PerspectiveBiochemistry

Ligand-Dependent and -Independent Regulation of PPARγ and Orphan Nuclear Receptors

See allHide authors and affiliations

Science Signaling  02 Dec 2008:
Vol. 1, Issue 48, pp. pe52
DOI: 10.1126/scisignal.148pe52

Abstract

Peroxisome proliferator–activated receptor γ (PPARγ) is considered to be a ligand-activated nuclear receptor with essential roles in adipogenesis, glucose and lipid homeostasis, and inflammatory responses. An unusually large ligand-binding pocket is a distinguishing feature of PPARγ and two related receptors, PPARα and PPARβ (also known as PPARδ), which allows these receptors to interact with diverse chemical ligands including various fatty acids, fibrates, and the thiazolidinedione class of antidiabetic drugs. However, the physiologically relevant ligand of PPARs remains unknown. A PPARγ mutant that is incapable of binding ligands unexpectedly directed adipogenesis when introduced into fibroblasts. This raises key issues regarding the existence and roles of the hypothetical physiological ligands for PPARγ, issues that may also apply to other “orphan” nuclear receptors lacking bona fide ligands. Identification of the physiological ligands of PPARs and orphan nuclear receptors will be crucial for understanding the biology of these receptors.

Nuclear receptors bind to small lipophilic ligands, such as steroid hormones, thyroid hormone, and vitamin A derivatives. The nuclear receptor and ligand complexes can then bind directly to DNA and activate the transcription of genes involved in a broad spectrum of physiological processes, ranging from development and differentiation to metabolism (1). Small-molecule ligands are critical for determining the specificity and activity of most nuclear receptors, because ligand binding induces receptor conformational changes that affect the recruitment of coactivators or co-repressors, a key mechanism for regulating target gene expression (2). This ligand-dependent regulation has become the focus of intense research, and over the past two decades, there have been tremendous efforts to identify nuclear receptor ligands and develop them for pharmaceutical purposes. In addition to therapeutic uses, nuclear receptor ligands have also been explored as chemical tools for probing the biology of nuclear receptors through a “reverse endocrinology” approach (3). For example, the identification of thiazolidinediones (TZDs), a group of antidiabetic drugs, as ligands for peroxisome proliferator–activated receptor γ (PPARγ) has linked this receptor to insulin resistance and glucose homeostasis. PPARγ therefore has emerged as an important drug target for the treatment of diabetes (4).

PPARγ and two related receptors, PPARα and PPARβ (also known as PPARδ), form a subfamily of nuclear receptors that play key roles in lipid and glucose metabolism (57). PPARγ was initially characterized as a master regulator of adipocyte gene expression and differentiation (8, 9). Overexpression of PPARγ in fibroblasts converts these cells to adipocytes, and the knockout of PPARγ in mice prevents the development of adipose tissue (10, 11). Moreover, the activation of PPARγ by synthetic ligands, such as TZDs, can also drive adipocyte differentiation (4).

The crystal structures of receptor and TZD complexes revealed that the PPARγ ligand-binding domain (LBD) is a three-layer α-helical sandwich, a feature common to all nuclear receptor structures determined to date (12, 13). The middle layer of helices is present only in the top half of the LBD and is absent in the bottom half, thus creating a large ligand-binding pocket (the white surface in Fig. 1A). Similar to the activation of other nuclear receptors, PPARγ activation depends on the conformation of its C-terminal activation function-2 (AF-2) helix (the red helix in Fig. 1B). The head group of TDZs forms a direct hydrogen bond with Y473 (14) in the end of the AF-2 helix and locks this helix into its active conformation (Fig. 1B). The interaction between AF-2 and TZDs is further stabilized by hydrogen bonds between the TZD head group and three PPARγ pocket residues: Q286, H323, and H449 (Fig. 1B; amino acid numbering is from the human system). The active conformation of the AF-2 helix forms a charge-clamp pocket that interacts with the LXXLL motif (where X is any amino acid) of coactivators, ultimately leading to gene activation (Fig. 1D). In addition to TZDs, PPARγ is also activated by diverse ligands with distinct chemical structures, such as oxidized and nitrated fatty acids or N-aryl tyrosine derivatives. This ligand-binding promiscuity may be due to the large size of the PPARγ pocket (~1500 Å3), which is three to five times larger than the steroid-binding pocket of estrogen or androgen receptors (15). The mechanism of PPARγ activation by its diverse chemical ligands is conserved; for example, nitrated fatty acids and the TZD drugs form a similar hydrogen bond network with the AF-2 helix (compare Fig. 1, B and C).

Fig. 1

The structural determinants of ligand binding and cofactor recruiting by PPARγ. (A) The ribbon structure of PPARγ, with its ligand-binding pocket shown as a white surface. The coactivator binding domain (AF-2) is indicated. (B) Q286 of helix 3 (H3) plays a crucial role in binding to a TZD head group. H323 of helix 5 (H5) and H449 of helix 10 (H10) are also involved in stabilizing the hydrogen bonds between the ligand and Y473 from the AF-2 helix (red). (C) These same residues that interact with TZD ligands contribute to binding a fatty-acid head group, (D) E471 from the AF-2 helix plays a crucial role in binding the LXXLL motif of the coactivator, SRC-1 (yellow helix). E471, together with K301 from helix 3 (H3), forms a charge clamp that caps both ends of the coactivator helix. The carbon atoms of ligands are yellow, and carbon atoms of PPARγ are green. Nitrogen and oxygen atoms are shown as blue and red, respectively.

In addition to ligand-activated function, PPARγ also has high basal ligand-independent activity (16, 17). This high basal activity has previously been attributed to endogenous fatty acids, which can activate PPARγ and are present in cells at relatively high concentrations (16, 17). Unexpectedly, a PPARγ mutant, Q286P, which does not respond to any known PPARγ ligands (18), can still initiate adipogenesis (19). Because Q286 forms conserved interactions with the head group of natural fatty acids (20, 21) or the TZD molecules (12, 13) (Fig. 1, B and C), the Q286P mutation should disrupt the ability of PPARγ to respond to ligands. However, the Q286P mutant exhibits basal transcriptional activity comparable to that of the wild-type receptor, suggesting that activation by endogenous fatty acids is not responsible for the high basal activity of PPARγ. Most remarkably, the Q286P mutant drives the adipogenesis program both in vitro and in vivo to a similar degree as the wild-type receptor (19), suggesting that ligand-dependent functions of PPARγ may not be required for adipogenesis. Nevertheless, it is possible that unidentified endogenous ligands bind and activate PPARγ in a mode that is distinct from that of TZDs or of fatty acid ligands, and thus ligand contacts with Q286 are not required. In contrast, a mutation in the charge clamp residue E471 of the AF-2 helix, which forms a crucial interaction with coactivators (Fig. 1D), abolishes both ligand-dependent and -independent activity of PPARγ and fails to activate PPARγ target genes or to promote adipogenesis. Thus, it is clear that the transcriptional activity of PPARγ is essential for adipogenesis, even if ligand activation is not.

What is the basis for the ligand-independent activity of PPARγ? The structure of ligand-free PPARγ assumes both active and inactive conformations (13), suggesting an equilibrium between these two functional states. The active conformation may be favored in the presence of PPARγ agonists or increased amounts of PPARγ coactivators, such as PPARγ coactivator 1α (PGC-1α). Indeed, purified PPARγ has a strong affinity for PGC-1α even in the absence of ligand (22), further supporting the notion that endogenous ligands may not be necessary for the high basal activity of PPARγ. Is this basal activity of PPARγ sufficient for adipogenesis? When the abundance of the Q286P mutant protein is the same as the abundance of the wild-type receptor in mature adipocytes, which is much higher than the amounts present in pre-adipocytes, then yes, it appears that ligand-independent PPARγ activity is sufficient. However, if PPARγ is present in pre-adipocyte amounts, it may be that unidentified endogenous ligands are required to initiate adipogenesis. Although naturally occurring compounds, such as oxidized and nitrated fatty acids, have been shown to bind to PPARγ and promote adipogenesis, the true endogenous and physiologically relevant ligand of PPARγ remains to be determined. Another possibility is that an unknown PPARγ ligand is only transiently produced during the early phase of adipogenesis (23), which, if true, poses a technical challenge for identifying the putative physiological ligand of PPARγ.

Despite the activity of PPARγ in the absence of ligand, many of its physiological and pharmacological actions—such as improving insulin sensitivity and countering inflammation—are ligand-dependent (24, 25). Indeed, the diverse pharmacological functions of PPARγ depend on specific ligands, which in turn dictate the recruitment of specific coactivators for regulating gene expression. For example, 9-fluorenylmethyl chloroformate-l-leucine (FMOC-l-leucine), a PPARγ ligand, enhances insulin sensitivity and reduces glucose levels without promoting the weight gain normally associated with the TZD class of PPARγ ligands. It does so by preferentially recruiting steroid receptor coactivator 1 (SRC-1) but not the related SRC-2 (26). Also, a constitutively active PPARγ-VP16 fusion mutant initiates a different adipogenic gene program from that produced by PPARγ bound to rosiglitazone, a TDZ-class drug (27). Hence, distinct ligands modulate PPARγ activity in a tissue- and target gene–specific manner. Given the large size of the PPARγ pocket, it is likely that PPARγ may have different physiological ligands in different tissues.

Besides PPARγ, there are a number of other nuclear receptors, including nuclear receptor related 1 (Nurr1) and steroidogenic factor 1 (SF-1) that exhibit constitutive activity in cell-based assays (28, 29). These receptors are orphan nuclear receptors, and their activities have been characterized as ligand-independent. Subsequent studies found that many so-called orphan receptors are regulated by ligands. For example, SF-1 activity is regulated by phospholipids in a manner dependent on the carbon length of phospholipid fatty acid tails (29, 30). Nurr1 does not contain a ligand-binding pocket (28), so it is surprising that Nur77, which is structurally similar to Nurr1, is activated by small-molecule ligands, such as cytosporone B (31, 32). Because nuclear receptors possess a great degree of structural plasticity around the ligand-binding pocket (33), it is possible that ligands induce conformational changes that open the Nur77 pocket. Another example of an orphan nuclear receptor that is no longer an orphan is chicken ovalbumin upstream promoter transcription factor II (COUP-TFII), which was subsequently revealed to be activated by retinoic acid (34). Together, these examples teach us that the constitutive activities observed for orphan nuclear receptors may not be truly ligand-independent and that ligand identification for the remaining orphan receptors may lead us to unexpected signaling mechanisms.

In conclusion, the ligand-binding domains of nuclear receptors assume a conserved three-layer helical bundle that is ideal for binding small molecules. Based on this common structural feature, nuclear receptors appear to be evolutionarily selected for binding and regulation by small-molecule ligands. Although overexpression of PPARγ in pre-adipocytes can drive adipogenesis in a ligand-independent manner, adipogenesis may still require endogenous and yet to be identified PPARγ ligands. A parallel conclusion may be made for orphan nuclear receptors that display constitutive activity. Pursuing knowledge about the physiological ligands of PPARγ and orphan receptors will continue to be one of the frontiers in the field of nuclear receptors.

Acknowledgments

We thank D. Nadziejka for editing the manuscript. This work was supported in part by the Jay and Betty Van Andel Foundation (H.E.X.), U.S. National Institutes of Health grants DK071662 and DK066202 (H.E.X.) and HL089301 (H.E.X. and Y.L.).

References and Notes

Stay Connected to Science Signaling

Navigate This Article