If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nuclear hormone receptor peroxisome proliferator-activated receptor γ (PPARγ (NR1C3)) plays a central role in adipogenesis and is the molecular target for the thiazolidinedione (TZD) class of antidiabetic drugs. In a search for novel non-TZD ligands for PPARγ, T0070907 was identified as a potent and selective PPARγ antagonist. With an apparent binding affinity (concentration at 50% inhibition of [3H]rosiglitazone binding or IC50) of 1 nm, T0070907 covalently modifies PPARγ on cysteine 313 in helix 3 of human PPARγ2. T0070907 blocked PPARγ function in both cell-based reporter gene and adipocyte differentiation assays. Consistent with its role as an antagonist of PPARγ, T0070907 blocked agonist-induced recruitment of coactivator-derived peptides to PPARγ in a homogeneous time-resolved fluorescence-based assay and promoted recruitment of the transcriptional corepressor NCoR to PPARγ in both glutathione S-transferase pull-down assays and a PPARγ/retinoid X receptor (RXR) α-dependent gel shift assay. Studies with mutant receptors suggest that T0070907 modulates the interaction of PPARγ with these cofactor proteins by affecting the conformation of helix 12 of the PPARγ ligand-binding domain. Interestingly, whereas the T0070907-induced NCoR recruitment to PPARγ/RXRα heterodimer can be almost completely reversed by the simultaneous treatment with RXRα agonist LGD1069, T0070907 treatment has only modest effects on LGD1069-induced coactivator recruitment to the PPARγ/RXRα heterodimer. These results suggest that the activity of PPARγ antagonists can be modulated by the availability and concentration of RXR agonists. T0070907 is a novel tool for the study of PPARγ/RXRα heterodimer function.
). PPARγ2 possesses 30 additional amino acids at its amino terminus. PPARγ1 is expressed broadly in many tissues, whereas PPARγ2 is expressed predominantly in adipose tissue. Both “gain of function” and “loss of function” studies strongly support a critical role for PPARγ in adipocyte gene expression and differentiation (
Like other members of the NHR superfamily, PPARγ binds to a DNA-response element (PPAR-response element or PPRE) upstream of the coding regions of target genes and forms a heterodimeric complex with one of the three retinoid X receptor (RXR) proteins (
). Binding of ligands to PPARγ causes conformational changes in the receptor, in particular to α-helix 12 (H12), which is located at the carboxyl-terminal end of the protein and forms part of the transcriptional activation function (AF-2). When agonists bind to PPARγ, H12 along with H3, H4, and H5 form a charge clamp and a hydrophobic pocket that allows the recruitment of coactivator protein complexes that are essential for transcriptional activation of PPARγ target genes (
). Although PPARγ, in isolation, is capable of binding to transcriptional corepressor proteins NCoR and SMRT in the absence of ligand, PPARγ does not interact with these corepressors in the context of the RXR heterodimer nor does the PPARγ/RXR heterodimer repress transcription of PPARγ target genes, unlike heterodimers of RXR with thyroid hormone receptor or retinoic acid receptor (
). Two explanations for the difference in PPARγ/corepressor interaction on and off DNA have been offered. The orientation of PPARγ and RXR on a PPRE could simply inhibit the binding of corepressor (
). Alternatively, PPARγ may be unable to stabilize a conformation of RXR that is permissive for corepressor interaction; unlike TR and retinoic acid receptor, PPARγ is unable to interact with H12 from RXR (
). Because other NHR antagonists stabilize the interaction of corepressors with their cognate receptors, PPARγ antagonists or inverse agonists would be useful tools to study PPARγ/corepressor interaction. One way to test these hypotheses is to study the effects of a PPARγ antagonist or inverse agonist on corepressor binding. This avenue has not yet been explored.
Both natural and synthetic ligands have been reported for PPARγ (reviewed in Ref.
). Naturally occurring compounds that have been reported to bind PPARγ include a number of fatty acids and eicosanoid derivatives such as 9- or 13-hydroxyoctadienoic acid and prostaglandin derivative 15-deoxy-Δ13,14-prostaglandin J2. The most widely used synthetic agonists of PPARγ are members of a class of antidiabetic agents known as thiazolidinediones (TZDs), including rosiglitazone, troglitazone, and pioglitazone. More recently, a series of tyrosine-based PPARγ agonists exemplified by GI262570 have been shown to be among the highest affinity PPARγ ligands described thus far. Synthetic partial agonists identified include GW0072 (
). However, relatively little is known about how these compounds affect PPARγ/RXR heterodimer function.
Here, we describe a novel, potent, and selective PPARγ ligand, T0070907. By using a variety of biochemical and cell-based assays, we demonstrate that T0070907 is a PPARγ antagonist. Our studies suggest that T0070907 modulates the interaction of PPARγ with cofactor proteins by affecting the conformation of helix 12 of the PPARγ ligand-binding domain (LBD). Finally, our studies reveal a functional asymmetry between the effects of PPARγ and RXR ligands on the activity of the permissive PPARγ/RXRα heterodimer.
EXPERIMENTAL PROCEDURES
Plasmid Construction
All PPARγ and RXRα proteins produced by in vitro translation and used in gel mobility shift assay were expressed from pET28b-based plasmids (Novagen, Madison, WI) containing inserts cloned into NcoI and NotI sites. For PPARγ, the following constructs were made: hPPARγ1, full-length human PPARγ1; hPPARγ1 ΔH12 (amino acids 1–461 of hPPARγ1); hRXRα, full-length human RXRα; hRXRα ΔH12 (amino acids 1–443 of hRXRα). The construct used to produce human NCoR (hNCoR) protein from the baculovirus expression system was made by inserting hNCoR DNA encoding amino acids 1948–2440 into the BamHI and NotI sites of the pFASTBAC HTa vector (Invitrogen). GST-PPARγ LBD was constructed by inserting hPPARγ1 DNA encoding amino acids 175–471 into the BamHI site of pGEX-2TK vector (sequence at the junction is 5′-GGATCCCATATG-hPPARγ1 amino acid 175).
Mass Spectrometry
After incubation with 10 μmT0070907 for 4 h at room temperature in 50 mm Tris, pH 7.9, 50 mm KCl, 1 mm EDTA, GST-PPARγ (12 μg) was purified on an SDS-polyacrylamide gel. The excised gel fragment containing PPARγ was digested with trypsin at 37 °C for 12 h without reduction and alkylation in 100 mmammonium bicarbonate using an enzyme/substrate ratio of 1:50 (w/w). Analysis of covalent binding of T0070907 to PPARγ was performed with a Voyager-DETM MALDI-TOF Mass Spectrometer (Perspective Biosystems, Framingham, MA) and an EsquireTMNano-electrospray Tandem Mass Spectrometer (Bruker Daltonik, Billerica, MA).
The MALDI matrix was prepared by mixing α-cyano-4-hydroxy-trans-cinnamic acid (40 mg/ml in acetone), nitrocellulose (20 mg/ml in acetone), and 2-propanol at a ratio of 2:1:1 (v/v/v). An aliquot (0.5 μl) of the sample/matrix was spotted and mixed on a MALDI sample plate. After drying completely, samples were washed with 3 μl of 5% formic acid and then with Milli-Q water (Millipore, Bedford, MA). The MALDI-TOF was operated in reflectron mode with delay extraction. The mass spectrometer was calibrated externally using des-Arg-bradykinin (m/z 904.4681), angiotensin I (m/z 1296.6853), and Glu-fibrinopeptide B (m/z 1570.6774) (
The in-gel tryptic digest was desalted and concentrated prior to analysis by nano-ESI-MS/MS. The in-gel digests were extracted twice with 10 μl of a 50% acetonitrile, 5% trifluoroacidic acid solution. All extracts were pooled and dried to 5 μl with a SpeedVac. An additional 15 μl of a 0.1% acidic acid solution was added, and a 10-μl sample was loaded into a ZiptipTM (Millipore, Bedford, MA) with C18 resin for desalting. After washing the column with 1% trifluoroacidic acid (in H2O), 5 μl of 50% acetonitrile, 0.1% acidic acid was used to elute the sample into a nanospray needle. On the basis of the MALDI-TOF mass analysis results, the tandem mass spectrometric sequencing was only acquired on selected precursor ions (
) with the following modifications. Reaction conditions were as follows: a 100-μl reaction volume contained 50 mm Tris, pH 7.9, 50 mm KCl, 1 mm EDTA, 0.5 mm 2-mercaptoethanol, 0.1 mg/ml bovine serum albumin, 800 ng/ml anti-GST-(Eu)K antibody (PerkinElmer Life Sciences), 1 ng/μl GST-PPARγ, 1.5 μg/ml streptavidin conjugated with allophycocyanin (Streptavidin-APC, PerkinElmer Life Sciences), 200 nm biotin-peptide, and 5 μl compound of interest in dimethyl sulfoxide (Me2SO) as indicated in the figure legends. GST-PPARγ/anti-GST-(Eu)K (20 μl) and biotin-peptide/streptavidin (20 μl) were incubated separately for 1 h at room temperature before being combined with the remaining components, and the complete mixture was incubated for an additional 1 h at room temperature. Reactions were carried out in 96-well plates (black polypropylene, Whatman), and fluorescence was measured on an LJL Analyst (LJL BioSystems, Sunnyvale, CA). Data were expressed as the ratio of the emission intensities at 665 and 620 nm multiplied by a factor of 1000.
Corepressor Recruitment Assay (Pull-down Assay)
Purified GST-PPARγ fusion protein (15 μg) was incubated with 10 μl of glutathione-Sepharose beads (50% slurry in GST binding buffer,Amersham Biosciences) in GST binding buffer (20 mm HEPES, pH 7.7, 100 mm KCl, 0.1 mm EDTA, 2.5 mm MgCl2, 0.01% Nonidet P-40, 2 mmdithiothreitol, 10% glycerol) for 90 min at room temperature. After washing 5 times (5 min each wash) with 1 ml of binding buffer, the bead-bound GST-PPARγ protein was incubated with 7.5 μl of [35S]methionine-labeled in vitro translated hNCoR protein (TNT T7 Rabbit Reticulocyte Lysate Translation System, Promega Corp., Madison, WI) and the indicated ligand concentration in a final volume of 300 μl at room temperature for 2 h. After washing with binding buffer as indicated above, the bound protein was eluted with 20 μl of 2× SDS buffer at 95 °C, separated on a 10% SDS-PAGE, and analyzed by autoradiography.
Ligand Binding Assay
To determine the binding affinity of T0070907 to the PPARs, scintillation proximity assay (SPA) was performed as described (
) with the following modifications. A 90-μl reaction contained SPA buffer (10 mmK2HPO4, 10 mmKH2PO4, 2 mm EDTA, 50 mm NaCl, 1 mm dithiothreitol, 2 mmCHAPS, 10% (v/v) glycerol, pH 7.1), 50 ng of GST-PPARγ (or 150 ng of GST-PPARα, GST-PPARδ), 5 nm3H-labeled radioligands, and 5 μl of T0070907 in Me2SO. After incubation for 1 h at room temperature, 10 μl of polylysine-coated SPA beads (at 20 mg/ml in SPA buffer) were added, and the mixture was incubated for 1 h before reading in Packard Topcount. [3H]Rosiglitazone was used for PPARγ, and [3H]GW2433 (
) with the following modifications. The sequence of the DNA probe used in the GMSAs was derived from the PPRE of the acyl-CoA oxidase gene (5′-AGCTGGACCAGGACAAAGGTCACGTTCAGCT-3′).In vitro translated PPARγ (0.5 μl) and RXRα (0.5 μl) were incubated in 20 mm Tris, pH 8.0, 1 mmEDTA, 50 mm KCl, 0.05% Nonidet P-40, 10% glycerol, 2 mm dithiothreitol, 50 μg/ml poly(dI-dC), labeled probe (typically 40,000 cpm per reaction), various ligands, and with or without baculovirus-expressed NCoR (5 μg) (final volume, 20 μl) for 30 min at room temperature. Reaction mixtures were loaded on a 5% (38:2)polyacrylamide nondenaturing gel in 1× TGE (50 mmTris, pH 8.5, 40 mm glycine, 2 mm EDTA) and separated in 1× TGE by electrophoresis at 4 °C. Gels were dried prior to autoradiography.
Transient Transfection and 3T3L1 Differentiation Assay
Luciferase reporter assays were carried out following transient transfection of HEK293 cells using GenePORTER2 reagent (GTS Inc., San Diego, CA) according to the manufacturer's protocol. 3T3-L1 preadipocytes were cultured and induced to differentiate as described (
) with the following modifications. 3T3-L1 cells were grown to confluence in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and induced to differentiate with 0.25 μmdexamethasone, 0.5 mm isobutylmethylxanthine, and 1 μg/ml insulin. Medium was replaced 2 days post-induction (and every 2–3 days thereafter) with Dulbecco's modified Eagle's medium, 10% fetal bovine serum supplemented with 1 μg/ml insulin.
RESULTS
T0070907 Is a Novel and Selective PPARγ Ligand
In a search for novel non-TZD ligands for PPARγ, T0070907 was identified to bind PPARγ with high affinity, capable of displacing [3H]rosiglitazone with an apparent Ki of 1 nm as shown in Fig. 1. Furthermore, T0070907 shows high selectivity among PPAR subtypes with a >800-fold preference for PPARγ over PPARα and PPARδ. In competition with the PPARα and PPARδ co-ligand [3H]GW2433 (
), T0070907 has an apparent Ki of 0.85 μm to PPARα and 1.8 μm to PPARδ (Fig. 1B).
Figure 1Chemical structure of T0070907 and analysis of its binding to PPARs.A, chemical structure of T0070907. B, the affinities of T0070907 for the various PPAR subtypes were determined with an SPA assay. [3H]GW2433 (5 nm) was used as a ligand for PPARα and PPARδ, and [3H]rosiglitazone (5 nm) was used for PPARγ.
T0070907 Is a Specific Potent PPARγ Antagonist in Transient Transfection Assays
The effect of T0070907 on the transcriptional activity of PPARγ in a cell-based reporter gene assay was examined. HEK293 cells were transiently transfected with an expression construct that contained the PPARγ LBD fused to the Gal4-DNA binding domain, together with a luciferase reporter gene under the transcriptional control of the Gal4 upstream activating sequence (Gal4-UAS). As shown in Fig. 2A, rosiglitazone activated transcription up to 20-fold, whereas T0070907 has no effect (or perhaps even a slight inhibitory effect) on basal transcription. In addition, T0070907 is a potent inhibitor (IC50 value in the nm range) of PPARγ transactivation in the presence of rosiglitazone (Fig. 2A). This inhibition is not due to cytotoxicity as the concentration required to kill 50% of cells is greater than 10 μm (data not shown). The specificity of T0070907 was also examined in cell-based reporter gene assays. HEK293 cells were transiently transfected with a GAL4-UAS reporter and expression constructs encoding the LBDs of PPARα, PPARδ, the farnesoid X receptor, the liver X receptor α, the liver X receptor β, or pregnane X receptor fused to the Gal4-DNA binding domain. As shown in Fig. 2A, T0070907 at 1 μm has no effect on the transcriptional activity of any other receptor besides PPARγ. These results demonstrate that T0070907 is a PPARγ-specific antagonist.
Figure 2T0070907 is a specific potent PPARγ antagonist.A, effects of T0070907 on transcriptional activities of the LBDs of PPARγ, PPARα, PPARδ, farnesoid X receptor, liver X receptor α, liver X receptor β, or pregnane X receptor γ fused with Gal4 protein on Gal4-UAS-luciferase reporter gene expression in HEK293 cells.B, effects of T0070907 on dexamethasone, 3-isobutyl-1-methylxanthine, and insulin-induced (DEX/MIX/INS) 3T3-L1 cell differentiation. Top panel, 3T3-L1 differentiation protocol. Cells were stained with Nile Red and photographed under light microscope. DMSO, Me2SO.
T0070907 Blocks Hormone-mediated Differentiation of the Adipogenic Cell Line 3T3-L1
We next investigated whether T0070907 could block the induction of adipogenesis by various treatments of the adipogenic cell line 3T3-L1. As shown in Fig. 2B, the standard treatment of dexamethasone, 3-isobutyl-1-methylxanthine, and insulin promoted lipid accumulation in 3T3-L1 cells. In contrast, lipid accumulation in these cells was completely inhibited when cells were treated with both 1 μm T0070907 and the differentiation mixture. Similar inhibitory effects of T0070907 were observed when adipogenesis was induced by treatment with the PPARγ agonist, rosiglitazone (data not shown).
T0070907 Covalently Modifies PPARγ on Cys313
To understand the mechanism by which T0070907 antagonizes PPARγ function, its binding properties were first examined. That rosiglitazone was unable to displace T0070907 prebound to PPARγ suggested that the binding of T0070907 was irreversible (data not shown). To verify the covalent nature of the interaction between T0070907 and PPARγ, and to identify the site of covalent attachment, we performed proteolytic mapping studies via mass spectrometry. The covalent binding of T0070907 (mass of 277.7 Da) to PPARγ would result in a mass change of the modified tryptic peptide(s) by 241.1 Da. By comparing the tryptic digests of PPARγ with and without T0070907 treatment, a candidate peptide containing the T0070907 attachment site (amino acids 272–279, IFQGCQFR, m/z998.49 Da) was identified based on its mass shift to m/z 1239.56 (data not shown). The precise binding site on this peptide was determined with ESI-tandem mass spectrometry (Fig. 3A). The calculated dominant y- and b-ion fragments of this peptide are shown at the top of Fig. 3A, with the ions observed in the mass spectrum underlined. The mass difference between the y3- and y4-ions (m/z344.1) identified Cys313 as the site of modification by T0070907. In addition, several double-charged y-ions and small internal fragment ions obtained also confirmed this conclusion.
Figure 3T0070907 covalently binds to PPARγ at Cys313.A, nano-ESI-MS/MS spectrum of tryptic fragment modified by T0070907. The calculated y- and b-ions of this peptide are shown at the top of the figure, with ions observed in the mass spectrum underlined. Cys313 is indicated by *. B, wt or C313S mutant PPARγ proteins were incubated with [3H]T0070907 and subsequently separated on an SDS-polyacrylamide gel, which was then subjected to autoradiography.Upper panel, autoradiography of the SDS-polyacrylamide gel. [3H]T0070907 bound covalently only to wt PPARγ.Lower panel, Coomassie Blue-stained same gel showing equal loading of wt and mutant PPARγ proteins. C, purified GST-PPARγ mixed with whole-cell extract (WCE) derived from HEK293 cells was incubated with 2 μm[3H]T0070907 and subsequently separated on an SDS-polyacrylamide gel, which was then subjected to autoradiography.Upper panel, autoradiography of the gel. Only GST-PPARγ was preferentially modified by [3H]T0070907. Lower panel, Coomassie Blue staining of the same gel.
To confirm the importance of Cys313 in T0070907 binding to PPARγ, a mutant PPARγ was constructed in which Cys313was converted to a serine residue, and the corresponding recombinant GST-PPARγ LBD (C313S) fusion protein was expressed and purified. [3H]T0070907 was first incubated with either wild (wt) type PPARγ or the C313S mutant as described, and the reaction mixtures were separated by SDS-PAGE. As shown in Fig. 3B, [3H]T0070907 was only able to modify the wild type protein (upper panel, autoradiograph), although equal amounts of wild type and C313S mutant PPARγ were added to each reaction (lower panel, Coomassie staining). Thus, Cys313 is necessary for the binding of T0070907 to PPARγ.
The specificity of the covalent modification was examined in a whole-cell extract (WCE) made from HEK293 cells. Purified GST-PPARγ LBD fusion protein (Fig. 3C) was mixed with the WCE and 2 μm [3H]T0070907 and then separated on an SDS-PAGE gel. As shown in Fig. 3C, the exogenously added PPARγ was preferentially modified by [3H]T0070907 (upper panel, autoradiograph), despite the presence of many other proteins in the WCE (lower panel, Coomassie staining).
T0070907 Behaves as an Inverse Agonist of PPARγ LBD in Vitro
By using the homogeneous time-resolved fluorescence (HTRF) technology, we developed an assay to study the effects of PPARγ ligands on the interaction of PPARγ with fragments of coactivator or corepressor proteins. Reporter peptides of ∼20 amino acids in length were synthesized from sequences derived from various coactivator and corepressor proteins (Table I) (
). The effects of various ligands on PPARγ binding to this collection of peptides are shown in Fig. 4A. The patterns that emerged from this peptide profiling have allowed us to distinguish between different functional classes of PPARγ ligands. First, known PPARγ agonists such as rosiglitazone, troglitazone, and the GSK compound GI262570 (
) showed very similar peptide profiles. Rosiglitazone and troglitazone, which both belong to the TZD chemical class, were more similar to each other than to tyrosine-based GI262570. GI262570 recruited additional peptides (peptides 11, 19, 23, and 27), suggesting perhaps that PPARγ adopted a slightly different conformation when bound to GI262570, compared with the conformation assumed by the receptor when bound to the TZD compounds. In contrast, the novel PPARγ ligand T0070907 shows a unique peptide profile (Fig. 4A), exclusively promoting recruitment of peptides derived from corepressor proteins NCoR and SMART (peptides 2 and 3, respectively). Furthermore, compared with the Me2SO control, T0070907 seems to suppress the basal interactions between PPARγ and coactivator-derived peptides.
Table IList of amino acid sequences and sources of peptides used in the HTRF peptide profiling assay
The L XXLL motif is in bold. Peptides are tagged on the N termini with biotin.
Figure 4T0070907 is an inverse agonist of the GST-PPARγ LBD in vitro.A, HTRF peptide profiling of T0070907, rosiglitazone, troglitazone, and GI262570. All compounds were tested at 1 μm, and peptides (x axis) were in numerical order as listed in Table I. Corepressor-derived peptides 2 and 3 are shown in black. DMSO, Me2SO. B, dose-response of rosiglitazone and T0070907 in the presence and absence of 1 μm rosiglitazone in an HTRF assay with GST-PPARγ LBD and peptide 1 (DRIP205). C, same as in B, except that the HTRF peptide is 2 (NCoR). D, dose-responses of rosiglitazone and T0070907 in a GST pull-down assay that measures the interaction between GST-PPARγ and NCoR protein. An arrow indicates the position of the NCoR protein.
In order to confirm these results, more extensive titration and competition experiments were carried out with two peptides derived from a representative coactivator and a representative corepressor (peptides 1 and 2). As shown in Fig. 4, B and C, rosiglitazone promoted the dose-dependent recruitment of peptide derived from coactivator DRIP205 to PPARγ, while suppressing the interaction between PPARγ and a peptide derived from corepressor NCoR. In contrast, T0070907 suppressed the interaction between PPARγ and the coactivator-derived peptide in the absence of ligand, while promoting the recruitment of the NCoR-derived peptide to PPARγ. T0070907 also effectively antagonized the effects of rosiglitazone in a dose-dependent manner.
To confirm independently these observations by using an alternative nonfluorescence-based format, the effects of T0070907 on PPARγ/NCoR interactions were examined using a GST pull-down assay. As shown in Fig. 4D, rosiglitazone suppresses the interaction between the GST-PPARγLBD and NCoR in a dose-dependent fashion, with an IC50 consistent with its binding affinity to PPARγ. On the other hand, T0070907 promoted a dramatic increase in NCoR binding to GST-PPARγ consistent with the results observed in the HTRF assay.
Effects of T0070907 and LGD1069 on PPARγ and RXRα Heterodimer in GMSAs
By having shown that T0070907 strongly promotes recruitment of NCoR to the PPARγ LBD in both HTRF and pull-down assays, we next used a GMSA to examine whether this could also occur in the context of the PPARγ/RXRα heterodimer. As shown in Fig. 5A, in vitro translated PPARγ and RXRα can bind simultaneously to a PPRE-containing DNA fragment derived from the promoter of acyl-CoA oxidase gene (lane 2). This shift in fragment mobility is absolutely dependent on the presence of both PPARγ and RXRα (data not shown), indicating the proper formation of a functional PPARγ/RXRα heterodimer under these conditions. Whereas NCoR could not bind efficiently to the PPARγ/RXRα heterodimer in the absence of ligand (compare lanes 2 and 3, and similar to Ref.
), T0070907 was able to promote a significant increase in the recruitment of NCoR to the heterodimeric complex (compare lanes 3 and 4).
Figure 5Effects of T0070907 on PPARγ/RXRα heterodimer in a GMSA. A, effects of T0070907 on the recruitment of NCoR to the wt PPARγ/wt RXRα and PPARγ ΔH12/wt RXRα heterodimers. B, effects of T0070907 on the recruitment of NCoR to the wt PPARγ/RXRα ΔH12 and PPARγ ΔH12/RXRα ΔH12 heterodimer complexes. C, effects of LGD1069 on T0070907-induced recruitment of NCoR to the wt PPARγ/RXRα ΔH12, PPARγ ΔH12/RXRα ΔH12 heterodimer complexes. D, effects of T0070907 on LGD1069-induced recruitment of SRC-1 to the PPARγ/RXRα heterodimer complex. Arrows indicate the positions of free probe and various complexes.
To ensure that this increased NCoR recruitment was the result of T0070907 binding to PPARγ and to understand the nature of the PPARγ conformational changes associated with T0070907 binding, we next investigated the effects of deleting H12 from both receptors on the binding of NCoR to the heterodimer. The deletion of PPARγ H12 (PPARγ ΔH12) increased the basal interaction of NCoR with the heterodimer; however, T0070907 did not provide further enhancement of binding (Fig. 5A, lanes 5–7). In contrast, the PPARγ wt/RXRαΔH12 heterodimer responded to T0070907 and almost all PPARγ wt/RXRα heterodimer could be super-shifted to form the PPARγ/RXRα/NCoR complex in the presence of the antagonist (Fig. 5B, lanes 2–4). Complexes containing H12 deletions in both PPARγ and RXRα interacted very efficiently with NCoR in the absence of ligand, but as in the case of the PPARγ wt/RXR ΔH12 complex, T0070907 had almost no effect on NCoR recruitment (Fig. 5B, lanes 5–7).
The allosteric effects between PPARγ and RXRα were studied next by examining the effects of simultaneous treatments of RXRα agonist and PPARγ antagonist on the recruitment of coactivator and corepressor proteins to the heterodimer. Because the RXRα ΔH12-containing heterodimers interacted much more strongly with NCoR than the wild type-containing heterodimers (Fig. 5B), we examined the effects of an RXRα agonist, LGD1069 (
), on T0070907-induced NCoR recruitment to PPARγ wt/RXRα ΔH12 and PPARγ ΔH12/RXRα ΔH12 complexes. Strikingly, the addition of LGD1069 dramatically inhibited NCoR binding to both pairs of heterodimer complexes (Fig. 5C, lanes 5–10 and lanes 14–19). Importantly, LGD1069 was not able to inhibit completely corepressor binding to the PPARγ ΔH12/RXRα ΔH12 complex as it was in the PPARγ wt/RXRΔH12 complex. To ensure that the effect of LGD1069 on NCoR recruitment was not due to blocking T0070907 from binding to the PPARγ/RXRα heterodimer, experiments were performed during which PPARγ was treated with T0070907 for an extended period prior to the addition of other components of the GMSA reaction mixture and LGD1069 to saturate all available binding sites on PPARγ. No difference on the recruitment of NCoR to the PPARγ/RXRα heterodimer was observed with or without preincubation with T0070907 (data not shown).
The effects of T0070907 binding to PPARγ on the ability of RXRα to interact with coactivators were also studied. In the presence of the RXRα agonist LGD1069, the coactivator protein SRC-1 can be recruited to the PPARγ/RXRα heterodimer, as evidenced by the super-shifted complex observed in GMSAs (Fig. 4D, lanes 5–10). This super-shifted complex was not observed in the presence of rosiglitazone (BRL, lane 4) or in reactions containing an RXR H12 mutant (
) suggesting that SRC-1 was specifically recruited to the RXRα subunit. When added to the GMSA reaction mixture together with LGD1069, T0070907 seems to have a modest effect on formation of the PPARγ/RXRα/SRC-1 super-shifted complex induced by LGD1069 (compare lanes 5–10 and lanes 12–17).
DISCUSSION
We have identified a specific, high affinity PPARγ antagonist, T0070907, that blocks PPARγ activity in both biochemical and cell-based assays. T0070907 is highly selective for PPARγ over PPARα, PPARδ, other NHRs, and proteins in an HEK293 WCE. Proteolytic mapping indicated that T0070907 irreversibly modifies PPARγ on Cys313 in helix 3 of the LBD, a residue that is conserved in all three PPAR subtypes. This indicates that other residues in the binding pocket confer the specific binding of T0070907 to PPARγ. Interestingly, this is also the site of covalent modification by L-764406, a PPARγ partial agonist that was described previously (
T0070907 functions as a PPARγ antagonist in cell-based assays. It effectively blocked TZD-induced transactivation by the GAL4-PPARγ LBD, as well as adipogenesis in 3T3-L1 cells treated with a differentiation mixture. Overall, these results further support the important role for PPARγ in fat cell differentiation. The antagonist properties of T0070907 were also demonstrated in a variety of in vitro biochemical assays using the PPARγ LBD. T0070907 suppressed agonist-induced interactions between the PPARγ LBD and coactivator-derived peptides and promoted recruitment of corepressor-derived peptide in HTRF assays (Fig. 4). The effect of T0070907 on assembly of corepressor NCoR/PPARγ complex was also observed in the pull-down assay (Fig. 4D).
) have suggested that corepressors bind to a hydrophobic groove on NHR LBDs formed by H3, H5, and H6. This binding site partially overlaps with that utilized by coactivators. NHR agonists disrupt the interaction between NHR LBDs and corepressors, and it is believed that the conformation of H12 stabilized by agonists partially occludes the corepressor binding site. In the unliganded state, H12 of NHR LBDs is thought to exist in multiple conformations, including the agonist-bound conformation. Consistent with this hypothesis, H12 is inhibitory for NCoR binding to most NHRs. Mutations and deletions of H12 from either PPARγ or RXRα significantly increase the recruitment of NCoR to the PPARγ/RXRα heterodimer (
) (Fig. 5). T0070907 can also promote NCoR recruitment to PPARγ/RXRα heterodimer, but T0070907 can only promote recruitment of NCoR to complexes containing wt PPARγ but not to complexes containing PPARγ ΔH12. These results suggest that the effect of T0070907 on the heterodimer is indeed mediated through PPARγ and that T0070907 induced NCoR recruitment requires H12. Indeed, T0070907 treatment or the deletion of H12 stabilize NCoR recruitment to comparable extents (Fig. 5) suggesting that T0070907 most likely acts on PPARγ by preventing H12 from adopting the agonist-bound conformation. The deletion of RXRα H12 domain has a synergistic effect with either T0070907 treatment or PPARγ ΔH12 on the recruitment of NCoR to PPARγ/RXRα heterodimer (Fig. 5B). Two NHR-binding motifs are present on NCoR protein (
), the deletion of H12 from RXRα together with either T0070907 treatment or the deletion of H12 from PPARγ perhaps allows the cooperative binding of both motifs to PPARγ/RXRα heterodimer.
In order to dissect the contributions of the PPARγ and RXRα to NCoR recruitment to the heterodimer, the effects of simultaneous treatment with T0070907 and LGD1069 were determined. Notably, LGD1069 dramatically inhibited the T0070907-mediated increase in NCoR recruitment to the PPARγ wt/RXRα ΔH12 and PPARγ ΔH12/RXRα ΔH12 heterodimers in a dose-dependent manner (Fig. 5C). Recent x-ray crystallographic studies of the apo-RXRα LBD (unliganded), the holo-RXRα LBD (agonist bound), and a PPARγ/RXRα heterodimer (each bound to agonist) suggest possible molecular mechanisms for these effects. The rosiglitazone-bound PPARγ/9-cis-retinoic acid-bound RXR heterodimer interface which is largely composed of residues from H10 and H11 of both receptors contains several important salt bridges. In particular, a salt bridge formed between the carboxylic acid of Tyr477 from PPARγ H12 and Lys431 from RXRα H10 stabilizes the positioning of H12 from PPARγ in the agonist-bound conformation (
). In addition, comparison of the apo- and holo-RXRα LBD structures reveal that ligand-binding triggers several large conformational changes. For example, H11, which partially fills the ligand binding pocket in the apo-RXRα structure, moves out of the binding pocket and rotates by ∼180° around its own axis upon binding of 9-cis-retinoic acid, allowing H10 and H11 to form an almost continuous helix (
). Although the structure of a PPARγ/apo-RXRα heterodimer has not yet been described, these structural results suggest that LGD1069 binding could lead to significant alterations in the PPARγ/RXRα heterodimer interface. Given that Lys431 is located near the site of the conformational changes involving H10 and H11, RXR agonists could also influence the stability of the Tyr477/Lys431salt bridge and hence the positioning of PPARγ H12. Thus, we suggest that LGD1069 binding inhibits binding of NCoR to the wild type heterodimer (with or without T0070907) by orienting PPARγ and RXRα such that binding of NCoR is disfavored and by stabilizing PPARγ H12 in the agonist-bound conformation. Consistent with this view, the effect of LGD1069 on T0070907-induced recruitment of NCoR is more potent on the PPARγ wt/RXRα ΔH12 heterodimer than on PPARγ ΔH12/RXRα ΔH12 heterodimer. In addition, a residual amount of NCoR remained on PPARγ ΔH12/RXRα ΔH12 heterodimer even at the highest LGD1069 concentrations (Fig. 5C), and LGD1069 was also ineffective in preventing NCoR binding to PPARγ ΔH12/RXRα wt heterodimer complexes (data not shown).
The effect of T0070907 on LGD1069-induced recruitment of coactivator SRC-1 was also tested. Whereas the T0070907-induced NCoR recruitment to PPARγ/RXRα heterodimer can be almost completely reversed by the simultaneous treatment with RXRα agonist LGD1069, the effects of T0070907 on LGD1069-induced coactivator recruitment to the PPARγ/RXRα heterodimer are more modest by comparison. These results suggest that RXRα agonists may have a greater influence on the conformation of the PPARγ/RXRα heterodimer than do PPARγ antagonists, and more importantly, PPARγ antagonist activity could be modulated by the availability and concentration of RXRα agonist. The in vivo relevance of these effects is the focus of our current and future studies.
ACKNOWLEDGEMENTS
We thank Merrill Ayres, Mitch Hull, Tim Hoey, Jonathan Houze, Jowell Jo, Xiao Hong Liu, Miki Rich, Heather Webb, and Haoda Xu for technical support and generous gifts of reagents. We also thank Andrew Shiau, Jurgen Lehmann, Hui Tian, and Zhulun Wang for helpful discussions, and Kelly LaMarco for editing this manuscript.
85028-5&id=gr1.jpg){kind=link}
85028-5&id=gr2.jpg){kind=link}
85028-5&id=gr3.jpg){kind=link}
85028-5&id=gr4.jpg){kind=link}
85028-5&id=gr5.jpg){kind=link}