The Peroxisome Proliferator-activated Receptor Regulates Expression of the Perilipin Gene in Adipocytes*

Recent studies have shown that lipid droplets are covered with a proteinaceous coat, although the functions and identities of the component proteins have not yet been well elucidated. The first identified lipid dropletspecific proteins are the perilipins, a family of proteins coating the surfaces of lipid droplets of adipocytes. The generation of perilipin-null mice has revealed that although they consume more food than control mice, they have normal body weight and are resistant to diet-induced obesity. In one study (Martinez-Botas, J., Anderson, J. B., Tessier, D., Lapillonne, A., Chang, B. H. J., Quast, M. J., Gorenstein, D., Chen, K. H., and Chan, L. (2000) Nat. Genet. 26, 474–479) it was reported that in an animal model obesity was reversible by breeding perilipin / alleles into Lepr db/db obese mice, ostensibly by increasing the metabolic rate of the mice. To understand the exact mechanisms that drive the exclusive expression of the perilipin gene in adipocytes, we analyzed the 5 -flanking region of the mouse gene. Treatment of differentiating 3T3-L1 adipocytes with an agonist of proliferator-activated receptor (PPAR) , the putative “master regulator” of adipocyte differentiation, significantly augmented perilipin gene expression. Reporter assays using the 2.0-kb promoter revealed that this region contains a functional PPAR -responsive element. Gel mobility shift and chromatin immunoprecipitation assays showed that endogenous PPAR protein binds to the perilipin promoter. PPAR 2, an isoform exclusively expressed in adipocytes, was found to be the most potent regulator from among the PPAR family members including PPAR and PPAR 1. These results make evident the fact that perilipin gene expression in differentiating adipocytes is crucially regulated by PPAR 2, providing new insights into the adipogenic action of PPAR 2 and adipose-specific gene expression, as well as potential anti-obesity pharmaceutical agents targeted to a reduction of the perilipin gene product.

Adipocytes are the major reservoir of energy stored in the form of triacylglycerol in the body. Triacylglycerol is stored within intracellular lipid droplets covered by a monolayer of phospholipids, free cholesterol, and proteins. In times of energy need, for example, as brought about by activities such as fast-ing and/or exercise, catecholamines rapidly activate cAMP-dependent protein kinase (PKA), 1 resulting in hydrolysis of triacylglycerol as catalyzed by activated hormone-sensitive lipase. The perilipins are a class of proteins found exclusively at the surface of lipid droplets in adipocytes and steroidogenic cells (1,2) and have been suggested to function as regulators of lipolysis, perhaps by acting as a barrier to lipase access (3,4). Upon chatecholamine stimulation, however, perilipins are hyperphosphorylated by PKA, and this hyperphosphorylation has been suggested to facilitate lipase access to lipid droplets (5,6).
Recently two reports have revealed that perilipin knockout mice have markedly reduced adipose tissue mass and constitutively high levels of basal lipolysis taking place in their isolated adipose cells (7,8). When fed a high fat diet, these mice are resistant to diet-induced obesity, in sharp contrast to normal mice. It has also been shown that in the absence of perilipin, the lipid droplets in adipose cells become coated with adipose differentiation-related protein, which is not phosphorylated by PKA, indicating that this related protein is not substitutable for perilipin as a protective barrier against lipolysis (8). Moreover, ectopic expression of perilipin in 3T3-L1 cells resulted in lipid droplets becoming perilipin-coated and increased the half-life of stored triacylglycerol deposits (9). These findings indicate that perilipins are required to maximize the storage of triacylglycerols in adipose tissue and serve the additional crucial function of controlling their release at times of energy demand.
Multiple perilipin isoforms are encoded by a single copy of a gene that gives rise to multiple mRNAs by alternative splicing mechanisms (2). The 5Ј end segments of the mRNAs of all of these isoforms are identical; therefore they are highly likely to be transcribed from the same promoter on the perilipin gene. In the present study, we undertook the characterization of the 5Ј-flanking region of the mouse perilipin gene and examined how it is expressed exclusively in adipocytes, its apparently exclusive site of activity. Treatment of 3T3-L1 adipocytes with an agonist for PPAR␥, typically regarded as "a master regulator of adipocyte differentiation," stimulates perilipin gene expression. Reporter assays reveal that the 5Ј-flanking region of the gene contains a functional PPAR␥-responsive element and that PPAR␥2, an isoform exclusively expressed in adipocytes, is a predominant regulator of the expression of this gene.
Plasmid Constructs-Expression plasmids for mouse PPAR␥1 (11), mouse PPAR␣ (11), human RXR␣ (12) human SREBP-1a (13), and rat C/EBP␣ (14) were described previously. An expression plasmid for mouse PPAR␥2 was made by inserting a fragment coding amino acids 1-169 of mouse PPAR␥2 into the HindIII sites of pCMX-PPAR␥1. To generate a perilipin promoter-reporter plasmid (pPlin-2.0), the BamHI-HindIII PCR fragments from mouse genomic DNA were ligated to the BglII-HindIII sites of a pGL3 basic vector (Promega). Mutant versions of perilipin promoter-reporter plasmids were constructed by a PCRassisted method using a site-directed mutagenesis kit following the instructions provided by the supplier (Stratagene).
Northern Blot Analysis-Total RNA was isolated using an RNA preparation kit (Isogen; Nippon Gene Corp.). The RNA was fractionated by electrophoresis in an 1% formaldehyde-agarose gel and transferred to nylon membranes (Hybond-N; Amersham Biosciences). Probes for perilipin, PPAR␥, aP2, human LPL, and 36B4 (15) were labeled with [␥-32 P]dCTP (3000 Ci/mmol; Amersham Biosciences) using a randomprimed DNA labeling kit (Megaprime DNA labeling system; Amersham Biosciences). The membrane was hybridized with radioactive cDNA probes, and the signals on the membrane were quantified using an image-analyzing system (FLA-3000; Fuji Film Inc.).
Luciferase Assay-3T3-L1 adipocytes cultured in a 12-well plate were transfected with 2 g of a reporter plasmid and 0.02 g of pRL-TK (Promega) complexed with LipofectAMINE TM (Invitrogen) according to the manufacturer's instructions. 3T3-L1 preadipocytes cultured in a 12-well plate were transfected with 1 g of a reporter plasmid, 0.01 g of pRL-TK and expression plasmids complexed with LipofectAMINE TM . The cells were incubated with a medium containing 10% charcoalstripped FBS and 10 M pioglitazone. Twenty-four hours later both firefly and Renilla luciferase activities were quantified using a Dual-Luciferase TM reporter system (Promega) according to the manufacturer's instructions (16,17). HEK293 cells were cultured and transfected as described previously (13,16).
Lipid Analysis-3T3-L1 cells were washed with cold phosphate-buffered saline, and lipids were extracted by chloroform/methanol (2:1, v/v). The lower organic phase was dried, and the lipids were dissolved in 2-propanol. Triacylglycerol content was determined using Triglyceride E-test Wako (Wako, Japan) according to the manufacturer's instructions.
Western Blot Analysis-3T3-L1 preadipocytes were differentiated into adipocytes by the addition of differentiation medium containing 10 M pioglitazone at day 0 and were cultured for 4 days. Total cellular proteins were fractionated by SDS-10% PAGE. Western blot analysis was carried out using anti-human perilipin or anti-mouse ␤-actin antibodies with ECL (Amersham Biosciences). The signals on the membrane were quantified with a LuminoImager (LAS-3000; Fuji Film Inc.).

A PPAR␥ Ligand Enhances Perilipin Gene Expression in
Adipocytes- Fig. 1A shows that perilipin gene expression was low to almost the zero point level in preadipocytes but was induced markedly at later stages of differentiation. This perilipin expression pattern was quite similar to those for PPAR␥, known to be a central regulator of adipocyte differentiation (18), and aP2, a target gene of PPAR␥. Therefore, we hypothesized that PPAR␥ participates in the induced expression of the perilipin gene during differentiation. As expected, treatment of 3T3-L1 cells for 48 h with a PPAR␥ agonist, pioglitazone, raised expression of the mRNA for perilipin and LPL, another PPAR␥ gene target (20), but only in adipocytes (Fig. 1B). This suggests that PPAR␥, which is highly expressed after adipocyte differ- entiation, is involved in perilipin gene expression.
When pioglitazone was added to the culture medium at an early stage of differentiation, the enhancing effect of the drug was significant (Fig. 2, day 2). In contrast, as a result of treatment of the cells with a PPAR␥ antagonist, GW9662, perilipin gene expression was inhibited for 4 days. Because a longer treatment with pioglitazone gradually suppressed PPAR␥ gene expression (21), the maximal drug effect was observed at day 2, and no effect was seen after day 6 (data not shown). Fig. 2B clearly shows that the perilipin protein level was significantly increased by treatment with pioglitazone on day 4. The increase in perilipin expression resulted in the higher intracellular accumulation of triacylglycerol (Fig. 2C), consistent with the finding of an increased number of lipid droplets in the cells in the presence of pioglitazone (data not shown).
PPAR␥ Directly Regulates the Transcription of the Perilipin Gene-To determine whether PPAR␥ directly regulates the transcription of the perilipin gene, we carried out luciferase assays with reporter genes, including various deletion versions of the mouse perilipin 5Ј-flanking region (Fig. 3A). 3T3-L1 preadipocyte and HEK293 cells were co-transfected with one of the reporter genes and expression plasmids for PPAR␥2 and RXR␣. The cells were treated with pioglitazone for the last 24 h. Fig. 3B clearly shows that forced expression of PPAR␥ in both adipogenic and nonadipogenic cells enhanced the luciferase activities ϳ40-fold as long as the reporter gene containing the segment around Ϫ2.0-kb was present (the transcription start site is position ϩ1). Deletion of ϳ100 and ϳ300 bp (pPlin-1.9 and pPlin-1.7, respectively) dramatically reduced this effect. We obtained the same results using nonmammalian cells, Drosophila SL2 cells (data not shown). It is likely that the 5and 10-fold induction of the promoter activity of pPlin-1.9 observed in 3T3-L1 preadipocyte and HEK293 cells, respectively, is largely attributable to overexpression of PPAR␥2 and RXR␣. We cannot rule out the possibility that the 1.9-kb promoter contains several weak PPAR␥ responsive elements. The luciferase activities of pPlin-2.0 were greatly stimulated only when both PPAR␥2 and RXR␣ were expressed in 3T3-L1 preadipocyte cells in the presence of the agonist (Fig. 3C). When only RXR␣ was expressed, no effect was observed, suggesting that the pioglitazone-mediated effect is not caused by a combination of RXR and some nuclear receptors, other than PPAR␥. The luciferase activities of pPlin-1.9 were slightly augmented by PPAR␥2/RXR␣. Next we performed further luciferase assays using pPlin-2.0 and pPlin-1.9 in differentiated 3T3-L1 adipocytes. Endogenous PPAR␥ activated by pioglitazone markedly enhanced the luciferase activities of pPlin-2.0, whereas deletion of ϳ100 bp completely abolished this effect (Fig. 3D). Even when PPAR␥2/RXR␣ were overexpressed in 3T3-L1 adipocytes, the Plin-1.9 promoter did not respond to pioglitazone. These results demonstrate that the 5Ј-flanking region of mouse peripilin gene contains PPRE(s), which responds to endogenous PPAR␥, located in a region of approximately Ϫ2.0-kb.
The Element Responsible for the PPAR␥-mediated Transcriptional Regulation of the Perilipin Gene-We identified two putative PPREs of approximately Ϫ2.0-kb (PPRE and PPRE-like in Fig. 3A). To confirm that these motifs are recognized by a PPAR␥/RXR␣ heterodimer, gel mobility shift assays were performed with recombinant PPAR␥/RXR␣. A single-shifted DNAprotein complex was observed in the presence of a control 32 P-labeled fragment corresponding to the PPRE of the aP2 gene and recombinant PPAR␥/RXR␣ (Fig. 4A, lane 2) and gradually disappeared after the addition of increasing amounts of the nonlabeled PPRE DNA fragment but not the PPRE-like fragment. Fig. 4B shows that a single-shifted perilipin PPRE-PPAR␥/RXR␣ complex was formed (lane 2). The band almost completely disappeared in the presence of an excess amount of an unlabeled wild type fragment but not a mutant fragment (lanes 3-6). These results clearly demonstrate that PPAR␥ is capable of binding to the TCACCTTTCACCC sequence in the perilipin promoter. As shown in Fig. 4C, the identical two repeats in the PPRE are conserved in the promoters of both the mouse and human genes. To determine whether this PPRE is involved in the PPAR␥-mediated regulation of perilipin gene expression, luciferase assays using wild type and mutant versions of reporter genes were carried out. Mutation of the PPRE sequence (Ϫ1977 to Ϫ1983 bp) resulted in a remarkable suppression of the PPAR␥-dependent induction of luciferase activities in preadipocyte cells (Fig. 4D). These results provide evidence that the PPRE motif (Ϫ1974 to Ϫ1986 bp) is crucially responsible for the transcriptional regulation of the perilipin gene during adipocyte differentiation.
To investigate whether PPAR␥ binds to the endogenous perilipin promoter, we performed chromatin immunoprecipitation. As shown in Fig. 5, endogenous PPAR␥ protein bound to the promoter of the perilipin gene as well as that of another PPAR␥ target, the aP2 gene. These data show that PPAR␥-mediated expression of the perilipin gene involves the binding of endogenous PPAR␥ to the PPRE in chromatin.

FIG. 2. Effect of pioglitazoneduring 3T3-L1 adipocyte differentiation.
A, 3T3-L1 preadipocytes were differentiated into adipocytes by the addition of a differentiation medium containing 10 M pioglitazone (PGZ) or 10 M GW9662 at day 0 and were cultured for 2 or 4 days. Total RNA (20 g/lane) was subjected to electrophoresis and blot hybridization with the indicated 32 P-labeled probe. B, 3T3-L1 preadipocytes were differentiated into adipocytes by the addition of differentiation medium containing 10 M pioglitazone at day 0 and were cultured for 4 days. Aliquots of total cellular proteins (20 g/lane) were subjected to SDS-PAGE and Western blotting using anti-perilipin and anti-␤actin antibodies. The results were normalized to the signal generated from ␤-actin. C, cellular lipids were extracted at the indicated times after treatment with a differentiation medium containing 10 M pioglitazone, and the triacylglycerol (TG) contents were quantified. The data are shown as the means Ϯ S.D.

PPAR␥2 Is a Major Regulator for Perilipin Gene Expression
in Adipocytes-To verify that PPAR␥ is a major regulator for perilipin gene expression in adipocyte cells, we established a stable NIH-3T3 cell line (NIH-␥2) expressing PPAR␥2, an isoform expressed exclusively in adipose tissue (10). When these PPAR␥2-expressing cells were cultured in a differentiation medium, more than 60% of the cells were stained with Oil Red O, whereas no lipid-containing cells were observed among the control NIH-vector cells (10). These two stable (NIH-vec and NIH-␥2) cells were cultured with a normal medium containing 10 M pioglitazone for 48h, and perilipin gene expression was analyzed by the reverse transcription-PCR method (Fig. 6A). Expression of adipogenic marker aP2, a PPAR␥-responsive gene, was markedly induced, but perilipin gene expression was only slightly enhanced in NIH-␥2 cells. Indeed, we performed an reverse transcription-PCR analysis to confirm perilipin gene expression after we failed to observe it in Northern blot analysis. Although perilipin expression in NIH-␥2 cells was induced at the detectable level during adipocyte differentiation, the expression level was much lower than that in differentiated 3T3-L1 adipocytes, probably because of an attenuated expression of the endogenous PPAR␥ gene (Fig. 6B).
Next we determined which isoform of PPAR␥ preferentially regulates the perilipin promoter activity. 3T3-L1 preadipocyte cells transfected with pPlin-2.0 and one of the expression plasmids for PPAR␥1 and ␥2 were cultured with various concentrations of pioglitazone for 24 h. Both PPAR␥1 and ␥2 expression levels were analyzed by immunoblotting. When equal FIG. 3. Activation of the perilipin promoter by PPAR␥2. A, mouse perilipin promoter sequence (31). The transcription start site is ϩ1. The PPRE is boxed. The PPRE-like sequence is underlined. The mutant sequence in the PPRE is shown by italic letters under the individual original sequence. B, 3T3-L1 preadipocytes or HEK293 cells were co-transfected with the indicated mouse perilipin promoter-luciferase reporter gene together with pRL-TK and expression plasmids for PPAR␥2 and RXR␣ (solid bar). pGL3 is a promoter-less reporter gene. The cells were incubated with a medium containing 10% charcoal-stripped FBS and 10 M pioglitazone for 24 h, and then luciferase assays were performed. Luciferase activities in the absence of PPAR␥2 and RXR␣ were considered as 1. The data are shown as the means Ϯ S.D. C, 3T3-L1 preadipocytes were co-transfected with the indicated reporter gene together with pRL-TK and 100 ng of expression plasmids for PPAR␥2 and/or RXR␣. The cells were incubated with a medium containing 10% charcoal-stripped FBS and 10 M pioglitazone (PGZ) for 24 h, and then luciferase assays were performed. Luciferase activities in the absence of PPAR␥2 and RXR␣ were considered as 1. The data are shown as the means Ϯ S.D. D, left panel, 3T3-L1 adipocytes were co-transfected with the indicated mouse perilipin promoter-luciferase reporter gene and pRL-TK. The cells were incubated with a medium containing 10% charcoalstripped FBS and 10 M pioglitazone (solid bar) or vehicle (open bar) for 24 h, and then luciferase assays were performed. Luciferase activities with vehicle were considered as 1. The data are shown as the means Ϯ S.D. All of the values given are the averages of data from three experiments performed in triplicate. Right panel, 3T3-L1 adipocytes were cotransfected with the indicated reporter gene together with pRL-TK and 500 ng of expression plasmids for PPAR␥2 and RXR␣ (solid bar). The cells were incubated with a medium containing 10% charcoal-stripped FBS and 10 M pioglitazone for 24 h, and then luciferase assays were performed. Luciferase activities in the absence of PPAR␥2 and RXR␣ were considered as 1. The data are shown as the means Ϯ S.D. All of the values given are the averages of data from three experiments performed in triplicate.
amounts of both isoforms were expressed (Fig. 7A, inset, 1.00  versus 1.08), PPAR␥2 was more potent than PPAR␥1 at every concentration of pioglitazone, suggesting that perilipin expression being exclusively limited to adipose tissue is partly attrib-utable to PPAR␥2 expression in this tissue. We then examined whether other transcription factors might be involved in perilipin expression during adipogenesis. Forced expression of C/EBP␣, SREBP1a, and LXR␣ (with a synthetic ligand), which are highly expressed and regulate the expression of several adipogenic genes during adipocyte differentiation (22)(23)(24), did not enhance the activities of the perilipin promoter (ϳ2.0 kb) (Fig. 7B). PPAR␣ together with a synthetic ligand slightly augmented perilipin promoter activity but was still a much weaker regulator as compared with PPAR␥2. Taken together,

FIG. 4. A PPAR␥-RXR␣complex binds to the PPRE in the perilipin promoter and regulates the gene expression.
A, a doublestranded DNA fragment corresponding to the PPRE of the aP2 gene (15) was 3Ј end-labeled with [␥-32 P] ATP. 32 P-Radiolabeled fragments were incubated with recombinant PPAR␥ and RXR␣. In competition assays, 40-, 80-, and 200-fold molar excesses of unlabeled doublestranded DNA fragments derived from the wild type (PPREwt) of perilipin PPRE (see Fig. 3A) or PPRE-like sequence (see Fig. 3A, underlined) were added. DNA-protein complexes were subjected to electrophoresis on 4% polyacrylamide gel. B, a double-stranded DNA fragment corresponding to the PPRE of the perilipin gene was 3Ј endlabeled with [␥-32 P]ATP. 32 P-Radiolabeled fragments were incubated with recombinant PPAR␥ and RXR␣. In competition assays, 40-, 80-, and 200-fold molar excesses of unlabeled double-stranded DNA fragments derived from the wild type (PPREwt) or the mutant (PPREmut) of perilipin PPRE (see Fig. 3A) were added. DNA-protein complexes were subjected to electrophoresis on 4% polyacrylamide gel. C, comparison of conserved PPRE sequences in the promoters of both the mouse and human (119460 -119472 in AC079075) genes. D, 3T3-L1 preadipocytes were co-transfected with either the wild type or mutant version of reporter gene (pPlin-2.0 or pPlin-mut) together with pRL-TK and the expression plasmids for PPAR␥2 and RXR␣ (solid bar). The cells were incubated with a medium containing 10% charcoal-stripped FBS and 10 M pioglitazone for 24 h, and then luciferase assays were performed. Luciferase activities in the absence of PPAR␥2 and RXR␣ were considered as 1. The data are shown as the means Ϯ S.D. it is likely that PPAR␥2 is a major regulator of perilipin gene expression in adipocyte cells. DISCUSSION The major finding of this study is that perilipin is a novel, adipocyte-specific target gene of PPAR␥. Perilipin mRNA in undifferentiated 3T3-L1 preadipocytes is not detectable but is remarkably increased during 3T3-L1 adipocyte differentiation. A recent paper showed that perilipin gene expression was very specifically regulated, the expression occurring concomitantly with changes in PPAR␥ gene expression induced by treatment of 3T3-L1 preadipocytes with a mitogen-activated protein kinase inhibitor and/or fibroblast growth factor-2 (25). The current finding can offer a mechanistic explanation of these observations.
We further report that the perilipin gene has a functional PPRE in its 5Ј-flanking region and that its expression in adipocytes is predominantly regulated by PPAR␥2 in adipocytes. Reporter assays revealed that the segment from Ϫ2.0 to Ϫ1.9 kb of the mouse perilipin gene is essential for PPAR␥-mediated gene expression (Figs. 3 and 4). There indeed exists a PPRE in this region, and the TCACCTTTCACCC sequence is both recognized by a PPAR␥/RXR␣ heterodimer and is indispensable for the activation of the promoter (Figs. 3 and 4). We confirmed a 74% nucleotide homology in the 300-bp promoter region (Ϫ2.0 to Ϫ1.7 kb for the mouse gene) and a 100% match of two direct repeats (TCACCT and TCACCC) in the PPRE in the mouse and human forms (Fig. 4), suggesting that a regulation likely occurs in humans highly similar to what has been found in the mouse model.
PPAR␥ is present in two major isoforms, PPAR␥1 and PPAR␥2, resulting from alternate promoter utilization (26,27). PPAR␥2 contains an additional 30 amino acids at the N-terminal end in comparison with PPAR␥1. PPAR␥2 expression is limited exclusively to adipose tissue, where it plays a pivotal role in adipogenesis, whereas PPAR␥1 is ubiquitously expressed in various tissues. Recently, two reports presented quite different data on the adipogenic action of the two isoforms (28,29). Ren et al. (28) demonstrated that PPAR␥2 was adipogenic and PPAR␥1 was not, whereas Mueller et al. (29) concluded that both PPAR␥ isoforms can drive the differentiation of fully functional fat cells. The data presented here show that both PPAR␥2 and PPAR␥1 are capable of enhancing perilipin promoter activity (Fig. 7), favoring the latter finding. However, a modest expression of PPAR␥2 in stable NIH-3T3 (NIH-␥2) cells was not sufficient for inducing perilipin gene expression, even in the presence of a PPAR␥ ligand, pioglitazone (Fig. 6). Once these cells were cultured with a differentiation medium containing pioglitazone, perilipin mRNA was enhanced concomitantly with the increase in endogenous PPAR␥, suggesting that a high expression of PPAR␥ including both isoforms is required for augmenting perilipin gene expression. Alternatively, other as yet unidentified factors activated during adipocyte differentiation might be involved in this event. During the period of this investigation, several transcription factors other than PPAR␥2 known to be involved in adipocyte differentiation, e.g. C/EBP␣, SREBP-1a, and LXR␣, were not able to stimulate perilipin promoter activity. It remains, therefore, to be precisely elucidated how perilipin gene expression is restricted to the adipose tissue. It is possible that the levels of PPAR␥1 and endogenous as yet unidentified ligands in extraadipose tissues are insufficient to induce perilipin gene expression, whereas in adipose tissues there are sufficiently high levels of the PPAR␥ isoforms, their ligands, and the several functional co-activators required for transcriptional activity (30). Further investigation should bring these issues into better focus.
The perilipin proteins are the most abundant PKA substrates in adipocytes (1,5,6) and play an important role in PKA-mediated lipolysis. Several reports have provided evidence that the perilipins protect stored triacylglycerol from hydrolysis by cellular lipases. Two reports describing the phenotype of the perilipin-null mouse also have demonstrated that these mice have both greatly diminished adipose stores and constitutively high levels of basal lipolysis (7,8). These findings suggest perilipin to be an attractive target for anti-obesity medications. Although one would expect reduced perilipin gene expression in adipocytes by the administration of a PPAR␥ antagonist, according to the findings described here, which strongly suggest that PPAR␥ plays a crucial role in the expression of perilipin, this approach might not be practical for reasons similar to the complications of side effects in the usage of PPAR␥ agonists against type 2 diabetes. As yet unidentified determinants of adipose-specific perilipin gene expression in addition to PPAR␥ could provide further targets for pharmaceutical intervention. Further analysis of the regulation of perilipin gene expression should make the elucidation of such targets possible, as well as provide greater insights into the underlying mechanisms controlling adipose-specific gene expression in general.