Identification of a functional peroxisome proliferator-responsive element in the murine fatty acid transport protein gene.

Fatty acid transport protein (FATP), a plasma membrane protein implicated in controlling adipocyte transmembrane fatty acid flux, is up-regulated as a consequence of adipocyte differentiation and down-regulated by insulin. Based upon the sequence of the FATP gene upstream region (Hui, T. Y., Frohnert, B. I., Smith, A. J., Schaffer, J. A., and Bernlohr, D. A. (1998) J. Biol. Chem. 273, 27420-27429) a putative peroxisome proliferator-activated receptor response element (PPRE) is present from -458 to -474. To determine whether the FATP PPRE was functional, and responded to lipid activators, transient transfection of FATP-luciferase reporter constructs into CV-1 and 3T3-L1 cells was carried out. In CV-1 cells, FATP-luciferase activity was up-regulated 4- and 5.5-fold, respectively, by PPARalpha and PPARgamma in the presence of their respective activators in a PPRE-dependent mechanism. PPARdelta, however, was unable to mediate transcriptional activation under any condition. In 3T3-L1 cells, the PPRE conferred a small but significant increase in expression in preadipocytes, as well as a more robust up-regulation of FATP expression in adipocytes. Furthermore, the PPRE conferred the ability for luciferase expression to be up-regulated by activators of both PPARgamma and retinoid X receptor alpha (RXRalpha) in a synergistic manner. PPARalpha and PPARdelta activators did not up-regulate FATP expression in 3T3-L1 adipocytes, however, suggesting that these two subtypes do not play a significant role in differentiation-dependent activation in fat cells. Electromobility shift assays showed that all three PPAR subtypes were able to bind specifically to the PPRE as heterodimers with RXRalpha. Nuclear extracts from 3T3-L1 adipocytes also showed a specific gel-shift complex with the FATP PPRE. To correlate the expression of FATP to its physiological function, treatment of 3T3-L1 adipocytes with PPARgamma and RXRalpha activators resulted in an increased uptake of oleate. Moreover, linoleic acid, a physiological ligand, up-regulated FATP expression 2-fold in a PPRE-dependent manner. These results demonstrate that the FATP gene possesses a functional PPRE and is up-regulated by activators of PPARalpha and PPARgamma, thereby linking the activity of the protein to the expression of its gene. Moreover, these results have implications for the mechanism by which certain PPARgamma activators such as the antidiabetic thiazolidinedione drugs affect adipose lipid metabolism.

Obesity, defined as an excessive accumulation of body fat, has become an increasingly common health concern in industrialized societies. Excessive adiposity has been linked to the pathogenesis of many diseases, including type 2 diabetes mellitus, coronary artery disease, and hypertension. The increased awareness of the detrimental effects of obesity contributes to the search for a greater understanding of the molecular mechanisms controlling the accumulation of adipose tissue and its metabolism.
A central issue in the function of fat tissue is the method by which adipocytes take up and release fatty acids. This process has been the source of considerable debate (1)(2)(3). Because free fatty acids are hydrophobic, they freely crossed membranes by passive diffusion. However, studies in adipocytes, hepatocytes, jejunal enterocytes, skeletal muscle, and heart myocytes support a saturable, protein-mediated mechanism for fatty acid transport (4 -8). Thus far, five putative mammalian fatty acid transporters have been identified: fatty acid-binding protein (plasma membrane) (9), 56-kDa renal fatty acid-binding protein (10), caveolin (11), fatty acid translocase (12), and fatty acid transport protein (FATP) 1 (13). FATP is an integral plasma membrane protein with four to six predicted membrane-spanning regions with the highest levels found in skeletal muscle, heart, and fat with lower levels in brain, kidney, lung, and liver. Although FATP mRNA is present at low levels in 3T3-L1 preadipocytes, it is up-regulated 5-7-fold as a consequence of adipose conversion (13,14). This increase is consistent with the increase in oleic acid uptake shown during preadipocyte differentiation (15).
The differentiation of adipose precursor cells into adipocytes has been shown to be mediated by three groups of transcription factors: peroxisome proliferator-activated receptor ␥ (PPAR␥ 1 and 2) the CCAATT enhancer-binding proteins (C/EBP), and the sterol-response element binding proteins (SREBP or ADD1) (16). The importance of PPARs for the development and maintenance of the adipocyte phenotype can be more directly shown by the existence of peroxisome proliferator response elements (PPREs) in the promoters of several genes whose protein products are critical for lipid metabolism and the development of the adipocyte phenotype such as lipoprotein lipase, phosphoenolpyruvate carboxykinase, acyl-CoA synthetase, malic enzyme, and adipocyte lipid-binding protein (ALBP or aP2) (16 -21).
PPARs constitute a subfamily of the steroid hormone receptor superfamily. PPAR␣ is predominantly expressed in liver, heart, kidney, and adipose tissue, whereas PPAR␦ (also known as NUC1 or FAAR) shows a similar expression, with the ex-ception of low levels in liver (22)(23)(24)(25). Two isoforms of mouse PPAR␥ have been cloned, ␥1 (22,24,26) and ␥2 (19), which are transcribed from the same gene and alternatively spliced (27,28). PPAR␥1 is expressed in liver, heart, and kidney, similar to PPAR␣, whereas PPAR␥2 is primarily expressed in adipose tissue (29). The two PPAR␥ isoforms do not appear to differ significantly in ligand binding affinities, ability to bind DNA response elements, or ability to activate transcription. The critical difference identified, thus far, is the distribution of expression.
FATP expression has been shown to be up-regulated in mouse liver by PPAR␣ activators and in white adipose tissues by activators of PPAR␥ (30,31). This evidence, together with the differentiation-dependent regulation of FATP in 3T3-L1 adipocytes, led us to investigate the upstream region of the recently cloned FATP gene (32) for a possible PPRE and to examine the role of the various PPAR subtypes in FATP expression. In this report we detail the regulation of the FATP gene, the identification of a functional PPRE, and its up-regulation by PPAR␣ and PPAR␥ agonists.
Plasmid Constructs-Polymerase chain reaction was used to introduce NheI sites into the upstream sequence of FATP. Reporter constructs pNH11 (Ϫ971/ϩ84), pNH13 (Ϫ556/ϩ84), and pNH15 (Ϫ160/ ϩ84) were constructed by ligating varying lengths of FATP upstream sequence into the NheI and HindIII sites of the pGL3-Basic luciferase expression vector (Promega). Construct pNH13⌬P was generated by single-stranded mutagenesis of construct pNH13 using the Muta-Gene in vitro mutagenesis kit (Bio-Rad, Hercules, CA) and the oligonucleotide: 5Ј-CTGGAACATCTCCTGAGTACTTCCTCCTCTCCC-3Ј. Sequence fidelity and the introduced mutation were verified by sequencing. The PPAR␥2 expression construct, pAH215, was made by cutting pBS-PPAR␥2 with KpnI and XbaI and ligating into pCDNA3.1 (Promega). The expression construct for RXR␣, pAH232, was subcloned from pRS-hRXR␣ into the EcoRI site of pCDNA3.1.
Cell Culture and Transfection-CV-1 cells were cultured in DMEM with 10% fetal bovine serum (FBS) until the day prior to transfection experiments, when the cells were plated in phenol red-free DMEM and 10% charcoal-treated FBS. CV-1 cells were transiently transfected using the calcium-phosphate method as described previously (33,34). Briefly, each well of a 12-well culture plate was transfected with 1 g of firefly (Photinus pyralis) luciferase reporter construct with or without 0.5 g of expression plasmid for mouse PPAR␥2, rat PPAR␣, or mouse PPAR␦. 20 ng of pRL-SV40 (Promega), a vector containing a sea pansy (Renilla reniformis) luciferase gene driven by an SV40 promoter, was cotransfected into each well as a transfection control. After approximately 20 h, the calcium phosphate/DNA precipitate was removed by washing three times with phosphate-buffered saline. Cells were then refed and treated for 48 h with either the appropriate PPAR activator or its vehicle, dimethyl sulfoxide (Me 2 SO).
3T3-L1 preadipocytes were grown to confluence and induced to differentiate to adipocytes as described previously (35). Briefly, preadipocytes were cultured in DMEM and 10% calf serum. Two days after reaching confluence, cells were differentiated by treatment with 174 nM insulin, 0.5 mM methylisobutylxanthine, and 0.25 M dexamethasone in DMEM containing 10% FBS. Cells were treated for an additional 2 days with insulin and maintained in DMEM with 10% FBS until they are fully mature adipocytes at day 8.
Transient transfection of 3T3-L1 preadipocytes was performed by calcium phosphate precipitation as described above. Each well was transfected with 1 g of firefly luciferase reporter construct and 10 ng of pRL-SV40 control vector. In the first set of transfections, cells were maintained in DMEM containing 10% calf serum until transfection mixture was washed off of cells. Cells were then treated with either linoleic acid or ethanol for 24 h in serum-free DMEM. For the second set of preadipocyte transfections, cells were maintained in DMEM with 10% calf serum for the duration of the experiments. After washing transfection mixture from cells, they were treated with either a PPAR activator or its vehicle (Me 2 SO) for 48 h.
3T3-L1 adipocytes grown in 12-well culture plates were transiently transfected as described previously (32) with 500 ng of firefly luciferase construct and 20 ng of pRL-SV40 control vector per well. Briefly, DNA in DMEM was complexed with LipofectAMINE (Life Technologies, Inc.) at the ratio of 1:8 (w/v) for 45 min before addition to the cells. After 4 h, an equal volume of 20% fetal bovine serum in DMEM was added. The transfection mixture was removed after 20 h. The cells were then refed with DMEM containing 10% FBS and treated with either PPAR activator or vehicle for 48 h.
All transiently transfected cells were washed two times with phosphate-buffered saline, lysed, and luciferase activities were determined using the Dual Luciferase Reporter Assay System (Promega) according to the manufacturer's protocol. Firefly luciferase activities were normalized to sea pansy luciferase activities to adjust for transfection efficiency. Activities were also normalized to the promoterless reporter construct, pGL3-Basic.
To study the binding of nuclear hormone receptors to the putative PPRE, a double-stranded oligonucleotide, PPREwt, spanning nucleotides Ϫ482 to Ϫ453 of the FATP upstream sequence was 32 P-labeled with polynucleotide kinase (Promega). A 15-l reaction containing and 0.5 ng of PPRE probe and 9 g of nuclear extract or 0.5-1 l of in vitro translation reaction was incubated for 20 min at 25°C and 15 min at 4°C in a buffer containing 20 mM HEPES (pH 8), 60 mM KCl, 1 mM dithiothreitol, 10% glycerol, and 2 g poly(dI-dC). The DNA-protein complexes were resolved from the free probe by electrophoresis at 4°C on a 4% polyacrylamide gel in 0.5 ϫ TBE buffer (1 ϫ TBE ϭ 9 mM Tris, 90 mM boric acid, 20 mM EDTA) (pH 8). Double-stranded oligonucleotides composed of the following sequences were used for competition analysis: PPREwt, 5Ј-GATCTAGAGGAGGAAGTGGGGCAAAGGGCA-CAGGA-3Ј; PPREmut, 5Ј-GATCTAGAGGAGGAAGTGGGGCtAtcGGC-ACAGGA-3. PPRE sequence is underlined. Mutated bases are shown in lowercase letters.
Fatty Acid Uptake-3T3-L1 adipocytes between days 7 and 9 were treated for 4 days with either 20 M troglitazone and 1 M 9-cisretinoic acid or their carrier, Me 2 SO. The adipocytes were then assayed for uptake of oleate essentially as described previously (37). Briefly, cells were first preincubated for 2-4 h in serum-free DMEM and then for 10 -30 min at 25°C in Krebs-Ringer phosphate solution with 2 mM glucose. Oleate uptake was measured at 25°C by incubating cells in a 100 M, 1: the cells and washing three times with ice-cold phosphate-buffered saline containing 0.1% albumin and 200 M phloretin. Cell lysate was then quantitated for radioactivity using a Beckman 3801 liquid scintillation counter.

RESULTS
The Mouse Fatty Acid Transport Protein Gene Contains a Functional PPRE-Previous studies of FATP regulation in mice indicated that transcription was activated in both liver and white adipose tissue by treatment of mice with activators of PPAR␣ or PPAR␥, respectively (30,31). Furthermore, FATP expression was observed to be up-regulated during adipose differentiation, a process known to be mediated in part by PPAR␥ (14). The upstream sequence of the recently cloned FATP gene (32) was therefore examined for a possible PPRE. A putative PPRE was identified, which is similar to the consensus sequence proposed by Palmer et al. (38) (Fig. 1, Table I).
In order to determine whether the putative PPRE identified in the FATP upstream sequence was able to mediate transcriptional activation, portions of the FATP 5Ј-flanking sequence were tested for their ability to mediate PPAR-activated transcription of a reporter gene. Four luciferase reporter constructs were made, each containing portions of the FATP 5Ј-flanking region linked to a promoterless firefly luciferase gene. Two of these plasmids, pNH11 and pNH13, contained the putative PPRE sequence, while the other two, pNH13⌬P and pNH15 did not (Fig. 2).
In the first set of experiments, two of these constructs, pNH13 and pNH13⌬P, as well as the promoterless pGL3-Basic control construct, were transfected into 3T3-L1 preadipocytes. Following transfection, the cells were maintained for 24 h in serum-free media and treated with 100 M linoleic acid or its   AACT AGGTCA A AGGTCA Tg g G carrier, ethanol. This experiment was performed in the absence of serum, since serum albumin binds fatty acids, leaving low levels of available fatty acid activator. Linoleic acid treatment activated transcription of the PPRE-containing construct approximately 2-fold over control-treated cells (Fig. 3), but did not affect transcription of the PPRE-deletion construct. Similar results were seen in transfection of 3T3-L1 adipocytes (data not shown). These experiments indicated that the putative PPRE identified in the FATP upstream sequence was indeed functional.
To further characterize the responsiveness of this PPRE to the various PPAR subtypes, all four reporter constructs were transfected into CV-1 cells and assayed for luciferase activity in the presence and absence of various PPARs and their activators. The activators used were troglitazone, a thiazolidinedione (PPAR␥ activator); WY14643, a fibrate drug (PPAR␣ activator); and the PPAR␦ activator, Delta Selective C. As shown in able to activate transcription in the PPRE-containing constructs, albeit to a lesser extent; transcription was increased 4and 2-fold in the presence and absence of activator, respectively (Fig. 4B). Finally, the PPAR␦ subtype was unable to positively regulate transcription of any of the reporter constructs, regardless of activator treatment (Fig. 4C). Deletion of the PPRE rendered the promoter unresponsive to any PPAR or agonist combination.
PPRE Involved in Differentiation-dependent Regulation of FATP-In order to determine the role of the FATP PPRE in the process of its gene regulation during the process of adipose differentiation, the luciferase reporter constructs were introduced into both 3T3-L1 preadipocytes as well as mature adipocytes. These experiments relied upon endogenous PPARs to mediate transcriptional activation. Cells were also treated with activators of the various PPAR subtypes to determine whether transcription could be further increased. In preadipocytes, which contain low levels of PPAR␥ as well as PPAR␦, the PPRE containing construct, pNH13, was expressed at about 1.6-fold the level of the PPRE-deletion construct, pNH13⌬P, when both were treated with Me 2 SO. Treatment with troglitazone had a small, but not statistically significant, effect on pNH13 expression. In adipocytes, the PPRE-containing construct was expressed at levels 5-fold higher than the PPRE-deletion construct (Fig. 5). Furthermore, troglitazone treatment resulted in a further 3-fold increase in expression. Neither WY14643 nor Delta Selective C caused any significant change in luciferase expression in either preadipocytes or adipocytes.
Synergistic Activation by PPAR␥ and RXR␣ Activators-Issemann et al. (39) showed that the RXR ligand, 9-cis-retinoic acid enhances PPAR action. To determine whether RXR␣ activation affected PPAR␥-dependent transactivation, 3T3-L1 adipocytes were treated with either troglitazone, 9-cis-retinoic acid, or both. Retinoic acid did not significantly affect transcription by itself; however, when added to cells in conjunction with troglitazone, it was able to produce an almost 2-fold increase in activity above that produced by tro-glitazone alone (Fig. 6). This result demonstrates the synergistic activation of the FATP gene in response to activation of both PPAR␥ and RXR␣.
PPARs and RXR␣ Bind as Heterodimers to the FATP PPRE-In order to determine whether PPARs bind to the PPRE as heterodimers with RXR␣, gel mobility shift assays were performed with a double-stranded oligonucleotide containing the FATP PPRE (Fig. 7A). The double-stranded probe, PPREwt, was end-labeled with 32 P and incubated with in vitro translated protein as well as 3T3-L1 nuclear extract. As shown in Fig. 7B, neither PPARs nor RXR␣ alone could bind to the PPRE; however, all three PPAR subtypes were able to bind as heterodimers with RXR␣ to the probe. Furthermore, nuclear proteins from 3T3-L1 adipocytes were able to form an in vitro complex with the PPRE (Fig. 7C). In order to test the specificity of the protein-DNA interactions, an excess of unlabeled oligonucleotide (PPREwt) was added to the reactions. The unlabeled oligonucleotide was able to compete for binding of all three PPAR-RXR␣-DNA complexes, as well as the nuclear protein-DNA complex. The introduction of 3-base pair substitutions (Fig. 7A) produced an oligonucleotide, PPREmut, which was no longer able to significantly compete for protein binding.
PPAR␥ and RXR␣ Activators Cause an Increase in Oleate Uptake-Finally, in order to correlate PPAR-mediated transcriptional activation with putative in vitro protein function, fatty acid uptake was analyzed in 3T3-L1 adipocytes. 3T3-L1 adipocytes were treated for 4 days with either troglitazone and retinoic acid or their carrier Me 2 SO. Cells were then incubated with [ 3 H]oleate:BSA mixture (1:1), washed, lysed at various time points, and assayed for radioactivity. As shown in Fig. 8, treatment with the activators of PPAR␥ and RXR␣ resulted in a significant increase in oleate uptake. This result is consistent with the up-regulation of FATP transcription by troglitazone and 9-cis-retinoic acid, as shown by the previously described transfection studies, and correlates well with the increase in FATP mRNA expression upon treatment of 3T3-L1 adipocytes with the PPAR␥ agonist BRL49653, as shown by Martin et al. (30). DISCUSSION The critical role of PPARs in the regulation of lipid metabolism has become increasingly apparent. Many genes whose products take part in some aspect of fatty acid catabolism, synthesis, or trafficking have been shown to contain functional PPREs (Table I). FATP, which has been argued to play a role fatty acid uptake, is a likely candidate for regulation by this group of nuclear hormone receptors. Indeed, two previous studies have shown that treatment of various cell types with PPAR activators leads to an increase in FATP mRNA levels (30,31). In this paper we have identified a sequence in the 5Ј region of the murine FATP gene, which is very similar (16/17) to the consensus sequence for previously identified PPREs (see Table  I).
By deletion analysis and mutation of this putative PPRE, we have demonstrated that this PPRE is necessary for the PPARmediated up-regulation of FATP expression. Transfection into 3T3-L1 preadipocytes showed that FATP transcription can indeed be activated by the naturally occurring compound, linoleic acid. This fatty acid has been shown previously to be able to activate transcription via both the ␣ and ␥ PPAR subtypes (40 -42). Further transfection experiments into CV-1 cells demonstrated that FATP transcription was activated by both PPAR␣ and PPAR␥. Both subtypes activate transcription upon treatment with synthetic activator; however, the receptors are also able to activate transcription in the absence of exogenous activator. This could be explained either by the presence of an endogenous activator, such as a fatty acid or its metabolite, or by a ligand-independent activity of these subtypes (43).
In addition to showing the functionality of the FATP PPRE, our experiments demonstrate a differential activation of gene expression by the various PPAR subtypes. Although both PPAR␣ and PPAR␥ are able to activate transcription, our studies show that PPAR␥ mediates a greater response both in the presence and absence of a synthetic activator. PPAR␦, in contrast, did not significantly activate transcription, either alone or upon treatment with activator. This difference has been demonstrated previously in other systems (44) and has been hypothesized to reflect the differing roles of the PPAR subtypes in the regulation of fatty acid metabolism. PPAR␣ has been best characterized in the liver, where it up-regulates many genes involved in the catabolism of fatty acids. PPAR␥ is chiefly active in the adipose tissue, where it contributes to lipid accumulation and the development of the adipose phenotype. In contrast to the other two subtypes, the role of PPAR␦ in wholebody fatty acid metabolism has not been well defined.
We and others have reported the up-regulation of FATP during adipose conversion (13,14). We hypothesized that this differentiation-dependent regulation of FATP is mediated by PPAR␥ and is dependent on the presence of the FATP PPRE. This was investigated in a further series of transfections, which Protein-DNA complexes were analyzed by electrophoretic mobility shift assay. C, labeled PPREwt probe was incubated with 3T3-L1 adipocyte nuclear extracts with or without competitor. Competitors PPREwt, PPREmut, and NS (a nonspecific competitor oligonucleotide) were added in 20-, 50-, and 200-fold molar excess. compared transcriptional activity in both 3T3-L1 preadipocytes and adipocytes, in both the presence and absence of synthetic activators. The PPRE conferred a 1.6-fold activation of FATP transcription in preadipocytes. At this point in the differentiation process, PPAR␥ is present in low levels relative to fully differentiated adipocytes. Exogenous activators of the three PPAR subtypes did not significantly increase transcription above that of untreated cells. In adipocytes, the PPRE conferred a 5-fold increase in FATP expression, and this activation was further stimulated by the PPAR␥ activator, troglitazone. Neither the activator of PPAR␣ nor PPAR␦ was able to increase FATP expression over untreated cells. This can be explained by the lack of significant PPAR␣ expression in adipocytes. PPAR␦, while present in adipocytes, has been demonstrated in the previously described studies to be unable to activate FATP expression.
Several previous studies have shown a synergistic effect of the RXR␣ activators, 9-cis-retinoic acid on PPAR-activated expression (18,19,21,(45)(46)(47)(48). This convergence of the PPAR and RXR signaling pathways was also demonstrated in transfection studies of FATP reporter constructs in 3T3-L1 adipocytes. Although the RXR␣ activator alone was unable to increase expression of FATP in 3T3-L1 adipocytes, it was able to enhance the troglitazone-mediated activation of expression.
The ability of the various PPAR subtypes to bind to the FATP PPRE in vitro was examined by electromobility shift assay. As has been shown in other systems, neither the PPARs nor RXR␣ were able to bind to the PPRE as homodimers (20,21,(45)(46)(47)(48)(49)(50); however, all three PPAR subtypes were able to bind as heterodimers. These protein-DNA complexes were sequence-specific, as shown by competition analysis, and were dependent upon the presence of an intact PPRE. Mutation of 3 base pairs of the PPRE abolished the formation of protein complexes on this element. It is interesting to note that PPAR␦, while unable to activate transcription, was able to form a heterodimer complex with the PPRE. This indicates that binding of the receptor heterodimers to an element is not equivalent with transcriptional activity.
In order to correlate FATP regulation by PPARs with its putative function, oleate uptake was measured in cells that were treated with troglitazone and 9-cis-retinoic acid. Uptake was significantly increased in activator-treated cells when compared with control-treated cells. This leads to an interesting model for positive feedback regulation of FATP. Increased FATP expression has been shown to result in increased fatty acid uptake (13). Fatty acids, in turn, are activators of PPAR␥ and PPAR␣, which are able up-regulate the expression of FATP.
We have reported previously the regulation of FATP by insulin via an insulin-responsive element (PLE3) in the upstream region of the FATP gene (Ϫ1353 to Ϫ1347). The downregulation of FATP by insulin, an anabolic hormone, seems counterintuitive, since fatty acid uptake would be expected to rise in response to insulin stimulation. It is important to note, however, that regulation of FATP at the transcriptional level is unlikely to be the result of the transient postprandial insulin peak, but rather a more chronic hyperinsulinemia, such as in type 2 diabetes mellitus. Furthermore, the majority of type 2 diabetics are obese, a condition associated with a down-regulation of PPAR␥ expression in adipose tissue (51). The combination of these two factors may contribute to the elevation in serum free fatty acid levels observed in type 2 diabetics. This leads to a possible mechanism for the antidiabetic effects of the drug, troglitazone; by reversing the effects of hyperinsulinemia and obesity on FATP regulation, troglitazone may enable adipose tissue improve fatty acid uptake. Further studies on the function of FATP and its regulation in the diabetic state may lead to insight into both normal and deranged fatty acid metabolism.