR?1, and RSV-RXR (9); retinoic acid receptor (RAR)?2TKLUC and RSV-RAR1 (10) constructs have been described previously. The PcDNA3-PPAR has been described previously (11) and was a gift from Dr. R. Vogel (Merck, Whitehouse Station, NJ). The hemagglutinin-tagged human PPAR- coactivator 1 (PGC1) plasmid was a gift from Dr. A. Kralli (University of Basel, Basel, Switzerland).
Transfection assays
HepG2 cells were maintained in DMEM containing 4.5 g glucose/liter (Sigma, Poole, UK) supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin. Cells were split into 24- or 96-well plates and transfected 8 h later in the same medium using Fugene (Roche, Stockholm, Sweden). A ratio of 3 μl Fugene to 1 μg DNA was used for all transfections. The transfection efficiency was estimated by cotransfecting the vector pRL-cytomegalovirus (CMV) (Promega, Southampton, UK) encoding the Renilla luciferase gene. Fifteen hours after transfection, the medium was removed, the cells were washed once with PBS, and fresh medium containing the relevant ligand added for 30 h. The cells were then harvested and both Firefly and Renilla luciferase activities were measured in 20 μl of lysate according to the manufacturer’s instructions (dual luciferase assay, Promega). The following specific ligands were used: PPAR-GW7647 (12) (Sigma), RXR-LG100268 (13) (a gift from Mark Leibowitz, Ligand Pharmaceuticals, San Diego, CA), PPAR-BRL49653 (14) (Alexis, San Diego, CA), and PPAR-L165041 (15) (a gift from Dr. D. Moller, Merck).
Recombinant adenovirus construction
The PPARmut recombinant adenovirus was created using the AdEAsy kit according to the supplier’s instructions (QBiogene, France). Briefly, the human PPARmut cDNA was cloned into pShuttle-CMV between KpnI and NotI restriction sites. The vector was linearized with PmeI and cotransformed with the pAdEasy-1 vector into BJ5183 competent cells by electroporation to effect recombination between the two vectors. Recombinants were then digested with BstXI. DH5 cells were transformed with one positive recombinant clone, and DNA was extracted using the large-construct maxiprep kit (Qiagen, West Sussex, UK). Direct sequencing was used to verify the integrity of the cDNA and the presence of the mutations. The DNA was linearized with PacI and QBI-293A cells plated in 6-well plates with agarose-DMEM overlay were infected with the viral DNA to produce viral particles. Plaques appeared after 14 d and were picked and transferred onto fresh QBI-293A cells. Cytopathic effect (CPE) started to appear after 8 d. A second round of amplification was performed and CPE was complete after 4 d. The viral particles were released by three cycles of freezing/thawing and large-scale amplification performed by QBiogene (Canada).
Use of recombinant adenovirus in cultured cells
Male Wistar rats, fed and housed as described previously (16), were used for the preparation of hepatocytes when they weighed between 200 and 300 g. Hepatocytes were prepared under sterile conditions (17) and were suspended [(0.75–1.0) x 106 cells/ml)] in Waymouth’s medium MB752/1 supplemented with amino acids, antibiotics, and 10% (vol/vol) heat-inactivated FBS. Then 3.0 ml of this suspension was added to culture dishes coated with rat-tail collagen, and after 3 h, the cells became attached as a monolayer. The medium was removed and the cells were washed twice with 3.0ml PBS before addition of 3.0 ml serum-free Waymouth’s medium supplemented with amino acids, antibiotics, 1 μM dexamethasone, 1 mM pyruvate, and 10 mM lactate (supplemented medium). To each plate was added adenoviral storage buffer, green fluorescent protein (GFP)-expressing adenovirus, or PPARmut-expressing adenovirus. Plaque-forming units (PFUs) of each virus (1.26 x 108) were used per plate, giving a multiplicity of infection (MOI) of approximately 60 PFU/cell, in line with previous publications. After 3 h of incubation in virus-containing medium, the cells were washed twice with PBS before addition of supplemented medium containing the relevant concentration of PPAR agonist GW7647 or dimethylsulfoxide. After 16 h RNA was extracted from the cells using the RNEasy kit (Qiagen).
HepG2 experiments were carried out in DMEM containing 4.5 g/liter glucose (Sigma) supplemented with 10% FBS and antibiotics in 12-well plates. Adenoviral vector or vehicle (5 x 106 PFU) was added per well (MOI around 20 PFU/cell), and cells were incubated for 12 h before washing with PBS and replacement with medium containing GW7647 or dimethylsulfoxide. At this stage, infection rates of greater than 95% for the GFP and GFP/PPARmut viruses were verified by fluorescence microscopy of the living cells. After 48 h RNA was extracted using the RNEasy kit.
3T3-L1 preadipocytes (American Type Culture Collection, Manassas, Va) were maintained at less than 70% confluence in DMEM containing 4.5 g/liter glucose supplemented with 10% newborn calf serum, penicillin/streptomycin, and 2 mM glutamine (all Sigma). For each experiment, cells were seeded into 12-well plates and grown until 2 d post confluence. Differentiation medium consisted of the FBS-containing medium supplemented with insulin (5 μg/ml), dexamethasone (0.1 μg/ml), and 3-isobutyl-1-methylxanthine (110 μg/ml). Where cells were to be fully differentiated, this was replaced after 3 d by the FBS-containing medium supplemented with insulin (5 μg/ml) alone and after a further 3 d by the FBS-containing medium alone. A pilot study to assess the validity of adipocyte fatty acid binding protein 4 (aP2) as a specific transcriptional target of PPAR in 3T3-L1 preadipocytes was carried out. The cells were induced to differentiate in the presence of BRL49653or vehicle and RT-PCR used to quantify mRNA levels aP2. In comparison with differentiation medium (DM) alone, BRL49653plus DM resulted in a more than 20-fold increase in aP2 mRNA levels at 48 h. Thus, BRL49653stimulation concomitant with the first 2 d of differentiation was chosen to test the effect of the mutant PPARs. For adenovirus experiments, cells at 2 d post confluence were exposed to maintenance medium containing either adenoviral storage buffer or 2 x 109 PFU of adenovirus per well (MOI around 5000 PFU/cell unless otherwise indicated) for 12 h with agitation. Adenovirus-expressing GFP alone (QBiogene) was used to assess infection rates in conjunction with fluorescence microscopy. The monolayers were then washed with PBS, and cells were incubated in DM containing different concentrations of BRL49653for 48 h before RNA extraction using the RNEasy kit (Qiagen). Oil Red O staining employing a standard protocol was used to assess differentiation in those experiments in which cells were differentiated for 12 d.
Real-time quantitative PCR
On the basis of published work, we identified peroxisomal fatty acyl-coenzyme A oxidase (pFAO), pyruvate dehydrogenase kinase (PDK) 4 and cytochrome IVA1 as robustly responsive targets of PPAR in rat liver and PDK4 and liver type carnitine palmitoyl transferase 1 (CPT1) as targets in human liver and derived cell lines. Primer Express software (version 1.0, PerkinElmer Applied Biosystems, Foster City, CA) was used to design the probes and primers for real-time quantitative PCR shown in Table 1. Five hundred nanograms of total RNA was reverse transcribed for 1 h at 37 C in a 20-μl reaction including 1 x reverse transcription buffer and 100 U Moloney murine leukemia virus reverse transcriptase, 250 ng random hexamers, 1.25 mmol/liter deoxynucleotide triphosphates (all Promega), and the solution was made up to 100 μl. Two microliters of the resulting cDNA were used in each 25 μl PCR, in which 300 nmol/liter of forward and reverse primers and 150 nmol/liter of fluorogenic probe were used. Reactions were carried out at least in duplicate for each sample on an ABI 7700 sequence detection system (PerkinElmer Biosystems) according to the manufacturer’s instructions, and target values were normalized to either glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or 18S rRNA as indicated (rodent and human reagents from PerkinElmer).
TABLE 1. Sequences of primers and probes used for quantitative real-time PCR
Results
Figure 1 shows the amino acid sequence comparison of the distal LBD of human, rat, and Xenopus PPARs. The leucine and glutamate residues that confer on human PPAR dominant-negative activity when mutated into alanine (5) are conserved in all homologs shown. Therefore, we decided to apply this mutational strategy to create a dominant-negative mutant of PPAR. Thus, wild-type PPAR cDNA was amplified and cloned from HepG2 cell cDNA, and Leu459 and Glu462 were replaced by alanine using site-directed mutagenesis to create PPARmut.
FIG. 1. C-terminal 141 amino acids of PPAR, -, and - from human (h), rat (r), and Xenopus (x). Human PPAR residues L459 and E462 (mutated to alanine) and their equivalents in the other PPARs are shaded, and residues conserved in all species are underlined.
The transcriptional activity of PPARmut was then investigated by cotransfecting HepG2 cells with expression vectors encoding RXR, PPARwt, or PPARmut and a reporter construct consisting of three copies of the PPRE of the acyl-coenzyme A oxidase gene coupled to the Firefly luciferase gene (PPRE3-TKLUC), with the Renilla luciferase vector as a transfection control. In the presence of 100 nM GW7647 and LG100268, PPARwt induced luciferase activity by 4-fold (Fig. 2A). The PPARmut, in contrast, exhibited intense suppression of both basal (Fig. 2A, inset) and ligand-stimulated transcription (Fig. 2A). Ligand dose response testing (Fig. 2B) confirmed the complete inability of PPARmut to mediate a transcriptional response to ligand. Thus, the introduction of the two mutations in the ligand-binding domain of PPAR was sufficient to abolish its transactivation capacity.
FIG. 2. Transcriptional activity of PPARwt and PPARmut. HepG2cells were transfected with 300 ng pcDNA3.1, pcDNA3-PPARwt, or pcDNA3-PPARmut, together with 300 ng RSV-hRXR, 60 ng PPRE3-TKLUC reporter, and 20 ng pRL-CMV. The activity is expressed relative to the wild-type maximum (100%): with 100 nM of both GW7647 and L165041 (A) or vehicle (open bars). Inset, Detail of baseline levels.*, P < 0.05; **, P < 0.01 (two-tailed Student’s t test). B, Dose response of reporter activity to equimolar GW7647 and L165041. Results are the mean ± SD of five experiments in duplicate.
Recruitment of transcriptional coregulators
To elucidate the mechanism of this loss of function, we went on to examine the interactions of PPARmut with a panel of transcriptional coregulator molecules. First, corepressor recruitment by PPARwt and PPARmut was assessed in a mammalian two-hybrid assay using the receptor-interacting domains of SMRT and NCoR fused to the Gal4 DNA binding domain (DBD) and the LBD of PPARwt or PPARmut fused to the activation domain of VP16 (Fig. 3). HepG2 cells were cotransfected with the Gal4-corepressor chimera, together with either the wild-type or mutant fusion construct and the UAS-TKLUC reporter. In the case of the wild-type receptor, recruitment of neither NCoR (Fig. 3A) nor SMRT (Fig. 3B) could be detected. However, in marked contrast, the PPARmut LBD was able to recruit either NCoR or SMRT in the unliganded state and release them in a dose-dependent manner after addition of GW7647.
FIG. 3. Mammalian two-hybrid assay with nuclear receptor interaction domains of NCoR (A) and SMRT (B). HepG2 cells were transfected with 100 ng of vectors encoding VP16 only, VP16 fused to the wild-type PPAR LBD (VP16-LBDwt), or VP16 fused to the double-mutant PPAR LBD (VP16-LBDmut) together with 100 ng of expression vector encoding the Gal4 DBD fused to the nuclear receptor interaction domains of NCoR (Gal4-NCoR) (A) or SMRT (Gal4-SMRT) (B) and 500 ng of the reporter construct UASTKLUC, with results normalized to Renilla luciferase activity as previously described. The dose response of reporter activity to GW7647 is shown. The activity is expressed relative to the maximum obtained for PPARmut in the absence of ligand (100%). Results are the mean ± SD of four experiments in duplicate.
Second, we explored the capacity of a known nuclear receptor coactivator to potentiate transcriptional activation by PPARwt and PPARmut. HepG2 cells were cotransfected with PPARwt or PPARmut plasmids, the PPRE3-TKLUC reporter plasmid, and incremental amounts of hemagglutinin-tagged hPGC1 expression vector and stimulated with ligand, with results normalized to Renilla luciferase activity. A clear dose response of reporter activity to increasing amounts of hPGC1 was observed both in the presence of the reporter plasmid alone (in which case endogenous PPAR expression presumably accounts for the transcriptional activity) and with overexpressed PPARwt, although it should be noted that basal activity was much higher in the presence of exogenous receptor (Fig. 4). Similar but enhanced transactivation profiles were recorded after the addition of GW7647. Strikingly, the presence of PPARmut virtually abolished transcriptional activity at all concentrations of PGC1 and ligand (Fig. 4).
FIG. 4. Suppression of PGC1-mediated transcriptional activation by PPARmut. HepG2 cells in 96-well plates were cotransfected with 8 ng PPRE3-TKLUC, 3 ng of either pcDNA3-PPARwt or pcDNA3-PPARmut, increasing amounts of plasmid encoding hemagglutinin-tagged human PGC1 as shown, and empty pcDNA3.1 (vector) to balance the total amounts of transfected DNA. Results were normalized to Renilla luciferase activity, and results are the mean ± SD of three experiments, expressed as percentages of the maximal recorded transcriptional activity. DMSO, Dimethylsulfoxide.
Dominant-negative activity of PPARmut
To evaluate the possible dominant-negative activity of the mutant, HepG2 cells were cotransfected with either PPARwt or PPARwt and PPARmut (Fig. 5). The addition of PPARmut reduced the ability of PPARwt to transactivate the luciferase gene by 50%, demonstrating that the mutant possesses dominant-negative activity (Fig. 5A). This effect shows dose responsiveness, with a 75% reduction of activity seen with a 10-fold excess of mutant.
FIG. 5. Dominant-negative activity of PPARmut over PPARwt (A), PPAR (B), and PPAR (C). A, HepG2 cells were transfected with 600 ng pcDNA3-PPARwt or 300 ng pcDNA3-PPARwt plus 300 ng or 3000 ng pcDNA3-PPARmut as indicated. In all cases 60 ng PPRE3-TKLUC and 20 ng pRL-CMV were cotransfected with pcDNA3.1 to equalize the total amounts of transfected DNA. Three hundred nanograms pcDNA3-PPARwt had previously been established to give the same transcriptional activity as 600 ng. Dose responses of reporter activity to GW7647 are shown, the activity being expressed relative to the wild-type maximum obtained with 100 nM GW7647 (100%). B is the same as for A but with pcDNA3-PPARwt, and BRL49653as ligand. C is the same as for A but with pcDNA3-PPAR, and L165041 as ligand. Results were normalized to Renilla luciferase activity, and results are the mean ± SD of five experiments (A and B) or three experiments (C), expressed as percentages of the transcriptional activity in response to 100 nM ligand.
We also tested the effect of the PPARmut on PPAR-mediated transcription. The presence of the thiazolidinedione ligand BRL49653at 100 nM resulted in an 8-fold activation of luciferase activity (Fig. 5B). However, addition of equimolar PPARmut was associated with a 75% decrease in PPAR-mediated transcriptional activity at 100 nM BRL49653 Further suppression was again seen with a 10-fold excess of mutant. Using the same reporter construct, we also demonstrated potent dominant-negative activity of the mutant PPAR over PPAR (Fig. 5C). Thus, in this cotransfection system, there appears to be no selectivity of the inhibitory effect of the PPARmut construct for PPAR-mediated over PPAR- and PPAR-mediated transcriptional activation, albeit in the context of an artificial, minimal PPRE-containing promoter.
To assess whether the dominant-negative activity of the mutant construct is selective or specific for PPAR-mediated responses, we also investigated the effect of coexpressing PPARmut with other nuclear hormone receptors in the presence and absence of their specific ligand. Expression of PPARmut either at the same level or in 5-fold excess had no significant effect on robust transcriptional responses mediated by RAR, thyroid hormone receptor ?1, or RXR in the presence of their cognate ligands, suggesting that its powerful dominant-negative activity is selective for PPAR-mediated responses (Fig. 6).
FIG. 6. Selectivity of dominant-negative activity of PPARmut in vitro. HepG2 cells in 96-well plates were transfected with reporter plasmid and nuclear receptor-encoding plasmid as shown with or without an equal amount or 5-fold excess of cotransfected pcDNA3-PPARmut. In all cases 3 ng pRL-CMV were cotransfected, and empty pcDNA3.1 was added to equalize amounts of transfected DNA. Reporter responses to specific ligand (as shown) were assayed, and results are the mean ± SD of three experiments, normalized to the baseline activity seen in the absence of overexpressed nuclear receptor in eachcase. A, RXR; B, thyroid hormone receptor (TR)?1; C, RAR1. ATRA, All-trans-retinoic acid.
Effect of PPARmut adenoviral expression on native PPAR-responsive genes
Because the PPAR-dependent promoter used hitherto was artificial and possibly unrepresentative of isoform-specific PPAR-driven responses in living cells, we went on to look at the effect of PPARmut expression on well-characterized canonical PPAR-mediated responses in primary rat hepatocytes, HepG2 cells, and murine 3T3-L1 preadipocytes. To this end, we created an adenovirus-based expression vector encoding the PPARmut cDNA and compared its ability to inhibit PPAR-mediated responses with the inhibitory activity of a previously described PPARmut/GFP-coexpressing adenovirus (5). Cells in culture were infected with vehicle, a GFP-expressing adenovirus, PPARmut-expressing adenovirus, or GFP/PPARmut-expressing adenovirus under conditions that have been optimized to permit high levels of infection. The infected cells were subsequently exposed to a range of ligand concentrations (PPAR-GW7647, PPAR-BRL49653, PPAR-L165041), and PPAR target gene mRNA expression was assessed by real-time quantitative PCR.
In rat hepatocytes, PDK4 showed a steep dose-dependent response to PPAR agonist (GW7647) in both uninfected cells and cells infected with GFP-expressing virus, with a 15- to 20-fold induction of mRNA levels (Fig. 7A). In the presence of the PPARmut, not only was induction of expression ablated, but there was also baseline suppression of PDK4 mRNA levels, and indeed they did not approach the uninfected baseline until very high concentrations of ligand were added (Fig. 7A). The PPARmut adenovirus also significantly attenuated the response to the PPAR agonist but to a lesser degree: no suppression of baseline activity was seen, and the peak stimulation of mRNA expression was around 3-fold. Similar patterns were seen for pFAO (fig 7B) and CYPIVA1 (data not shown), although the degree of induction of mRNA was less (8- to 10-fold).
FIG. 7. Effect of PPARmut or PPARmut on endogenous PPAR target genes in primary rat hepatocytes. Primary rat hepatocytes were infected with adenovirus expressing GFP (Ad-GFP), PPARmut (Ad-mut), or PPARmut and GFP (Ad-mut) before stimulation by 16 h of exposure to different concentrations of GW7647, after which mRNA levels of PDK4 (A) or pFAO (B) were measured by RT-PCR and normalized to GAPDH mRNA levels. Representative experiments from three are shown, with error bars representing the SD of duplicates of RT-PCR.
In HepG2 cells, addition of PPAR agonist GW7647 resulted in modest induction of CPT1 (Fig. 8A) and strong induction of PDK4 (Fig. 8B). In both cases, PPARmut and PPARmut attenuated this, most markedly in the case of PDK4 with, once again, a more robust effect seen with the PPARmut-expressing virus.
FIG. 8. Effect of PPARmut or PPARmut on endogenous PPAR target genes in HepG2 cells. HepG2 cells were infected with Ad-mut, Ad-GFP, or Ad-mut and then exposed to GW7647 for 48 h before measurement of PDK4 (A) and CPT1 (B) mRNA levels by quantitative RT-PCR, normalized to GAPDH. The means of three experiments are shown, with error bars representing the SD.
Next, to determine whether the adenoviral constructs were equally capable of attenuating the responsiveness of selective PPAR target genes, we turned to the murine 3T3-L1 preadipocyte cell line. Preliminary experiments established that addition of 100 nM BRL49653to standard proadipogenic DM for 48 h resulted in a 20-fold greater increase in mRNA levels of aP2 than with DM alone, which we thus regarded as a robust PPAR-specific response against which to test the effects of the mutant constructs. This response was dramatically attenuated in the presence of either the PPARmut or PPARmut, with the PPARmut once again exhibiting a greater effect (Fig. 9A). A correlate of this was seen in the failure of 3T3-L1 cells infected with either construct to accumulate lipid in response to standard differentiation conditions (Fig. 9C) (5).
FIG. 9. Effect of PPARmut or PPARmut on endogenous PPAR and PPAR target genes. A, Two-day postconfluent 3T3-L1 preadipocytes were infected with Ad-mut, Ad-GFP, or Ad-mut for 12 h before 48 h of exposure to adipogenic DM and different concentrations of BRL49653 aP2 mRNA levels were then determined by quantitative real-time PCR and normalized to 18S rRNA levels. B, HepG2 cells were infected with Ad-mut, Ad-GFP, or Ad-mut and then exposed to GW7647 for 48 h before measurement of PDK4 mRNA levels by quantitative RT-PCR, normalized to GAPDH. The means of three experiments are shown, with error bars representing the SD. C, Two-day postconfluent 3T3-L1 preadipocytes were infected with no adenovirus, Ad-GFP, or Ad-mut at the range of MOIs shown before adipogenic differentiation as described using a standard protocol. Lipid accumulation was assessed by staining with oil Red O at d 10. Representative images are shown. Maximal infection rates of 90% were achieved at an MOI of 109 Ad-GFP PFU per well (not shown).
Because PDK4 has been shown previously to be a robustly responsive gene to PPAR stimulation as well as PPAR stimulation (18), we finally tested the effect of adenoviral overexpression of either PPARmut or PPARmut in HepG2 cells on this response. Both mutant receptors were found to inhibit the response of PDK4 to stimulation with the specific PPAR agonist L165041 (Fig. 9B), suggesting that the dominant-negative activity of these agents extends also to the action of PPAR.
Discussion
The PPAR family consists of three structurally related nuclear receptors, , , and , which are intimately involved in the sensing of, and metabolic response to, nutritional status. Whereas PPAR is best established as a pivotal mediator of adipocyte differentiation and the cellular trapping and accumulation of lipid, PPAR is principally involved in the control of fatty acid oxidation, lipoprotein assembly, and amino acid catabolism and plays a central role in the metabolic response to fasting. The physiological roles of PPAR, which is widely expressed, are, as yet, less well defined. Although all three isoforms are capable of binding and responding to a wide array of lipid mediators, the question of whether each isoform has a specific high-affinity endogenous ligand is still unresolved.
Murine genetic knockout models have contributed substantially to our understanding of the biology of the PPARs (19, 20). Human genetics has also made important contributions: most notably, four different germline mutations of PPAR resulting in amino acid substitutions in the LBD of the receptor have been described in association with a stereotyped clinical syndrome of severe insulin resistance, dyslipidemia, hypertension, hepatic steatosis, and partial lipodystrophy. Two of these (Pro467Leu and Val290Met) have been shown to exert potent dominant-negative activity over their wild-type counterpart (3, 4). In contrast, Phe388Leu reportedly failed to exhibit dominant-negative activity (21). However, subsequent studies in our own laboratory with both the Phe388Leu and previously uncharacterized Arg425Cys mutants have revealed comparable ability to interfere with wild-type signaling (Gurnell, M., and V. K. K. Chatterjee, unpublished data). The importance of dominant negativity, as opposed to haploinsufficiency, is supported by the absence of a marked metabolic phenotype in human subjects carrying a PPAR frameshift/premature stop mutation, which produces a transcriptionally inactive truncated receptor that is unable to act as a dominant negative (22). It is further attested to by the discovery that a proportion of thyroid follicular carcinomas express a fusion protein consisting of the DBD of the transcription factor paired box transcription factor 8 and the LBD of PPAR. This chimeric species not only fails to transactivate on PPAR-responsive promoters but also suppresses the transactivation function of wild-type PPAR on such promoters, an activity that has been implicated in the pathogenesis of these neoplasia (23).
In contrast, no pathogenic mutations in PPAR or PPAR have been described to date in human subjects. A human PPAR splice variant (PPARtr) lacking exon 6 (resulting in deletion of part of the hinge region and the entire LBD) has been reported, comprising 20–50% of native PPAR in some human cells. Although this splice variant was unable to bind DNA, it was able to exert dominant-negative activity after forced nuclear localization (24).
To explore PPAR-mediated dominant-negative activity further, we exploited a strategy previously successfully employed for PPAR (5) to create a powerful artificial dominant-negative PPAR (PPARmut) by mutagenesis to alanine of two residues in the activation function 2 domain that critically regulate coactivator recruitment. Analysis of the structure of liganded PPAR bound to coactivator peptide (PDB 1K7L) (25) (Fig. 10) provides support for this strategy as applied to PPAR. Thus, glutamate 462 makes two hydrogen bonds to backbone amides in the coactivator helix and complements the N-terminal-positive charge of the helix dipole. This charge clamp in the wild-type molecule anchors the amphipathic, helical LXXLL motifs of many coactivators, thereby orienting the helices and apposing hydrophobic interaction surfaces on coregulatory molecules and nuclear receptor (26). Second, leucine 459 forms part of closely packed nonpolar interaction with helices 3 and 12 and the helical LXXLL coactivator motif. Interrupting both the charge clamp and hydrophobic coactivator-binding pocket would be predicted to have a marked destabilizing influence on the PPAR-coactivator interaction.
FIG. 10. Location of L459 and E462 in the three-dimensional crystal structure of PPAR complexed to a chemical agonist and coactivator peptide (PDB 1K7L).
Both the previously described double-mutant PPAR (PPARmut) and PPARmut act as constitutive repressors in the basal state, exhibit strong dominant-negative inhibition of transcriptional activation by their wild-type counterpart in response to a specific ligand, and recruit NCoR and SMRT avidly. However, the corepressor association was significantly more attenuated by ligand in the case of PPARmut than PPARmut. In the same system, wild-type PPAR interacted with both SMRT and NCoR (5), whereas no association was seen in this study for wild-type PPAR. This suggests that, whereas the double mutation in each case enhances corepressor interaction, the baseline levels of corepressor recruitment of the two receptors differ. Whether this is due to intrinsic differences in corepressor binding of the unliganded receptors or rather due to different degrees of receptor occupancy by endogenous ligand remains undetermined (5). The failure of PGC1 to augment transcriptional activation by cotransfected PPARmut is consistent with structural predictions (Fig. 10) and provides evidence that, as for PPARmut, the dominant-negative activity of PPARmut is due to impaired coactivator recruitment in the liganded state as well as enhanced corepressor association.
NCoR and SMRT possess a bipartite nuclear receptor interaction domain, which permits binding to specific residues in the hydrophobic pocket of nuclear receptors. Ligand binding, by establishing the charge clamp, introduces steric constraints on the length of helix that can be accommodated in the hydrophobic pocket, resulting in dissociation of corepressors with their longer helical interaction domain and binding of the smaller LXXLL-containing domain of coactivators (27, 28, 29). Not only has the crystal structure of a PPAR-agonist-coactivator complex been described but, uniquely to date among nuclear hormone receptors, the PPAR structure in complex with a chemical antagonist and a corepressor peptide from SMRT (30). However, how these well-delineated interactions relate to transcriptional regulation by PPARs in a cellular context is less clear. Although unliganded nuclear receptors such as thyroid hormone receptor and RAR are known to recruit corepressors in the absence of exogenous ligand, whether PPARs do so in vivo is contentious. Ligand-dissociable interaction between PPAR and NCoR has been shown in a yeast two-hybrid assay, and cotransfection studies in human embryonic kidney 293 cells have shown NCoR to suppress the constitutive reporter activity seen in the presence of PPAR and RXR, this repression being alleviated by specific ligand dose dependently (31). However, consistent with our results, PPAR-Gal4 fusion proteins have been shown to attenuate transcriptional activation by Gal4 in yeast but not mammalian cells (32), and it has not proved possible to detect either DNA-bound PPAR-RXR-NCoR or PPAR-RXR-SMRT complexes (32). Our failure to detect an interaction between NCoR or SMRT and the wild-type PPAR LBD in a mammalian two-hybrid system adds to these conflicting data.
It is possible that these discrepant results are explained by variable levels of endogenous PPAR ligand in different cell types. It may be that in some cells the generation of low levels of intracellular ligand means that native PPAR is never truly in the unliganded state, so that helix 12 is always in an active conformation, precluding corepressor interaction. Leu459Ala and Glu462Ala double mutation may mimic the binding of a chemical antagonist, destabilizing helix 12, and strongly promoting corepressor interaction, even in the presence of endogenous ligand. However, as shown in Fig. 3, potent exogenous ligand retains the ability to promote dissociation of PPARmut from corepressors in the two-hybrid system.
The dominant-negative activity of PPARmut was shown on an artificial PPRE and native robustly PPAR-responsive promoters in both HepG2 cells and rat primary hepatocytes, establishing that PPARmut is a potent dominant-negative inhibitor of PPARwt on endogenous promoters, even in primary culture in which PPAR expression levels are high. Similarly, concerns that cross-inhibition of PPAR and PPAR by PPARmut was due only to the use of a minimum artificial PPRE and that, in a more natural promoter context, selectivity would be exhibited were addressed by examining the effect of PPARmut expression on lipid accumulation and the induction of aP2 expression during differentiation of 3T3-L1 preadipocytes in the presence of a potent PPAR ligand (Fig. 9). Despite reports that the aP2 enhancer PPRE, which mediates activation by PPAR, is the most PPAR-selective of a raft of PPREs examined in vitro (33), expression of PPARmut was sufficient to block both aP2 transcriptional up-regulation and lipid accumulation. Together with our demonstration that PPARmut also severely attenuates the transcriptional response to a PPAR ligand in HepG2 cells, this suggests that the lack of selectivity of the mutant receptor’s inhibitory activity may be generalizable to specific responses mediated by all three PPARs. Moreover reexamination of the properties of PPARmut showed that it, too, is not PPAR selective with regard to suppression of PPAR-mediated transcriptional responses from endogenous promoters, albeit to a lesser extent than PPARmut. Although the lack of specificity using adenoviral delivery of the mutant PPARs could be attributed to high levels of expression from the constitutive CMV promoter in the viral vectors, a similar lack of specificity was seen in cotransfection experiments with stoichiometric levels of expression of wild-type and mutant receptors.
The lack of selectivity of the artificial PPAR mutants described here may also have pathophysiological relevance for human subjects harboring dominant-negative PPAR mutations, which are believed to exhibit mechanistically similar, albeit less potent, dominant-negative effects (3, 4). Indeed, before loss of PPAR signaling alone can be deemed to be the cause of the severe metabolic syndrome seen in subjects harboring dominant-negative PPAR mutations, it must be established whether dominant-negative inhibition by these naturally occurring mutants can spill over onto other PPARs. Evidence for such inhibitory cross-talk is also provided by a recent study of wild-type PPAR, which, in the unliganded state, was shown to inhibit PPAR and PPAR action, probably due to occupancy of the relevant PPAR response elements allied to its greater affinity for SMRT and perhaps other corepressors (32). Thus, it is conceivable that in some or all of the tissues in which the receptor isoforms are coexpressed, cross-inhibition of PPAR and/or PPAR signaling by a mutant PPAR with an aberrant affinity for corepressor may occur.
In summary, we have generated and characterized a double-mutant PPAR with potent dominant-negative activity toward its wild-type counterpart due to enhanced corepressor recruitment and impaired interaction with coactivators. The mutant receptor exhibits dominant-negative activity both in cotransfection reporter assays and when tested against endogenous PPAR responses in cell culture, and this inhibition is manifest against all three PPAR isoforms. This is also true of the previously reported analogous dominant-negative PPAR mutant, which may have implications for the pathogenesis of the metabolic syndrome seen in human subjects harboring dominant-negative PPAR mutations.
References
Savage DB, Tan GD, Acerini CL, Jebb SA, Agostini M, Gurnell M, Williams RL, Umpleby AM, Thomas EL, Bell JD, Dixon AK, Dunne F, Boiani R, Cinti S, Vidal-Puig A, Karpe F, Chatterjee VK, O’Rahilly S 2003 Human metabolic syndrome resulting from dominant-negative mutations in the nuclear receptor peroxisome proliferator-activated receptor-. Diabetes 52:910–917
Gurnell M, Savage DB, Chatterjee VK, O’Rahilly S 2003 The metabolic syndrome: peroxisome proliferator-activated receptor and its therapeutic modulation. J Clin Endocrinol Metab 88:2412–2421
Agostini M, Gurnell M, Savage DB, Wood EM, Smith AG, Rajanayagam O, Garnes KT, Levinson SH, Xu HE, Schwabe JW, Willson TM, O’Rahilly S, Chatterjee VK 2004 Tyrosine agonists reverse the molecular defects associated with dominant-negative mutations in human peroxisome proliferator-activated receptor . Endocrinology 145:1527–1538
Barroso I, Gurnell M, Crowley VE, Agostini M, Schwabe JW, Soos MA, Maslen GL, Williams TD, Lewis H, Schafer AJ, Chatterjee VK, O’Rahilly S 1999 Dominant-negative mutations in human PPAR associated with severe insulin resistance, diabetes mellitus and hypertension. Nature 402:880–883
Gurnell M, Wentworth JM, Agostini M, Adams M, Collingwood TN, Provenzano C, Browne PO, Rajanayagam O, Burris TP, Schwabe JW, Lazar MA, Chatterjee VK 2000 A dominant-negative peroxisome proliferator-activated receptor (PPAR) mutant is a constitutive repressor and inhibits PPAR-mediated adipogenesis. J Biol Chem 275:5754–5759
Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM 1995 15-Deoxy- 12, 14-prostaglandin J2 is a ligand for the adipocyte determination factor PPAR . Cell 83:803–812
Tone Y, Collingwood TN, Adams M, Chatterjee VK 1994 Functional analysis of a transactivation domain in the thyroid hormone ? receptor. J Biol Chem 269:31157–31161
Heyman RA, Mangelsdorf DJ, Dyck JA, Stein RB, Eichele G, Evans RM, Thaller C 1992 9-Cis retinoic acid is a high affinity ligand for the retinoid X receptor. Cell 68:397–406
Collingwood TN, Adams M, Tone Y, Chatterjee VK 1994 Spectrum of transcriptional, dimerization, and dominant-negative properties of twenty different mutant thyroid hormone ?-receptors in thyroid hormone resistance syndrome. Mol Endocrinol 8:1262–1277
Wentworth JM, Schoenfeld V, Meek S, Elgar G, Brenner S, Chatterjee VK 1999 Isolation and characterisation of the retinoic acid receptor- gene in the Japanese pufferfish, F. rubripes. Gene 236:315–323
Schmidt A, Endo N, Rutledge SJ, Vogel R, Shinar D, Rodan GA 1992 Identification of a new member of the steroid hormone receptor superfamily that is activated by a peroxisome proliferator and fatty acids. Mol Endocrinol 6:1634–1641
Brown PJ, Stuart LW, Hurley KP, Lewis MC, Winegar DA, Wilson JG, Wilkison WO, Ittoop OR, Willson TM 2001 Identification of a subtype selective human PPAR agonist through parallel-array synthesis. Bioorg Med Chem Lett 11:1225–1227
Boehm MF, Zhang L, Zhi L, McClurg MR, Berger E, Wagoner M, Mais DE, Suto CM, Davies JA, Heyman RA 1995 Design and synthesis of potent retinoid X receptor selective ligands that induce apoptosis in leukemia cells. J Med Chem 38:3146–3155
Lehmann JM, Moore LB, Smith-Oliver TA, Wilkison WO, Willson TM, Kliewer SA 1995 An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor (PPAR). J Biol Chem 270:12953–12956
Berger J, Leibowitz MD, Doebber TW, Elbrecht A, Zhang B, Zhou G, Biswas C, Cullinan CA, Hayes NS, Li Y, Tanen M, Ventre J, Wu MS, Berger GD, Mosley R, Marquis R, Santini C, Sahoo SP, Tolman RL, Smith RG, Moller DE 1999 Novel peroxisome proliferator-activated receptor (PPAR) and PPAR ligands produce distinct biological effects. J Biol Chem 274:6718–6725
Duerden JM, Bartlett SM, Gibbons GF 1989 Long-term maintenance of high rates of very-low-density-lipoprotein secretion in hepatocyte cultures. A model for studying the direct effects of insulin and insulin deficiency in vitro. Biochem J 263:937–943
Hebbachi AM, Brown AM, Gibbons GF 1999 Suppression of cytosolic triacylglycerol recruitment for very low density lipoprotein assembly by inactivation of microsomal triglyceride transfer protein results in a delayed removal of apoB-48 and apoB-100 from microsomal and Golgi membranes of primary rat hepatocytes. J Lipid Res 40:1758–1768
Muoio DM, MacLean PS, Lang DB, Li S, Houmard JA, Way JM, Winegar DA, Corton JC, Dohm GL, Kraus WE 2002 Fatty acid homeostasis and induction of lipid regulatory genes in skeletal muscles of peroxisome proliferator-activated receptor (PPAR) knock-out mice. Evidence for compensatory regulation by PPAR. J Biol Chem 277:26089–26097
Miles PD, Barak Y, Evans RM, Olefsky JM 2003 Effect of heterozygous PPAR deficiency and TZD treatment on insulin resistance associated with age and high-fat feeding. Am J Physiol Endocrinol Metab 284:E618–E626
Wang YX, Lee CH, Tiep S, Yu RT, Ham J, Kang H, Evans RM 2003 Peroxisome-proliferator-activated receptor activates fat metabolism to prevent obesity. Cell 113:159–170
Hegele RA, Cao H, Frankowski C, Mathews ST, Leff T 2002 PPARG F388L, a transactivation-deficient mutant, in familial partial lipodystrophy. Diabetes 51:3586–3590
Savage DB, Agostini M, Barroso I, Gurnell M, Luan J, Meirhaeghe A, Harding AH, Ihrke G, Rajanayagam O, Soos MA, George S, Berger D, Thomas EL, Bell JD, Meeran K, Ross RJ, Vidal-Puig A, Wareham NJ, O’Rahilly S, Chatterjee VK, Schafer AJ 2002 Digenic inheritance of severe insulin resistance in a human pedigree. Nat Genet 31:379–384
Kroll TG, Sarraf P, Pecciarini L, Chen CJ, Mueller E, Spiegelman BM, Fletcher JA 2000 PAX8-PPAR1 fusion oncogene in human thyroid carcinoma. Science 289:1357–1360
Gervois P, Torra IP, Chinetti G, Grotzinger T, Dubois G, Fruchart JC, Fruchart-Najib J, Leitersdorf E, Staels B 1999 A truncated human peroxisome proliferator-activated receptor splice variant with dominant-negative activity. Mol Endocrinol 13:1535–1549
Xu HE, Lambert MH, Montana VG, Plunket KD, Moore LB, Collins JL, Oplinger JA, Kliewer SA, Gampe Jr RT, McKee DD, Moore JT, Willson TM 2001 Structural determinants of ligand binding selectivity between the peroxisome proliferator-activated receptors. Proc Natl Acad Sci USA 98:13919–13924
Heery DM, Kalkhoven E, Hoare S, Parker MG 1997 A signature motif in transcriptional coactivators mediates binding to nuclear receptors. Nature 387:733–736
Perissi V, Staszewski LM, McInerney EM, Kurokawa R, Krones A, Rose DW, Lambert MH, Milburn MV, Glass CK, Rosenfeld MG 1999 Molecular determinants of nuclear receptor-corepressor interaction. Genes Dev 13:3198–3208
Hu X, Lazar MA 1999 The CoRNR motif controls the recruitment of corepressors by nuclear hormone receptors. Nature 402:93–96
Nagy L, Kao HY, Love JD, Li C, Banayo E, Gooch JT, Krishna V, Chatterjee K, Evans RM, Schwabe JW 1999 Mechanism of corepressor binding and release from nuclear hormone receptors. Genes Dev 13:3209–3216
Xu HE, Stanley TB, Montana VG, Lambert MH, Shearer BG, Cobb JE, McKee DD, Galardi CM, Plunket KD, Nolte RT, Parks DJ, Moore JT, Kliewer SA, Willson TM, Stimmel JB 2002 Structural basis for antagonist-mediated recruitment of nuclear co-repressors by PPAR. Nature 415:813–817
Dowell P, Ishmael JE, Avram D, Peterson VJ, Nevrivy DJ, Leid M 1999 Identification of nuclear receptor corepressor as a peroxisome proliferator-activated receptor interacting protein. J Biol Chem 274:15901–15907
Shi Y, Hon M, Evans RM 2002 The peroxisome proliferator-activated receptor, an integrator of transcriptional repression and nuclear receptor signaling. Proc Natl Acad Sci USA 99:2613–2618
Juge-Aubry C, Pernin A, Favez T, Burger AG, Wahli W, Meier CA, Desvergne B 1997 DNA binding properties of peroxisome proliferator-activated receptor subtypes on various natural peroxisome proliferator response elements. Importance of the 5'-flanking region. J Biol Chem 272:25252–25259