Regulation of the Transcriptional Activity of the Peroxisome Proliferator-activated Receptor α by Phosphorylation of a Ligand-independent trans-Activating Domain*

The peroxisome proliferator-activated receptors (PPARs) are a subgroup of nuclear receptors activated by fatty acids and eicosanoids. In addition, they are subject to phosphorylation by insulin, resulting in the activation of PPARα, while inhibiting PPARγ under certain conditions. However, it was hitherto unclear whether the stimulatory effect of insulin on PPARα was direct and by which mechanism it occurs. We now demonstrate that amino acids 1–92 of hPPARα contain an activation function (AF)-1-like domain, which is further activated by insulin through a pathway involving the mitogen-activated protein kinases p42 and p44. Further analysis of the amino-terminal region of PPARα revealed that the insulin-induced trans-activation occurs through the phosphorylation of two mitogen-activated protein kinase sites at positions 12 and 21, both of which are conserved across evolution. The characterization of a strong AF-1 region in PPARα, stimulating transcription one-fourth as strongly as the viral protein VP16, is compatible with the marked basal transcriptional activity of this isoform in transfection experiments. However, it is intriguing that the activity of this AF-1 region is modulated by the phosphorylation of two serine residues, both of which must be phosphorylated in order to activate transcription. This is in contrast to PPARγ2, which was previously shown to be phosphorylated at a single site in a motif that is not homologous to the sites now described in PPARα. Although the molecular details involved in the phosphorylation-dependent enhancement of the transcriptional activity of PPARα remain to be elucidated, we demonstrate that the effect of insulin on the AF-1 region of PPARα can be mimicked by the addition of triiodothyronine receptor β1, a strong binder of corepressor proteins. In addition, a triiodothyronine receptor β1 mutant deficient in interacting with corepressors is unable to activate PPARα. These observations suggest that the AF-1 region of PPARα is partially silenced by corepressor proteins, which might interact in a phosphorylation-dependent manner.

The peroxisome proliferator-activated receptors (PPARs) 1 are nuclear hormone receptors, the transcriptional activity of which is mostly thought to be regulated by the binding of ligand, such as certain fatty acids and eicosanoids (1,2). However, the activity of some nuclear receptors is also modulated by phosphorylation, usually resulting in transcriptional enhancement, e.g. through the increased binding to ligand or DNA (3)(4)(5)(6)(7)(8)(9). Similarly, PPARs were previously shown to be phosphorylated, resulting in altered transcriptional activity depending on the isoform and cellular systems examined (10 -14). Whereas the transcriptional capacity of the mostly adipocyte-specific PPAR␥ isoform is inhibited in certain cell types by the mitogen-activated protein kinase (MAP-K)-dependent phosphorylation of a serine residue in its amino terminus (Ser-112) (11,14), we have shown recently that PPAR␣ is phosphorylated in response to insulin, resulting in an enhanced transcriptional response (10). However, it remained unclear whether this represented a direct effect of insulin and by which mechanism insulin stimulated the activity of PPAR␣.
The PPAR␣ isoform is highly expressed in liver and brown adipose tissue, both of which are target tissues for insulin action (15). 2 PPAR␣ binds and is activated by fibrates, certain fatty acids, arachidonic acid analogs (e.g. 5,8,11,14-eicosatetraenoic acid), and leukotriene B4 (2,15). However, in transient transfection systems, PPAR␣ exhibits strong transcriptional activity even in the absence of exogenous ligands (16,17). Although this might be due to the presence of an endogenous stimulator, this activity could also be mediated through a constitutive trans-activating domain, similar to the activation function (AF)-1 region present in some other members of the nuclear receptor family, such as the retinoic acid and estrogen receptors (18 -20). Because we have previously demonstrated that PPAR␣ is phosphorylated in response to insulin, resulting in the stimulation of basal as well as ligand-dependent transcriptional activity, we hypothesized that PPAR␣ might contain a phosphorylation-regulated trans-activation domain (10).
In the present study, we examined the hypothesis that PPAR␣ contains a ligand-independent transcriptional activation domain that might be subject to regulation by phosphorylation. We describe a novel AF-1 function within the A/B domain of PPAR␣, which harbors two consensus MAP-K sites, which are phosphorylated in response to insulin and both of which are necessary to mediate the insulin-induced transcriptional acti-vation of PPAR␣. Intriguingly, these MAP-K sites are distinct from the inhibitory MAP-K site described for PPAR␥. Moreover, our results demonstrate that the transcriptional activation of PPAR␣ by insulin may involve the dissociation from co-repressor proteins, which may be related to the nuclear receptor co-repressor (NCoR) and the silencing mediator for RAR and TR (SMRT) (21,22). Hence, the insulin-induced phosphorylation of PPAR␣ and ␥ is mediated through distinct phosphorylation motifs, resulting in transcriptional activation or repression, respectively.

MATERIALS AND METHODS
Plasmid Constructs and Mutagenesis-A fusion construct linking the GAL4-DNA-binding domain (amino acids 1-147) upstream of the amino-terminal region (amino acids 1-92) of hPPAR␣ was constructed by cloning the cDNA corresponding to the first 92 amino acids of hPPAR␣ amplified by polymerase chain reaction using the Pfu DNA polymerase (Life Technologies, Inc.) into the BamHI and PstI sites of the pM-GAL4-BD expression plasmid (CLONTECH, Stehelin AG, Basel, Switzerland), resulting in the vector pM-GAL4BD-hPPAR␣. The expression plasmid pSG5-hPPAR␣ was kindly provided by Dr. F. Gonzalez (NCI, National Institutes of Health, Bethesda, MD) (16). The final construct was verified by automated sequencing (ABI 373, Perkin-Elmer).
Site-directed mutagenesis was performed by polymerase chain reaction amplification of the entire pM-GAL4BD-hPPAR␣1-92 or pSG5-hPPAR␣ plasmids with Pfu DNA using sense and antisense primers containing the desired mutation: S77A mutation (AGC 3 GCC), S76A (TCG 3 GCG), S12A (TCC 3 GCC), S21A (AGC 3 GCC) according to the manufacturer's protocol (QuickChange site-directed mutagenesis, Stratagene, Basel, Switzerland). The presence of the desired mutation and the absence of spurious mutations in the amplified cDNA was assessed by automated sequencing.
Transfection Experiments-The human HepG2 hepatoma cell line was cotransfected using the calcium phosphate method as described previously with 4.2 g of the expression plasmid for the parental vector pM-GAL4BD or the fusion protein pM-GAL4BD-hPPAR␣1-92, together with 0.42 g of the reporter plasmid pG5-CAT (17,24). Twenty-four hours before cell harvesting, cells were treated with 1 M insulin and/or 30 M of the MAP-K p42/p44 inhibitor PD98059 (Alexis Corp., Lä ufelfingen, Switzerland) dissolved as described in the manufacturer instructions. CAT activity was determined as described by determining the incorporation of [ 14 C]chloramphenicol (Hartmann Analytic, Braunschweig, Germany) into acetyl-CoA as analyzed by thin layer chromatography (24). The results were quantitated by a PhosphorImager (Molecular Dynamics), and CAT activity was normalized to the protein concentration as measured by the Bradford method (Bio-Rad).
Western Blotting-As described above, HepG2 cells were transfected with 4.6 g of pM-GAL4BD or the appropriate pM-GAL4BD-hPPAR␣ constructs. After 48 h, cells were lysed in radioimmune precipitation buffer, and the fusion proteins were immunoprecipitated by using 1 g of rabbit polyclonal anti-GAL4BD antibody. After electrophoresis, the proteins were electrotransferred to a 0.45-m nitrocellulose membrane (Bio-Rad) for 45 min at 100 V (4°C) in 25 mM Tris base, 190 mM glycine, 20% (v/v) methanol. Nonspecific binding was blocked by incubating the membrane for 2 h in 5% powdered milk in TBST (20 mM Tris, 150 mM NaCl, 0.1% Tween, pH 7.6). A monoclonal anti-GAL4BD antibody (CLONTECH, Stehelin AG, Basel, Switzerland) was added to a solution of 1% powdered milk/TBST for 12 h. The membrane was washed twice with TBST, followed by five rinses of 15 min each in TBST. The second antibody (anti-mouse IgG horseradish peroxidase (Santa-Cruz Biotechnology, Inc., Dr. Glaser AG, Basel, Switzerland)) was diluted (1:1000 v/v) and was added for 2 h at 4°C. After incubation, the membrane was washed as described above. Immunolabeled proteins were revealed by chemiluminescence as described by the manufacturer (Amersham Pharmacia Biotech, Zurich, Switzerland).

The Amino Terminus of hPPAR␣ (amino acids 1-92) Contains a Ligand-independent Activation Domain (AF-1), Which
Is Regulated by Phosphorylation-In order to explore the possibility that the previously described significant constitutive activity of the full-length hPPAR␣ is due to an AF-1-like function in the amino terminus of this protein, we created a fusion construct linking the GAL4 DNA-binding domain to the first 92 amino acids of hPPAR␣ (pM-GAL4BD-hPPAR␣). The transcriptional response of this construct was assessed in transient transfection experiments in HepG2 cells using a UAS-CAT reporter (pG5-CAT). Fig. 1 shows that the addition of the first 92 amino acids of hPPAR␣ to the GAL4BD enhances the constitutive transcriptional activity over 20-fold, as compared with an 80-fold stimulation by the known strong transcriptional activator VP16.
To examine whether the previously described stimulation of the basal transcriptional activity of PPAR␣ by insulin might involve this novel AF-1 domain, similar experiments were performed in the presence of 1 M insulin for 24 h. As shown in Fig.  1, treatment with insulin enhances the activity of the amino terminus of PPAR␣ over 4-fold, whereas no effect was seen on the control constructs GAL4BD and VP16. Because we have shown recently that insulin rapidly stimulates the phosphorylation of PPAR␣ in adipose tissue and transfected cells, we also assessed the effect of insulin on PPAR␣ in the presence of the specific inhibitor of the p42/44 MAP-K pathway PD98059 (25). Compatible with the hypothesis that the stimulation of the amino terminus of PPAR␣ by insulin involves a MAP-K-mediated phosphorylation event, PD95059 completely abolished this effect (Fig. 1).
The Amino Terminus of hPPAR␣ Contains Several Putative MAP-K Sites-Because the results described above suggest that insulin modulates the basal transcriptional activity of PPAR␣ through the phosphorylation of a MAP-K site in its amino terminus, we examined the known sequences of PPAR␣ from various species for conserved MAP-K motifs (PX n SP) (Fig.  2). Whereas several MAP-K motifs are present within the first 92 amino acids of hPPAR␣, three putative sites were highly conserved across evolution: serine 12, serine 21, and serine 77. Interestingly, the latter residue corresponds to a MAP-K motif described in mPPAR␥2, which has been shown to be phosphorylated by MAP-K and results in an inhibition of transcription. Hence, we created the corresponding mutants of the pM-GAL4BD-hPPAR␣ construct for the functional analysis of these sites (pM-GAL4BD-S12A, pM-GAL4BD-S21A, and pM-GAL4BD-S76A/S77A).
Serines 12 and 21, but Not 76/77, Mediate the Insulin Stimulation of the AF-1 Domain of PPAR␣-Using the various mutants of the putative MAP-K sites within the amino terminus of PPAR␣, we performed transient transfection experiments in HepG2 cells to assess their functional relevance. Fig. 3 shows that the serine to alanine mutations at positions 12 and 21 were both able to abolish the stimulatory effect of insulin, whereas similar mutations of another putative MAP-K site (Ser-76/77) did not alter the fold transcriptional stimulation by insulin. The lower transcriptional activity of the S76A/S77A mutant correlates with its lower expression as assessed in Western blot experiments (data not shown).
These results are thus compatible with a model in which insulin activates the AF-1 region of PPAR␣ through the MAP-K-mediated phosphorylation of serines 12 and 21.
Serines 12 and 21 of hPPAR␣ Are Phosphorylated in Response to Insulin-In order to test the hypothesis that insulin mediates the phosphorylation of serine residues 12 and 21, we performed phosphorylation experiments in HepG2 cells transfected with the appropriate wild-type and mutant pM-GAL4BD-PPAR␣ constructs. Shown in Fig. 4 are the phosphorylated fusion proteins after precipitation with an anti-GAL4BD antibody, resulting in a main band at 39 kDa, corresponding to the apparent molecular mass of the GAL4BD-PPAR␣ fusion protein (CLONTECH document 5399 -1), as well as a smaller band, which is likely to correspond to a proteolytic fragment. These experiments demonstrated a 2.8-fold increased phosphorylation of wild-type GAL4BD-PPAR␣ in response to insulin, which is quantitatively comparable to the 4.5-fold enhancement in trans-activation observed above. However, whereas the single mutants S12A and S21A were still phosphorylated by insulin, the double mutant S12A/S21A did not show an in-crease in its phosphorylation status after treatment with insulin. These data demonstrate that serines 12 and 21 are both phosphorylated by insulin and that no other residues contribute to the insulin-enhanced phosphorylation of the amino terminus of PPAR␣.
Expression of GAL4BD-PPAR␣ and Its Mutants at the Protein Level-In order to ascertain that the absent response of the GAL4BD-PPAR␣-S12A and S21A mutants was not due to the altered expression of these proteins, we performed Western blot analyses with an anti-GAL4BD antibody. As shown in Fig.  5, the S12A and S12A/S21A mutants were expressed at similar levels as the wild-type protein, whereas the S21A mutant was somewhat less abundant.
Serines 12 and 21 Are Required for the Insulin-mediated Increase in Transcriptional Activity of the Full-length PPAR␣-In order to examine the role of the Ser-12 and Ser-21 residues in the context of the intact hPPAR␣, both serines were mutated to alanine (pSG5-hPPAR␣-S12A/S21A). As shown in Fig. 6, the wild-type intact hPPAR␣ was activated by insulin in a dose-dependent manner (p ϭ 0.03, one-way analysis of variance), whereas the double-mutant S12A/S21A exhibited only a minimal and statistically not significant response to insulin (p Ͼ 0.2, one-way analysis of variance).

Phosphorylation of the AF-1 Region of PPAR␣ May Result in the Dissociation of Co-repressor Proteins-Multiple scenarios
can possibly explain how the phosphorylation of serines 12 and 21 activates transcription, such as the recruitment or dissociation of putative coactivator or corepressor proteins. Therefore, we hypothesized specifically that the phosphorylation event might result in the dissociation of a corepressor protein, such as NCoR or SMRT, thereby resulting in transcriptional derepression and hence activation. The testable prediction of this model is that the addition of a strong binder of such co-repressor proteins should exert an insulin-like stimulation on the AF-1 region of PPAR␣. In order to examine this hypothesis, we transfected HepG2 cells with GAL4BD-PPAR␣ (amino acids 1-92) either alone or in the presence of increasing amounts of

FIG. 3. Transcriptional activity of various mutants of the putative MAP-K sites in the amino-terminal region of PPAR␣.
HepG2 cells were transiently transfected with the pG5-CAT reporter gene together with either wild-type pM-GAL4BD-hPPAR␣ (1-92) or the mutated constructs containing the mutations S12A, S21A, or S76A/ S77A. Cells were treated with either vehicle or insulin (1 M) 24 h prior harvesting. Transcriptional activity is expressed as relative CAT activity, normalized to protein content. Values are the mean Ϯ S.E. of three separate experiments, each performed in triplicate. TR␤1, which is known to strongly interact with corepressor proteins (21,22). Interestingly, the addition of TR␤1 was able to increase the transcriptional activity of the AF-1 domain of PPAR␣ in a dose-dependent manner and to a level corresponding to that observed with insulin (Fig. 7A). However, to exclude the possibility that TR␤1 might squelch other proteins nonspecifically, a similar experiment was performed with a mutant TR␤1, known to be deficient in its interaction with corepressor proteins, such as NCoR and SMRT (TR␤1-⌬NCoR box). As shown in Fig. 7B, this mutant has lost its capacity to activate PPAR␣, compatible with the model that TR␤1 competes for corepressor proteins, resulting in a derepression of the AF-1 domain of PPAR␣.

DISCUSSION
The human PPAR␣ was previously shown to exhibit significant ligand-independent transcriptional activity, which can be modulated by insulin (10). The observation that insulin also enhances the phosphorylation of PPAR in adipocytes as well as in transfected cells raised the possibility that insulin might directly phosphorylate PPAR␣, thereby enhancing its transcriptional capacity. In the present paper, (i) we demonstrate that the amino-terminal 92 amino acids of PPAR␣ contain a AF-1-like trans-activation domain, which is further activated by insulin, (ii) we demonstrate that the effect of insulin on the AF-1 region is dependent on MAP-K p42/44 and involves the phosphorylation of two serine residues at positions 12 and 21, and (iii) we present data indicating that the phosphorylation of the amino terminus of PPAR␣ may result in the dissociation of corepressor proteins, which may then result in a further transcriptional enhancement of this domain.
The concept that nuclear hormone receptors contain a ligand-independent (AF-1) and a ligand-dependent (AF-2) transactivation domain has been well demonstrated for the steroid and retinoid receptors (18,20,26). However, whereas the activity of the AF-2 domain is regulated by the binding of ligand resulting in the recruitment of coactivator proteins, the AF-1 region is conventionally thought to be constitutively active by molecular mechanisms that are poorly understood (27). The presence of a strong AF-1 region in the A/B domains of hPPAR␣ shown in the present paper explains, at least in part, the basal activity observed with this PPAR isoform in transient transfection experiments. Interestingly, the present data demonstrate that this activity is subject to regulation through phosphorylation of two MAP-K sites. Because we have previously shown that PPAR␣ is a phosphoprotein in primary adipocyte cultures stimulated with insulin, it is quite likely that the FIG. 4. Effect of insulin on the level of phophorylation of the GAL4BD-hPPAR␣ (1-92) wild type (WT) and the mutated constructs S12A, S21A, and S12A/S21A. HepG2 cells were transfected with pM-GAL4BD, pM-GAL4BD-hPPAR␣ (1-92) wild-type, or the appropriate mutants. Twenty-four hours before harvesting, cells were placed in serum-free and phosphate-free medium, followed by labeling with [ 32 P]orthophosphate for 2 h, and stimulated with insulin (1 M) for 20 min. After cell lysis, fusion proteins were immunoprecipitated with anti-GAL4BD antibodies and submitted to SDS-polyacrylamide gel electrophoresis. Dried gels were autoradiographed and exposed on a PhosphorImager.

FIG. 5. Level of expression of the wild-type and mutant GAL4BD-hPPAR␣ constructs assessed by Western blotting.
HepG2 cells were transfected with pM-GAL4BD, or the wild-type or mutant pM-GAL4BD-hPPAR␣ (amino acids 1-92) constructs. After harvesting and lysis, the fusion proteins were immunoprecipitated with a polyclonal anti-GAL4BD antibody and submitted to SDS-polyacrylamide gel electrophoresis. Proteins were electrotransferred to a nitrocellulose membrane and blotted with a monoclonal anti-GAL4BD antibodies and revealed with an anti-mouse IgG-horseradish peroxidase antibody. Immunolabeled proteins were visualized by chemiluminescence.
FIG. 6. Effect of insulin on the transcriptional activity of intact wild-type hPPAR␣ and the mutated hPPAR␣-S12A/S21A. HepG2 cells were transfected with either pSG5-hPPAR␣ or pSG5-hPPAR␣-S12A/S21A together with the reporter construct pBL-1xPPRE-MEp-CAT8ϩ. After transfection, the cells were exposed to vehicle or insulin (2 or 10 M) for 24 h. Transcriptional activity is expressed as relative CAT activity, normalized to protein content. Values are the mean Ϯ S.E. (n ϭ 3). The insulin dose-response curves were analyzed by oneway analysis of variance. Whereas insulin had a significant effect on the transcriptional activity of wild-type PPAR␣ (p ϭ 0.03), the response of the S12A/S21A mutant was not statistically significant.
cross-talk between these two signaling pathways is of functional relevance. This notion is supported by our observation that the full-length PPAR␣ is also activated by insulin and that this ligand-independent trans-activation requires serines 12 and 21, compatible with the results obtained with the GAL4-PPAR␣ 1-92 fusion construct. Although not statistically significant, the full-length S12A/S21A-PPAR␣ retains a minimal residual response to insulin. However, this small effect is not mediated by the amino-terminal 92 amino acids, as supported by experiments with a PPAR␣ deletion mutant lacking the amino terminus (data not shown). This residual stimulation by insulin may thus be mediated either through other regions of PPAR␣ or through insulin effects on, for example, co-activators or the general transcriptional machinery. It is also of interest to note that treatment of primary hepatocytes with insulin for 3 days decreases the expression of the PPAR␣ gene, suggesting that short-term exposure to insulin would increase trans-acti-vation by PPAR␣ (via the phosphorylation of PPAR␣), whereas a longer exposition to this hormone might result in the downregulation of PPAR␣ activity in liver (via the decreased expression of PPAR␣ mRNA) (28).
In contrast to PPAR␣, the PPAR␥2 isoform, which was recently shown to be phosphorylated in its amino terminus in response to activators of the MAP-K pathway, is inhibited after phosphorylation of a serine residue at position 112 as assessed by its adipogenic potential (11). Nevertheless, when examined in different experimental systems or in the context of GAL4fusion constructs, the amino terminus of PPAR␥ can also be stimulated by phosphorylation. Hence, it appears that at least in adipose tissue, which coexpresses the PPAR␣ and ␥ isoforms at high levels, the phosphorylation by insulin has opposite effects on PPAR␣ and ␥, the former being activated, whereas the latter is inhibited. This observation, together with their different ligand binding spectrum and differences in DNA binding, provides a novel mechanism for generating isoform-specific responses of PPAR (17).
The phosphorylation sites in PPAR␣ are highly conserved across species, and they are distinct from the homologous MAP-K site shown to be phosphorylated in ␥2 (see Fig. 2). Hence, the functionally opposite effects of insulin on PPAR␣ and ␥2 are reflected by the presence of distinct phosphorylation sites. The mechanisms mediating phosphorylation-dependent alterations in transcription rates are still poorly understood, as is the case for the PPARs. Because the AF-1 domain as well as the insulin-stimulated phosphorylation sites colocalize within the same domain of PPAR␣, it is tempting to speculate that the altered phosphorylation results in the differential recruitment of coactivator and/or corepressor proteins. However, the currently known coactivator proteins interact preferentially in a ligand-dependent manner with the AF-2 domain of nuclear receptors, although some cooperation with AF-1 may occur (29,30). However, the corepressor proteins, such as NCoR and SMRT, interact with different regions of the nuclear receptors in a ligand-independent manner. Therefore, our finding that TR␤1 can activate the amino terminus of PPAR␣ is of interest, because the TRs are known to strongly interact with NCoR and SMRT, thereby resulting in a putative intranuclear sink for corepressor proteins. Because this effect is abolished by a mutation in the NCoR box of TR␤1, it is tempting to speculate that the amino terminus of PPAR␣ interacts with a NCoR-like corepressor protein that is able to partially silence the activity of AF-1. However, whether the phosphorylation of serines 12 and 21 derepresses AF-1 by inducing the dissociation of corepressors is a matter of conjecture. Although PPAR␥ was shown to interact with NCoR in solution, this interaction involved the hinge region of PPAR, rather than the amino terminus (31). Nevertheless, some recent data demonstrate that an anti-NCoR antibody blocks the transcriptional repression of PPAR␥ by insulin, suggesting at least indirectly a functional interaction between the phosphorylated amino terminus of PPAR␥ and NCoR (32). Preliminary experiments utilizing gel mobility shift and glutathione S-transferase pull-down assays examining the possibility of direct interactions of NCoR with the amino terminus of PPAR␣ were inconclusive. 3 These findings indicate that either the interactions between the AF-1 region of PPAR␣ and NCoR are weak under the conditions used, or alternatively, novel corepressor proteins are involved, which have the potential to interact with the AF-1 region as well as the conventional NCoR box motif of nuclear receptors.
In summary, we have demonstrated that PPAR␣ contains a strong ligand-independent AF-1 domain, which can be further 3 C. Juge and C. A. Meier, unpublished data. FIG. 7. TR␤1, but not TR␤1-⌬NcoR box, mimics the effect of insulin on the transcriptional activity of GAL4BD-hPPAR␣ (amino acids 1-92). HepG2 cells were transiently transfected with the pG5-CAT reporter gene and either pM-GAL4BD or pM-GAL4BD-hPPAR␣ (1-92), with or without wild-type TR␤1 or TR␤1-⌬NcoR box. Transcriptional activity is expressed as relative CAT activity, normalized to protein content. Values are the mean Ϯ S.E. of a representative experiment performed in triplicate. activated through the MAP-K-mediated phosphorylation of two serine residues, which are distinct from the inhibitory phosphorylation site present in PPAR␥2. In addition, the AF-1 region of PPAR␣ can be equally well activated by squelching corepressor proteins, suggesting that such proteins might silence the amino terminus of PPAR␣. It can thus be speculated that the phosphorylation of the AF-1 domain of PPAR␣ might result in the dissociation of such corepressor proteins, thereby resulting in transcriptional activation.