Minireview: The PGC-1 Coactivator Networks: Chromatin-Remodeling and Mitochondrial Energy Metabolism

Abstract

Transcriptional coactivators and corepressors are emerging as important regulators of energy metabolism and other biological processes. These factors exert their effects on the transcription of target genes through interaction with selective transcription factors and the recruitment of chromatin-remodeling complexes. Recent genetic and biochemical analyses of the peroxisomal proliferator-activated receptor-γ coactivator 1 networks provide novel mechanistic insights regarding their role in the control of mitochondrial oxidative metabolism. These coactivators integrate tissue metabolic functions in response to nutritional signals as well as circadian timing cues. In contrast to coactivators, transcriptional corepressors have been demonstrated to play an opposite role in the control of mitochondrial biogenesis and respiration. The balance of these coactivator and corepressor proteins and, more importantly, their access to specific transcriptional partners are predicted to dictate the epigenetic states of target genes as well as the metabolic phenotype of the cells. This review highlights the biological role and mechanistic basis of the peroxisomal proliferator-activated receptor-γ coactivator 1 networks in the regulation of chromatin-remodeling and mitochondrial oxidative metabolism.

PGC-1 coactivators control mitochondrial metabolism through chromatin-remodeling.

Mitochondrial oxidative phosphorylation (OXPHOS) serves a central role for energy homeostasis in mammals. Mitochondrial volume density and bioenergetic properties are regulated by physiological signals as well as environmental stimuli to meet the demand for energy in tissues. Not surprisingly, mitochondrial mass is elevated in tissues with high ATP consumption, such as cardiac and skeletal muscle, the central nervous system, and brown adipose tissue. Further, the coupling status and fuel preference of mitochondria are influenced by nutritional status in the body. Impaired mitochondrial OXPHOS contributes to the pathogenesis of a wide range of disease conditions, including metabolic disorders, neurodegeneration, and heart failure. Genetic control of mitochondrial biogenesis and function has been an active area of research in recent years, in particular with regard to the role of the peroxisomal proliferator-activated receptor (PPAR) coactivator 1 (PGC-1) family of transcriptional coactivators (1,2,3,4,5). This review focuses on recent studies that highlight the role of PGC-1α and PGC-1β in mitochondrial energy metabolism, its integration with other biological processes, and the molecular analyses of the PGC-1 transcriptional networks.

Regulation of Mitochondrial Energy Metabolism by the PGC-1 Family of Transcriptional Coactivators

PGC-1α was originally identified as a transcriptional coactivator for nuclear hormone receptors that is highly inducible by cold exposure in brown fat (6,7). Overexpression of PGC-1α in cultured adipocytes activates the expression of mitochondrial genes as well as uncoupling protein 1 (UCP1), an important regulator of adaptive thermogenesis. Subsequent analyses in cultured cells and in transgenic mice have established the biological function of PGC-1α in the regulation of mitochondrial oxidative metabolism in diverse cell types. PGC-1α stimulates mitochondrial biogenesis in C2C12 myotubes, which results in increased mitochondrial mass and respiration (8). PGC-1β shares similar molecular structure and function with PGC-1α, including nuclear receptor binding and transcriptional activation as well as several conserved domains. Like PGC-1α, PGC-1β also activates mitochondrial biogenesis; however, PGC-1β increases a larger fraction of coupled respiration (9). These findings suggest that PGC-1α and PGC-1β may have distinct functions in fine tuning mitochondrial metabolism. The expression of PGC-1 and mitochondrial OXPHOS genes is significantly reduced in skeletal muscle from diabetic patients, implicating a potential role for these factors in the pathogenesis of insulin resistance (10,11). In addition to their role in mitochondrial biology, the PGC-1 coactivators also control diverse biological processes, including hepatic gluconeogenesis and lipoprotein metabolism, skeletal muscle fiber determination, circadian clock function, and angiogenesis, as well as macrophage polarization (Fig. 1).

Figure 1.

Metabolic regulation by the PGC-1 coactivators. PGC-1α and PGC-1β are highly responsive to nutritional and circadian signals as well as environmental stimuli. They enhance mitochondrial biogenesis and, in some cases, the remodeling of the bioenergetic properties of mitochondria. These coactivators also regulate glucose and lipid metabolism as well as other biological processes, serving as a molecular integrator of energy metabolism and tissue function.

Several transgenic mouse models have been generated to investigate the metabolic functions of PGC-1 in vivo. Transgenic expression of PGC-1α in skeletal muscle or heart results in robust activation of mitochondrial biogenesis (12,13). Remarkably, muscle-specific transgenic expression of PGC-1α, to the endogenous levels of PGC-1α present in oxidative muscle fibers, drives the formation of slow-twitch muscle fibers (13). Compared with fast-twitch muscle fibers, slow-twitch muscle fibers have high mitochondrial content and can sustain prolonged muscle contraction through oxidation of glucose and fatty acids. Transgenic muscles acquire functional characteristics of slow-twitch muscle and are more resistant to contraction-induced fatigue. PGC-1α expression is induced in skeletal muscle by exercise in rodents as well as in humans (14,15,16). These studies strongly suggest that PGC-1α likely mediate, to a significant degree, the benefits of exercise, particularly with regard to the effects of physical activity on skeletal muscle fiber composition and metabolic characteristics. The induction of PGC-1α in skeletal muscle is regulated by calcium signaling through calcium/calmodulin-dependent protein kinase IV and the myocyte enhancer factor 2 family of transcription factors (17,18). Remarkably, transgenic expression of calmodulin-dependent protein kinase IV in skeletal muscle leads to the induction of PGC-1α and slow-twitch muscle fiber program (18). Recent studies have demonstrated that PGC-1α protects skeletal muscle from starvation- and denervation-induced atrophy (19,20). Consistently, PGC-1α significantly improves muscle function in the context of muscle dystrophy. These beneficial effects of PGC-1α on muscle biology are likely mediated through improved bioenergetic function as well as the amelioration of inflammation in skeletal muscle (2). Transgenic expression of PGC-1β in skeletal muscle also stimulates mitochondrial biogenesis, whereas it activates type IIx muscle fiber markers (21).

The physiological significance of PGC-1α and PGC-1β in mitochondrial energy metabolism has been demonstrated in mouse strains lacking PGC-1α, PGC-1β, or both. PGC-1α null mice have reduced expression of mitochondrial genes in multiple tissues, particularly brown fat, skeletal and cardiac muscles, and the brain (22,23). Their ability to mount adaptive metabolic responses to stresses, such as cold exposure and starvation, is severely impaired in mice lacking PGC-1α. These defects underscore the role of PGC-1α in the control of mitochondrial OXPHOS as well as other metabolic pathways, such as hepatic gluconeogenesis, adaptive thermogenesis, and heme biosynthesis. Subsequent analyses of liver and skeletal muscle-specific PGC-1α null mice further support these conclusions (24,25,26). PGC-1α deficiency profoundly perturbs cardiac energy homeostasis and leads to impaired contractile function and heart failure after aortic constriction (22,27,28). An intriguing finding in these studies was that PGC-1α null mice develop spongioform neurodegeneration in selective brain areas. Given that only a subset of neurons is affected by PGC-1α deficiency, it is likely that neurons have different sensitivity to the deficiency of mitochondrial OXPHOS. Alternatively, PGC-1β may be sufficient to maintain the energetic function of mitochondria in some, but not all, neuronal populations. The significance of PGC-1α in neuronal energy metabolism was further illustrated in mouse models of Huntington’s disease and Parkinson’s disease (29,30).

In PGC-1β null mice, mitochondrial gene expression is also decreased in several tissues, including the liver and brown fat as well as skeletal and cardiac muscle (31,32,33). RNA interference (RNAi) knockdown of PGC-1β in PGC-1α null brown adipocytes further reduces mitochondrial gene expression, suggesting that both factors contribute to normal expression of mitochondrial genes (34). Recent work with mice deficient in both PGC-1α and PGC-1β clearly demonstrate the essential role of these factors in the control of mitochondrial biogenesis (35). Cardiac-specific PGC-1 deficiency causes an arrest in the induction of mitochondrial oxidative program and growth inhibition during the late fetal stage. These results illustrate that both PGC-1 coactivators are required for maintaining normal mitochondrial OXPHOS, potentially in response to distinct physiological signals. In fact, activation of cAMP signaling has been demonstrated to induce PGC-1α expression in brown adipocytes and hepatocytes as well as muscle cells (17,36,37). In contrast, PGC-1β expression is stimulated by certain fatty acids and cytokines (38,39,40).

Several recent studies have focused on identifying novel molecular pathways and chemical compounds that activate the PGC-1α pathway and mitochondrial OXPHOS. TORC2 (transducer of regulated CREB activity 2) has been demonstrated to be an important regulator of PGC-1α expression through its coactivation of cAMP response element binding protein transcription factor (41,42). TORC2 overexpression in C2C12 myotubes induces mitochondrial biogenesis in a PGC-1α-dependent manner (43). Interestingly, uncoupling of mitochondrial respiratory chain leads to the induction of PGC-1α expression, which contributes to the maintenance of ATP homeostasis in cells (44). Thus, partial mitochondrial uncoupling may enhance fuel oxidation through proton leak as well as the induction of mitochondrial biogenesis by PGC-1α. Unexpectedly, large-scale chemical screens have identified several classes of compounds that regulate PGC-1α expression and mitochondrial function, including microtubule inhibitors (45,46). The mechanistic details of how these compounds activate PGC-1α expression and mitochondrial biogenesis remain unknown.

Transcriptional Coactivators in the Integration of Diverse Biological Processes

Transcriptional coregulators are emerging as physiological integrators of energy metabolism and other biological processes (47,48). The expression of PGC-1α and PGC-1β is regulated by nutritional and hormonal signals as well as by circadian pacemakers. A common theme of these coactivator regulatory networks is that, by coactivating multiple transcription factor targets, they are able to coordinate diverse biological processes that constitute biological responses (3,49). For example, the liver adapts to short-term starvation by activating metabolic functions that are critical for systemic nutrient and energy homeostasis. After starvation, genes involved in hepatic gluconeogenic and fatty acid β-oxidation are induced to enhance glucose production and fat oxidation, respectively. These apparently independent metabolic pathways are controlled by different transcription factors. PGC-1α augments the transcriptional function of hepatocyte nuclear factor 4α, FOXO1, and glucocorticoid receptor, which are known to regulate gluconeogenic genes, whereas it coactivates PPARα in the stimulation of peroxisomal and mitochondrial fat oxidation (37,50,51). Interestingly, the function of PGC-1α in this context is negatively regulated by insulin signaling pathway and requires leucine-rich protein 130 kDa (LRP130), a binding partner for PGC-1α (52,53).

On the other hand, PGC-1β regulates triglyceride synthesis and lipoprotein secretion by the liver in response to dietary fat intake. Hepatic expression of PGC-1β using recombinant adenoviral vectors enhances triglyceride synthesis and significantly elevates plasma triglyceride levels (38,54). Through its coactivation of sterol regulatory element-binding protein and liver X receptor, as well as Foxa2 transcription factors, PGC-1β stimulates the program of hepatic very low-density lipoprotein secretion (38,55). Consistent with a role in lipoprotein secretion, PGC-1β null mice have increased triglyceride accumulation in the liver (31,32,33). Recent studies indicate that PGC-1β is a target of cytokine signaling and is induced in response to interferon and IL-4 through the canonical cytokine-signaling pathway (39,40). In macrophages, PGC-1β activates the alternative pathway that enables macrophages to acquire antiinflammatory properties (40). This functional specification of macrophage phenotype is accompanied by the stimulation of oxidative metabolic program. The molecular partners for PGC-1β in the regulation of macrophage phenotype, however, remain to be explored. In addition to its role in macrophages, PGC-1β also cooperates with estrogen-related receptor (ERR)α and regulates downstream genes that are critical for pathogen clearance (39).

In addition to metabolic regulation by nutrient and hormonal signals, the PGC-1 coactivators play a role in the temporal organization of metabolic functions by circadian pacemakers (56,57). The phenomenon of diurnal metabolic rhythm was observed several decades ago (58,59). In addition to the oscillations of circulating nutrients and hormones, the enzymatic activity and flux through major pathways, such as hepatic glucose production and lipogenesis, exhibit robust daily cycles. The restriction of metabolic functions to specific times during the day is believed to provide unique advantages for organisms as they anticipate and synchronize their feeding and activity cycles to the environment. In mammals, the circadian rhythms of physiology and behaviors are coordinated by biological clocks residing in the brain and also in the peripheral tissues (60,61,62,63,64,65). These molecular clocks consist of transcriptional activators and repressors assembled into autoregulatory loops that generate cyclic transcriptional activation with a period of approximately 24 h. Large-scale transcription profiling has revealed extensive transcriptional rhythms of the genes involved in glucose, lipid, and mitochondrial metabolism (66,67,68). Remarkably, the expression of PGC-1α and PGC-1β is under the regulation of circadian clocks. Moreover, rhythmic PGC-1α activation is required for normal circadian rhythms of locomotor activity, body temperature, and metabolic rate (57). Consistently, PGC-1α null mice have aberrant temporal regulation of metabolic gene expression in the liver and skeletal muscle.

Mechanistically, PGC-1α coordinates energy metabolism and the pacemaker functions through direct regulation of the expression of core clock genes (57). PGC-1α stimulates the expression of Bmal1 and Reverbα through its coactivation of the ROR family of orphan nuclear receptors (Fig. 2A). In this case, PGC-1α functions as a component of the clock oscillator through its direct regulation of clock gene transcription. Interestingly, Reverbα suppresses the induction of Bmal1 transcription by PGC-1α, thereby serving as a negative feedback mechanism to turn off the stimulatory effects of PGC-1α. Because PGC-1α expression itself is diurnally regulated, it is possible that the circadian pacemaker also signals to PGC-1α. The molecular mechanisms that transduce the circadian timing cues to PGC-1 expression and/or activity are currently unknown. Interestingly, two groups have recently demonstrated that Sirtuin 1 (SIRT1), an NAD(+)-dependent histone deacetylase, plays an important role in the regulation of clock gene expression (69,70). Because SIRT1 deacetylates PGC-1α and augments its transcription activity (4,71), it is possible that the cyclic induction of SIRT1 may lead to rhythmic activation of PGC-1α. By linking the metabolic and clock transcriptional programs, PGC-1α serves as a molecular switch that synchronizes metabolic functions to the circadian timing cues in mammalian tissues. Mice deficient in PGC-1β have altered locomotor activity patterns (32), suggesting that it may also play a role in the regulation of circadian biological rhythms.

Figure 2.

Transcriptional networks in the control of mitochondrial gene expression. A, PGC-1α and the regulation of molecular clock. Note that PGC-1α activates the expression of Bmal1 and Reverbα, which function as positive and negative regulators of the clock transcriptional network, respectively. B, The PGC-1 coactivators induce mitochondrial gene expression through docking transcription factors and recruiting chromatin-remodeling complexes. The balance between coactivators and corepressors determines the transcriptional output of mitochondrial genes and tissue metabolic phenotype. CBP, cAMP response element binding protein-binding protein; ROR, retinoid-related orphan receptor; Per, period; Cry, chryptochrome.

Transcription Factor Partners for PGC-1α and PGC-1β in Mitochondrial Gene Expression

The transcriptional partners for PGC-1, particularly those that recruit these coactivators to the promoters of mitochondrial genes, have been the focus of recent studies (5,72). PGC-1α interacts with nuclear respiratory factor 1 (NRF1) and NRF2, which bind to their respective sites present on numerous nuclear-encoded mitochondrial genes. Computational analyses of the proximal promoter sequences of mitochondrial genes have identified several novel motifs that may regulate their expression (73,74,75). Among these, binding sites for the ERR family of nuclear receptors, NRF2, and Yin Yang-1 (YY1) were highly enriched for the PGC-1α target genes (74). ERRα expression is strongly induced by PGC-1α, which also functions as a target for PGC-1α coactivation (74,76,77). Overexpression of ERRα in primary rat neonatal cardiac myocytes induces the expression of genes involved in fatty acid uptake, oxidation, and mitochondrial respiration (78). In this context, PPARα also serves as a downstream target gene of ERRα signaling and is required for the induction of the oxidative program by ERRα.

A critical role for ERRα in maintaining mitochondria bioenergetics in the heart was demonstrated using ERRα null mice. ERRα-deficient hearts have reduced OXPHOS gene expression, ATP synthesis rate, and rapid depletion of phosphocreatine after hemodynamic stress (79). Mice lacking ERRα develop signatures of heart failure in response to left ventricular pressure overload. Recently, genome-wide binding and expression analyses elegantly demonstrate a direct role for ERRα as well as ERRγ in orchestrating a broader program of cardiac energy metabolism and contractile functions (80). In C2C12 myotubes, PGC-1α stimulates the expression of PDK4 through coactivation of ERRα, leading to a switch from glucose to fatty acid oxidation (81). PDK4 expression in skeletal muscle and the liver exhibits robust diurnal rhythms in a PGC-1α-dependent manner (57). Hence, ERRα may also mediate the circadian regulation of PDK4 expression by PGC-1α in these tissues. The physical and functional interaction between PGC-1α and ERRα appears to go beyond mitochondrial oxidative metabolism. Recent studies indicate that PGC-1α induces the expression of vascular endothelial growth factor through the ERRα bindings sites present on the vascular endothelial growth factor promoter (82). The induction of angiogenesis by PGC-1α could, in principle, couple oxygen delivery and mitochondrial oxidative metabolism and protects skeletal muscle from ischemic injury.

NRF2 and YY1 are also predicted to regulate the expression of nuclear-encoded mitochondrial genes. The binding sites for these two transcription factors are significantly overrepresented on the promoters of mitochondrial genes, and interestingly, YY1 appears to serve as a common target for the PGC-1α and mammalian target of rapamycin signaling pathways (73,83). Reporter gene assays indicate that YY1 directly activates the promoters of mitochondrial genes, including Ndufs8 and subunits of the mitochondrial F0F1-adenosine triphosphatase (84,85,86). Further, RNAi knockdown of YY1 reduces mitochondrial gene expression as well as respiratory function. These studies implicate YY1 as a key component of the regulatory pathway that mediates the effects of nutrient status on mitochondrial energy metabolism. In the context of mitochondrial gene regulation by growth and nutrient signaling, the E2F4 transcriptional repressor complex appears to play a role in the suppression of mitochondrial metabolism during growth arrest (87). Global chromatin immunoprecipitation studies indicate that E2F4 binding sites are frequently adjacent to the NRF1 sites on mitochondrial genes, suggesting potential functional cross talk between these factors.

In addition to the transcription factors that regulate core mitochondrial gene expression and organelle biogenesis, several transcription factors have been demonstrated to modulate the bioenergetic properties of mitochondria. For example, PGC-1α induces UCP1 expression through coactivating PPARγ in brown adipocytes in response to cold exposure (7). Similarly, PGC-1α induction of Alas1, a rate-limiting enzyme for heme synthesis that resides in the mitochondrial matrix, enhances hepatic heme biosynthesis during starvation (26). The induction of Alas1 by PGC-1α is mediated, at least in part, through the coactivation of FoxO1 in the liver. Along this line, short-term starvation increases the expression of pyruvate dehydrogenase kinase 4 (PDK4) in skeletal muscle and the liver, in part, through the coactivation of ERRα by PGC-1α (81). PDK4 serves as a gatekeeper for the entry of pyruvate into mitochondria for oxidation. The induction of PDK4 results in a switch from glucose to fatty acid oxidation in the mitochondria. Hence, elevated PDK4 is expected to suppress the oxidation of glucose, a physiological response that is necessary for maintaining glucose homeostasis during starvation. The ability to control mitochondrial bioenergetics without affecting mitochondrial mass is emerging as an important mechanism that regulates the metabolic properties of this organelle.

Role of PGC-1α in the Recruitment of Chromatin-Remodeling Complexes

PGC-1α and PGC-1β function through recruiting chromatin-remodeling complexes, including the enzymatic complexes responsible for histone modification and nucleosome remodeling. PGC-1α physically interacts with p300/cAMP response element-binding protein-binding protein and GCN5 histone acetyltransferases (36,88). The recruitment of these histone acetyltransferase complexes to the proximity of PGC-1α target genes increases histone acetylation, a key step in transcriptional activation. Interestingly, GCN5 also acetylates PGC-1α and represses its transcriptional activity, potentially serving as a feedback mechanism. PGC-1α contains multiple lysine residues that undergo acetylation, although the biological function of individual acetylation sites remains to be defined. In contrast, the SirT1 histone deacetylase plays the opposite role and deacetylates PGC-1α in response to cellular nutrient status. Deacetylation of PGC-1α by SirT1 at these lysine residues enhances its function in the regulation of hepatic gluconeogenic and mitochondrial gene expression (4,71,88). In the context of metabolic regulation, PGC-1α appears to serve as an important target of SirT1 and the beneficial effects of resveratrol, an activator compound for SirT1 (89,90).

Recent studies demonstrate that PGC-1α increases histone 3 lysine 4 trimethylation, a hallmark for transcriptional activation, whereas at the same time it reduces H3K9 dimethylation, a marker associated with transcriptional repression (57). The exact histone methyltransferase and demethylase complexes that function in cooperation with PGC-1α remain to be identified. However, both PGC-1 coactivators interact with host cell factor 1, a protein that associates with the mixed lineage leukemia (MLL) family of histone 3 lysine 4 methyltransferase complexes (91,92). It is possible that host cell factor 1 may direct histone methyltransferases to the PGC-1α target loci.

In a genome-wide coactivation study, PGC-1α was found to interact with BRG1/BRM-associated factor (BAF60a), a subunit of the SWI/SNF nucleosome-remodeling complex (93). Adenoviral-mediated expression of BAF60a in primary hepatocytes activates the expression of peroxisomal and mitochondrial fatty acid oxidation genes and lowers hepatic triglyceride levels in mouse models of hepatic steatosis. The BAF60a/PGC-1α complex is required for the transcriptional function of PPARα in the context of hepatic lipid metabolism. Interestingly, BAF60a and its homolog BAF60c also interact with several other nuclear hormone receptors (94,95). As such, BAF60a may function as a targeting subunit for the SWI/SNF complexes through its interaction with PGC-1α and PPARα, leading to the stimulation of specific transcriptional programs. The induction of hepatic fat oxidation genes by PGC-1α is further modulated by Lipin 1, a gene mutated in fatty liver dystrophy (96). BAF60a also induces the expression of PEPCK as well as several clock genes, raising the possibility that the BAF60a/PGC-1α interaction may play a more general role in the transcriptional regulation by PGC-1α.

Transcriptional Repressors in the Regulation of Mitochondrial Oxidative Metabolism

Several transcriptional corepressors, most notably receptor-interacting protein 140 (RIP140) (NRIP1), have been implicated in the regulation of mitochondrial biogenesis (97). RIP140 was initially identified as a nuclear receptor interacting protein that has repressor activities (98,99). Although RIP140 is ubiquitously present in many tissues, its expression is regulated by various hormones and strongly induced during adipocyte differentiation (100,101). Moreover, RIP140 is abundantly expressed in EDL and gastrocnemius, which contain glycolytic and mixed muscle fibers, respectively. In contrast, the mRNA levels of RIP140 are significantly lower in oxidative muscles, such as soleus and diaphragm (102). Recent studies with RIP140 null mice indicate that RIP140 deficiency results in increased mitochondrial gene expression and oxidative capacity. The most striking effects were observed in the EDL muscle with a strong increase of succinate dehydrogenase activity in RIP140 null mice. These metabolic changes are accompanied by increased abundance of oxidative type IIa and IIx muscle fibers in the EDL. Mice deficient in RIP140 are lean and resistant to diet-induced obesity due to increased energy expenditure (100). Consistent with these loss of function analyses, transgenic expression of RIP140 in skeletal muscle has the opposite effects on mitochondrial and muscle fiber gene expression. Elevated RIP140 levels result in more glycolytic muscle phenotype (102). These findings indicate RIP140 and PGC-1α have opposing effects on the metabolic and contractile properties of skeletal myofibers.

Similar effects of RIP140 in the regulation of mitochondrial oxidative metabolism were observed in adipocytes (103,104). RIP140 null adipocytes have increased mitochondrial biogenesis and fatty acid oxidation rate as well as UCP1 gene expression (103). RNAi knockdown of RIP140 in adipocytes enhances insulin-stimulated glucose uptake, likely due to increased glucose oxidation (104). In contrast, reexpression of RIP140 in RIP140 null adipocytes suppresses the expression of a large number of mitochondrial genes and decreases cellular respiratory function. These studies clearly demonstrate that RIP140 regulates the oxidative metabolic program in a cell-autonomous manner. RIP140 contains four repression domains as well as multiple LXXLL motifs that mediate its interaction with nuclear receptors. The repressor activity of RIP140 is mediated, at least in part, through its interaction with histone deacetylase complexes (105,106). In reporter gene assays, RIP140 inhibits the transcriptional activity of all ERR members (107). ERRα also appears to mediate the derepression of UCP1 in RIP140-deficient adipocytes (108). Because ERRα and ERRγ interact with PGC-1 coactivators and are required for normal mitochondrial gene expression, a possible mechanism may involve competitive interaction of ERRα with RIP140 and the PGC-1 coactivators. Similar antagonistic action may occur in the context of other nuclear receptors, particularly PPAR and thyroid hormone receptor, which also regulate the expression of certain mitochondrial genes.

Because both RIP140 and PGC-1α interact with their nuclear receptor partners in a ligand-dependent manner, it remains unclear how the recruitment of these factors respond to hormonal signals. The stoichiometry of coactivators and corepressors in the cell may play an important role in determining the biological outcome. Deficiency of Sin3A, a core component of a multiprotein corepressor complex, significantly increases the expression of genes involved in oxidative metabolism (109,110). The role of Sin3A corepressor in the regulation of mitochondrial genes appears to be conserved in Drosophila and in mice. How the Sin3A transcriptional complex controls mitochondrial gene expression remains unclear. Nevertheless, Sin3A may modulate nuclear receptor function through its interaction with the nuclear receptor corepressor and silencing mediator of retinoid and thyroid hormone receptor corepressor complexes (111,112,113). Together, these data strongly suggest that the relative level and/or activity of coactivators and corepressors may serve an important function in maintaining cellular oxidative metabolism (Fig. 2B).

In summary, the regulatory networks that control mitochondrial oxidative metabolism in mammals comprise transcription factors and coregulators that act in concert to bring about organelle biogenesis as well as fine tuning of its bioenergetic properties. It is somewhat unexpected that transcriptional coactivators and corepressors are emerging as core regulators of the program of oxidative metabolism. A potential advantage here is that coregulators have the capability of modulating the activity of multiple target transcription factors, thereby coordinating different aspects of complex biological programs. In the case of PGC-1 coactivators, it has become clear that several transcription factors are involved, including NRF1, NRF2, ERRs, and YY1. In addition, certain transcription factors appear to mediate the regulation of unique mitochondrial functions in response to physiological signals, such as adaptive thermogenesis and heme biosynthesis. In contrast to PGC-1, RIP140 and Sin3A corepressor complexes inhibit the expression of mitochondrial genes, raising an intriguing possibility that the balance of coactivators and corepressors may determine cellular metabolic phenotype with regard to mitochondrial OXPHOS. Hence, molecular and/or chemical approaches that alter this balance are predicted to be an effective strategy for targeting mitochondrial metabolism and systemic physiology.