SRC-1 and TIF2 Control Energy Balance between White and Brown Adipose Tissues

Cell, Vol. 111, 931–941, December 27, 2002, Copyright 2002 by Cell Press

SRC-1 and TIF2 Control Energy Balance between White and Brown Adipose Tissues Fre´de´ric Picard,1 Martine Ge´hin,1,5 Jean-Se´bastien Annicotte,1,5 Ste´phane Rocchi,1,5 Marie-France Champy,2 Bert W. O’Malley,3 Pierre Chambon,1,2 and Johan Auwerx1,2,4 1 Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire CNRS/INSERM/ULP 67404 Illkirch France 2 Institut Clinique de la Souris Ge´nopole Strasbourg 67404 Illkirch France 3 Baylor College of Medicine One Baylor Plaza Houston, Texas 77030

Summary We have explored the effects of two members of the p160 coregulator family on energy homeostasis. TIF2⫺/⫺ mice are protected against obesity and display enhanced adaptive thermogenesis, whereas SRC-1⫺/⫺ mice are prone to obesity due to reduced energy expenditure. In white adipose tissue, lack of TIF2 decreases PPAR␥ activity and reduces fat accumulation, whereas in brown adipose tissue it facilitates the interaction between SRC-1 and PGC-1␣, which induces PGC-1␣’s thermogenic activity. Interestingly, a high-fat diet increases the TIF2/SRC-1 expression ratio, which may contribute to weight gain. These results reveal that the relative level of TIF2/SRC-1 can modulate energy metabolism. Introduction Chromatin structure creates effective barriers interfering with transcriptional activation. Coregulators assist transcription factors and the basal transcription machinery to overcome this barrier (Glass and Rosenfeld, 2000). Current knowledge on coregulators is to a large extent based on their interaction with nuclear receptors (NR). Members of the p160 family, (steroid receptor coactivator-1 [SRC-1, NcoA-1], transcriptional intermediary factor-2 [TIF2, GRIP-1, SRC-2, NcoA-2] and SRC-3 [p/CIP, ACTR, RAC-3, AIB-1, TRAM-1]), were among the first coregulators to be cloned, based on their ligand-dependent recruitment to NR ligand binding domain (LBD) through three ␣-helical LXXLL motifs located in their N-terminal region (Aranda and Pascual, 2001; Glass and Rosenfeld, 2000). Coregulators of the p160 family also contain conserved leucine-rich motifs in their C-terminal domain that mediate interactions with additional coregulators (Aranda and Pascual, 2001). Although a wealth of evidence suggests that p160 4 5

Correspondence: [email protected] These authors contributed equally to this work.

coregulators modulate NR function in vitro, their role in vivo has not yet been fully explored. Mice deficient for individual members of the p160 family are characterized by partial hormone resistance (Ge´hin et al., 2002; Qi et al., 1999; Wang et al., 2000; Xu et al., 2000, 1998). Decreased growth of organs affected by steroids (testis, uterus, and mammary gland), and reduced reproductive function were observed either in TIF2⫺/⫺ (Ge´hin et al., 2002) and/or SRC-1⫺/⫺ (Xu et al., 1998) mice, whereas defective growth hormone signaling was reported in SRC-3⫺/⫺ mice (Wang et al., 2000; Xu et al., 2000). These findings suggest that p160 coregulators possess some functional specificity. Such a specificity was also demonstrated for PPAR␥, where ligand-specific coregulator recruitment resulted in major differences in PPAR␥ transactivation (Kodera et al., 2000; McInerney et al., 1998; Shao et al., 2000) and in alterations in glucose homeostasis, adipogenesis, and body weight regulation (Reginato et al., 1998; Rocchi et al., 2001). Thus, different p160 coregulators appear to have specific functions in energy metabolism. To validate this possibility, we evaluated the metabolic consequences of TIF2 and SRC-1 gene inactivation in mice. Protection against Obesity and Increased Insulin Sensitivity in TIF2⫺/⫺ Mice After weaning, TIF2⫹/⫹ and ⫺/⫺ mice gained weight similarly when fed a regular chow diet (Figure 1A). TIF2⫺/⫺ animals were, however, protected against obesity induced either by high-fat feeding or by neonatal injection of monosodium glutamate (MSG) (Figure 1A), a drug inducing hypothalamic lesions that lead to hyperphagia (Olney, 1969). The differences in body weight occurred despite higher food intake in TIF2⫺/⫺ mice in the two obesity models (Figure 1B). These studies suggested that TIF2 modified energy balance, most likely through enhanced energy expenditure. TIF2⫺/⫺ animals had lower fasting glycemia than their wild-type littermates (Figure 1C). One hour postprandial, insulin levels of TIF2⫺/⫺ mice were much lower than those of wild-type animals despite similar glucose levels (Figure 1C). The rate of glucose clearance upon insulin injection was higher in TIF2⫺/⫺ than in TIF2⫹/⫹ mice fed a chow diet (Figure 1D). Glucose uptake was also higher in primary isolated adipocytes from TIF2⫺/⫺ mice compared to that in TIF2⫹/⫹ adipocytes (Figure 1E). Furthermore, TIF2⫺/⫺ mice cleared glucose more effectively after intraperitoneal glucose injection than TIF2⫹/⫹ animals. This was observed both in mice fed a chow or a high-fat diet (Figure 1F). Thus, whole-body insulin sensitivity appears to be enhanced in the absence of TIF2. Another indication of an improved metabolic profile is provided by the lower fasting triglyceride levels in TIF2⫺/⫺ mice (0.72 ⫾ 0.17 versus 1.80 ⫾ 0.26 mM, P ⫽ 0.0135). As TIF2⫺/⫺ mice were protected against obesity, we tested the impact of this mutation on fat mass accretion. No difference was observed in total body fat content between chow-fed TIF2⫹/⫹ and ⫺/⫺ mice as measured

Cell 932

Figure 1. Body Weight Regulation and Glucose Homeostasis in TIF2⫺/⫺ Mice (A–B) TIF2⫺/⫺ mice have lower body weight despite higher food intake. Body weights were recorded weekly in TIF2⫹/⫹ and ⫺/⫺ animals either fed a regular chow diet, a high-fat diet, or in mice treated neonataly with MSG (2 mg/kg/day, for 7 days) (A). Food intake per mouse was measured every second day for 15 days (B). (C–F) TIF2⫺/⫺ mice are insulin sensitized. (C) Plasma glucose and insulin levels after a 6 hr fast or 1 hr following a standard meal. (D) Plasma glucose levels after an acute injection of insulin (0.75 UI/kg) in overnight fasted mice (E) Glucose uptake in primary isolated adipocytes from TIF2⫹/⫹ and ⫺/⫺ mice, as evaluated in cells 30 min after exposure to radiolabeled 14 C-deoxyglucose. (F) Plasma glucose levels after a glucose load (2 g/kg) in chow fed or high fat fed, overnight fasted mice. (G) Lack of TIF2 protects against fat mass accretion and obesity. Body fat mass was evaluated by DEXA scanning and epididymal WAT weight in TIF2⫹/⫹ and ⫺/⫺ mice fed a chow diet with or without neonatal exposure to MSG. (H–I) TIF2⫺/⫺ adipocytes are smaller. (H) Adipose tissue morphology (Top images). Adipose tissue sections were either stained with hematoxylin and eosin or subjected to scanning electron microscopy. Magnification for histology, 20⫻, and for scanning microscopy, 500⫻. (I) Cell size and cell count in a representative population of white adipocytes from the epididymal depot of TIF2⫹/⫹ and ⫺/⫺ mice. Data were analyzed on 3 slides for each animal, 5 animals per genotype. Indicates significant differences between ⫹/⫹ and ⫺/⫺ genotypes. In each experiment, n ⱖ 5. Black symbols represent TIF2⫺/⫺ mice, whereas white symbols represent their wild-type littermates.

by dual energy X-ray absorptiometry (Figure 1G). In contrast, after neonatal exposure to MSG, TIF2⫺/⫺ mice had a significantly lower total body fat content than wild-type animals (Figure 1G). These differences were mirrored by a lower weight of the epididymal white adipose tissue (WAT) in TIF2⫺/⫺ mice upon MSG treatment (Figure 1G). Compared to those of TIF2⫹/⫹ mice, the size of the white adipocytes from TIF2⫺/⫺ animals was reduced by an average of 70% even under a regular chow diet (1565 ⫾ 48 versus 5081 ⫾ 263 ␮M2), as assessed by both classical histology and scanning electron microscopy (Figures 1H–1I). Since TIF2⫺/⫺ mice have a similar fat pad weight yet smaller adipocytes than TIF2⫹/⫹ animals, the cell number was increased in the fat depot (Figure 1I), suggesting that increased cell number compensates for decreased adipocyte size. Both Lipolysis and Fat Storage Are Affected by the Lack of TIF2 We evaluated glycerol release to measure lipolysis in primary cultures of adipocytes isolated from TIF2⫺/⫺

and wild-type mice. Lipolysis was higher in TIF2⫺/⫺ cells both under basal conditions and following exposure to adrenaline (Figure 2A). Perilipins are anti-lipolytic lipid droplet binding proteins whose absence in mice has been associated with resistance to obesity (Martinez-Botas et al., 2000; Tansey et al., 2001). Consistent with increased lipolysis, white adipocytes from TIF2⫺/⫺ mice had lower levels of perilipin A and B protein (Figure 2B) and mRNA (Figure 2C) compared to those of TIF2⫹/⫹ animals. The expression of genes involved in fatty acid uptake and trapping, such as fatty acid binding protein (aP2), lipoprotein lipase (LPL), and PPAR␥ were also lower in TIF2⫺/⫺ mice (Figure 2C). In contrast, leptin mRNA levels in adipose tissue (Figure 2C) and plasma leptin concentrations (Figure 2D) were higher in TIF2⫺/⫺ mice. These changes likely contributed, at least in part, to the improvement in insulin sensitivity and the resistance to obesity in mice lacking TIF2. On the other hand, plasma adiponectin levels remained unchanged in TIF2⫺/⫺ mice (13.3 ⫾ 2.9 versus 15.4 ⫾ 2.9 ␮g/mL, P ⫽ 0.46), indicating that the absence of TIF2

SRC-1 and TIF2 Control Energy Metabolism 933

Figure 2. Effects of Lack of TIF2 in White Adipocytes (A–B) Enhanced lipolysis in TIF2⫺/⫺ adipocytes. (A) Isolated adipocytes from TIF2⫹/⫹ and ⫺/⫺ mice were treated with 0, 10⫺7 M, and 10⫺5 M adrenaline and the release of glycerol was analyzed 30 min later. See Figure 1 for statistical significance. Protein levels of perilipin in WAT of TIF2⫹/⫹ and ⫺/⫺ mice either visualized by immunofluorescence (B, upper image) or by Western blotting (B, lower image). Two perilipin isoforms were detected, (A and B), and loading was controlled with ␤-actin levels. (C–D) Lack of TIF2 decreases the expression of genes involved in fatty acid trapping but upregulates leptin. (C) mRNA levels of the indicated genes in WAT were evaluated by Northern blotting with 32P-labeled probes. (D) Plasma leptin levels in non-fasted TIF2⫹/⫹ and ⫺/⫺ mice. (E–G) Importance of TIF2 in fat storage. (E) TIF2 and SRC-1 mRNA and protein levels in 3T3-L1 cells induced to differentiate into adipocytes. (F) Oil Red O staining in TIF2⫹/⫹ and ⫺/⫺ MEFs stimulated to store lipids by exposure to differentiation mix with or without rosiglitazone (10–7 M). (G) Oil Red O staining in 3T3-L1 and NIH-3T3 cells overexpressing TIF2 or SRC-1 after retroviral infection and induced to store lipids as in (F). The number in each image corresponds to the intensity of red staining. (H–J) PPAR␥ interaction with p300 is influenced by p160 coregulators. (H) GST pull-down assays were performed with His-tagPPAR␥2(DE203-477) or in vitro 35S translated pCMX-ER␣ and increasing amounts (0.5, 1.5, and 3 ␮g) of GST-p300, GST-SRC-1 or GST-TIF2 proteins prepared in E. coli. 293T cells were transfected with 500 ng of pCMV-p300, pcDNA-TIF2, pcDNA-SRC-1, and either BDGal4-hPPAR␥ DE and pGL3-(Gal4)5TKLuc reporter construct (I), or pSG5-hPPAR␥2 encoding fulllength receptor and LPL promoter-AN-CAT reporter construct (J). Cells were then grown for 24 hr with rosiglitazone (10⫺7 M). Luciferase activity is expressed as units/␤-galactosidase activity. Each point was performed in triplicate. Bars not sharing similar superscripts are significantly different. nd: not determined.

alters the expression of some, but not all adipose tissuederived signaling molecules. In general, the altered gene expression pattern due to the absence of TIF2 appears to reduce the potential for fatty acid storage. Both mRNA and protein levels of TIF2 and SRC-1 were induced during the differentiation of 3T3-L1 preadipocytes into adipocytes (Figure 2E). This prompted us to compare the capacity of mouse embryonic fibroblasts (MEFs) from TIF2⫹/⫹ and ⫺/⫺ embryos to accumulate lipids when stimulated with insulin, dexamethasone, and isobutylmethylxanthine. MEFs lacking TIF2 had a reduced fat storage potential, as evaluated by Oil Red O lipid staining and lower level of aP2 mRNA (Figure 2F). The differences in fat accumulation between TIF2 ⫹/⫹ and ⫺/⫺ MEFs were more pronounced in the presence of rosiglitazone, as expected from the well-established

ligand-dependent interaction of p160 coregulators with PPAR␥ (McInerney et al., 1998). We then evaluated the role of overexpression of TIF2 and SRC-1 on fat storage in different cell lines by retroviral infection. In the presence of insulin, dexamethasone, isobutylmethylxanthine, and rosiglitazone, 3T3-L1 cells infected with a retrovirus expressing TIF2 accumulated approximately 2-fold more lipids compared to cells infected either with control or SRC-1-expressing retrovirus, as evaluated by visualization of entire culture plates (data not shown) and microscopically representative fields (Figure 2G). In rosiglitazone-treated NIH-3T3 cells, which lack PPAR␥ (Tontonoz et al., 1994), neither overexpression of TIF2 nor SRC-1 induced fat storage (Figure 2G), indicating that TIF2 modulates mainly the PPAR␥-mediated fat uptake and storage pathway.

Cell 934

The affinity of SRC-1, TIF2, and p300 for rosiglitazonebound PPAR␥ and estradiol-bound estrogen receptor ␣ (ER␣) was compared. GST pull-down assays using increasing amounts of GST-coregulator fusion proteins and a fixed amount of the LBD of these NRs were performed. Recruitment of TIF2 by rosiglitazone-bound PPAR␥ increased when higher amounts of TIF2 were added (Figure 2H). SRC-1 was much less effectively recruited by PPAR␥ than TIF2 (Figure 2H), suggesting that TIF2 is the preferred p160 coregulator of PPAR␥. Estradiol-bound ER␣ did not show a preferential recruitment of the coregulators tested (Figure 2H). To further support these data, we assessed the activity of these coregulators on PPAR␥ by cotransfecting a Gal4-PPAR␥ LBD chimera, an UAS-driven luciferase reporter, and expression vectors for TIF2, SRC-1, and p300. Based on a dose-response study of p300’s ability to enhance PPAR␥ activity, we selected 500 ng of p300 in all further transfection assays (Supplemental Figure S1A available at http://www.cell.com/cgi/content/full/111/7/931/ DC1). With this fixed amount of p300, TIF2 was clearly more effective than SRC-1 at enhancing the activity of Gal4-PPAR␥ (Supplemental Figure S1B available at above website). PPAR␥ activity was reduced by the absence of TIF2 but not by the lack of SRC-1 when a fixed amount (500 ng) of each coregulator was used (Figure 2I and Supplemental Figure S1B available at above website). Similar findings were observed in transfection assays using full-length PPAR␥ and the human LPL promoter instead of the UAS-Gal4 reporter (Figure 2J). These results demonstrated that in the absence of TIF2, but not SRC-1, the p300/PPAR␥ complex was less active. This was probably caused by a weaker capacity of SRC-1, relative to TIF2, to interact with other coregulators such as p300/CBP and TRAP220, which have been shown to be involved in adipogenesis (Ge et al., 2002; Takahashi et al., 2002; Yamauchi et al., 2002). Together, these data illustrate that the absence of TIF2 enhances lipolysis and impairs fat uptake and storage most likely through the reduction of PPAR␥ activity, and hence protects against fat mass accretion under conditions favoring obesity. Since classical full PPAR␥ agonists, such as the thiazolidinediones, are being used for the treatment of type 2 diabetes—a condition often associated with obesity—the specific recruitment of TIF2 over SRC-1 after activation of PPAR␥ by these ligands might explain the fat accretion occurring upon treatment. Adaptive Thermogenesis Is Enhanced in TIF2⫺/⫺ Brown Adipose Tissue Adaptive thermogenesis is an important component of energy balance and a metabolic defense against obesity, through stimulation of shivering and non-shivering thermogenesis in skeletal muscle and brown adipose tissue (BAT), respectively. BAT is the main thermogenic tissue in rodents, where fatty acid oxidation stimulated by the sympathetic nervous system generates heat through the induction of uncoupling protein-1 (UCP-1) (Matthias et al., 2000; Ricquier et al., 1979). Interestingly, mice with a genetic ablation of BAT cannot adapt to cold or high-fat diets and are prone to obesity (Lowell et al., 1993). Thermogenesis is coordinated by the coregulator PPAR␥-coactivator-1 ␣ (PGC-1␣), which trans-

mits the effects of NRs, such as the thyroid receptor, PPAR␥, and PPAR␣, on the transcriptional process of adaptive thermogenesis and fatty acid oxidation (Puigserver et al., 1998; Vega et al., 2000). Since PGC-1␣, however, requires additional cofactors to activate transcription (Puigserver et al., 1999), we tested the hypothesis that the ablation of TIF2 would affect PGC-1-mediated adaptive thermogenesis. Macroscopically, interscapular BAT of TIF2⫺/⫺ mice looked darker and had less WAT surrounding it than that of TIF2⫹/⫹ animals (Figure 3A). Intracellular lipid accumulation was decreased in TIF2⫺/⫺ brown adipocytes, accounting at least in part for the darker aspect of BAT (Figure 3A). TIF2⫺/⫺ brown adipocytes seemed collapsed and less rounded. In addition, enlarged mitochondria with more cristae (Figure 3A bottom, arrows) and much smaller lipid droplets (Figure 3A bottom, asterisks) were observed in TIF2⫺/⫺ BAT. Oxygen consumption was initially similar in TIF2⫹/⫹ and ⫺/⫺ animals but became significantly higher in TIF2⫺/⫺ mice after ⵑ6 hr, a time point concomitant with the beginning of the impact of fasting (Figure 3B, top). MSG-treated TIF2⫺/⫺ mice had significantly higher oxygen consumption throughout the test (Figure 3B, bottom), which was consistent with their lower body weight despite higher caloric intake. Since oxygen consumption is determined by mitochondrial activity and fatty acid oxidation, it was logical that steady state mRNA levels of UCP-1, PGC-1␣, and acetylCoA oxidase (AOX) were all higher in BAT of TIF2⫺/⫺ mice (Figure 3C). SRC-1 mRNA levels were similar between the two genotypes. These findings indicated that the absence of TIF2 results in higher energy expenditure through enhanced fatty acid oxidation and uncoupling of respiration. To further support these data, mice were placed under conditions in which adaptive thermogenesis was challenged. Whereas no difference in rectal temperature was observed at 23⬚C (control), TIF2⫺/⫺ mice had a higher temperature than TIF2⫹/⫹ animals when exposed to 4⬚C or fasted overnight (Figure 3D). Feeding-induced thermogenesis was also more pronounced in TIF2⫺/⫺ mice (Figure 3D). In line with the higher lipolytic activity in white adipocytes lacking TIF2, plasma free fatty acid levels were more increased in TIF2⫺/⫺ mice upon cold exposure and fasting (Figure 3E). In cold-challenged BAT, oxidation of free fatty acids provides substrates for heat production. As an index of this oxidation, we measured circulating ketone bodies in mice challenges with cold exposure or fasting. Whereas the liver is the major source of ketone body production during fasting, other tissues, such as BAT, can contribute to ketogenesis (Guezennec et al., 1988). Upon cold exposure, circulating ketone bodies were higher in TIF2⫺/⫺ mice (Figure 3E), suggesting an enhanced fatty acid oxidation in this group. To evaluate the contribution of fatty acids released from WAT as energy substrate for heat production in BAT, mice were injected with nicotinic acid, an antilipolytic agent (Larsen and Illingworth, 1993), 1 hr before being placed at 4⬚C. Under these conditions, the elevation in plasma free fatty acid levels was blunted (Figure 3G), and differences in rectal temperature and circulating ketone bodies were no longer observed between the genotypes (Figures 3F–3G). Although nicotinic acid

SRC-1 and TIF2 Control Energy Metabolism 935

Figure 3. Higher Adaptive Thermogenesis in TIF2⫺/⫺ Mice (A–C) TIF2 alters brown adipose tissue. (A) Morphology of BAT in TIF2⫺/⫺ mice. Adipose tissue was analyzed by hematoxylin and eosin staining and scanning and transmission electron microscopy. Magnification 40⫻, 4000, and 100000⫻, respectively. Asterisks indicate lipid droplets and arrows show mitochondria. (B) Oxygen consumption (VO2) analyzed by indirect calorimetry in TIF2⫹/⫹ and ⫺/⫺ mice raised under normal conditions (upper image) or after neonatal exposure to MSG (lower image). (C) mRNA levels in BAT of TIF2⫹/⫹ and ⫺/⫺ mice as evaluated by Northern blotting with the indicated 32P-labeled probes. (D–G) TIF2 knockout mice display higher adaptive thermogenesis. Rectal temperature (D) and plasma levels of free fatty acids and ketone bodies (E) were evaluated in chow-fed TIF2⫹/⫹ and ⫺/⫺ mice either in a normal state, after 6 hr of cold exposure (4⬚C), after an overnight fast, or 1 hr following subsequent refeeding. Rectal temperature (F) and plasma levels of free fatty acids and ketone bodies (G) in TIF2⫹/⫹ and ⫺/⫺ mice injected with either saline (Sal) or nicotinic acid (Nic; 150 mg/kg) and exposed to cold 1 hr later. See Figure 1 for statistical significance. † indicates significant differences between saline and nicotinic acid-treated animals within the same genotype.

might have additional pharmacological effects than blocking fatty acid release, these data suggest that the increased lipolytic capacity in WAT of TIF2⫺/⫺ mice likely contributes, at least in part, to the enhanced thermogenesis. It is furthermore likely that the increase in leptin (Figures 2C–2D) also contributed to this phenotype in TIF2⫺/⫺ mice, as it is well established that leptin stimulates thermogenesis (Trayhurn and Rayner, 1996). The relative contribution of leptin over that of other adipose-derived molecules, such as fatty acids, to the enhanced thermogenesis in TIF2⫺/⫺ mice is, however, difficult to determine at present. Adaptive thermogenesis can also occur in skeletal muscle, where induction of shivering and respiration substantially contribute to whole-body thermogenic control (Puigserver et al., 1998). Expression levels of both PGC-1␣ and AOX mRNA were similar in skeletal muscle of TIF2⫹/⫹ and ⫺/⫺ mice (Supplemental Figure S2A available at http://www.cell.com/cgi/content/full/ 111/7/931/DC1). Likewise, we observed no histological (Supplemental Figure S2B available at above website) or quantitative (Supplemental Figure S2C available at above website) differences in triglyceride accumulation in skeletal muscle. These results suggest that the improved insulin sensitivity in TIF2⫺/⫺ mice is not due to changes in muscle triglyceride content. These findings furthermore exclude a major contribution of the central nervous system in the upregulation of adaptive thermo-

genesis in TIF2⫺/⫺ mice since both BAT and skeletal muscle are under similar neuronal control (Richard et al., 1992). In addition, several hormones such as glucocorticoids, thyroid, and sex hormones can modulate adaptive thermogenesis. Consistent with a previous report (Weiss et al., 2002), we detected no difference in the concentrations of T3, T4, and TSH between TIF2⫹/⫹ and ⫺/⫺ animals (data not shown). Furthermore, circulating corticosterone levels were also similar between the two genotypes (86 ⫾ 22 versus 99 ⫾ 33 ng/mL in TIF2 ⫹/⫹ and ⫺/⫺ mice, respectively, n ⫽ 9). Finally, it is unlikely that the partial sex hormone resistance in TIF2⫺/⫺ mice (Ge´hin et al., 2002) mediates the increased thermogenesis, as this condition is usually associated with decreased energy expenditure (Richard et al., 2002). In general, these data suggest that the lack of TIF2 mainly affected thermogenesis through modulation of BAT activity, and that the increased adaptive thermogenesis is not the consequence of a general neuroendocrine signaling pathway aimed at increasing body temperature. TIF2 and SRC-1 Modulate PGC-1␣ Transactivation Properties PGC-1␣ binds to the DNA binding domain (DBD) of PPAR␥ and other transcription factors to induce the expression of key enzymes of the mitochondrial respira-

Cell 936

Figure 4. Control of PGC-1␣/PPAR␥ Interactions by SRC-1 and TIF2 (A) Sequence alignment of the C-terminal region of p160 coregulators responsible for the interaction between PGC-1␣. Sequences were found for human (h), mouse (m), rat (r), Xenopus (x), zebrafish (dr), and Coturnix japonica (cj). Numbers in brackets correspond to the amino acid fragment studied. Boxed areas represent well-conserved leucine-rich motifs amongst the p160 family members. The blue shaded area corresponds to the first leucine-rich motif, which is not conserved in TIF2. (B–D) TIF2 decreases the SRC-1-mediated interaction between PPAR␥ and PGC-1␣. (B) Equal amounts of full-length PPAR␥ and pSVSPORT-PGC-1␣ were cotransfected in COS cells with different amounts of SRC-1 and TIF2. Cells were treated for 6 hr with rosiglitazone (10⫺7 M) before being harvested for coimmunoprecipitation assays. (C) 293T cells were transfected with expression vectors for BDGal4-hPPAR␥ DE, pGL3-(Gal4)5TKLuc reporter construct, and increasing amounts of expression vector for PGC-1␣ (top) or 500 ng of PGC-1␣ and increasing amounts of SRC-1, TIF2, or both. Cells were then grown for 24 hr with rosiglitazone (10⫺7 M). (D) Transfection assay as described above but with a fixed amount (500 ng) of coregulator DNA transfected. Each point was performed in triplicate. (E) Schematic representation of the interaction between PPAR␥ and PGC-1␣. PGC-1␣ has been shown to interact with PPAR␥ DBD (Puigserver et al., 1998), whereas p160 coregulators bind to PPAR␥ LBD (McInerney et al., 1998). We hypothesize that the first C-terminal leucine-rich motif out of four present in SRC-1 and SRC-3, but not in TIF2, reinforces the interaction with PGC-1␣.

tory chain, stimulating adaptive thermogenesis (Puigserver et al., 1998). PGC-1␣ has little transcriptional activity on its own, but becomes activated through interaction with SRC-1 (Puigserver et al., 1999). The SRC-1 domain responsible for binding PGC-1␣ has been mapped between aa 782 to 1139 in its carboxy-terminal region (Puigserver et al., 1999). The C-terminal domains of the p160 coregulators domain were therefore aligned in different species. Four highly conserved leucine-rich motifs were found in SRC-1 and SRC-3 (Demarest et al., 2002) (Figure 4A). In contrast, only the three C-terminal, but not the first, leucine-rich motifs were well conserved in the five TIF2 sequences analyzed. The absence of the first leucine-rich motif could therefore potentially account for functional differences between the p160 coregulators. We compared the ability of SRC-1 and TIF2 to stabilize PGC-1␣ binding to PPAR␥. Coimmunoprecipitations of coregulator complexes from rosiglitazone-treated cells transfected with PPAR␥ and various coregulators were performed. As expected, PPAR␥ and PGC-1␣ interacted

in the presence of SRC-1 (Figure 4B). TIF2 dose dependently attenuated this interaction, suggesting that it competes with SRC-1 for the formation of PGC-1␣/ PPAR␥ complexes. To functionally support these data, we performed transient transfection assays using a UAS-driven luciferase reporter and a Gal4-PPAR␥ LBD fusion protein. PGC-1␣ slightly increased PPAR␥ activity on its own (Figure 4C, top). Based on this, we selected a 500 ng dose of PGC-1␣ in all further assays. Interestingly, SRC-1, but not TIF2, strongly enhanced PGC1␣-mediated PPAR␥ transactivation over a wide range of DNA concentrations (Figure 4C). Moreover, when compared to the condition in which the three coregulators were present at a fixed amount of 500 ng, absence of SRC-1 decreased, whereas lack of TIF2 increased PPAR␥ activity (Figure 4D). Together, these results further support that TIF2 and SRC-1 compete for PGC-1␣ and that SRC-1 is required for PGC-1␣ activation (Figure 4E) (Puigserver et al., 1999). These molecular mechanisms, along with the increased PGC-1␣ expression in BAT of TIF2⫺/⫺ mice, could constitute possible molecu-

SRC-1 and TIF2 Control Energy Metabolism 937

Figure 5. Lack of SRC-1 Impairs Adaptive Thermogenesis (A) Lipid infiltration in BAT of SRC-1⫺/⫺ Mice. Brown adipose tissue was analyzed by hematoxylin and eosin staining in mice fed a highfat diet for 8 weeks. Magnification 20⫻. Arrows indicate white adipocyte infiltration. (B–C) SRC-1⫺/⫺ mice have a reduced adaptive thermogenesis. (B) Rectal temperature evaluated in high-fat fed SRC-1⫹/⫹ and ⫺/⫺ mice in a normal state (23⬚C), after cold (4⬚C) exposure for 6 hr or after an overnight fast. (C) Oxygen consumption (VO2) and respiratory quotient (RQ) analyzed by indirect calorimetry in non-fasted animals. (D–E) SRC-1⫺/⫺ mice are prone to obesity. (D) Body weight changes and evaluation of fat mass accretion by DEXA scanning and tissue weight in SRC-1⫹/⫹ and ⫺/⫺ mice fed a high-fat diet for 8 weeks. (E) mRNA levels of genes involved in BAT thermogenesis as evaluated by Northern blotting in SRC-1⫹/⫹ and ⫺/⫺ mice used in D. See Figure 1 for statistical significance.

lar explanations to the biological process of higher adaptive thermogenesis observed in these mice. SRC-1⫺/⫺ Mice Have Reduced Energy Expenditure and Are Prone to Obesity The above observations indicate that the formation of a more active SRC-1/PGC-1␣ complex is favored in the absence of TIF2, accounting, at least in part, for the higher thermogenesis in TIF2⫺/⫺ mice. Thus, the absence of SRC-1 should result in the formation of a less active TIF2/PGC-1␣ complex and in a reduction of energy expenditure and thermogenesis. SRC-1⫹/⫹ and ⫺/⫺ mice were therefore fed a high-fat diet for 8 weeks to stimulate energy expenditure. After this period, we observed more lipid infiltration in BAT of SRC-1⫺/⫺ mice (Figure 5A), a characteristic usually associated with decreased BAT thermogenic activity. Likewise, rectal temperature of SRC-1⫺/⫺ mice was significantly lower upon cold exposure (4⬚C for 6 hr) and overnight fasting than that of SRC-1⫹/⫹ animals (Figure 5B). These differences in SRC-1⫺/⫺ mice were associated with similar plasma fatty acid levels but lower ketone body concentrations (data not shown), suggesting that lipolysis in WAT was normal but that fatty acid oxidation in BAT was reduced in the absence of SRC-1. Moreover, SRC-1⫺/⫺ mice consumed significantly less oxygen and had higher RQ (CO2 formed/O2 used) than their wild-type littermates over a 6 hr period (Figure 5C), indicative of a lower energy expenditure and an attenuated fatty acid oxidation relative to that of glucose. The diminished fatty acid oxidation and adaptive thermogenesis in mice lacking SRC-1 led to a significantly higher body weight gain (⫹10%) (Figure 5D). This was mainly due to WAT accumulation (Figure 5D), as assessed by both DEXA scanning (⫹46% versus SRC-1⫹/⫹ mice) and the weight of the epididymal WAT (⫹41% versus SRC-1⫹/⫹ mice). Consistent with the increased fat mass, plasma leptin levels were higher in SRC-1⫺/⫺ than in ⫹/⫹ mice (5.4 ⫾ 1.7 versus 12.8 ⫾

2.2, P ⬍ 0.002). The mRNA levels of UCP-1, PGC-1␣, and AOX were all reduced in BAT of SRC-1⫺/⫺ mice (Figure 5E), indicating that the thermogenic machinery in BAT was diminished in the absence of SRC-1. It is likely that lower PGC-1␣ levels, along with the less stable PGC-1␣/TIF2 transactivation complex imposed by the absence of SRC-1, contributed to the decreased thermogenic potential of BAT in SRC-1⫺/⫺ mice. Overall, these results suggest that mice lacking SRC-1 are prone to obesity as a result of their reduced capacity for energy expenditure and fatty acid oxidation upon high-fat feeding. TIF2 and SRC-1 Are Modulated in Diet-Induced Obese Mice Chronic ingestion of high-fat diets leads to altered energy balance, resulting in obesity, and insulin resistance. We tested the effects of a high-fat diet for four months on the expression of TIF2 and SRC-1 in C57BL/6J mice. Cellular levels of TIF2 protein in both WAT and BAT were markedly increased upon high-fat feeding, an effect further underscored by determination of tissue DNA content (Figure 6). In comparison, the changes in SRC-1 levels were much more modest in both WAT and BAT. In contrast, in other tissues that contribute to energy homeostasis (hypothalamus, liver, skeletal muscle, and pancreas), we observed no significant alterations in expression levels of both the TIF2 and SRC-1 proteins upon high-fat feeding (Figure 6). These results indicate that a high-fat diet modulates the expression ratio of p160 coregulators specifically in WAT and BAT. Since increased lipid accretion in BAT, possibly stimulated by the induction in TIF2 expression, reduces BAT’s efficacy to compensate for the deleterious effects of high-fat diets on energy balance and because of the positive effect of TIF2 on white adipose tissue accumulation, these findings suggest that a change in TIF2/SRC-1 ratio could participate in the metabolic alterations that ultimately lead to obesity and insulin resistance.

Cell 938

Figure 6. Tissue-Specific Modulation of SRC-1 TIF2 in Diet-Induced Obesity TIF2 and SRC-1 protein levels and DNA concentrations were determined in homogenates from WAT, BAT, brain (hypothalamus), liver, gastrocnemius skeletal muscle, and pancreas of normal C57BL/6J mice fed either a regular chow diet or a high-fat diet for 4 months. ␤-actin levels are shown as loading control.

Conclusions Our findings demonstrate that coregulators of the p160 family are involved in the control of energy balance mainly by affecting both WAT and BAT homeostasis. Based on the present data, we propose the scheme summarized in Figure 7. Upon challenges such as cold exposure or fasting, lipolysis in WAT of wild-type animals occurs. The free fatty acids released into the circulation are taken up by thermogenic tissues such as BAT, where they serve as fuel for adaptive thermogenesis. Upon similar challenges, lipolysis in WAT is enhanced in the absence of TIF2, releasing more substrates for fatty acid oxidation in the hypertrophic mitochondria of BAT. This stimulated energy expenditure protects against excessive fat accumulation under conditions promoting obesity and favors insulin sensitization. In contrast, lack of SRC-1 reduces oxygen consumption and adaptive thermogenesis, which renders mice prone to obesity upon high-fat feeding. These effects reflect that TIF2 appears to be the preferred p160 partner of PPAR␥ in WAT, its absence leading to decreased PPAR␥ activity. The lack of TIF2 in BAT suppresses its competition with SRC-1 for PGC-1␣ and promotes the formation of stable and more active SRC-1/PGC-1␣ complexes (Figure 4E). In contrast, the absence of SRC-1 in BAT favors the formation of less active TIF2/PGC-1␣ complexes, which reduce energy expenditure. Thus, an altered composition of coregulator complexes, secondary to changes in the ratio of TIF2 and SRC-1, modifies the

transcriptional control of fat storage and thermogenesis. Our findings also support the conclusion that TIF2 and SRC-1 can exhibit specific and non-overlapping activities (Ge´hin et al., 2002; Qi et al., 1999; Reginato et al., 1998; Rocchi et al., 2001; Wang et al., 2000; Xu et al., 2000, 1998). Since our model is established on PPAR␥, it will be important to identify whether these p160 cofactors also modulate the role of other nuclear receptors such as PPAR␣, thyroid hormone receptor, etc. that are involved in the energy balance equation. Although we cannot at present fully exclude a contribution of tissues beyond WAT and BAT to the metabolic alterations in TIF2⫺/⫺ mice, several arguments support the concept that WAT and BAT play a predominant role. First, cell autonomous effects of TIF2 in both MEFs and 3T3-L1 cells were observed on lipid storage. These observations were corroborated by changes in glucose uptake and lipolysis in primary adipocytes isolated from TIF2⫺/⫺ mice. Second, thermogenic effectors such as PGC-1␣ and AOX were not increased in skeletal muscle of TIF2⫺/⫺ mice, which, like BAT, is a thermogenic tissue controlled by the central nervous system. Third, the absence in TIF2⫺/⫺ mice of major changes in nonadipose-derived endocrine modulators of adaptive thermogenesis and fat accumulation suggests that these metabolic alterations are not due to neuroendocrine defects. Finally, ingestion of a chronic high-fat diet induces changes in the TIF2/SRC-1 ratio only in WAT and BAT and not in other organs involved in metabolism (brain, skeletal muscle, and liver). Nevertheless, tissue-specific

Figure 7. Role of p160 Nuclear Coregulators in Energy Metabolism Scheme illustrating the hypothetical role of TIF2 and SRC-1 in obesity and insulin resistance. See discussion for details.

SRC-1 and TIF2 Control Energy Metabolism 939

conditional knockout mouse models will be necessary to fully exclude an eventual contribution of each of these tissues to the TIF2- and SRC-1-mediated changes in energy balance. The altered TIF2/SRC-1 ratio in WAT and BAT during high-fat feeding may be relevant to the pathogenesis of obesity. Under conditions of high-energy intake, such as consumption of a high-fat diet, the increase in TIF2 in WAT leads to fat accumulation. In contrast, the increased TIF2/SRC-1 ratio in BAT reduces the thermogenic potential. Under these conditions, energy expenditure cannot compensate for higher energy intake, thus promoting the development of obesity and insulin resistance (Figure 7). Interestingly, mice heterozygous for a mutation of another coregulator, CBP, are characterized by lipodystrophy and insulin sensitization (Yamauchi et al., 2002). Together with our present findings, this underscores the importance of coregulators as targets for preventive and therapeutic strategies in metabolic disorders such as obesity, type 2 diabetes, and associated complications. Experimental Procedures Animal Experiments The generation of the SRC-1⫺/⫺ (Xu et al., 1998) and TIF2⫺/⫺ (Ge´hin et al., 2002) mouse lines has been described previously. These mice were maintained in a hybrid C57BL/6J/129 SV background (93.75 %/6.25 %, respectively). C57BL/6J mice used in the chronic high-fat feeding study were purchased from Jackson Laboratories (IffaCredo, Lyon, France). Only male, age matched (10–16 weeks old) mice were used. Animals were maintained in a temperature-controlled (23⬚C) facility with a 12 hr light/dark cycle according to the EU guide for use of laboratory animals. Mice had ad libitum access to water and either regular rodent chow (DO4, UAR, France) or high-fat diet (D12327, Research Diet, New Brunswick, NJ). Monosodium glutamate (MSG) was injected subcutaneously (2 mg/kg/ day) in newborn mice for 7 days as described (Imai et al., 2001). Body weight was recorded every week and food intake was measured every second day for 15 consecutive days. Oxygen consumption was measured using the Oxymax apparatus (Columbus Instruments, Columbus, OH), and body fat mass was evaluated in anesthetized mice by dual energy X-ray absorptiometry (PIXIMUS, GE Medical Systems, Buc, France). Unless otherwise stated, blood was collected from the retroorbital sinus after an overnight fast. Blood was kept on ice until centrifugation (1500 g, 15 min at 4⬚C), and the plasma was stored at ⫺20⬚C until analysis. All animals were killed by decapitation. Tissues were harvested in overnight fasted animals, weighed, quickly frozen in liquid nitrogen, and stored at ⫺70⬚C until further processing. Clinical Biochemistry and Evaluation of Glucose and Lipid Homeostasis Intraperitoneal glucose tolerance tests (2 g/kg) were performed as described (Rocchi et al., 2001). Insulin sensitivity experiments were done in overnight fasted animals after which insulin was then injected at a dose of 0.75 U/kg. For the meal tolerance test, food was given ad libitum after an overnight fast. Food intake was recorded and blood was collected 1 hr after the beginning of the meal. Glucose quantification was done with the Maxi Kit Glucometer 4 (Bayer Diagnostic, Puteaux, France). Plasma insulin, adiponectin, and leptin concentrations were measured using ELISA kits (Cristal Chem Inc., IL). Nonesterified fatty acids, triglycerides, ketone bodies and HDL, and total cholesterol were determined by enzymatic assays (Boehringer-Mannheim, Germany). RNA and Protein Preparation and Analysis Total RNA from mouse tissues or cultured cells were extracted and analyzed by Northern blots as described (Rocchi et al., 2001). 36B4 mRNA and 28S levels were determined as a control for loading

(Laborda, 1991). 32P-labeled specific probes used were those for PPAR␥ (Fajas et al., 1999), lipoprotein lipase (Lefebvre et al., 1997), aP2 (Graves et al., 1992), SRC-1 (Xu et al., 1998), TIF2 (Voegel et al., 1996), UCP1 (Arvaniti et al., 1998), PGC-1␣ (Puigserver et al., 1998), and acetylCoA oxydase (Miyazawa et al., 1989). Proteins from mouse tissues were extracted in a solution of pH 7.4 containing 20 mM HEPES, 250 mM sucrose, 4 mM EDTA, and protease inhibitor cocktail. Nuclear extracts from cultured 3T3-L1 cells were performed as described (Fajas et al., 1999). Western blotting was done as described (Gelman et al., 1999) with antibodies directed against SRC-1 (Santa Cruz, CA), TIF2 (Voegel et al., 1998), HA (Wasylyk et al., 1997), or perilipin A and B epitopes (Research Diagnostics Inc., Flanders, NJ). Morphological Studies Pieces of mouse tissues were fixed in Bouin’s solution, dehydrated in ethanol, embedded in paraffin, and cut at a thickness of 5 ␮m. Sections were deparaffinized, rehydrated, and stained with haematoxylin and eosin or used for immunofluorescence studies with a fluorescence-labeled antibody directed against perilipin (BlanchetteMackie et al., 1995). Scanning and transmission electron microscopy were performed as described (Choi et al., 1997). Cell Culture Retroviral Infection and Transfection Studies 3T3-L1, NIH-3T3, gp293, COS, and 293T cells (ATCC, Rockville, MD) were cultured in Dulbecco’s modified Eagle’s medium with 10% FCS, and antibiotics. Mouse embryos were carefully collected between day E13.5 and E14.5 and mouse embryo fibroblasts (MEFs) were prepared as described (Hansen et al., 1999). Retroviral infection was performed as described (Tontonoz et al., 1994). gp293 cells were transfected with either pBABE empty vector (control), pBABE-SRC-1, or pBABE-TIF2 using lipofectamine (Invitrogen SARL, Cergy Pontoise, France). After 48 hr of transfection, the medium containing retroviruses was collected, filtered, treated by polybrene (1 mg/mL), and transferred on 3T3-L1 or NIH-3T3 target cells. Infected cells were selected with puromycin (2.5 mg/mL) for 7 days. To stimulate the accumulation of lipids in cells, medium was supplemented with 2 ␮M insulin, 1 ␮M dexamethasone, and 0.25 mM isobuthyl methyl xanthine for 2 days. The cells were then incubated with insulin and 10⫺7 M rosiglitazone for another 4 days, changing the medium every second day. Fat accumulation was scored by determination of mRNA and protein levels of adipocyte-specific genes and by staining of lipids with Oil Red O (Chawla and Lazar, 1994). Primary adipocytes were prepared as described previously (Marette et al., 1991). Cells were incubated for 30 min at 37⬚C with or without (⫺)-adrenaline (Sigma-Aldrich SARL, St-Quentin Fallavier, France) at 10⫺7 and 10⫺5 M. For glucose uptake experiments, adipocytes were incubated with 100 mM of radiolabeled 14C-deoxyglucose for 30 min. After that period, the cells were washed three times with PBS and lysed in NaOH 0.1 M. Transfections of all cell lines with luciferase or CAT reporter constructs were carried out as described (Rocchi et al., 2001; Schoonjans et al., 1996). All data were normalized for transfection efficiency. The pGL3-(J3wt)-TKLuc and pGL3-(Gal4)5TKLuc reporter constructs were described elsewhere (Rocchi et al., 2001). The following expression vectors were used: pSG5-hPPAR␥2 (Gelman et al., 1999); pcDNA3-BDGal4-hPPAR␥ DE, encoding a chimeric protein composed of the Gal4 DNA binding domain fused to the hPPAR␥ LBD (Gelman et al., 1999; Oberfield et al., 1999); pCMV-p300-CHA (Gelman et al., 1999); pcDNA-TIF2 (Voegel et al., 1998); pcDNA-SRC-1 (Spencer et al., 1997); pSVSPORT-PGC-1␣ (Puigserver et al., 1998); and pCMV-␤-Galactosidase. Production of Proteins, Pull-Down Experiments, and Coimmunoprecipitation Assays The production and purification of SRC-1-GST, p300Nt-GST, TIF2GST, and polyhistidine-tagged hPPAR␥ LBD (His-tagPPAR␥2DE203– 477) has been described (Rocchi et al., 2001). GST pull-down assays using the His-tagPPAR␥2DE203–477 or in vitro translated ER␣ proteins were performed in the presence of rosiglitazone (10⫺4 M) and increasing doses of the coactivator fusion proteins p300Nt-GST, TIF2-

Cell 940

GST, or SRC-1-GST. The samples were then processed by Western blotting as described (Rocchi et al., 2001). Coimmunoprecipitation assays were done as described (Puigserver et al., 1998). Statistical Analysis Data are presented as means ⫾ SEM. The main and interactive effects were analyzed by ANOVA factorial or repeated measures when appropriate. When justified by the ANOVA analysis, differences between individual group means were analyzed by Fisher’s PLSD test. Differences were considered statistically significant at p ⬍ 0.05. Acknowledgments We thank B. Spiegelman for the PGC-1␣ clone and S. Miard and N. Messaddeq for technical assistance. This work was supported by grants from CNRS, INSERM, Hopitaux Universitaires de Strasbourg, European Union (QLG1-CT-1999-00674 and QLRT-2001-00930), and NIH (1-P01-DK59820-01). F.P. held postdoctoral fellowships from the Canadian Institutes of Health Research and the French Atherosclerosis Society. Received: September 5, 2002 Revised: October 28, 2002 References Aranda, A., and Pascual, A. (2001). Nuclear hormone receptors and gene expression. Physiol. Rev. 81, 1269–1302. Arvaniti, K., Ricquier, D., Champigny, O., and Richard, D. (1998). Leptin and corticosterone have opposite effects on food intake and the expression of UCP1 mRNA in brown adipose tissue of lep(ob)/ lep(ob) mice. Endocrinology 139, 4000–4003. Blanchette-Mackie, E.J., Dwyer, N.K., Barber, T., Coxey, R.A., Takeda, T., Rondinone, C.M., Theodorakis, J.L., Greenberg, A.S., and Londos, C. (1995). Perilipin is located on the surface layer of intracellular lipid droplets in adipocytes. J. Lipid Res. 36, 1211–1226. Chawla, A., and Lazar, M.A. (1994). Peroxisome proliferator and retinoid signaling pathways co-regulate preadipocyte phenotype and survival. Proc. Natl. Acad. Sci. USA 91, 1786–1790. Choi, D.S., Ward, S.J., Messaddeq, N., Launay, J.M., and Maroteaux, L. (1997). 5–HT2B receptor-mediated serotonin morphogenetic functions in mouse cranial neural crest and myocardiac cells. Development 124, 1745–1755. Demarest, S.J., Martinez-Yamout, M., Chung, J., Chen, H., Xu, W., Dyson, J., Evans, R.M., and Wright, P.E. (2002). Mutual synergistic folding in recruitment of CBP/p300 by p160 nuclear receptor coactivators. Nature 415, 549–553. Fajas, L., Schoonjans, K., Gelman, L., Kim, J.B., Najib, J., Martin, G., Fruchart, J.C., Briggs, M., Spiegelman, B.M., and Auwerx, J. (1999). Regulation of PPAR␥ expression by ADD-1/SREBP-1: implications for adipocyte differentiation and metabolism. Mol. Cell. Biol. 19, 5495–5503. Ge, K., Guermah, M., Yuan, C.X., Ito, M., Wallberg, A.E., Spiegelman, B.M., and Roeder, R.G. (2002). Transcription coactivator TRAP220 is required for PPAR␥2-stimulated adipogenesis. Nature 417, 563–567. Ge´hin, M., Mark, M., Dennefeld, C., Dierich, A., Gronemeyer, H., and Chambrier, C. (2002). The function of TIF2/GRIP1 in mouse reproduction is distinct from those of SRC-1 and p/CIP. Mol. Cell. Biol. 22, 5923–5937. Gelman, L., Zhou, G., Fajas, L., Raspe, E., Fruchart, J.C., and Auwerx, J. (1999). p300 interacts with the N- and C-terminal part of PPAR␥2 in a ligand-independent and -dependent manner respectively. J. Biol. Chem. 274, 7681–7688. Glass, C.K., and Rosenfeld, M.G. (2000). The coregulator exchange in transcriptional functions of nuclear receptors. Genes Dev. 14, 121–141. Graves, R.A., Tontonoz, P., and Spiegelman, B.M. (1992). Analysis of a tissue-specific enhancer: ARF6 regulates adipogenic gene expression. Mol. Cell. Biol. 12, 1202–1208. Guezennec, C.Y., Nonglaton, J., Serrurier, B., Merino, D., and Defer,

G. (1988). Hormonal and metabolic response to physical exercise, fasting and cold exposure in the rat. Effects on ketogenesis in isolated hepatocytes. Eur J Appl Physiol Occup Physiol 57, 114–119. Hansen, J.B., Petersen, R.K., Larsen, B.M., Bartkova, J., Alsner, J., and Kristiansen, K. (1999). Activation of peroxisome proliferator activated receptor ␥ bypasses the function of the retinoblastoma protein in adipocyte differentiation. J. Biol. Chem. 274, 2386–2393. Imai, T., Jiang, M., Chambon, P., and Metzger, D. (2001). Impaired adipogenesis and lipolysis in the mouse upon selective ablation of the retinoid X receptor alpha mediated by a tamoxifen-inducible chimeric Cre recombinase (Cre-ERT2) in adipocytes. Proc. Natl. Acad. Sci. USA 98, 224–228. Kodera, Y., Takeyama, K., Murayama, A., Suzawa, M., Masuhiro, Y., and Kato, S. (2000). Ligand type-specific interactions of peroxisome proliferator-activated receptor ␥ with transcriptional coactivators. J. Biol. Chem. 275, 33201–33204. Laborda, J. (1991). 36B4 cDNA used as an estradiol-independent mRNA control is the cDNA for human acidic ribosomal phosphoprotein PO. Nucleic Acids Res. 19, 3998. Larsen, M.L., and Illingworth, D.R. (1993). Triglyceride-lowering agents: fibrates and nicotinic acid. Curr. Opin. Lipidol. 4, 34–40. Lefebvre, A.-M., Peinado-Onsurbe, J., Leitersdorf, I., Briggs, M.R., Paterniti, J.R., Fruchart, J.-C., Fievet, C., Auwerx, J., and Staels, B. (1997). Regulation of lipoprotein metabolism by thiazolidinediones occurs through a distinct, but complementary mechanism relative to fibrates. Arterioscler. Thromb. Vasc. Biol. 17, 1756–1764. Lowell, B.B., S.-Susulic, V., Hamann, A., Lawitts, J., Himms-Hagen, J., Boyer, B.B., Kozak, L.P., and Flier, J.S. (1993). Development of obesity in transgenic mice after genetic ablation of brown adipose tissue. Nature 366, 740–742. Marette, A., Deshaies, Y., Collet, A.J., Tulp, O., and Bukowiecki, L.J. (1991). Major thermogenic defect associated with insulin resistance in brown adipose tissue of obese diabetic SHR/N-cp rats. Am. J. Physiol. 261, E204–213. Martinez-Botas, J., Anderson, J.B., Tessier, D., Lapillonne, A., Chang, B.H.-J., Quast, M.J., Gorenstein, D., Chen, K.-H., and Chan, L. (2000). Absence of perilipin results in leanness and reverses obesity in Leprdb/db mice. Nat. Genet. 26, 474–479. Matthias, A., Ohlson, K.B., Fredriksson, J.M., Jacobsson, A., Nedergaard, J., and Cannon, B. (2000). Thermogenic responses in brown fat cells are fully UCP1-dependent. UCP2 or UCP3 do not substitute for UCP1 in adrenergically or fatty scid-induced thermogenesis. J. Biol. Chem. 275, 25073–25081. McInerney, E.M., Rose, D.W., Flynn, S.E., Westin, S., Mullen, T.M., Krones, A., Inostroza, J., Torchia, J., Nolte, R.T., Assa-Munt, N., et al. (1998). Determinants of coactivator LXXLL motif specificity in nuclear receptor transcriptional activation. Genes Dev. 12, 3357– 3368. Miyazawa, S., Osumi, T., Hashimoto, T., Ohno, K., Miura, S., and Fujiki, Y. (1989). Peroxisome targeting signal of rat liver acyl-coenzyme A oxidase resides at the carboxy terminus. Mol. Cell. Biol. 9, 83–91. Oberfield, J.L., Collins, J.L., Holmes, C.P., Goreham, D.M., Cooper, J.P., Cobb, J.E., Lenhard, J.M., Hull-Ryde, E.A., Mohr, C.P., Blanchard, S.G., et al. (1999). A peroxisome proliferator-activated receptor ␥ ligand inhibits adipocyte differentiation. Proc. Natl. Acad. Sci. USA 96, 6102–6106. Olney, J.W. (1969). Brain lesions, obesity, and other disturbances in mice treated with monosodium glutamate. Science 164, 719–721. Puigserver, P., Wu, Z., Park, C.W., Graves, R., Wright, M., and Spiegelman, B.M. (1998). A cold-inducible coactivator of nuclear receptors linked to adaptive thermogenesis. Cell 92, 829–839. Puigserver, P., Adelmant, G., Wu, Z., Fan, M., Xu, J., O’Malley, B., and Spiegelman, B.M. (1999). Activation of PPAR␥ coactivator-1 through transcription factor docking. Science 286, 1368–1371. Qi, C., Zhu, Y., Pan, J., Yeldandi, A.V., Rao, M.S., Maeda, N., Subbarao, V., Pulikuri, S., Hashimoto, T., and Reddy, J.K. (1999). Mouse steroid receptor coactivator-1 is not essential for peroxisome prolif-

SRC-1 and TIF2 Control Energy Metabolism 941

erator-activated receptor alpha-regulated gene expression. Proc. Natl. Acad. Sci. USA 96, 1585–1590. Reginato, M.J., Bailey, S.T., Krakow, S.L., Minami, C., Ishii, S., Tanaka, H., and Lazar, M.A. (1998). A potent antidiabetic thiazolidinedione with unique peroxisome proliferator-activated receptor ␥-activating properties. J. Biol. Chem. 273, 32679–32684. Richard, D., Labrie, A., and Rivest, S. (1992). Tissue specificity of SNS response to exercise in mice exposed to low temperatures. Am. J. Physiol. 262, R921–R925. Richard, D., Picard, F., Lemieux, C., Lalonde, J., Samson, P., and Deshaies, Y. (2002). The effects of topiramate and sex hormones on energy balance of male and female rats. Int. J. Obes. Relat. Metab. Disord. 26, 344–353. Ricquier, D., Gervais, C., Kader, J.C., and Hemon, P. (1979). Partial purification by guanosine-5⬘-diphosphate—agarose affinity chromatography of the 32,000 molecular weight polypeptide from mitochondria of brown adipose tissue. FEBS Lett. 101, 35–38. Rocchi, S., Picard, F., Vamecq, J., Gelman, L., Potier, N., Zeyer, D., Dubuquoy, L., Bac, P., Champy, M.F., Plunket, K.D., et al. (2001). A unique PPAR␥ ligand with potent insulin-sensitizing yet weak adipogenic activity. Mol. Cell 8, 737–747. Schoonjans, K., Peinado-Onsurbe, J., Lefebvre, A.M., Heyman, R., Briggs, M., Deeb, S., Staels, B., and Auwerx, J. (1996). PPAR␣ and PPAR␥ activators direct a tissue-specific transcriptional response via a PPRE in the lipoprotein lipase gene. EMBO J. 15, 5336–5348. Shao, G., Heyman, R.A., and Schulman, I.G. (2000). Three amino acids specify coactivator choice by retinoid X receptors. Mol. Endocrinol. 14, 1198–1209. Spencer, T.E.J., Jenster, G., Burcin, M.M., Allis, C.D., Zhou, J., Mizzen, C.A., McKenna, N.J., Onate, S.A., Tsai, S.Y., Tsai, M.J., and O’Malley, B.W. (1997). Steroid receptor coactivator-1 is a histone acetyltransferase. Nature 389, 194–198. Takahashi, N., Kawada, T., Yamamoto, T., Goto, T., Taimatsu, A., Aoki, N., Kawasaki, H., Taira, K., Yokoyama, K.K., Kamei, Y., and Fushiki, T. (2002). Overexpression and ribozyme-mediated targeting of transcriptional coactivators, CREB binding protein (CBP) and p300, revealed their indispensable roles in adipocyte differentiation through the regulation of peroxisome proliferator-activated receptor-␥ (PPAR-␥). J. Biol. Chem. 277, 16906–16912. Tansey, J.T., Sztalryd, C., Gruia-Gray, J., Roush, D.L., Zee, J.V., Gavrilova, O., Reitman, M.L., Deng, C.X., Li, C., Kimmel, A.R., and Londos, C. (2001). Perilipin ablation results in a lean mouse with aberrant adipocyte lipolysis, enhanced leptin production, and resistance to diet-induced obesity. Proc. Natl. Acad. Sci. USA 98, 6494– 6499. Tontonoz, P., Hu, E., and Spiegelman, B.M. (1994). Stimulation of adipogenesis in fibroblasts by PPAR␥2, a lipid-activated transcription factor. Cell 79, 1147–1156. Trayhurn, P., and Rayner, D.V. (1996). Hormones and the ob gene product (leptin) in the control of energy balance. Biochem. Soc. Trans. 24, 565–570. Vega, R.B., Huss, J.M., and Kelly, D.P. (2000). The coactivator PGC-1 cooperates with peroxisome proliferator-activated receptor ␣ in transcriptional control of nuclear genes encoding mitochondrial fatty acid oxidation enzymes. Mol. Cell. Biol. 20, 1868–1876. Voegel, J.J., Heine, M.J.S., Zechel, C., Chambon, P., and Gronemeyer, H. (1996). TIF2, a 160 kDa transcriptional mediator for the ligand-dependent activation function AF-2 of nuclear receptors. EMBO J. 15, 3667–3675. Voegel, J.J., Heine, M.J., Tini, M., Vivat, V., Chambon, P., and Gronemeyer, H. (1998). The coactivator TIF2 contains three nuclear receptor binding motifs and mediates transactivation through CBP binding-dependent and -independent pathways. EMBO J. 17, 507–519. Wang, Z., Rose, D.W., Hermanson, O., Liu, F., Herman, T., Wu, W., Szeto, D., Gleiberman, A., Krones, A., Pratt, K., et al. (2000). Regulation of sematic growth by the p160 coactivator p/CIP. Proc. Natl. Acad. Sci. USA 97, 13549–13554. Wasylyk, C., Bradford, A.P., Gutierrez-Hartmann, A., and Wasylyk,

B. (1997). Conserved mechanisms of Ras regulation of evolutionary related transcription factors, Ets1 and Pointed P2. Oncogene 14, 899–913. Weiss, R.E., Ge´hin, M., Xu, J., Sadow, P.M., O’Malley, B.W., Chambon, P., and Refetoff, S. (2002). Thyroid function in mice with compound heterozygous and homozygous disruptions of SRC-1 and TIF-2 coactivators: evidence for haploinsufficiency. Endocrinology 143, 1554–1557. Xu, J., Qiu, Y., DeMayo, F.J., Tsai, S.Y., Tsai, M.J., and O’Malley, B.W. (1998). Partial hormone resistance in mice with disruption of the steroid receptor coactivator-1 (SRC-1) gene. Science 279, 1922– 1925. Xu, J., Liao, L., Ning, G., Yoshida-Komiya, H., Deng, C., and O’Malley, B.W. (2000). The steroid receptor coactivator SRC-3 (p/CIP/RAC3/ AIB1/ACTR/TRAM-1) is required for normal growth, puberty, female reproductive function, and mammary gland development. Proc. Natl. Acad. Sci. USA 97, 6379–6384. Yamauchi, T., Oike, Y., Kamon, J., Waki, H., Komeda, K., Tsuchida, A., Date, Y., Li, M.X., Miki, H., Akanuma, Y., et al. (2002). Increased insulin sensitivity despite lipodystrophy in Crebbp heterozygous mice. Nat. Genet. 30, 221–226.