HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 18, 15703-15711, May 3, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/18/15703    most recent M110158200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Sato, O. Articles by Motojima, K. Search for Related Content PubMed PubMed Citation Articles by Sato, O. Articles by Motojima, K. Social Bookmarking                      What's this?

Dual Promoter Structure of Mouse and Human Fatty Acid Translocase/CD36 Genes and Unique Transcriptional Activation by Peroxisome Proliferator-activated Receptor  and  Ligands^*

Osamu Sato^§, Chikako Kuriki, Yuka Fukui^§, and Kiyoto Motojima^§¶

From the  Department of Biochemistry, School of Pharmaceutical Sciences, Toho University, Funabashi, Chiba 274-8510 and the ^§ Department of Biochemistry, Meiji Pharmaceutical University, Kiyose, Tokyo 204-8588, Japan

Received for publication, October 22, 2001, and in revised form, January 23, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Fatty acid translocase (FAT)/CD36 is a glycoprotein involved in multiple membrane functions including uptake of long-chain fatty acids and oxidized low density lipoprotein. In mice, expression of the gene is regulated by peroxisome proliferator-activated receptor (PPAR)  in the liver and by PPAR in the adipose tissues (Motojima, K., Passilly, P. P., Peters, J. M., Gonzalez, F. J., and Latruffe, N. (1998) J. Biol. Chem. 273, 16710-16714). However, the time course of PPAR ligand-induced expression of FAT/CD36 in the liver, and also in the cultured hepatoma cells, is significantly slower than those of other PPAR target genes. To study the molecular mechanism of the slow transcriptional activation of the gene by a PPAR ligand, we first cloned the 5' ends of the mRNA and then the mouse gene promoter region from a genomic bacterial artificial chromosome library. Sequencing analyses showed that transcription of the gene starts at two initiation sites 16 kb apart and splicing occurs alternatively, producing at least three mRNA species with different 5'-noncoding regions. The PPAR ligand-responsive promoter in the liver was identified as the new upstream promoter where we found several possible binding sites for lipid metabolism-related transcriptional factors but not for PPAR. Neither promoter responded to a PPAR ligand in the in vitro or in vivo reporter assays using cultured hepatoma cells and the liver of living mice. We also have cloned the human FAT/CD36 gene from a bacterial artificial chromosome library and identified a new independent promoter that is located 13 kb upstream of the previously reported promoter. Only the upstream promoter responded to PPAR and PPAR ligands in a cell type-specific manner. The absence of PPRE in the responding upstream promoter region, the delayed activation by the ligand, and the results of the reporter assays all suggested that transcriptional activation of the FAT/CD36 gene by PPAR ligands is indirectly dependent on PPAR.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

FAT¹/CD36 is a glycosylated membrane protein normally expressed on the surface of monocyte-macrophage lineage cells (1), platelets (2), microvascular endothelial cells (3), and adipocytes (4). FAT was identified as a fatty acid translocase abundantly expressed on differentiated adipocytes (4), and CD36 was identified as an oxidized low density protein receptor that does not recognize acetylated low density lipoprotein on macrophages (5). Sequence comparison revealed that the two independently identified glycoproteins are the same. FAT/CD36 has been shown also to be involved in foam cell formation in early atherosclerosis (6, 7), in apoptosis (8), and in angiogenesis (9) by recognizing a variety of ligands such as long-chain fatty acid (10), anionic phospholipids (11), apoptotic cells (7), thrombospondin (12), and collagen (13).

Studies using the FAT/CD36-deficient rodents suggested that this molecule is involved in the action of insulin, long-chain fatty acid, and lipid metabolism (14-16). A case control study on people with FAT/CD36 deficiency also supported the results obtained in rodents (17). However, not all the results of rodents and human studies are consistent. The reported phenotypes of FAT/CD36-deficient rodents and of human patients are considerably variable between studies except the increased nonesterified fatty acid concentrations and the reciprocally decreased glucose concentrations in serum (18). These results suggest that other genetic and/or environmental factors, in addition to FAT/CD36 deficiency, are also strongly associated with the overall phenotypes and the onset of several diseases, but they also suggest that FAT/CD36 is actually a fatty acid transporter in adipocytes and muscle cells. This was further supported by the demonstration that a defect in human myocardial long-chain fatty acid uptake is caused by FAT/CD36 mutations (19).

Expression of the FAT/CD36 gene should be properly regulated to ensure its unique functions in various cell types. FAT/CD36 expression has been studied most extensively on macrophages and is known to be regulated by cytokines (20, 21) and by ligands of the peroxisome proliferator-activated receptor (PPAR)  (6, 7). We studied the effects of PPAR and PPAR ligands on the expression levels of FAT/CD36 mRNA in mouse tissues and showed that the mRNA is induced by PPAR in the liver and by PPAR ligands in adipose tissues (22). The time course of mRNA induction by a PPAR ligand in the liver is significantly slower than those of other typical PPAR target genes, although the essential role of PPAR in transcriptional activation was evident because of its disappearance in PPAR-null mice (22). These differences suggested that the same mechanism of transcriptional activation is not operating on the FAT/CD36 gene as on other PPAR target genes. However, extensive studies on the PPAR ligand-induced expression of the human FAT/CD36 gene in monocyte-macrophages showed that the human gene promoter binds PPAR and is activated by its ligands in a normal fashion (6). The PPRE identified by its DR1 sequence (7) is not typical, although Tontonoz et al. (7) showed its capability to bind PPAR, and the transactivation by the receptor and its ligand conducted in cultured CV-1 cells using the minimal length of the promoter sequence was not as large as that of the endogenous gene in monocyte-macrophages (7). They did not fully describe the physiological responses of the human FAT/CD36 gene because the structural analysis of the gene was not sufficient. Therefore, we decided to characterize the detailed structures of the FAT/CD36 genes of mouse and then of human.

The results obtained here by analysis of the mouse and human FAT/CD36 genes demonstrated that the promoter described previously is the proximal promoter that does not respond to PPAR ligands. Both genes have another independent promoter located far upstream, which indirectly responds to the ligands.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- (4-Chloro-6-(2,3-xylidino)-2-pyrimidinyl-thio)acetic acid (Wy14,643) was purchased from Tokyo-Kasei (Tokyo, Japan), and bezafibrate was from Sigma. All the thiazolidindiones were synthesized at Nippon Chemiphar (Misato, Japan). Troglitazone, pioglitazone-HCl (pioglitazone), and BRL-49653 were synthesized as described (23). Dual Luciferase reporter assay system and pGL reporter plasmids were purchased from Promega. KOD^TM polymerase and restriction enzymes were obtained from Toyobo (Osaka, Japan). [-³²P]dCTP (3000 Ci/mmol) and a random primer labeling kit were purchased from Muromachi Kagaku (Tokyo, Japan) and Takara (Osaka, Japan), respectively. Fluorescein isothiocyanate-labeled anti-human FAT/CD36 was purchased from Ylem (Rome, Italy).

Animals and Treatment-- Normal male NZB mice (5-6 weeks old) were kept under a 12-h light-dark cycle and provided with food and water ad libitum. Mice were fed either a control diet (CE7, Clea Japan) or that containing 0.05% Wy14,643 for 1-7 days. In vivo electroporation of plasmid DNA into the liver of mice was performed as described.² Briefly, plasmid DNA in 15 µl of phosphate-buffered saline was injected and square electric pulses were applied 10 times at 99-ms intervals at 25 V using an ElectroSquarePorator T820 (BTX, a division of Genetronics Inc., San Diego, CA). After injection, mice were fed control diet (CE7) or a diet containing Wy14,643 (0.05%) for 7 days. The extract of the liver was prepared, and the luciferase activities were measured.

Cell Culture and DNA Transfection-- Fao cells, a subclone of rat hepatoma HIIE cells, were cultured in Ham's F-12 medium under the conditions reported previously (22). Human hepatoblastoma HepG2 cells and human monocyte-macrophages HL-60 and THP-1 (kind gifts from Prof. Y. Kobayashi, Toho University, Chiba, Japan) were cultured in standard medium (Dulbecco's modified Eagle's medium containing 25 mM glucose, 2 mM glutamine, and 10% bovine serum). 3T3L1 pre-adipocytes were cultured in standard medium. Adipocyte differentiation was induced as described previously (23). A PPAR ligand was added to the medium as a concentrated solution in Me₂SO dissolved completely by sonication. The final concentrations of the ligands were: Wy14,643, 50 µM; bezafibrate, 200 µM; troglitazone, 1 µM; pioglitazone, 1 µM; BRL-49653, 1 µM. The same concentrations were used in the cell culture studies and reporter assays.

Transfection was performed in 24-well plates with SuperFect^TM (Qiagen) for Fao cells and TransFast^TM (Promega) for 3T3L1 cells according to the manufacturers' protocols. For transactivation studies, we used 0.02 µg/well pRL-TK DNA as an internal control (Promega) and 0.98 µg/well reporter DNA. The transfected cells were cultured in growth media in the presence or absence of PPAR ligands. The luciferase activities were measured using the Dual Luciferase assay system (Promega) according to the manufacturer's protocol.

RNA Preparation and Northern Blotting-- Total RNA was prepared from the mouse liver and cultured cells by the acid guanidinium isothiocyanate-phenol-chloroform extraction method (24). Northern blotting analysis was carried out essentially as described previously (22), except that hybridization was performed in Express Hyb^TM hybridization solution (CLONTECH) at 68 °C for 3 h. After hybridization, the membranes were washed and autoradiographed with an intensifying screen at 80 °C. For quantitative analysis, a BAS2000 image analyzer (Fujix, Tokyo, Japan) was used. The cDNAs to detect individual isoforms of the mouse FAT/CD36 mRNA were prepared by amplification of total cDNA using following the exon 1A- or 1B-specific primers: 1A-left, 5'-CTTGTGGCCTTGCACTCTCTCATCG; 1A-right, 5'-TCAGAAGGCAGTACACAGAAG; 1B-left, 5'-TGAGGACTTTTCTTCTTCACAGCTGCC; 1B-right, 5'-TCAGAAGGCAGCTGTGAAGAAG. After PCR, the fragments were cloned and their identities were confirmed by sequencing. The cDNA probes were prepared by PCR of the plasmid DNAs. Isolation and identification of other cDNAs used as probes have been described previously (22).

RT-PCR Analysis of the FAT/CD36 mRNA Isoforms-- For the mouse mRNA analysis, total RNA was prepared from the liver, intestine, muscle, and fat of mice fed either a control diet or that containing Wy14,643 for 7 days as described above. The cDNAs were synthesized from 5 µg of total RNA using oligo(dT) primer and reverse transcriptase (CLONTECH). The FAT/CD36 isoform cDNAs were amplified by 35 cycles of PCR using the primer combinations of exon 1A- or exon 1B-specific left primer and exon 2-specific common right primer. For the human mRNA analysis, total RNA was prepared from HepG2 cells and THP-1 cells treated with PPAR ligands and the cDNAs were synthesized from 5 µg of total RNAs using oligo(dT) primer or random primers and reverse transcriptase (CLONTECH). The FAT/CD36 isoform cDNAs were amplified by 35 cycles of PCR using the combinations of an exon 1A-specific left primer (5'-ACACAACCAGAGTCTTCGGTGTTAA) or exon 1B-specific left primer (5'-GCATTTGATTGAAAAATCCTTCTTAGCC) and an exon 2-specific common right primer (5'-GGTCTCCAACTGGCATTAGAATACC). Reaction products were separated by agarose gel electrophoresis and stained with ethidium bromide. All the major DNA fragments were cloned into the HincII site of pUC18 after treatment with KOD polymerase to make their ends blunt and sequenced for their identification.

5'-RACE Cloning of FAT/CD36 mRNA-- The cDNAs containing the 5' end sequences of mouse and human FAT/CD36 mRNAs were obtained by a 5'-RACE method (22). The first strand cDNA was synthesized from 2 µg of poly(A) RNA isolated from the livers of mice fed Wy14,643 for 5 days or from 10 µg of total RNA isolated from HepG2 cells. The cDNA was divided into three tubes and tailed with 10 units of terminal transferase (Takara) and 0.2 mM dCTP at 37 °C for 3, 10, or 30 min. The mixture of the tailed cDNAs was amplified by PCR using oligo(dG) and a mouse FAT/CD36-specific primer, 5'-TCCAGTAATGAGCCCACAGTTCCGAT, corresponding to nucleotides 84-119 of the published rat sequence (4) (GenBank^TM accession no. NM031561), or a human FAT/CD36-specific primer, 5'-GGTCTCCAACTGGCATTAGAATACC, corresponding to the sequence in the exon 2a (25-27) in Fig. 4. After PCR, KOD polymerase was added to the PCR mixture to make the ends of the amplified DNA fragments blunt. The DNA fragments were purified using a column (Qiagen) and cloned into the HincII site of pUC18. The transformants on Escherichia coli JM109 were randomly selected and sequenced to identify FAT/CD36 cDNA clones.

Cloning of a New Promoter of the Mouse and Human FAT/CD36 Gene by PCR-- A possible new promoter sequence of the mouse FAT/CD36 gene was isolated by degenerate PCR using a mouse GenomeWalker^TM kit (CLONTECH) according to the manufacturer's protocol. The primer of the following sequence was designed from the new cDNA sequence obtained above: 5'-ATAACGGCTTCCAGTAATGATCCCACAGT.

After a second PCR using the same primers as described above and the adapter primer 2 supplied with the kit, the blunt-ended PCR products were purified on agarose gels and cloned into the HincII site of pUC18. Similarly, the previously reported promoter region was cloned by PCR using the following primers (4): R1, 5'-ATAACGGCTCCAGTAATGAGCCCACAGT; R2, 5'-TCCAGTAATGAGCCCACAGTTCCGAT.

For a possible new and the previously reported promoter sequences of the human FAT/CD36 gene, the same method as above was employed. Primers of the following sequences were designed from the new cDNA sequence obtained as described above and the published sequence (27): MR1, 5'-AGAAGTCCGATGAAAGAGCACAAGGC; MR2, 5'-GGATGCTAGTTGGGAAGACCAGGGAA; VR1, 5'-GAAGGATTTTTCAATCAAATGCTCCAACA; VR2, 5'-CAATTCTTAATAGGATCAAATCTTCCTGT.

After nested PCR, the blunt-ended PCR products were purified on agarose gels and cloned into the HincII site of pUC18.

Isolation of the Genomic BAC Clones for the Mouse and Human FAT/CD36 Genes-- High density filters of BAC Mouse ES (Rel. II) genomic library (IncyteGenomics, Inc.) were screened using the mouse cDNA for the coding region of FAT/CD36 (4, 22) under the conditions recommended by the manufacturer. Six positive BAC clones were purchased and analyzed by Southern blotting to choose a clone containing a long promoter region. For the human genomic clones, high density filters of BAC Human (Rel. II) genomic library (IncyteGenomics, Inc.) were screened using the cDNA fragment obtained by RT-PCR as described above under the conditions recommended by the manufacturer. Four positive BAC clones were purchased and analyzed by Southern blotting to choose a clone containing a long promoter region.

Sequencing-- For sequencing of the plasmids containing long inserts, a series of plasmids containing a transposon at a random site were constructed using GPS^TM-1 genome priming system (BioLabs). A SequiTherm EXCEL II Long-Read^TM DNA sequencing kit (Epicentre Technologies) was used for the reaction, and the fragments were analyzed by a DNA sequencer (LI-COR model 4000).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Induction of FAT/CD36 mRNA Expression by a PPAR Ligand, Wy14,643, Is Delayed in Mouse Liver-- We have shown previously that the PPAR ligand Wy14,643 induces expression of FAT/CD36 mRNA in the mouse liver. We also noticed a longer lag time before induction of the mRNA than other PPAR target gene mRNAs (22). To verify the delayed induction of FAT/CD36 mRNA by the ligand, we carried out a detailed time-course study to quantify the changes in the levels of FAT/CD36 and typical PPAR ligand-inducible mRNAs such as peroxisomal bifunctional enzyme (HD), liver fatty acid-binding protein (L-FABP), and mitochondrial long-chain acyl-CoA synthetase (LACS) (28) (Fig. 1, A and B). Upon intraperitoneal injection of Wy14,643, all except FAT/CD36 mRNA were markedly induced in the mouse liver after a time lag of a few hours, as evidenced by Northern blotting. Feeding a diet containing Wy14,643 also largely elevated the levels of the mRNAs in a day, and the induced levels were maintained for several days. In contrast, the FAT/CD36 mRNA remained at a low level for 1 day and gradually increased, reaching the maximum level at day 3-4. Thus, it is clear that the PPAR ligand-induced expression of FAT/CD36 mRNA was significantly delayed as compared with other typical PPAR-regulated mRNAs.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Time course of FAT/CD36 mRNA induction by Wy14,643 in the mouse liver (A and B) and in rat hepatoma Fao cells (C). For short time-course study (hours), Wy14,643 was injected intraperitoneally, and for long time-course study (days), a diet containing Wy14,643 was given to NZB mice. Total RNA (5 µg) isolated from individual livers was subjected to Northern blotting analysis using the cDNAs for FAT/CD36 (coding region), HD, LACS, L-FABP, and apoE (control for RNA loading). A and B, autoradiograms (A) and relative changes (B) of the mRNA levels after quantification using a BAS2000 image analyzer are shown. C, Wy14,643 was added to the medium of rat hepatoma Fao cells in the confluent stage at time 0 and the cells were collected at the times indicated for total RNA isolation. The levels of mRNA for FAT/CD36, Aox, HD, L-FABP, LACS, and apoE were measured by Northern blotting.

PPAR Ligand-induced Expression of FAT/CD36 mRNA Is Also Delayed in Cultured Hepatoma Cells-- To examine whether the delayed induction of FAT/CD36 mRNA by the PPAR ligand in the liver was caused by secondary effects of the ligand-induced changes in other tissues, we next measured changes in the levels of the mRNAs in the PPAR-responsive hepatoma Fao cells (Fig. 1C). The mRNAs for the established PPAR-target genes such as AOx, HD, L-FABP, and LACS were induced by the addition of Wy14,643 after a lag of 4-6 h, whereas FAT/CD36 mRNA was induced after a much longer lag of 10-12 h. Thus, the mechanism of the PPAR-dependent transcriptional activation of the FAT/CD36 gene is not exactly the same as that of other typical PPAR target genes. Replication of the in vivo delayed response in the cultured cells indicated that metabolism or signaling events in extrahepatic tissues do not play an essential role in the induction mechanism.

Mouse FAT/CD36 Gene Has a New Upstream Promoter-- To study the molecular mechanism of the delayed induction of mouse FAT/CD36 gene expression by a PPAR ligand, we analyzed the promoter structure of the gene. cDNA cloning of the 5' end of the mRNA from the liver of Wy14,643-fed mice by the 5'-RACE method yielded a clone containing a 5' end sequence different from the published sequence. BLAST search of the unique 72-bp 5' end sequence showed good homology with a rat FAT/CD36 cDNA clone (29) (Fig. 2A), suggesting that the sequence was not a cloning artifact but was transcribed from the first alternative exon and that it has some physiological significance.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   The structure of the new mouse FAT/CD36 cDNA (A) and the corresponding region of the gene (B). A, alignment of sequences of the new mouse FAT/CD36 cDNA, previously reported cDNA and a rat cDNA isoform. The nucleotide sequence of the new mouse FAT/CD36 cDNA cloned by the 5'-RACE method is aligned with those of previously published mouse cDNA (5) and a rat isoform cloned from taste buds (29). The dots indicate identical nucleotides. The numbers on the right refer to nucleotide position from the 5' end of the new clone. B, genome organization around the dual promoter region and the sequenced PCR clones and BAC subclones. The relevant restriction sites are shown: B, BamHI; E, EcoRV; H, HindIII; S, SmaI; Sp, SphI.

To determine the organization of the mouse FAT/CD36 gene, we first cloned the previously reported promoter and the possible adjacent upstream sequences of the new 72-bp sequence using a GenomeWalker^TM kit. After confirming the cloned genomic fragments by sequencing, a BAC mouse genomic library was screened using the cDNA fragment for the coding region (4). The DNAs from six positive BAC clones were analyzed by Southern blotting for the presence of both the new 72-bp sequence and the published FAT/CD36 cDNA sequence, again confirming that both sequences were included on each BAC insert. All the BAC inserts also included the two possible promoter sequences. These results excluded the possibility that the new 72-bp sequence was derived from a different place from the FAT/CD36 gene by a cloning artifact. One BAC clone was chosen and further analyzed by subcloning and sequencing, as summarized in Fig. 2B. The sequences of the BAC subclones from a single clone completely overlapped with those of the GenomeWalker^TM fragments, and the distance between the two exons was estimated by Southern blotting analysis as 16 kb.

To further confirm that our new cDNA and the published one arose from the same gene, we next analyzed RT-PCR cDNA clones by sequencing all the major bands shown in Fig. 3B. As summarized in Fig. 3A, the cDNA fragments containing exon 3 and exon 4 sequence contained either exon 1A or exon 1B sequence and no clones contained both of the exons. Thus, at least three mRNA species with different 5'-noncoding regions were identified by cloning and sequencing. These results indicated that the mouse FAT/CD36 gene has two alternative and independent first exons, and the new first exon is located 16 kb upstream from that published previously (5).

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Summary of the promoter region of the mouse FAT/CD36 gene (A) and RT-PCR analyses of the FAT/CD36 mRNA isoforms (B). A, the 5' end structures of FAT/CD36 mRNA isoforms produced by alternative promoter selection and alternative splicing. The structures of the three mRNA isoforms were determined by cloning and sequencing the three major RT-PCR products shown in B. B, RT-PCR analyses. cDNAs were synthesized from total RNA isolated from the tissues of mice fed either a control diet or that containing Wy14,643. Each isoform cDNA was amplified by 35 cycles of PCR using exon 1A- or exon 1B-specific left primer and a common right primer (in exon 2) (shown in A). The reaction products were separated in a 2.0% agarose gel and stained with ethidium bromide. All the major products were cloned and sequenced for their identities. The products with asterisks were not related to the FAT/CD36 sequence.

During preparation of this report, we performed BLAST search again in GenBank^TM data base and found a rat genomic sequence (GenBank^TM accession no. AF317787) corresponding to the upstream promoter of the mouse FAT/CD36 gene. Comparison of the sequences revealed that several dispersed regions in the promoters from position 1 to 1,800 relative to the transcriptional start site are conserved well. In particular, more than 90% of the nucleotides in the regions from position 1 to 160 are identical between the two genes, further confirming the unique structures of the FAT/CD36 genes.

Human FAT/CD36 Gene Also Has a New Upstream Promoter-- To examine whether the dual promoter structure of the FAT/CD36 gene found in the mouse is conserved in the human gene, we first analyzed the 5' end structure of the FAT/CD36 mRNA by the 5'-RACE method. The cDNA synthesized from the total RNA from HepG2 cells was amplified by PCR using a primer designed according to the published sequence of exon 2A (27) (see Fig. 4A), and DNA fragments 150-200 bp in length were analyzed after cloning. Sequencing of 30 clones revealed that one clone contained the composite sequence of exon 2A and a new exon. Based on this new sequence, we cloned the possible new promoter sequence using a GenomeWalker^TM kit and then the genomic BAC clones from the genomic library as described for the mouse clones.

View larger version (49K): [in this window] [in a new window]   Fig. 4.   Identification of human FAT/CD36 mRNA isoform induced by PPAR ligands by RT-PCR analysis. cDNAs were synthesized from total RNA isolated from control HepG2 and THP-1 cells or those treated with the PPAR ligand bezafibrate or PPAR ligand pioglitazone. Each isoform cDNA was amplified by 35 cycles of PCR using an exon 1A- or exon 1B-specific left primer and a common right primer (in exon 2 or exon 3) as schematically shown (A). Results of semiquantitative RT-PCR analysis of RNA from HepG2 cells (B) and that from THP-1 cells (D) are shown. Absence of mRNA containing exon 1B sequence in HepG2 cells was confirmed by extensive PCR (C). The reaction products were separated in a 2.0% agarose gel and stained with ethidium bromide. All the major products were cloned and sequenced for their identities.

The results of Southern blotting analysis and sequencing of ~4.5 kb around the new exon and ~5.3 kb around the previously reported first exon are summarized in Fig. 5, together with the mouse gene. The human FAT/CD36 gene has two promoters separated by a distance of 13 kb, as estimated by Southern blotting analysis. These results confirmed that the dual promoter structure is conserved at least in mouse and human, suggesting some physiological significance of the unique structure (see below).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Conserved dual promoter structures of the mouse and human FAT/CD36 genes. The dual promoter structures of the mouse (A) and human (B) FAT/CD36 genes are shown in scale. The horizontal bars below the promoters show the regions where the sequences in kilobases were determined. The distances between the two promoters were estimated by Southern blotting for the mouse gene or determined by comparison with the published sequences for the human genome.

After completion of this study, we found these sequences in the data bases of the Human Genome Project. Comparison of the two sets of sequences showed that our sequence around the proximal promoter was 99-100% identical to that in the public data base and that of the distal promoter was 98-100% identical. The identities of the intron sequences were 98-100, except a small region of ~100 bp in the middle of the first intron with 87% identity. The distance between the two promoters based on the published sequences is ~14 kb instead of the 13 kb suggested by Southern blotting, also indicating that the two promoter sequences come from a single gene but not from two closely related genes.

Only the New Distal Promoter of Mouse FAT/CD36 Gene Responds to PPAR and PPAR Ligands-- Previous analyses of the induction of FAT/CD36 mRNA used the cDNAs for the common coding region as probes and did not distinguish between the two alternative first exons. To determine which promoter responded to PPAR and PPAR ligands, we used alternative exon-specific cDNA fragments for Northern blotting and the oligonucleotides for RT-PCR analysis.

We first performed Northern blotting analysis of total RNA isolated from the liver and intestine of mice fed a diet containing a PPAR ligand, Wy14,643, for up to 5 days using exon-specific cDNA probes (Fig. 6A). The level of FAT/CD36 mRNA detected by the coding region increased gradually both in the liver and intestine, whereas that of the mRNA detected by the exon 1A-specific probe gradually increased only in the liver and was not detected by the exon 1B-specific probe. Similar analysis to examine the responsiveness of the FAT/CD36 gene to PPAR ligands was carried out using the total RNA isolated from 3T3L1 cells treated with PPAR or PPAR ligands (Fig. 6B). FAT/CD36 mRNA as a whole was induced by both PPAR and PPAR ligands. Hybridization using the exon-specific probes revealed that only the mRNA containing the sequence of 1A was detected as increased.

View larger version (52K): [in this window] [in a new window]   Fig. 6.   Identification of the FAT/CD36 mRNA isoforms induced by the PPAR ligands in the mouse tissues (A) and 3T3L1 cells (B) by Northern blotting. A, total RNA isolated from the livers (Liver) and proximal (Intest. 1) and distal (Intest. 2) intestine of mice fed either a control diet or that containing 0.05% Wy14,643 for 1-5 days as indicated was subjected to Northern blotting analysis using the cDNAs for peroxisomal HD, L-FABP, the coding region of FAT/CD36, exon 1A of FAT/CD36 (CD36(IA)), exon 1B (CD36(IB)), and apoAI (RNA loading control). B, total RNA isolated from the differentiated 3T3L1 cells at time 0 (Start) and after 48 h of culture in the absence (None) or in the presence of a PPAR ligand as indicated (Wy, Wy14,643; Bez., bezafibrate; Piog., pioglitazone; Trog., troglitazone; BRL, BRL-49653) was subjected to Northern blotting as described above.

As we could not detect the mRNA containing the sequence of exon 1B with the exon-specific cDNA probe by Northern blotting, perhaps because of the limited length of the cDNA with low specific activities of ³²P, we carried out RT-PCR analysis to confirm that the exon 1B mRNA was not induced by the PPAR ligands. RT-PCR analysis of the mRNA from the liver, intestine, muscle, and fat of mice fed a control diet or that containing Wy14,643 showed that the induced FAT/CD36 mRNA species in the liver contained exon 1A but not exon 1B sequences and that neither of these exons were transcribed in the intestine (Fig. 3B). These results indicated that both promoters of the FAT/CD36 gene are active in the liver, muscle, and fat but only the distal promoter responds to Wy14,643 at least in the liver and that a completely different mechanism is operating in the intestine.

Both PPAR and PPAR Ligands Induce FAT/CD36 mRNA from the Distal Promoter of the Human Gene in THP-1 but Not in HepG2 Cells-- It has been established that the human FAT/CD36 mRNA is induced by PPAR ligands in monocyte-macrophages (30). The proposed direct mechanism (7) was based on the incomplete genome structure, and we found a dual promoter structure in the human gene as in that of the mouse. To examine which promoter responds to PPAR ligands and the effects of PPAR ligands, we analyzed the expression of FAT/CD36 at both the protein and mRNA level.

We first examined the effects of PPAR ligands on the expression of FAT/CD36 mRNA in two human monocyte-macrophage-derived cell lines, HL60 and THP-1, by Northern blotting. Total RNA from the cells treated with the PPAR ligand bezafibrate or PPAR ligand pioglitazone was probed with the FAT/CD36 cDNA for the coding region (Fig. 7A). No signal was detected in the RNA from HL60 cells under any conditions, whereas strong induction of the mRNA by either PPAR or PPAR ligand was observed in THP-1 cells. We next confirmed the cell surface expression of FAT/CD36 protein in THP-1 cells by flow cytometry using a monoclonal antibody to FAT/CD36 (Fig. 7B). A significant and similar increase in level of immunoreactive FAT/CD36 was detected on THP-1 cells treated with the PPAR agonist Wy14,643 or PPAR ligand pioglitazone. Thus, both PPAR and PPAR were equally effective in induction of FAT/CD36 mRNA and protein in the human monocytic cell line THP-1.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   PPAR ligand-induced expression of FAT/CD36 in human monocyte-macrophage THP-1 cells. A, Northern blotting analysis of FAT/CD36 expression in two human monocyte-macrophage cell lines, HL-60 and THP-1 cells. Cells were treated with the PPAR ligand bezafibrate (Bez) or PPAR ligand pioglitazone (Pio) for 24-48 h. Total RNA was isolated and analyzed by Northern blotting using the cDNA probes for human FAT/CD36 (FAT/CD36) and mouse ribosomal S14 protein (S14, RNA loading control). B, flow cytometric analysis of the expression of FAT/CD36 protein on THP-1 cells. After treatment of THP-1 cells with the PPAR ligand Wy14,643 or with PPAR ligand pioglitazone for 48 or 72 h, the cell-surface expression of FAT/CD36 was determined by flow cytometry using an anti-human FAT/CD36 antibody. Fluorescence distribution of the counted cell population is shown.

To determine which promoter responds to PPAR ligands, we carried out RT-PCR analysis using the alternative exon-specific PCR primer and a common primer, as shown in Fig. 4A. As we cloned the cDNA with a new 5' end from HepG2 cells, we analyzed RNA isolated from HepG2 and THP-1 cells. In HepG2 cells, no amplification of the DNA fragment was observed when primers were chosen to detect the mRNA containing alternative exon 1b, but amplification was detected using the primers for exon 1a (Fig. 4B). No increase in level of the mRNA by bezafibrate was observed in HepG2 cells even with excess amplification cycles (Fig. 4C), suggesting that only the distal promoter is active but not responsive to PPAR ligands in HepG2 cells under the conditions used. In THP-1 cells, in contrast, both promoters were active and only the distal promoter responded to both PPAR and PPAR ligands (Fig. 4D).

Two Promoters of the FAT/CD36 Gene Are Independently Active but Did Not Respond to PPAR Ligands in Reporter Assay-- To directly examine whether the two 16-kb separated promoters of the mouse FAT/CD36 gene are independent and only the distal promoter responds to PPAR ligands as in the body, a reporter assay of the promoters in PPAR-responsive Fao cells and in PPAR-responsive 3T3L1 cells was carried out (Fig. 8). We subcloned the long promoter sequences from a BAC clone and prepared a series of distal deletion promoters because long promoter sequences sometimes contain negative elements to suppress the in vitro promoter activities (31). Basal activities of the two series of promoters of the mouse gene were higher than those of TK promoter in both Fao and 3T3L1 cells, indicating that the two promoters are independently active in these cells. Addition of the ligands for PPAR or - showed no effect on the promoter activities of any construct in Fao or 3T3L1 cells, although the ligands significantly activated the artificial PPRE tandem promoter in the cells. We further examined the PPAR ligand responsiveness of the promoter constructs by an in vivo gene transfer method using the livers of living mice,² but no ligand-induced transcriptional activation of the reporter gene was observed (data not shown). Thus, we concluded that the two cloned promoter sequences of the mouse FAT/CD36 gene did not contain the cis elements sufficient for transcriptional activation by the PPARs and their ligands. We also tried to analyze the two promoter sequences of the human FAT/CD36 gene by a reporter assay using the PPAR ligand-responsive cell line THP-1. However, no transfection reagents gave good results. By an electroporation method, we confirmed that both promoters of the human gene were independently active in THP-1 cells without co-transfection of nuclear receptor-expressing vectors, although the efficiency of transfection and expression of the reporter gene was quite low (data not shown). Further improvement of the method will be necessary for quantitative analysis of PPAR ligand responsiveness of the two promoters in THP-1 cells.

View larger version (19K): [in this window] [in a new window]   Fig. 8.   Effects of PPAR ligands on expression levels of the proximal and distal promoter-driven reporter genes. Fao cells (A) and 3T3L1 cells (B) were transfected with the indicated reporter plasmids and control plasmids. The cells were cultured with or without the indicated PPAR ligands for 48 h. Dual LuciferaseTM system was used for normalization of the activities. Relative promoter activities of each construct in the presence of a PPAR ligand to those in the absence of the ligand are shown. The values given are the averages of data from more than three experiments performed in duplicate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

FAT/CD36 is a major receptor responsible for the uptake of oxidized low density lipoprotein in macrophages and a transporter for long-chain fatty acids in the muscles and is thus relevant to a number of diseases, including atherosclerosis and myocardiopathy. Regulation of its expression is an important issue, but previous studies have been based on the incomplete promoter structure of the FAT/CD36 gene. The results presented here demonstrated that the mouse and human genes have two independent promoters and only the newly identified upstream promoters respond to the PPAR ligands, although not directly.

As a unique feature of FAT/CD36 gene expression, we noticed that the mRNA is induced by PPAR ligands after a long lag in the liver and hepatoma cells (22, 32, 33). The simple PPAR-RXR heterodimer model (33) alone cannot explain the difference in the time course of induction, but the mechanism of PPAR ligand-induced expression of the human FAT/CD36 gene in monocyte-macrophages was explained previously by a simple model (7). Although the possibility that this discrepancy could be caused by differences in species and/or cell types could not be excluded, we read the data to support the human model as insufficient. The DR-1 sequence taken as PPRE in this previous study is not the typical sequence, and the promoter sequence does not contain an important PPRE adjacent sequence proposed by Juge-Aubry et al. (34). Data indicating binding of the naked DR-1 DNA fragment and PPAR in extracts prepared not from monocyte-macrophages but from receptor-abundant cells do not always indicate that the DR-1 in the promoter sequence binds the receptor in vivo. Recently, it was reported that binding of transcription factors and DNA differs between naked DNA fragment binding assay and chromatin-based assay (35). Furthermore, the reported transactivation ratio of the value with the receptors and their ligands to that without them was too low to explain the response of the endogenous gene (7).

The significance of the complicated promoter structure of the FAT/CD36 gene can be interpreted in two ways. First, multiple independent promoters should make it possible to control transcription of the gene by diverse signals in a cell type-specific manner, exhibiting multiple functions in various cells. Second, the differences in the 5'-untranslated region may affect the stability of the mRNA and/or the efficiency of translation, making post-transcriptional control possible. Recently, it was proposed that the 5'-untranslated region of the human FAT/CD36 mRNA forms a complex folded pattern with numerous hairpin loops and is important for glucose-mediated control of FAT/CD36 translation (36). We examined the changes in the secondary structures of the mouse and human FAT/CD36 5'-untranslated regions. The first upstream open reading frame, which was shown to be essential for translational control to endow glucose responsiveness, was lost in the transcript using the upstream promoter. This may be related to the mechanism to independently control FAT/CD36 expression at the translational level of the downstream mRNA by a glucose signal and at the transcriptional level of the upstream transcript by a PPAR ligand. However, the mouse may not use the same mechanism as the human gene because the changes in the mouse 5'-untranslated regions induced by the alternative promoters are not similar to those in human. The physiological significance of the unique structures of the FAT/CD36 genes must await further analysis.

The mouse FAT/CD36 mRNA is induced by PPAR ligands in a PPAR subtype- and tissue-specific manner (22). The activation mechanism, however, is not the same as that of typical PPAR target genes. First, the time course of the induction is significantly delayed as compared with other mRNAs. The gene has a complex promoter structure in which a consensus and functional PPRE was not found. Furthermore, a reporter assay using the two cloned promoters of the mouse gene could not replicate the in vivo response. These results suggested that transcriptional activation of the mouse CD36/FAT gene by the PPAR ligands, at least in the liver, is indirectly dependent on PPAR.

On the human FAT/CD36 gene, we also found that the previously described proximal promoter does not respond to PPAR ligands, in contrast to the published results (7). Induction of the mRNA by the PPAR ligands was confirmed and supplemented by the observation that the PPAR ligands are also effective. We feel that the data of reporter and a gel shift assays to support the model to explain the induction mechanism by direct activation of the proximal promoter are not conclusive as discussed above. Furthermore, their data indicating that the DR-1 motif in the promoter binds PPAR but not PPAR is contradictory to our observation that the mRNA was induced by both PPAR and PPAR ligands at concentrations that showed PPAR subtype specificities. We showed that the distal promoter of the human gene responded to PPAR ligands by RT-PCR analysis of the 5' end of the mRNA. We could not find a typical PPRE (34) in the distal promoter sequence of the human gene as in the mouse gene. From these similar results of studies of the structure and expression of the mouse and human genes, we inferred that the expression of the human FAT/CD36 gene is indirectly regulated by the PPARs.

If the indirect mechanism is the case, detailed analysis of the structure and function of the distal promoter found in this study will be of great importance because an indirect mechanism should involve as yet unidentified steps that may be largely dependent on the cell type. The roles of PPAR and FAT/CD36 in cardiovascular biology have attracted considerable attention because the initial reports of the positive effects of PPAR ligands on the expression of FAT/CD36 in monocyte-macrophages (6, 7), and the regulatory mechanisms have been shown to be complex. An anti-atherogenic role of PPAR was demonstrated from several aspects (37-40), although the increased expression of FAT/CD36 is considered to be involved in induction of foam cell formation (7). The opposite effect of the PPAR ligands on the expression of FAT/CD36 and SR-A, another scavenger receptor for oxidized low density lipoprotein (41), is also paradoxical as both scavenger receptors are induced on macrophages by cytokines (37, 42). Several attempts have been made to explain the mechanism, including a balancing model between influx by FAT/CD36 and SR-A and efflux by ABCA1 of lipid (39, 40), and an alternative signaling model for the anti-inflammatory effects of the PPAR ligands (43). All of the models proposed previously assumed that the expression of FAT/CD36 itself is directly activated by PPAR and its ligands. However, our observations in the present study suggested an indirect mechanism, which may provide insight to resolve the complex regulatory mechanism.

The mechanism of indirect activation by PPAR remains unclear, and further experimentation is required. However, as the PPAR ligand-responsive promoter, we identified the upstream promoters of the mouse and human genes, where we found several possible binding sites for lipid metabolism-related transcriptional factors including SRE-binding protein and Sp-1 (Fig. 9). Involvement of these factors in an indirect activation mechanism is an interesting possibility of coordination of a homeostatic balance between fatty acids and sterols at the transport process.

View larger version (65K): [in this window] [in a new window]   Fig. 9.   The mouse and human promoter sequences of the FAT/CD36 genes. The putative transcriptional start sites are +1. The putative SRE sites and Sp-1 sites are underlined.

There are increasing number of reports on the diverse functions of PPARs. However, most previous studies on characterization of genes that are transcriptionally activated by PPAR ligands have been limited to identification of PPRE in the promoter regions by in vitro analysis and to presentation of the responsiveness to the ligands by reporter assays that are usually accompanied by co-transfection of the receptors in non-differentiated cells. It is noteworthy that the responsiveness of the typical PPAR-responsive peroxisomal genes that contributed to the establishment of the PPAR-RXR model is atypical and observed only in the rodent liver (44). Hsu et al. (45) showed that the forced high level expression of PPAR in HepG2 cells endowed a few but not all of the PPAR-responsive genes in the rodent liver with responsiveness to the ligand in human cells. These results strongly suggested that some factors in addition to PPAR and RXR play essential roles even in the PPAR ligand-dependent direct transcriptional activation of several genes. It will be necessary not only to classify the genes by stereotypic experiments but more importantly to understand the PPAR mediated physiological responses.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EBI Data Bank with accession number(s) MU511016 AF434765 and MU501026 AF434766.

^¶ To whom correspondence should be addressed: Dept. of Biochemistry, Meiji Pharmaceutical University, Noshio 2-522-1, Kiyose, Tokyo 204-8588, Japan. Tel./Fax: 81-424-95-8474; E-mail: motojima@my-pharm.ac.jp.

Published, JBC Papers in Press, February 26, 2002, DOI 10.1074/jbc.M110158200

² Fujishiro, K., Fukui, Y., Sato, O., Kawabe, K., Seto, K., and Motojima, K. (2002) Mol. Cell. Biochem. (in press).

	ABBREVIATIONS

The abbreviations used are: FAT, fatty acid translocase; BAC, bacterial artificial chromosome; Aox, acyl-CoA oxidase; PPAR, peroxisome proliferator-activated receptor; PPRE, peroxisome proliferator-activated receptor-responsive element; RXR, retinoid X receptor; SRE, sterol-responsive element; RT, reverse transcription; HD, hydratase-dehydrogenase; L-FABP, liver fatty acid-binding protein; LACS, long-chain acyl-CoA synthetase; apo, apolipoprotein.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES