Review: A Cold-Inducible Coactivator of Nuclear Receptors Linked to Adaptive Thermogenesis

Pere Puigserver, Zhidan Wu, Cheol Won Park, Reed Graves, Margaret Wright, and Bruce M. Spiegelman. Cell, Vol. 92, 829-839, March 20, 1998.

Background: Obesity is the result of a consistently greater intake of energy, i.e. food, than output of energy. To shift the energy equilibrium back toward a balanced state, energy dissipation can be increased to match energy intake. Energy expenditure occurs through physical activity, basal metabolic rate, or adaptive thermogenesis. Adaptive thermogenesis is a part of the sympathetic nervous system's attempt at maintaining homeostasis during changing environmental conditions. Typically in response to cold exposure or excessive caloric intake, adaptive thermogenesis increases the production of mitochondria and the rate of ATP production in order to create more heat.

Vital to this elevated energy dissipation are the mitochondrial uncoupling proteins (UCP)s. UCPs facilitate proton transport within the mitochondrial membrane, by reducing the hydrogen ion gradient associated with oxidation occurring in the electron transport chain. UCP-1, one of three UCPs, is located in brown atipose tissue (BAT), whose cells function primarily to dissipate energy as heat in response cold or excessive caloric intake. Sufficient stores of BAT are not found in humans, but UCP-1 expression can been seen in the white adipose tissue (WAT) of humans, indicating that WAT may contain unrealized BAT.

Thyroid hormone receptor (TR) and PPAR are two nuclear hormone receptors which play essential roles in UCP-1 gene expression and BAT cell differention. PPAR's ability to promote BAT cell differentiation, stems from the binding of a PPAR/RXR heterodimer to an enhancer sequence in the Ucp-1 gene. This binding, however, is specific to brown fat cells only. This indicates that PPAR is regulated by a cofactor which allows it to only regulate thermogenic genes within BAT cells.

This paper describes a coactivator of PPAR and other nuclear receptors from brown fat cells. Called PPAR Gamma Coactivator (PGC)-1, this cofactor shows induction in BAT cells and skeletal muscle in response to cold exposure and regulation of mitochondrial genes which contribute to the process of adaptive thermogenesis.

Figure 1: The opening figure contains the amino acid sequence of PGC-1. The yeast two-hybrid system was used to locate the PGC-1 gene. Once located the gene was cloned and characterized. PGC-1 is a 795 amino acid protein of a predicted molecular weight of 92kda. Identified within the sequence are particular regions demonstrating a specific activity. The sequence indicates an RNA-binding region (aa 677-709), three sites for phosphorylating kinase A, a couple of SR rich domains (aa 565-598 and 617-631), and a LXXLL motif (aa 142-146). Databank searches indicated that PGC-1 represented a novel protein.

Figure 2: This figure addresses the interaction between PGC-1 and PPAR. A fusion protein of GST and PGC-1 was made and coupled to an immobilized matrix and incubated with radiolabeled PPAR and other nuclear receptors. Bound protein were eluted a resolved by SDS-PAGE and visualized by autoradiography. To test if PGC-1's binding to a nuclear receptor protein was ligand dependent, a condition was set such that each nuclear receptor had its appropriate ligand or vehicle. GST without PGC-1 was used for a negative control; there was no positive control.

PGC-1 showed nonligand-dependent binding with PPAR. Thyroid hormone receptor (TR) showed primarily nonligand-dependent binding though there was some increase in band intensity after the addition of TR's ligand. Strong ligand-dependent binding was seen in with retinoic acid receptor (RAR) and the estrogen receptor (ER). No binding was seen, with retinoid X receptor (RXR). These results indicate that PGC-1 binds to many different nuclear receptor proteins, with and without ligand dependence.

Figure 2 also shows PPAR interaction with PGC-1 within mammalian cells. Using transfected COS cells an immunoprecipitation was conducted which showed an nonligand-dependent interaction between PGC-1 and PPAR. PGC-1 was tagged with HA and anti-HA-PGC-1 antibodies were used to detect PGC-1 and anti-PPAR was used to detect PPAR. The final part of figure 2 shows PGC-1 expression entirely in the nucleus of COS cells. PGC-1 was coupled with green fluorescent protein and expressed in COS cells. A phase contrast image was provided to localize the nucleus before fluorescence.

Figure 3: The objective to this set of images was to demonstrate PGC-1 interaction with PPAR does not use the typical C-terminal AF-2 domain indicative of coactivators with an LXXLL motif. The first image shows the C-terminal deletions made to PGC-1 proteins. The fragments were created using particular restriction enzymes and translation of the truncated mRNA was done in the presence of radiolabeled methionine.

The various radiolabeled C-terminal PGC-1 fragments were incubated with GST, serving as a negative control, and a GST-PPAR fusion protein. The percentage of binding was quantified using phosphorimage analysis. The results showed that PPAR bound to PGC-1 effectively with as many as 505 of 797 total amino acids subtracted from the C-terminal. The only instance in which PPAR did not bind was when the protein was cut down to 292 amino acids. This region still contained the LXXLL motif indicating that it was not necessary for PGC-1 interaction with PPAR. This confirms the claim made that PGC-1 does not use the LXXLL motif to bind as in typical coactivator fashion.

Figure 4: The aim of these data were to demonstrate that PGC-1 does not bind at the C-terminal AF-2 domain of PPAR. To do this deletions of PPAR were made and incubated an uncut, radiolabeled PGC-1 protein. PGC-1 showed binding up until a fragment with 228 amino acids deleted from the amino terminus. This truncated PPAR protein did show binding with a SRC-1, a coactivator which has the LXXLL motif and binds at the C-terminal AF-2 domain, but the data were not shown.

Figure 5: The aim of figure 5 was to demonstrate that PGC-1 expression is induced by exposure to cold. A northern blot showed that at ambient temperatures (24 degrees) Pgc-1 shows expression in mouse heart, kidney, brain, and brown fat. Ethidium bromide-stained 28sRNA was shown in all lanes as a positive control for quality of RNA. In a second northern blot, mice were exposed to a cold environment (4 degrees) for 3 or 12 hours. PGC-1 showed increased band intensity in brown fat. PGC-1 did not show expression in white fat at either ambient or cold temperatures.

In the final northern blot of figure 5, cold induced expression of other mitochondrial proteins were examined in various mouse tissues in addition to the expression of PGC-1. The northern showed PGC-1 production in skeletal muscle, indicating PGC-1 expression occurs in skeletal muscle only in a cold environment. Many of the mitochondrial proteins essential for adaptive thermogenesis were produced in all tissue types. The data from figure 5 demonstrate that when exposed to cold PGC-1 is expressed in skeletal muscle and its expression is increased in brown fat. Figure 5 also shows that cold induced expression of PGC-1 parallels the expression of many mitochondrial proteins used in adaptive thermogenesis.

Figure 6: The aim of this figure was to show if PGC-1 expression would be increased in brown fat cells upon exposure to a B-adrenergic agonist. A B-adrenergic agonist, when introduced to brown fat cells, mimics the effect of cold exposure. HlB1B brown fat cells were treated with isoproterenol, a B-adrenergic agonist, for 10 hours. A northern blot showed bands of higher intensity for both UPC-1 and PGC-1 than their respective brown fat cell production without the presence of isoproternol. Ethidium bromide-stained 28S RNA was shown, as in the northerns of figure 5, as a control for the quality of RNA.

Figure 7: The aim of this figure was to demonstrate that PGC-1 can function as a transcriptional coactivator of TR and PPAR. To accomplish this a CAT assay was used. A CAT reporter gene was linked to the UCP-1 promoter and tranfected into a cell along with expression vectors for PPAR, RXR, and PGC-1. B-galactosidase was used as the reporter gene in the CAT assay to determine transfection efficiency. As mentioned earlier, PPAR and RXR form a heterodimer which recognizes the UCP-1 promoter. In fibroblasts, however, this recognition results in poor transactivation. To improve transactivation in the fibroblasts a cocktail was added containing ligands, which when used in brown fat cells increased transactivation of UCP-1. This also showed little improvement of transactivation. The addition of PGC-1 to this entire mixture, however, showed dramatic improvement in the transactivation. PGC-1 acts as a coactivator because without the presence of the other additives it has little effect on the transactivation activity. These same results were found using the TR/RXR heterodimer.

The third image in figure 7 demonstrates that PGC-1 shows its grteatest transactivation activity in the presence of a host of different ligands. The fourth image in figure 7 shows that PGC-1 can transactvite independently of a nuclear receptor protein such as PPAR or TR. BY coupling PGC-1 to the DNA-binding domain (DBD) of GAL4 and combining it with a GAL4 DNA binding target sequence transactivation occurs. This opposes the behavior of another nuclear receptor coactivator, SRC-1, which when coupled to GAL4 DBD shows very little transactivation. These data indicate that the nuclear receptor proteins serve only to bring PGC-1 to the proper DNA sites.

Figure 8: The aim of figure 8 was to assess the ability of PGC-1 to induce the expression of genes for adaptive thermogenesis. To do this a retroviral vector containing pBabe, serving as a control, and a retroviral vector containing PGC-1 were added to 3T3F442A preadipose cells. RNA produced by the control set showed little PGC-1 production, whereas the PGC-1 vector showed substantial amounts of PGC-1 RNA. This ectopic PGC-1 enhanced the expression of all the mitochondrial proteins analyzed. Finally a Southern blot of mitochondrial DNA production during the presence and absence of PGC-1 was performed. The result showed greater band intensity in the presence of the retroviral vector containing PGC-1. This indicates that ectopic expression of PGC-1 can induce the production of mitochondria.

Future research areas: The data presented in the paper make a strong case for PGC-1 as a novel coactivator in the process of adaptive thermogenesis. The research is sound except in a few areas. Many of the figures presented lacked positive controls. Though, this does not disprove any of the claims made it does weaken them somewhat. The practice of not showing data pertinent to a claim weakens its validity, as in the figure 4 for example. Finally the differing band intensities, especially in figure 8, need to be quantified for claims to be more believable. Other than these flaws the paper proved its claims rather well.

The discussion at the end of the paper brings forth some questions for future research. The next step, as the paper alludes, that should be taken in assessing the part of PGC-1 in adaptive thermogenesis would be experiments with genetic gain-of-function and loss-of-function in mice. Both these scenarios could be played out using gene targeting by homologous recombination. The paper also leaves PGC-1 transactivation mechanism up in the air. The paper mentions the tendency of proteins showing a close proximity of an SR region and a RNA binding-domain to bind to the C-terminal domain of RNA polymerase II. This could be tested using the same C-terminal deletion assay performed in figures 3 and 4. This would indicate if binding occurs and if so what parts of the respective proteins need to be present.

To return to the DLC Homepage click here.