The coactivator PGC-1 cooperates with peroxisome proliferator-activated receptor alpha in transcriptional control of nuclear genes encoding mitochondrial fatty acid oxidation enzymes

Abstract

Peroxisome proliferator-activated receptor alpha (PPARalpha) plays a key role in the transcriptional control of genes encoding mitochondrial fatty acid beta-oxidation (FAO) enzymes. In this study we sought to determine whether the recently identified PPAR gamma coactivator 1 (PGC-1) is capable of coactivating PPARalpha in the transcriptional control of genes encoding FAO enzymes. Mammalian cell cotransfection experiments demonstrated that PGC-1 enhanced PPARalpha-mediated transcriptional activation of reporter plasmids containing PPARalpha target elements. PGC-1 also enhanced the transactivation activity of a PPARalpha-Gal4 DNA binding domain fusion protein. Retroviral vector-mediated expression studies performed in 3T3-L1 cells demonstrated that PPARalpha and PGC-1 cooperatively induced the expression of PPARalpha target genes and increased cellular palmitate oxidation rates. Glutathione S-transferase "pulldown" studies revealed that in contrast to the previously reported ligand-independent interaction with PPARgamma, PGC-1 binds PPARalpha in a ligand-influenced manner. Protein-protein interaction studies and mammalian cell hybrid experiments demonstrated that the PGC-1-PPARalpha interaction involves an LXXLL domain in PGC-1 and the PPARalpha AF2 region, consistent with the observed ligand influence. Last, the PGC-1 transactivation domain was mapped to within the NH(2)-terminal 120 amino acids of the PGC-1 molecule, a region distinct from the PPARalpha interacting domains. These results identify PGC-1 as a coactivator of PPARalpha in the transcriptional control of mitochondrial FAO capacity, define separable PPARalpha interaction and transactivation domains within the PGC-1 molecule, and demonstrate that certain features of the PPARalpha-PGC-1 interaction are distinct from that of PPARgamma-PGC-1.

FIG. 1

PGC-1 enhances PPARα-mediated transactivation. The heterologous promoter reporter construct (PPRE)₃TKLuc (A) or the homologous promoter reporter MCPT.Luc.781 (B) was transiently transfected into CV-1 cells. Expression constructs encoding RXRα/PPARα and/or PGC-1, were cotransfected in the presence of BSA vehicle or oleic acid (250 μM) complexed to BSA as indicated. Bars represent mean (± standard error) relative luciferase units (RLU) normalized (=1.0) to the activity of (PPRE)₃TKLuc (A) or MCPT.Luc.781 (B) cotransfected with expression vector backbone in the absence of ligand. All transfection data represent the means of at least three independent experiments.

FIG. 2

PPARα and PGC-1 cooperate to induce PPARα gene target markers of the mitochondrial FAO pathway. Autoradiographs of Northern blot analysis performed with total RNA (15 μg) isolated from 3T3-L1 preadipocytes infected with recombinant retroviral particles encoding LacZ, PPARα, PGC-1, or PPARα and PGC-1 as indicated at the top are shown. Cells were grown to confluence and induced to differentiate as described in Materials and Methods. Six days after addition of differentiation media, ETYA (+) or vehicle control (−) was added. RNA was isolated 48 h after addition of ligand or vehicle. The blot was hybridized with the radiolabeled cDNA probes indicated on the right. The ethidium bromide (EtBr)-stained RNA is included as a control for loading and RNA integrity.

FIG. 3

PPARα and PGC-1 increase cellular palmitate oxidation rates. Palmitate oxidation studies were performed on 3T3-L1 preadipocytes in culture infected with retroviral vectors expressing LacZ (control), PPARα, PGC-1, or PPARα and PGC-1 as described in Materials and Methods. Following incubation of the cells with [1-¹⁴C]palmitate for 6 h, the amount of ¹⁴CO₂ liberated from the cells was measured by scintillation counting. The bars represent mean (± standard error) ¹⁴CO₂ (in counts per minute) normalized (=1.0) to the mean value obtained with the LacZ-infected control cells. An asterisk denotes a significant difference (P < 0.01) compared to the control value. The double dagger denotes a significant difference (P < 0.01) between the values for PPARα without and with PGC-1.

FIG. 4

PGC-1 interacts with PPARα. (A) The GST–PGC-1 fusion proteins used for PPARα pulldown assays are shown schematically at the top with numbers corresponding to the amino acids within the PGC-1 molecule (31). The locations of the domain necessary for PPARγ binding (31) and the single LXXLL domain are also shown. Autoradiographs depicting the results of GST pulldown assays performed with ³⁵S-labeled PPARα and several GST–PGC-1 fusion proteins or GST alone in the presence of the DMSO vehicle (−) or the PPARα ligand ETYA (+) are shown at the bottom of each panel. The numbers below each pulldown product shown in the autoradiographs indicate the percent total input as determined by phosphorimager analysis. 25% of the input is shown for comparison. (B) Pulldown studies using ³⁵S-labeled PPARα deletion mutant proteins (shown at the top) and GST-PGC₂₈₄. (C) The results of pulldown studies performed with the GST.PGC₁₉₀ fusion protein and the LXXLL mutant, GST.PGC_LXXFF.

FIG. 5

Coactivation of PPARα by PGC-1 requires intact AF2 and LXXLL motifs. To examine functional correlates of the GST pulldown interaction studies, a mammalian cell hybrid system was employed (shown schematically at the top). PPARα or PPARΔAF2 was fused to the Gal4 DNA binding domain (DBD) and cotransfected with an expression plasmid encoding PGC-1 or a mutant PGC-1 in which the LXXLL motif was mutated (PGC_LXXFF). A plasmid containing the Gal4 upstream activating sequence (UAS) multimerized upstream of TK luciferase [(UAS)₃TKLuc] was used as a reporter in these experiments. Transfections were performed in the presence of the PPARα ligands (oleic acid or ETYA) or vehicle controls. Bars represent mean RLU normalized (=1.0) to the value obtained with Gal4DBD cotransfected with expression plasmid backbone in the presence of vehicle.

FIG. 6

The NH₂-terminal region of PGC-1 is required for transactivation function. (A) Schematic representations of PGC-1 deletion mutants used in the transactivation studies shown in Fig. 6B and 7. aa, amino acids. A region homologous with known RNA binding domains and a serine-arginine (SR)-rich domain are shown. (B) Gal4-PPARα was cotransfected with expression vectors into CV-1 cells for each of the PGC-1 deletion mutants shown in panel A in the presence or absence of oleic acid. Bars represent RLU normalized (=1.0) to the activity of the (UAS)₃TKLuc reporter cotransfected with Gal4DBD and empty PGC-1 expression vector. (C) Expression vectors encoding PGC-1 deletion mutants (Fig. 6A), fused to the Gal4DBD, were cotransfected with the (UAS)₃TKLuc reporter plasmid (one-hybrid assay). The values represent RLU normalized (=1.0) to that of the Gal4DBD alone.