Regulation of human delta-6 desaturase gene transcription: identification of a functional direct repeat-1 element.

The rate-limiting step in 20:4(n-6) and 22:6(n-3) synthesis is the desaturation of 18:2(n-6) and 18:3(n-3) by Delta-6 desaturase. In this report, we demonstrate that n-6 and n-3 PUFAs suppressed the hepatic expression of rodent Delta-6 desaturase by inhibiting the rate of Delta-6 desaturase gene transcription. In contrast, consumption of the peroxisome proliferator-activated receptor (PPAR)alpha activator WY 14,643 significantly enhanced the transcription of hepatic Delta-6 desaturase by more than 500%. Transfection reporter assays with HepG2 cells revealed that the PUFA response region for the human Delta-6 desaturase gene involved the proximal promoter region of -283/+1 human Delta-6 desaturase gene, while the WY 14,643 response element (RE) was identified as an imperfect direct repeat (DR-1) located at -385/-373. The WY 14,643 induction of the human Delta-6 desaturase promoter activity was dependent upon the expression of PPARalpha. Electrophoretic mobility shift assays revealed that nuclear proteins extracted from HepG2 cells expressing PPARalpha specifically interacted with the -385/-373 DR-1 sequence of the human Delta-6 desaturase gene. The interaction was eliminated by the unlabeled PPARalpha RE of the rat acyl-CoA oxidase gene, and the protein-DNA complex was super-shifted by treatment with anti-PPARalpha. The -385/-373 sequence also interacted with a mixture of in vitro translated PPARalpha-retinoic acid receptor X (RXR)alpha, but by themselves neither PPARalpha nor RXRalpha could bind to the Delta-6 desaturase DR-1. These data indicate that the 5'-flanking region of the human Delta-6 desaturase gene contains a DR-1 that functions in the regulation of human Delta-6 desaturase gene transcription, and thereby plays a role in the synthesis of 20- and 22-carbon polyenoic fatty acids.

The mechanisms by which dietary PUFA and PPAR ␣ activators regulate hepatic ⌬ -6 desaturase and ⌬ -5 desaturase gene expression are unknown. In this report, we demon-strate that PUFAs suppress and non-PUFA PPAR ␣ ligand activators induce transcription for both the rat and human liver ⌬ -6 desaturase gene. Moreover, we have determined that the human ⌬ -6 desaturase gene contains an imperfect direct repeat-1 (DR-1) at Ϫ 385/ Ϫ 373 that imparts PPAR ␣ responsiveness to the ⌬ -6 desaturase promoter.

Dietary study
Male Sprague Dawley rats (Harlan Sprague-Dawley) were housed in a temperature-and light-controlled environment and The hepatic transcription of ⌬ -6 desaturase, fatty acid synthase (FAS), and acyl-CoA oxidase (AOX) genes in rats fed a fat-free (FF) diet or the FF diet supplemented with 0.1% WY 14,643 (WY), 10% (w/w) fish oil (FO), safflower oil (SO), or triolein oil (TO) was determined by nuclear run-on assay. Transcription activities (dpm/transcript per 10 6 dpm total RNA) were corrected for nonspecific hybridization to the pBS vector, and data are expressed as means Ϯ SE; n ϭ 4 rats/treatment. ND, nondetectable.
a Indicates a significant difference from the FF values ( P Ͻ 0.05).

Fig. 1.
Transcriptional regulation of rat liver ⌬-6 desaturase by WY 14,643 (WY) and PUFA. Nuclear run-on assays were conducted using nuclei isolated from rats fed the fat-free (FF) diet or the FF diet supplemented with 0.1% WY, 10% fish oil (FO), safflower oil (SO), or triolein (TO). FAS, fatty acid synthase; AOX, acyl-CoA oxidase. A quantitative summary of the nuclear run-on output is presented in Table 1.  Downloaded from adapted to a 3 h per day (9-12) meal-feeding regimen using a high-glucose, fat-free diet (Dyets, Bethlehem, PA) (23). After a 7-day adaptation period, the rats were randomly assigned to one of the following dietary treatments (n ϭ 5 rats per group) and fed for an additional 5 days: a high glucose, fat-free diet; the fat-free diet plus 0.1% (w/w) WY 14,643 (Chemsyn Science Labs, Lenexa, KS); or the fat-free diet supplemented with 10% fat (w/w) as triolein [99% 18:1(n-9)], safflower oil [65% 18:2(n-6)], or sterol-free, menhaden fish oil [35% 20:5 and 22:6(n-3)]. Nuclei for nuclear run-on assays, and RNA for measures of transcript abundance were isolated from the rats immediately after the last 3 h meal (see below).

RNA analysis and gene transcription
The abundance of a variety of hepatic transcripts described in the figures was determined by Northern blot analysis using total RNA extracted by the phenol-guanidinium isothiocyanate method (24). The abundance of specific transcripts of interest were quantified following hybridization with cDNA probes labeled with [␣-32 P]dCTP (Amersham, Arlington Heights, IL) using polymerase chain reaction radiolabeling or random prime la-beling (Life Technologies, Baltimore, MD) (11,16). The impact of WY 14,643 and various dietary fats on the in vivo transcription of rat liver ⌬-6 desaturase, fatty acid synthase, and acyl-CoA oxidase (AOX) was determined using the nuclear run-on assay procedure (11,25). Equivalent counts of nuclear RNA labeled with [␣-32 P]UTP (Amersham) were hybridized for 72 h at 40ЊC to filter-bound cDNAs specific for ⌬-6 desaturase, FAS, and AOX. After hybridization and washing, the membranes were exposed to X-ray film (X-OMAT-AR, Kodak, Rochester, NY). Each RNA hybrid was cut out and counted by liquid scintillation counting. Transcription and mRNA abundance data were subjected to oneway ANOVA, and treatment effects (P Ͻ 0.05) were determined as differences from the fat-free group.

Genomic cloning and reporter vector construction for human ⌬-6 desaturase
Human ⌬-6 desaturase and ⌬-5 desaturase cDNA sequences were used to BLAST search the human genomic database. Clone PAC AC004228 corresponding to the region of human chromosome 11q12.2-13.1 was found to contain all of the exons for ⌬-6 desaturase and ⌬-5 desaturase, as well as the entire region spanning the distance between the two genes. A KpnI-AviII fragment representing the sequence of Ϫ6,249 to ϩ279 was cut from the human clone AC004228. A luciferase (LUC) reporter construct containing the ⌬-6 desaturase proximal promoter region of Ϫ118/ϩ132 was prepared by cutting the Ϫ6,249/ϩ279 fragment with SacI and NaeI, and subsequently inserting the Ϫ118/ ϩ132 sequence into the SacI and SmaI sites of pGL3.LUC basic vector (Promega, Madison, WI). The Ϫ1,749/ϩ132p⌬-6 desaturase.LUC construct was prepared by removing the SacI fragment Ϫ1,749/Ϫ118 from Ϫ6,249/ϩ279 and inserting it into the SacI site of the Ϫ118/ϩ132p⌬-6 desaturase.LUC construct. The reporter Ϫ6,249/ϩ132p⌬-6 desaturase.LUC was prepared by linking the KpnI-PstI fragment of Ϫ6,249/Ϫ1,581 with the corresponding sites located in Ϫ1,749/ϩ132p⌬-6 desaturase.LUC. Constructs Ϫ417/ϩ132p⌬-6 desaturase.LUC and Ϫ283/ ϩ132p⌬-6 desaturase.LUC were generated by 5Ј-digestion of Ϫ1,749/ϩ132p⌬-6 desaturase.LUC using exonuclease III and mung bean nuclease. Mutation of the DR-1 located at Ϫ385/ Ϫ373 of the human ⌬-6 desaturase gene was accomplished using the vector Ϫ417/ϩ132p⌬-6 desaturase.LUC as the template in a polymerase chain reaction procedure that employed 5Ј-CCTC-CGGTACCCGGGGCCGGAGAG TGGGGGAGtGAGGcGaTCG-GACACG-3Ј as the forward primer (mismatched bases indicated by lower-case letters). The polymerase chain reaction DNA product was then digested with KpnI and NaeI, and ligated to the KpnI and SmaI sites of the pGL 3 basic vector (i.e., Ϫ417mp ⌬-6 desaturase.LUC). Sequence fidelity and the introduced mutations were verified by sequencing.

Site of transcription initiation
The start site of transcription for ⌬-6 desaturase was mapped by a modification of the S1 nuclease method (26). A 5Ј-end labeled, single-strand 220 nucleotide (nt) DNA fragment corresponding to the 12-231 nts upstream of the translation start codon for human ⌬-6 desaturase was synthesized using Klenow fragment. The labeled fragment was purified by electrophoresis in a 7% acrylamide gel. Total RNA (100 g) extracted from HepG2 or glioma cells was mixed with 8 ng (15,000 dpm) of single-strand probe, and the mixture dried under vacuum. The pellet was resuspended in 25 l of 80% formamide, 40 mM Hepes (pH 6.4), 1 mM EDTA, and 0.4 M NaCl, and incubated at 90ЊC for 5 min and then at 50ЊC overnight. The sample was subsequently digested with 800 units of S1 nuclease for 1 h at room temperature. The S1 nuclease-digested products were precipi- tated with ethanol. The precipitate was dried and resuspended in 1ϫ Tris-EDTA (pH 7.5), boiled with formamide loading dye, and the resulting fragments were separated by electrophoresis in a 5% polyacrylamide, 7 M urea denaturing gel.

Cell culture and transfection
HepG2 cells and CV-1 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (11,27,28). The impact of 18:1(n-9) and 20:4(n-6) on endogenous HepG2 expression of ⌬-6 desaturase was determined by quantifying ⌬-6 desaturase mRNA abundance in confluent HepG2 cells that had been incubated for 36 h with 0 M, 10 M, 50 M, or 200 M albumin-bound (fatty acid-albumin ratio of 4:1, w/w) 18:1(n-9) or 20:4(n-6) (11). The influence of the PPAR␣-specific ligand activator, WY 14,643, on endogenous ⌬-6 desaturase expression was determined by measuring the ⌬-6 desaturase mRNA abundance in HepG2 cells that had been transfected with a PPAR␣ expression vector (0.2 g pSG5.mPPAR␣) and treated for 36 h with 100 M WY 14,643 dissolved in dimethylsulfoxide or with vehicle alone. The influence of fatty acids on human ⌬-6 desaturase promoter activity was determined by transfecting HepG2 cells in 6-well plates with 1.8 g of Ϫ1,749, Ϫ668, Ϫ417, Ϫ417m, or Ϫ283p⌬-6 desaturase.LUC and incubating the cells for 36 h with 0.4% fatty acid-free BSA or 100 M albuminbound 18:1(n-9) or 20:4(n-6) (11). The effect of PPAR␣ activation on human ⌬-6 desaturase promoter activity was evaluated by Fig. 4. Identification of the transcription initiation sites for the human ⌬-6 desaturase gene. A: Transcription initiation sites for the human ⌬-6 desaturase were identified using S1 nuclease analysis and 100 g total RNA extracted from HepG2 cells or human glioma cells. Lanes 1 and 2 are adenine and thymidine DNA sequencing ladders, respectively. Lanes 3 and 4 depict the transcription initiation sites (TSS) for the ⌬-6 desaturase gene in human glioma cells and the human hepatoma HepG2 cells, respectively. TSS 1 and TSS 2 are located 177 and 143 nucleotides (nts) upstream from the translation initiation codon, AUG. Lane 5 is the 220 nt probe subjected to S1 nuclease in the absence of RNA, and Lane 6 is the free radiolabeled 220 nt probe. B: Depicts the nt sequence for the Ϫ309/ϩ141 region of human ⌬-6 desaturase gene. The gray oval denotes a candidate consensus CCATT box, and the gray rectangles denote candidate Sp1 binding sites. The two major transcription start sites are marked as TSS 1 and TSS 2, and the translation start site for the transcript is noted at ϩ180. transfecting HepG2 cells in 6-well plates with 1.8 g of Ϫ1,749, Ϫ668, Ϫ417, Ϫ417m, Ϫ283, or Ϫ118 p⌬-6 desaturase.LUC and treating the cells with 100 M WY 14,643. The dependence of ⌬-6 desaturase promoter activity on PPAR␣ was further evaluated in CV-1 cells that had been cotransfected with 0.2 g pSG5.mPPAR␣ plus 0.2 g pSG5.retinoic acid receptor X (RXR)␣. The specificity of the PPAR␣ effect was further evaluated by transfecting HepG2 and CV-1 cells with pSG5 vector lacking the open reading frame for PPAR␣, and by treating cells with WY 14,643 that had not been transfected with either pSG5 or pSG5.mPPAR, but had been treated with transfection reagent. The influence of hepatic nuclear factor 4 (HNF-4) on human ⌬-6 desaturase promoter activity was examined by transfecting CV-1 cells with the expression vector pCMV.HNF-4. Transfection of HepG2 and CV-1 cells was conducted using cells seeded onto 6-well plates and grown to 65-75% confluence. At this point, the cells were transfected with the respective vector(s) by incubating them for 12 h in a serum-free transfection medium containing the lipofectamine (Life Technology, Rockville, MD) (28). After a 12 h incubation period, the transfection medium was removed and replaced with a serum-free medium containing 10 ؊7 M insulin and dexamethasone, 10 g/ml ␣-tocopherol plus the fatty acid, or WY 14,643. Cells were harvested after 36 h treatment using lysis buffer (Promega). Luciferase activity was quantified and is expressed as relative light units (RLUs) per g protein (28). Transfection efficiency was evaluated by cotransfection with 0.2 g/well pCMV.␤gal and determining the activity of ␤-galactosidase.

Electrophoretic mobility shift assay
Nuclear proteins for use in electrophoretic mobility shift assays (EMSAs) were extracted from HepG2 cells transfected with pSG5.mPPAR␣, empty pSG5, or no vector, and treated with or without 100 M WY 14,643 for 36 h (28,29). Briefly, cells were washed twice in ice-cold PBS, scraped into 1.5 ml microfuge tubes, and centrifuged at 500 g for 20 s in a microcentrifuge. The cell pellet was then resuspended in 1 ml ice-cold buffer A [10 mM Hepes (pH 7.9), 1 mM EDTA, 10 M KCl, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 5 g/ml pepstain A, 10 g/ml leupeptin, and 2 g/ml aprotinin]. After incubation on ice for 20 min, Nonidet P-40 was added to a final concentration of 0.5%. After vigorous vortexing for 20 s, the cell suspension was centrifuged at 15,000 g for 30 s to collect the nuclei. The nuclei were resuspended in 10 vol of buffer B [10 mM Hepes (pH 7.9), 1 mM EDTA, 0.42 M NaCl, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, 5 g/ml pepstain A, 10 g/ml leupeptin, and 2 g/ml aprotinin], incubated on ice for 20 min, and centrifuged at 15,000 g for 10 min at 4ЊC. The resulting supernatants were stored at Ϫ80ЊC. In vitro translated mouse PPAR␣ and rat RXR␣ were synthesized using the TNTcoupled reticulocyte lysate system (Promega). Double-strand oligonucleotides composed of the following sequences were used for EMSAs and competition analyses: human ⌬-6 desaturase DR-1 (Ϫ385/Ϫ373), 5ЈGTGGGGGAGGGAGGAGGTCGGACA-CGG-3Ј; mutated ⌬-6 desaturase DR-1, 5Ј-GTGGGGGAGtGA-GGcGaTCGGACACGGTA-3Ј; rat AOX DR-1/PPAR-response element (RE), GGGGACCAGGACAAAGGTCAAGCAGCCAT. The DR-1/PPRE sequences are underlined, and mutated bases are shown in lowercase letters. Annealed oligonucleotides were endlabeled with [␥ Ϫ32 p]ATP (Amersham) using T4 polynucleotide kinase. A 15 l reaction containing 0.5-1.0 ng (50,000 cpm) of labeled DR-1/PPAR-RE and 4 g of nuclear extract or in vitro translated 2 l of PPAR␣ and/or RXR␣ were incubated for 30 min on ice in a buffer containing 20 mM Hepes (pH 8.0), 60 mM KCl, 1 mM dithiothreitol, 10% glycerol, and 0.2 g poly(dI-dC). In the super-shift analyses, 2 g of H-98 anti-PPAR␣ (Santa Cruz Biotechnology, Santa Cruz, CA) was incubated with nuclear protein extracts on ice for 12 h before labeled probe was added. After incubation, DNA-protein complexes were separated by electrophoresis on 5% polyacrylamide gel in Tris-glycine buffer at 4ЊC and visualized by autoradiography (28).

Transcriptional regulation of ⌬-6 desaturase by PUFA and WY 14,643
Feeding rats or mice a high-carbohydrate, fat-free diet supplemented with n-6 and n-3 PUFA lowers the hepatic mRNA abundance and enzymatic activity of ⌬-6 desaturase and ⌬-5 desaturase (15,16). On the other hand, feeding the PPAR␣-specific activator WY 14,643 leads to a marked increase in the mRNA abundance and enzymatic activity of ⌬-6 desaturase and ⌬-5 desaturase (19). Nuclear run-on assays revealed that the ingestion of safflower oil rich in 18:2(n-6) or dietary fish oil rich in 20-and 22-carbon n-3 fatty acids reduced the hepatic abundance of ⌬-6 desaturase mRNA by inhibiting the rate of ⌬-6 desaturase gene transcription 60% and greater than 95%, respectively (Table 1; Figs. 1, 2). The extent of inhibition by dietary PUFA was comparable to that of the fatty acid synthase gene, a gene whose transcription is well recognized as being inhibited by dietary PUFA (9). The inhibition of ⌬-6 desaturase gene expression was specific for n-6 and n-3 PUFA, because feeding comparable amounts of triolein [i.e., 18:1(n-9)] did not lower the rate of ⌬-6 desaturase gene transcription or the hepatic abundance of ⌬-6 desaturase and ⌬-5 desaturase mRNA (Table 1; Figs. 1, 2). In contrast to the effects of dietary PUFA, ingestion of WY 14,643 increased the level of rat liver ⌬-6 desaturase mRNA by inducing the rate of ⌬-6 desaturase gene tran-  Fig. 1). Unfortunately, the impact of dietary PUFA and WY 14,643 on the rate of ⌬-5 desaturase gene transcription could not be determined because the hybridization signal was below the level of reliable detection. Expression of the human ⌬-6 desaturase gene was also inhibited by PUFA and induced by PPAR␣ activators (Fig.  3). Specifically, treating HepG2 cells with 20:4(n-6) resulted in a dose-dependent reduction in the cellular abundance of ⌬-6 desaturase mRNA (Fig. 3), and supplementing the media with 100 M WY 14,643 significantly increased the cellular abundance of ⌬-6 desaturase in HepG2 cells that expressed PPAR␣ (Fig. 3).

Identification of the start of transcription for the human ⌬-6 desaturase
A 100 Kbp human Chr 11 fragment (Chr 11q12.2 PAC clone AC 004228) was determined to contain the entire transcribed sequences for the ⌬-6 desaturase and ⌬-5 desaturase genes, as well as containing the 11.2 Kbp intervening sequence lying between the two genes. S1 nuclease analysis revealed that transcription for the human ⌬-6 desaturase gene was initiated at multiple sites in both human glioma and liver cells (Fig. 4). The two major points for transcription initiation were located at Ϫ177 and Ϫ143 nt from the ATG codon (Fig. 4A). The presence of multiple transcription initiation sites is consistent with the fact that the human ⌬-6 desaturase gene does not appear to contain a classic TATA box (Fig. 4B). For the purpose of describing the location of cis-acting elements in the 5Ј-flanking sequence of the human ⌬-6 desaturase gene, the Ϫ177 start point is considered ϩ1. Although the human ⌬-6 desaturase gene lacks a TATA-box, the G/C rich region between Ϫ280 and ϩ1 contains several candidate binding sites for Sp1 (Fig.  4B). Moreover, a CCAAT-box motif is located at Ϫ269/ Ϫ265, and the Ϫ289/Ϫ200 region contains recognition sequences for the enhancer factors, sterol regulatory element binding protein-1 (SREBP-1) and NF-Y (22).
In this report, we demonstrate that PUFAs lower the hepatic abundance of ⌬-6 desaturase mRNA by inhibiting the rate of ⌬-6 desaturase gene transcription. This inhibi-tion of ⌬-6 desaturase gene transcription applies to both the rat and human ⌬-6 desaturase genes (Tables 1, 2). Transfection reporter assays with HepG2 cells revealed that the PUFA response sequences for the human ⌬-6 desaturase gene resided within the proximal promoter region of Ϫ283/ϩ1 (Table 2). Recently, Nara et al. (22) reported that the E-box-like sterol RE located at Ϫ222/ Ϫ231 and the NF-Y recognition site at Ϫ273/Ϫ268 are both required for PUFA suppression of the human ⌬-6 desaturase promoter. SREBP-1 and NF-Y have been implicated in the PUFA inhibition of transcription of other lipogenic genes, including FAS and stearoyl-CoA desaturase-1 (33,34). Dietary PUFAs exert their inhibitory influence by lowering the nuclear content of mature SREBP-1 protein and by interfering with the transactivation action of NF-Y (11,(33)(34)(35)(36)(37). PUFAs decrease the nuclear content of SREBP-1 in two ways. First, they inhibit the proteolytic release of mature SREBP-1 from its membrane-anchored precursor (35,37). Second, PUFAs accelerate the decay of SREBP-1 mRNA and consequently lower the abundance of SREBP-1 mRNA, which in turn leads to a reduction in the amount of membraneanchored precursor SREBP-1 protein (36). The mechanism by which PUFAs interfere with NF-Y action is unclear, but it may involve a posttranslational modification of NF-Y (34).
PPAR␣ modulates the transcription of a gene by interacting with its heterodimer partner RXR␣ and subsequently binding to a hexameric (AGGTCA) DR with a single nt spacer. The 5Ј-flanking sequence of the human ⌬-6 desaturase gene was found to contain an imperfect DR-1 ( Ϫ385 AGGGAGgAGGTCT Ϫ373 ). Binding of PPAR␣ to the Ϫ385/Ϫ373 ⌬-6 desaturase DR-1 required RXR␣ (Fig. 7), and transfection reporter analyses demonstrated that the DR-1 imparted PPAR␣ responsiveness to the human ⌬-6 desaturase promoter, i.e., expression of PPAR␣ in HepG2 and CV-1 cells significantly enhanced human ⌬-6 desaturase promoter activity in response to WY 14,643 (Fig. 5). Nevertheless, the physiological role of the DR-1 in human ⌬-6 desaturase gene transcription remains unclear because human tissues contain low amounts of PPAR␣ (27), and because the imperfect DR-1 may be a possible recognition sequence for several tran-scription factors including HNF-4, PPAR␥, PPAR␦, farnesoid X receptor, and chicken ovalbumin upstream promoter transcription factor (30,(38)(39)(40). In this regard, it is noteworthy that expression of HNF-4 in CV-1 cells did not alter ⌬-6 desaturase promoter activity, but this does not eliminate the possibility that transcription factors other than HNF-4 or PPAR␣ may recognize the DR-1 site from the human ⌬-6 desaturase gene. Nevertheless, our data strongly suggest that the Ϫ385/Ϫ373 DR-1 of the human ⌬-6 desaturase is a functional response element that plays a role in the expression of ⌬-6 desaturase and ultimately the synthesis of 20-and 22-carbon PUFA.