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

J. Biol. Chem., Vol. 277, Issue 37, 34589-34600, September 13, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/37/34589    most recent M108866200v1 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 Rim, J. S. Articles by Kozak, L. P. Search for Related Content PubMed PubMed Citation Articles by Rim, J. S. Articles by Kozak, L. P. Social Bookmarking                      What's this?

Regulatory Motifs for CREB-binding Protein and Nfe2l2 Transcription Factors in the Upstream Enhancer of the Mitochondrial Uncoupling Protein 1 Gene^*

Jong S. Rim and Leslie P. Kozak

From the Pennington Biomedical Research Center, Baton Rouge, Louisana 70808

Received for publication, September 13, 2001, and in revised form, June 20, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Thermogenesis against cold exposure in mammals occurs in brown adipose tissue (BAT) through mitochondrial uncoupling protein (UCP1). Expression of the Ucp1 gene is unique in brown adipocytes and is regulated tightly. The 5'-flanking region of the mouse Ucp1 gene contains cis-acting elements including PPRE, TRE, and four half-site cAMP-responsive elements (CRE) with BAT-specific enhancer elements. In the course of analyzing how these half-site CREs are involved in Ucp1 expression, we found that a DNA regulatory element for NF-E2 overlaps CRE2. Electrophoretic mobility shift assay and competition assays with the CRE2 element indicates that nuclear proteins from BAT, inguinal fat, and retroperitoneal fat tissue interact with the CRE2 motif (CGTCA) in a specific manner. A supershift assay using an antibody against the CRE-binding protein (CREB) shows specific affinity to the complex from CRE2 and nuclear extract of BAT. Additionally, Western blot analysis for phospho-CREB/ATF1 shows an increase in phosphorylation of CREB/ATF1 in HIB-1B cells after norepinephrine treatment. Transient transfection assay using luciferase reporter constructs also indicates that the two half-site CREs are involved in transcriptional regulation of Ucp1 in response to norepinephrine and cAMP. We also show that a second DNA regulatory element for NF-E2 is located upstream of the CRE2 region. This element, which is found in a similar location in the 5'-flanking region of the human and rodent Ucp1 genes, shows specific binding to rat and human NF-E2 by electrophoretic mobility shift assay with nuclear extracts from brown fat. Co-transfections with an Nfe2l2 expression vector and a luciferase reporter construct of the Ucp1 enhancer region provide additional evidence that Nfe2l2 is involved in the regulation of Ucp1 by cAMP-mediated signaling.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Adaptive thermogenesis can be induced by cold exposure (1, 2) and/or a high fat diet (3, 4) in brown adipose tissue (BAT)¹ through the mitochondrial uncoupling protein (UCP1). Although four homologues of Ucp1 have been identified (5-9), definitive proof establishing that uncoupling proteins are essential for thermogenesis has been shown only for UCP1 (10). UCP1 is located in the inner membrane of mitochondria, where it reduces the mitochondrial membrane potential to generate heat instead of ATP synthesis during oxidative phosphorylation (11). Overexpression of Ucp1 can be achieved pharmacologically by administration of thermogenic 3 agonists (12, 13) or genetically by using tissue-specific gene promoters (14, 15) to drive expression in transgenic mice or by the increase of UCP1 because of increased protein kinase A activity in protein kinase A RII knockout mice (16). Each of these animals with increased UCP1 has increased brown fat activity, energy expenditure, and reduced adiposity. Accordingly, determining mechanisms to increase UCP1 has practical applications to the problem of obesity.

There are two aspects of Ucp1 expression that require explanation, one is the molecular basis of its unique expression in BAT (5) and the other is its tightly controlled regulation by the hypothalamus via the sympathetic nervous system (17) in response to cold and possibly diet. A considerable body of information has accumulated showing that a 200-bp enhancer, located ~2.5 kb upstream of the transcription start site (18, 19) that contains cis-acting elements that play a critical role in the regulation of Ucp1. These elements include PPRE (20), TRE/RARE (21, 22), and cAMP responsive elements (CRE) (19). Recently, it has been shown that synergism between retinoids, isoproterenol, and thiazolidinedione regulate human Ucp1 transcription in an enhancer region located 3.5 kb upstream of the gene (23). The brown adipocyte-specific expression of Ucp1 almost certainly involves the interaction of PPAR, RXR, and PGC1 via the PPRE site. Additional regulatory elements and transcription factors are likely to be involved. The strong evidence that induction is initiated by norepinephrine action on G protein-coupled 1 and 3 adrenergic receptors (24-26) suggests that cyclic AMP (cAMP) directly regulates the expression of Ucp1 through the interaction of CREB with putative CREs in the 5'-flanking region of the Ucp1 gene. An alternative mechanism, suggested by Spiegelman and co-workers (27, 28), postulates that the adrenergic regulation of Ucp1 does not involve CREB binding to Ucp1, rather CREB activates Pgc1 expression by the protein kinase A pathway and Ucp1 is subsequently induced by the coactivation of PPAR by PGC1. However, there is no evidence that CREB is involved in the activation of Pgc1. We only know that Pgc1 mRNA levels are increased in BAT in response to cold exposure (27). It has also been reported that thyroid hormones (29), retinoids (30-32), and thiazolidinediones (33, 34) increase transcription of the Ucp1 in rodent, in vivo and in vitro.

Previous transient transfection analyses utilizing primary cell cultures from a SV40 t-antigen-induced brown adipocyte tumor showed that mutations to two of four half-site CREs in a chloramphenicol acetyltransferase-reporter construct carrying 3 kb of the 5'-flanking region almost completely abolished expression. These sites, CRE2 and CRE4, were located in the enhancer region and just 5' of the TATA box region, respectively. Mutations to CRE1 and CRE3 showed only slight reductions in reporter activity. However, it was not established whether these essential half-site CREs bind homodimers of serine 133-phosphorylated CRE-binding protein (CREB) or whether they interacted with heterodimers formed between CREB and novel transcription factors. In this study, we have demonstrated that CRE2 in the enhancer region interacted with CREB using electrophoretic mobility shift assays (EMSA). Furthermore, transient transfection assays of luciferase reporter constructs and site-directed mutagenesis indicates that CREs are involved in transcriptional regulation of the Ucp1 through interaction with phosphorylated CREB in response to cold exposure or administration of norepinephrine. We also show that two NF-E2 regulatory motifs, one of which overlaps with the CRE2 motif, bind to Nfe2l2 in a cAMP-dependent manner to control transcription of Ucp1.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Cell Culture and Transfection-- HIB-1B cells were maintained in Dulbecco's modified Eagle's medium (4,500 mg/liter D-glucose, 584 mg/liter L-glutamine, and 15 mg/liter phenol red, Invitrogen) supplemented with 10% fetal bovine serum (Invitrogen) and 0.1 mM nonessential amino acids. Medium was changed every 2 days. Reporter constructs were transiently transfected into HIB-1B cells using LipofectAMINE PLUS reagent (Invitrogen) according to the manufacturer's protocol. The day before transfection, 2 × 10⁵ cells were seeded into a 24-well cluster dish (Corning). Briefly, 0.5 µg of reporter construct was transfected with 50 ng of pRL/SV40 (Promega), a plasmid containing Renilla luciferase gene under control of SV40 promoter, in a mixture of PLUS and LipofectAMINE reagent. For the co-expression experiment, each 0.3 µg of reporter construct and expression vector were transfected with 50 ng of pRL/SV40. Transfected cells were cultured in the medium in the presence or absence 1 µM norepinephrine (Sigma) or 0.5 mM 8-bromo-cAMP (Calbiochem) for 16 h. Cell extracts were prepared, and the activity of Photinus and Renilla luciferase were determined using the dual-luciferase reporter assay system (Promega). For each construct, the activity of the Photinus luciferase was divided by the activity of the Renilla luciferase to correct for transfection efficiency. Under each treatment, the corrected activity was again divided by activity from pGL3/basic (Promega), the empty vector, to estimate the -fold increase for each construct. The -fold increase for the overexpression experiment was obtained by dividing the corrected activity by the empty vector (pCMV/tag). Each experiment was performed in duplicate dishes.

Subcloning of the 5'-Flanking Region of the Mouse Ucp1 Gene and Reporter Constructs-- The 3.1 kb of the 5'-flanking region containing the four CRE and the 220-bp BAT-specific enhancer of the mouse Ucp1 gene were obtained by PCR amplification. The 3.85-kb BglI fragment in pGEM, which was previously used in our characterization of Ucp1 (19) (note that nucleotide positions correspond to those in the Ucp1 gene as described in GenBank^TM U63418), was used as a template with forward and reverse primers, 5'-ggggtaCCGTGCACACTGCCAAATCATCTC (4379/4355, a new KpnI site is underlined) and 5'-gggagCTCCTGCAGAGCCACCTGGGCTAGG (7514/7538, a new SacI site is underlined), respectively, and subcloned into pGL3/basic using the KpnI and SacI restriction enzyme sites. To obtain the Ucp1 promoter with or without CRE4, forward primers 5'-ggggatccGAGTGACGCGCGGCTGGG (nucleotide sequences for CRE4 are shown as bold and a new BamHI site is underlined, 7261/7278) or 5'-ggggatcCGGCTGGGAGGCTTGCGCA (a new BamHI site is underlined, 7271/7289) and reverse primer 5'-gggaagcttGGGCTAGGTAGTGCCAG (a new HindIII site is underlined, 7504/7520) were used for PCR amplification and subcloned into pGL3/basic using BglII and HindIII restriction enzyme sites. For the 220 bp of BAT-specific enhancer region, the 3.85-kb BglI fragment was PCR amplified using primers 5'-ggggagCTCCTCTACAGCGTCACAGAGG (SacI site is underlined, 4841/4862) and 5'-gggctcgagAGTCTGAGGAAAGGGTTGA (a new XhoI site is underlined, 5025/5045) and subcloned into the luciferase reporter construct containing the Ucp1 promoter (give nucleotide sequences). For the rat Ucp1 enhancer region, genomic DNA from rat liver was amplified by PCR using primers 5'-gtgaaccttgctgccgctcctttgc (forward primer, the putative NF-E2 site is underlined, 2519/2494) and 5'-tgtgatgtcagctcaagacagggag (reverse primer, 2283/2308) and subcloned into the luciferase reporter construct containing the Ucp1 promoter. To generate the mutations in NF-E2 site, primer 5'-gtgaacctgtaggccgctcctttgc (forward primer, the putative NF-E2 site is underlined with mutations shown italic, 2519/2494) and 5'-tgtgatgtcagctcaagacagggag (reverse primer, 2283/2308) were used for PCR amplification. The structure of each fragment was verified by DNA sequencing. Nfe2l2 cDNA was kindly provided by Dr. Paul Ney (St. Jude Children's Research Hospital). A Nfe2l2 expression vector was made by cloning a NotI fragment into the pCMV/tag1 (Stratagene).

Site-directed Mutagenesis for CRE2 and CRE3-- CRE2 and CRE3 sequences in the 220 bp of the BAT-specific enhancer region were mutated using PCR and subcloned into the luciferase reporter plasmid, pGL3/basic. For CRE3 the forward primer was 5'-ggggagCTCCTCTACAGCtgaACAGAGG (CRE3 shown in bold with lowercase italic letters that represent mutations; a new SacI site is underlined, 4841/4862) and the reverse primer was 5'-gggctcgagAGTCTGAGGAAAGGGTTGA (a new XhoI site is underlined, 5025/5045). To mutate CRE2, two pairs of primers were required in separate amplifications. The first pair was 5'-ggggagCTCCTCTACAGCGTCACAGAGG (forward primer, a new SacI site is underlined, 4841/4862) and the 5'-AGTGGAAAGGTtcaGACTAGTTCAG (reverse primer, CRE2 is shown in bold with lowercase italic letters representing mutations, 4883/4907). The second pair was 5'-CTGAACTAGTCtgaACCTTTCCACT (forward primer, CRE2 is shown in bold with lowercase italic letters representing mutations, 4883/4907) and 5'-gggctcgagAGTCTGAGGAAAGGGTTGA (reverse primer, a new XhoI site is underlined, 5025/5045). To generate the 220-bp enhancer region with mutations in CRE2, aliquots (1 µl of each 50 µl PCR reactions) of the two PCR products were mixed and subjected to PCR amplification using primer pairs for intact the 220-bp BAT-specific enhancer region. The resulting mutations were confirmed by sequencing. To mutate both CRE2 and CRE3, the 220-bp fragment, which contains the mutation in CRE2, was subjected to PCR amplification using primer pairs 5'-ggggagCTCCTCTACAGCtgaACAGAGG (CRE3 shown in bold with lowercase italic letters that represent mutated sites; a new SacI site is underlined, 4841/4862) and 5'-gggctcgagAGTCTGAGGAAAGGGTTGA (a new XhoI site is underlined, 5025/5045). After the mutations were verified by sequencing, the DNA fragments containing the mutated sites in CRE2 and/or CRE3 were subcloned into luciferase reporter plasmid containing Ucp1 promoter with or without CRE4.

Preparation of Nuclear Extracts and EMSA-- Nuclear extracts from various tissues of A/J or C57BL/6J (B6) mice, and HIB-1B cells were prepared as described (35), except that phosphatase inhibitor mixtures 1 and 2 (Sigma) were added. The protein concentration was determined by the Lowry method (36) using bovine serum albumin as a standard. To prepare probes for EMSA, single-stranded oligonucleotides were synthesized and purified (Operon). 200 pmol each of the complementary oligonucleotides were annealed in 100 µl containing 100 mM NaCl to obtain a double-stranded probe. Five µg of nuclear extract (or in vitro translated CREB, Nfe2l2, and p18) were incubated initially for 10 min at room temperature in 29 µl containing 20 mM HEPES (pH 7.9), 100 mM KCl, 0.1 mM EDTA, 10% glycerol, 1 mM dithiothreitol, 1.5 µg of poly(dA-dT), and 5 mM MgCl₂. The mixture was then incubated for an additional 20 min after adding ³²P-labeled probe (4 × 10⁵ cpm/µl) with or without an unlabeled competitor or antibody for supershift. The antibodies were purchased from Santa Cruz Biotechnology. The reaction was electrophoresed on a 6% polyacrylamide gel (Bio-Rad) in 0.5× TBE buffer. The gel was then dried and exposed to a phosphorimage screen. The radioactivity was visualized and quantified using PhosphorImager and ImageQuant software (Amersham Biosciences).

Western Blot Analysis-- Western blot analyses were performed as described by Laemmli (37) and Towbin et al. (38) with little modification. Cell lysates from HIB-1B cells were prepared by adding SDS sample buffer containing 62.5 mM Tris-Cl (pH 6.8), 2% (w/v) SDS, 10% glycerol, 50 mM dithiothreitol, 0.1% (w/v) bromphenol blue with 1% (v/v) phosphatase inhibitor mixtures 1 and 2 (Sigma). After cell lysates were separated on 8% SDS-polyacrylamide gels, protein was transferred onto nitrocellulose membrane (Millipore). The blots were then incubated with antibody against CREB (1:1,000 dilution, Santa Cruz), or phospho-CREB (Ser¹³³, 1:1,000 dilution, New England Biolabs), overnight at 4 °C with gentle agitation, followed by incubation with anti-rabbit IgG as a secondary antibody (horseradish peroxide conjugated, Amersham Biosciences). Bands were visualized by using the enhanced chemiluminescence reagent (Amersham Biosciences) and exposed to X-Omat film (Kodak).

In Vitro Translation and DNase I Footprinting Analysis-- In vitro protein translation was performed using the TNT coupled reticulocyte lysate system (Promega) according to the manufacturers protocol. The cDNA for CREB and Nfe2l2 was subcloned into pBluescript II (Stratagene) and pCITE4a (Novagen) vectors, respectively, then used for template. The cDNA for p18 was provided by Dr. Paul Ney (St. Jude Children's Research Hospital). To generate ³²P-end-labeled probe for the DNase I footprinting assay, pGL3/3.1 kb (20 µg) was digested with the BstEII restriction enzyme. After cleaning up with phenol-chloroform extraction, the 5' overhang was filled-in with [³²P]dCTP (Amersham Biosciences) and Klenow fragment polymerase (New England Biolabs), and residual nucleotides were removed using the Qiagen nucleotide removal kit. A 286-bp of ³²P-end-labeled probe containing NF-E2 as well as CRE2 and CRE3 was isolated from DNA digested with XbaI and separated on the 0.8% agarose gel. Eluted DNA was subjected to further purification using Elutip-d (Schleicher & Schüll). Nuclear extracts or in vitro translated CREB were incubated at room temperature for 1 h with a probe (15,000 cpm per reaction) in 180 µl of assay buffer containing 10 mM Tris-Cl (pH 8.0), 5 mM MgCl₂, 1 mM CaCl₂, 2 mM dithiothreitol, 50 µg/ml bovine serum albumin, 2 ng/ml calf thymus DNA, and 100 mM KCl. 0.05 unit of DNase I (Roche Molecular Biochemicals) was added, then incubated another 2 min at 37 °C for DNase digestion. DNA was precipitated by adding 700 µl of DNase I stop solution and separated on sequencing gel with the ³²P-end-labeled size marker.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Identification of CRE Sequences for the Binding of CREB-- Four potential CRE sites were located in the 5'-flanking region of Ucp1 (Table I). All four CREs have half-site consensus sequences (CGTCA) and evidence that these sites are involved in the regulation of Ucp1 is limited to loss of chloramphenicol acetyltransferase reporter activity in transient expression assays in a BAT cell line (19). From this analysis CRE2 and CRE4 appeared to be essential; mutations to CRE1 showed no loss of expression and mutations to CRE3 only slightly reduced expression. This study will largely focus on evaluating the function of CRE2 located in the upstream enhancer (Fig. 1). A CRE2 probe for EMSA was made with 5 bp of half-site CRE2 (CGTCA) flanked by 14 bp of 5'- and 3'-flanking sequences as shown in Table I. Nuclear extracts, prepared from BAT of newborn mice maintained at room temperature, and BAT, retroperitoneal fat tissue, inguinal fat tissue, and liver of mice kept in the cold (4 °C) overnight, showed a major retarded band that was eliminated by competition with a 20-fold excess of cold CRE2 (specific shifted bands are shown with the dark arrow in Fig. 2A). However, probes prepared from the region just downstream of the CRE2 motif failed to form a similar retarded band (data not shown). The complex from liver was ~10 times stronger than that of other fat tissues (loading for liver was 1/10^th of the reaction as indicated in the legend; the second retarded band in liver is nonspecific and can be seen with other probes, data not shown). It is of great interest that the binding activity of nuclear extracts of newborn mice is much greater than that of adult mice (Fig. 2C).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Nucleotide sequence of the 221-bp (4828/5048) BAT-specific region of the Ucp1 gene. Enhancer elements are shown with boxes and the NF-E2-binding site is shown underlined with bold letters.

View larger version (80K): [in this window] [in a new window]   Fig. 2.   Binding of half-site CRE sequences to nuclear extracts from tissues of mice. A, autoradiogram of an EMSA using 32P-end-labeled CRE2 (0.1 pmol) from the mouse Ucp1 gene. Each lane (except for liver which was 1/10th of the reaction) was loaded with a binding reaction containing 5 µg of nuclear extracts on 6% nondenaturing acrylamide gel. To verify specificity of binding, competitors included 2 pmol of cold probe or various antibodies (1 µl) as indicated. Slowly migrating bands representing CREB binding or free probes at the gel front are indicated by arrowheads. B, competitive binding activity of CRE2 with half-site CREs from the mouse Ucp1 gene and a palindromic CRE from the somatostatin gene. Autoradiogram of an EMSA showing only the CREB bands. Each lane was loaded with a binding reaction containing 0.1 pmol of 32P-end-labeled CRE2, 5 µg of nuclear extracts from the BAT of A/J mice that were exposed to cold (4 °C, overnight), and 0.4 pmol of cold competitors as indicated on the top. Percent competition of 32P-end-labeled CRE2 to CREB by CRE sequences from the mouse Ucp1 and somatostatin (named CRE) genes was calculated from the radioactivity of the slow migrating bands in the lane without (first lane) and the lane with the individual competitors as shown at the bottom. C, autoradiogram of an EMSA using 32P-end-labeled CRE2 from the mouse Ucp1 gene. Each lane was loaded with a binding reaction containing 5 µg of nuclear extracts from adult (4-week-old) or 0-day-old B6 mice. Only the CREB bands are shown.

To further characterize the binding sites of CRE2, we have performed competitive binding assays with the same mutation, GTC to TGA, in two contiguous locations in the sequence. For the m1CRE2 probe, the mutation occurs in the middle of the half-site CRE motif, whereas the m2CRE2 mutation only overlaps the first C in the CRE motif (Table I). In competitive EMSA, the m1CRE2, but not the m2CRE2 mutant oligonucleotide, has lost the ability to compete with the labeled CRE2 probe (Fig. 2A). This suggests that the half-site CRE motif, but not the flanking 5'-region, is active in binding the specific factor(s). To identify the nuclear factor(s) that bind to CRE2, we applied specific antibodies against Fos, Jun, CBP, or CREB/ATF1 in an EMSA reaction. Because of the sequence similarity of CRE and AP-1-binding sites for the Jun/Fos heterodimer (palindromic CRE, TGACGTCA; palindromic AP-1, TGA(C/G)TCA; half-site CRE sequences are shown underlined) and the known interaction between CREB and CBP, we have tested their antibodies in the supershift assay. The data in Fig. 2A demonstrates that the factors that bind to CRE2 are part of the CREB/ATF1 family. It suggests that CREB/ATF1 does not interact with either jun and fos or CBP. To quantify binding of the four half-site CREs to CREB/ATF1, we measured the ability of each CRE to compete with the CRE2 probe that binds to CREB/ATF1 as described in Table I and in the legend for Fig. 2B. Under these conditions, palindromic CRE from the somatostatin gene (39) competes better than CRE2 itself (percent competition of 58.1% versus 29.1% in Fig. 2B) as we expected. As shown in Fig. 2B, all the half-site CREs show competition to CRE2 binding. Most of the labeled CRE2 probe complexed with proteins in nuclear extracts (as illustrated in Fig. 2A) disappeared with a 40-fold excess (4 pmol) of cold probe (data not shown). This competition data together with the interference on probe binding upon addition of anti-CREB antibody suggests that CRE2 is a high affinity binding site and that this interaction probably involves CREB/ATF.

Changes in CREB/ATF1 Phosphorylation and Binding to CRE2 in Response to Norepinephrine-- HIB-1B is an immortalized brown adipose cell from hibernoma tissue (40) that expresses Ucp1 in response to retinoic acid (27, 32) and -adrenergic agonists such as norepinephrine and isoproterenol (27, 41). We found that treatment of HIB-1B cells with 1 µM norepinephrine significantly increased phosphorylation of both CREB and ATF1 significantly over a 60-min time course, whereas only a modest increase occurred with a change of culture medium (Fig. 3A). We have confirmed and quantified CRE2 binding to CREB/ATF1 factors from HIB-1B cells with EMSA using nuclear extracts from HIB-1B cells treated with 1 µM norepinephrine for 0, 10, and 60 min. As shown in Fig. 3B, nuclear extracts from HIB-1B cells showed at least four retarded bands that were specifically removed with an excess of cold CRE2 (shown with arrows at right). The major thick band migrated to the same position on the gel as the single band from the nuclear extract of cold-exposed BAT of A/J mice (right lane). Treatment of HIB-1B cells with norepinephrine (1 µM) increased the intensity of the four retarded bands 20.6 and 24.2% after 10 and 60 min, respectively (from mean of two experiments). These results suggest that norepinephrine induces phosphorylation and binding of CREB/ATF1 proteins to CRE2.

View larger version (53K): [in this window] [in a new window]   Fig. 3.   Effects of norepinephrine on CREB/ATF1 phosphorylation and CRE2 binding to nuclear proteins in HIB-1B cells. A, Western blot analysis for norepinephrine-induced phosphorylation of CREB/ATF1 in HIB-1B cells. HIB-1B cells were treated with fresh medium (control) or medium containing 1 µM norepinephrine (NE) for 0, 5, 10, 20, 30, and 60 min. Cell lysates were prepared and analyzed by a Western blot with phospho-CREB (Ser133) specific antibody. Arrows on the right indicate the location of phosphorylated CREB (pCREB) and ATF1 (pATF1) with molecular weights of 43,000 and 35,000, respectively. B, EMSA of binding activity of the CRE2 motif is increased in HIB-1B cells by NE treatment. Nuclear extracts were isolated from HIB-1B cells treated with 1 µM NE for 0, 10, and 60 min. Each lane was loaded with a binding reaction containing 5 µg of nuclear extracts and 0.1 pmol of 32P-end-labeled CRE2 as indicated on the top. The lane for the nuclear extracts from BAT of the A/J mouse (cold, overnight) was added to compare intensity (lane BAT). Cold probe (2 pmol) was added in the reaction for the competition (lane CRE2). Only the CREB bands are shown with arrows.

Functional Characterization of CRE2-- To further characterize the functionality of CRE1-4, we performed a transient transfection assay using luciferase reporter constructs and site-directed mutagenesis. The same site-directed mutations present in the probes used in competitive EMSA were introduced into CRE2 and CRE3 in the 220-bp BAT-specific enhancer region (Fig. 2A), because changes from GTC to TGA (m1CRE2 probe in Fig. 2A) eliminated the capacity of the oligonucleotide to compete with CRE2 probe. It implies that this mutated sequence was no longer able to interact with CREB. The promoter without CRE4 (pGL3/pro) had low basal promoter activity (Fig. 4). Addition of CRE4 (pGL3/CRE4pro) to the promoter construct showed about a 3-fold increase in luciferase activity in response to NE and cAMP. This level of transient expression was similar to the promoter construct containing 220 bp of the BAT-specific enhancer region, but without CRE4 (pGL3/pro/220). Importantly, the 220-bp BAT-specific enhancer region together with CRE4 (pGL3/CRE4pro/220) showed a level of expression activity similar to 3.1 kb of 5'-flanking region of Ucp1 (pGL3/3.1 kb). This data suggests that CRE4 cooperates with the 220-bp BAT-specific enhancer region in determining the response to NE and cAMP. When CRE2 or CRE3 were mutated independently or together in pGL3/CRE4pro/220 to evaluate the contribution of CRE2 and CRE3 to the enhancer activity, the expression was diminished in assays carrying the mutant constructs, but the relative differences were much less than we had previously observed with a more differentiated BAT cell line (19). One interpretation is that other regulatory elements in the enhancer can mediate NE or cAMP-induced expression. This could be PGC1-mediated expression or NFE2l2 as described below. An important conclusion that emerges from this analysis is the essential requirement for interactions between the distal enhancer with CRE4 in the proximal promoter to confer high levels of expression. Overall, the results show a role for CRE2 in the enhancer activity, but quantification of the effect will require in vivo analysis or a highly differentiated cell culture system.

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Transient expression analysis of luciferase reporter constructs to determine the function of individual CREs. Luciferase reporter constructs (named in the left) were generated by subcloning the various fragments from the 5'-flanking region of the mouse Ucp1 gene into the pGL3/basic vector (Promega). DNA fragments from the mouse Ucp1 gene are shown as thick lines with position of the individual CREs indicated as ovals. Mutations of CRE2 and/or CRE3 by mutating key nucleotide residues as described under "Experimental Procedures" are indicated as open ovals with X marks. Each construct was transfected into HIB-1B cells with pRL/SV40 vector (Promega), and the cells were cultured under the medium containing 1 µM norepinephrine (NE) or 0.5 mM 8-bromo-cAMP (cAMP) another 16 h. Luciferase activity was measured from cell lysates using the Dual Luciferase assay system (Promega), and -fold increase of luciferase activity by NE or cAMP was calculated. Data is presented as the mean ± S.D. of -fold increase from three experiments. The restriction map is shown at the top left. H, HindIII; X, XbaI; B, BglI.

Nfe2l2 Binds NF-E2-binding Sites in the Upstream Enhancer of Mouse Ucp1-- We have identified a consensus NF-E2 binding motif, ACTAGTCGT, that partially overlaps the CRE2 half-site and is located 6 bp downstream of the peroxisomal proliferator activator receptor binding motif (PPRE) (Fig. 1). A probe containing 10 bp of the NF-E2 binding motif with 3 bp of nonspecific flanking sequence (CCC) were synthesized and incubated with nuclear extracts from HIB-1B cells (Fig. 5A). Nuclear extracts from HIB-1B cells interacted with probes to the NF-E2 binding site from the mouse Ucp1 gene to generate shifted bands that were eliminated in a competition assay with a 20-fold excess of cold probe and with antibody against Nfe2l2, but not with anti-Nfe2l1 or anti-NF-E2 p45 (Fig. 5A). Nuclear extracts from the HIB-1B cells treated with 1 µM norepinephrine for 30 min increased the intensity of the complex (Fig. 5A) as did nuclear extracts prepared from brown adipose tissue of cold exposure adult mice (Fig. 5D). The strongest binding to the NF-E2 probe was present in BAT of the 19-day-old fetus (Fig. 5B) and this binding capacity decreased during post-natal development until the signal with the nuclear extracts from the 4-month-old mice was less than 3% of the binding activity of the fetus. A similar reduction in binding capacity has been found for virtually all transcription factors associated with Ucp1 transcription.² The binding activity to nuclear extracts from newborn mice was reduced by competition with cold probe and antibody to Nfe2l2, but not with antibody to PPAR, Jun, Fos, CBP, or CREB (Fig. 5C). Accordingly, protein-protein interactions between Nfe2l2 and other bZIP proteins with putative roles in Ucp1 regulation, including CREB, Fos, Jun, and PPAR (42-45), did not participate in Nfe2l2 binding to Ucp1 NF-E2. Despite this large reduction in binding capacity during post-natal development, an induction in binding activity still occurs when adult mice were exposed to the cold (Fig. 5D).

View larger version (65K): [in this window] [in a new window]   Fig. 5.   Binding of NF-E2 sequences to nuclear extracts from HIB-1B and BAT of A/J mice. A, binding of NF-E2 sequences to nuclear extracts from HIB-1B cells. Autoradiogram of an EMSA using 32P-end-labeled NF-E2 (0.1 pmol) from the mouse Ucp1 gene. Nuclear extracts were prepared from HIB-1B cells with (+) or without () NE (1 µM, 60 min). Each lane was loaded with a binding reaction containing 5 µg of nuclear extracts with 2 pmol of cold probe or antibody (1 µl) as indicated on 6% nondenaturing acrylamide gels. Slowly migrating bands representing NFE2l2 are indicated with an arrow on the right. NE, norepinephrine. B, autoradiogram of an EMSA using 32P-end-labeled NF-E2 from the mouse Ucp1 gene. Each lane was loaded with a binding reaction containing 5 µg of nuclear extract from mice ranging in age from 19 days of gestation to 4 months. Only the NFE2l2 bands are shown. C, binding of NF-E2 sequences to nuclear extracts from BAT of B6 mice (0 day old). Each lane was loaded with a binding reaction of 32P-end-labeled NF-E2 (0.1 pmol) incubated with 5 µg of nuclear extracts from BAT of B6 mice (0 day old) with 2 pmol of cold probe or antibody (0.5 µl) as indicated on 6% nondenaturing acrylamide gels. Only the NFE2l2 bands are shown. D, binding activity of the NF-E2 sequence increased by cold exposure in BAT of the B6 mouse. Nuclear extracts were isolated from BAT of the B6 mice kept at room temperature or in the cold (4 °C) for 1 and 7 days. 5 µg of nuclear extracts were incubated with 32P-end-labeled NF-E2 probe (0.1 pmol), separated on a 6% nondenaturing acrylamide gel. Only the NFE2l2 bands are shown.

Competition between Nfe2l2 and CREB-- An overlap of the binding motif of NF-E2 with the half-site CRE2 suggests that competition for binding may exist between Nfe2l2 and CREB. To test this we designed a 19-bp oligonucleotide probe, NFCRE, which covered both NF-E2 and CRE2, for a gel shift and supershift assay (Fig. 6A). The band shifts with the NFCRE probe were very similar to the pattern observed for CRE2, whereas the band shift with the NF-E2 probe migrates slightly faster. The NF-E2 band that should have been formed with the NFCRE probe was not detected. Both cold CRE2 and NFCRE can compete with the NFCRE probe. NF-E2 cannot compete away the band shifts with either CRE2 or NFCRE probe, but it can with the NF-E2 probe. We interpret these findings as indicating that CREB binds to CRE2 with high affinity and its binding to its own motif will interfere in a competitive manner with the binding of NFE2l2 to the NF-E2 motif.

View larger version (67K): [in this window] [in a new window]   Fig. 6.   Competition between Nfe2l2 and CREB. A, nucleotide sequences for NFCRE containing both NF-E2 and CRE2 binding sites. Enhancer elements are shown with underlined bold letters (NF-E2) and a box (CRE2). Nucleotide sequences for cold probes are shown with underlined bold letters and 3 bp of nonspecific flanking sequences. B, binding activity of NFCRE, NF-E2, and CRE2 with nuclear extracts from HIB-1B cells. Each lane was loaded with a binding reaction containing 5 µg of nuclear extracts (HIB-1B cells, 1 µM NE for 60 min for treatment) with different concentrations of cold probe (2 or 0.2 pmol) as indicated on the 6% nondenaturing acrylamide gels. Slowly migrating bands representing CREB and Nfe2l2 complex with the 32P-end-labeled probe are shown.

To evaluate whether both CREB and Nfe2l2 occupy CRE2 and the overlapping NF-E2 binding sites, DNase I footprinting was performed with CREB and Nfe2l2/p18 synthesized by in vitro transcription and translation. Because NF-E2 transcription factors belong to the cap"n"collar-type basic leucine zipper family and bind as heterodimers (46), it was necessary to generate a heterodimer by in vitro synthesis. In vitro synthesized CREB formed a complex with the CRE2 probe that could be removed by cold CRE2 probe (Fig. 7A). On the other hand, the Nfe2l2/p18 heterodimer formed a complex with the -globin NF-E2 binding motif, but not with the Ucp1 NF-E2 motif, as evidenced by the lack of competition with cold Ucp1 NF-E2 probe. As shown in Fig. 7D, these binding sites differ from each other by one base. Martin et al. (47) had previously observed the extreme sensitivity of Nfe2l2/p18 binding to NF-E2 motifs with single base changes. These results indicate that p18 is not the appropriate partner to enable Nfe2l2 binding to Ucp1 NF-E2. Nfe2l2 synthesized in vitro, without p18, showed no binding to either the Ucp1 or -globin probes (data not shown). DNase I footprinting with nuclear extracts from livers of mice clearly show that both NF-E2 and CRE2 binding sites were protected with proteins from nuclear extracts (Fig. 7C). However, when DNase I footprinting was done with in vitro synthesized CREB, only the CRE2 site was protected. A comparable test with in vitro synthesized Nfe2l2 cannot be performed until the appropriate heterodimeric partner can be identified.

View larger version (92K): [in this window] [in a new window]   Fig. 7.   Binding of in vitro translated CREB and Nfe2l2 and DNase I footprinting. A, autoradiogram of an EMSA using 32P-end-labeled CRE2. Each lane was loaded with a binding reaction containing 1 µl of in vitro CREB translation reaction (50 µl of reaction volume), or translation reaction without CREB cDNA template (Ctl) on a 6% nondenaturing acrylamide gel. 2 pmol of cold probe (+) was added in the reaction for competition. CREB and nonspecific binding (N.S.) are shown with an arrow. B, autoradiogram of an EMSA using NF-E2. Each 2 or 1 µl of in vitro translation reactions for Nfe2l2 and p18 were incubated with 32P-end-labeled NF-E2 probe from the -globin gene. 2 pmol of cold probe (UCP1 and -globin) was added in the reaction for competition. Heterodimers from Nfe2l2 and p18 are shown with arrows. C, DNase I footprinting assay. An end-labeled probe containing binding motifs for NF-E2 as well as CRE2 and CRE3 was incubated with CREB synthesized by in vitro translation (1 and 2 µl as indicated) or nuclear extract from liver (2.5 and 5 µg as indicated), then incubated with 0.05 units of DNase I, 2 min at 37 °C. Zero, 0.01, and 0.005 units of DNase I were incubated with probe alone as indicated on the top of the gel. Binding sites for CRE2, NF-E2, and CRE3 are shown with vertical lines as indicated. N.E. is nuclear extract. D, comparison of NF-E2 binding sequences. The consensus sequence for NF-E2 binding are shown. PBGD, porphobilinogen deaminase; ALAS-E, -aminolaevulinate synthase.

Human and Rat Ucp1 Gene Contains NF-E2 binding Sites-- Recently, the human Ucp1 gene was cloned and evidence describing key elements controlling its transcriptional regulation were obtained (23). A 350-bp hormone-sensitive region of the human gene showed significant similarity with the mouse (60.1%) and rat (62.5%) BAT-specific enhancer element. This region in the human gene was able to bind the nuclear factors, RARs, RXRs, CREB/ATF, and PPAR indicating that transcriptional regulation of the Ucp1 gene between rodents and human have mechanisms in common. Comparison of 100 bp of the human, rat, and mouse Ucp1 enhancer sequence is shown in Fig. 8. A sequence similarity search indicates that a NF-E2-binding site (TGCTGYCNCT) in the mouse, human, and rat is located in a comparable location (Fig. 8A). However, unlike the mouse neither the rat nor the human genes contain the downstream NF-E2 site that overlaps with CRE2. This provides an opportunity to evaluate the effects of selective mutations to the single NF-E2 motif. Binding activity of the human and rat NF-E2 sites was assayed with nuclear extracts from BAT. Nuclear extracts from cold-exposed mice (1 or 7 days at 4 °C) showed increased binding activity for both human and rat NF-E2 (Fig. 8B), a similar binding activity as the mouse NF-E2 probe (Fig. 5D). Human and rat NF-E2-binding sites (hNF-E2 and rNF-E2 in Fig. 8B) showed comparable competition with the mouse NF-E2-binding site (mNF-E2) overlapping CRE2, and antibody against Nfe2l2 (mouse, rat, and human reactive) interacts to inhibit complex formation in the supershift assay. To further characterize the functionality of the NF-E2 site, we performed transient transfection assays using luciferase reporter constructs and site-directed mutagenesis. A 263-bp region of enhancer element from rat Ucp1 located between bp 2519 to 2283 containing putative NF-E2, CRE, and PPRE was subcloned into pGL3/basic reporter constructs driven by either a mouse Ucp1 promoter or a SV40 promoter. Transient transfection analysis demonstrated that 260 bp of rat Ucp1 enhancer (pGL3/cre4pro/Rat221) activated the mouse Ucp1 promoter in response to both cAMP and troglitazone (Fig. 8C). Importantly, a mutation in the NF-E2 site (pGL3/cre4pro/mRat221) significantly decreased the cAMP-stimulated Ucp1 promoter activity but not the troglitazone-dependent activity. This suggests that the NF-E2 site in the rat Ucp1 enhancer contributes to the induction of Ucp1 in response to cAMP or -adrenergic stimulation. In addition, transient transfection analysis with reporter constructs containing SV40 promoter (pGL3/sv40/Rat221 and pGL3/sv40/mRat221) confirms the results obtained with the mouse Ucp1 promoter (Fig. 8C). To further test the function of Nfe2l2 on the rat Ucp1 enhancer, luciferase reporter constructs containing the rat Ucp1 enhancer with or without mutations to the NF-E2 site (pGL3/cre4pro/Rat221 and pGL3/cre4pro/mRat221) were co-expressed with a cytomegalovirus promoter-controlled Nfe2l2 expression vector in HIB-1B cells. As shown in Fig. 8D, overexpression of Nfe2l2 increased Ucp1 promoter activity but only with a wild-type NF-E2 element. Taken together, this result suggests that NF-E2 sites in the human and rodent Ucp1 enhancer elements participate in the transcriptional activation of Ucp1 in response to cAMP.

View larger version (37K): [in this window] [in a new window]   Fig. 8.   Characterization of the Ucp1 NF-E2 element. A, comparison of human, mouse, and rat Ucp1 enhancer sequences. 100 bp (3762/3662) of nucleotide sequences from the human 350-bp enhancer and corresponding mouse and rat enhancer sequences were compared (23). Half-sites for ATF/CREB (CRE2 and CRE3), putative NF-E2-binding site, and PPRE are shown within the boxes. Bold letters represent bases that were matched between the three species. B, binding of human and rat NF-E2 sequences to nuclear extracts from BAT. Nuclear extracts were isolated from BAT of A/J and B6 mice kept at room temperature or cold (4 °C) for the indicated times. 5 µg of nuclear extracts were incubated with 32P-end-labeled NF-E2 probe (0.1 pmol) corresponding to human (ACTTGCTGCCACTCCT) and rat (CCTTGCTGCCTCTCCT) Ucp1 genes, and separated on 6% nondenaturing acrylamide gels. 2 pmol of cold probe or antibody for Nfe2l2 (1 µl, Santa Cruz Biotechnology) were added for competition and supershift assays, respectively. Only the Nfe2l2 bands are shown. C, transient transfection analysis of rat Ucp1 enhancer and effects of mutations in the NF-E2 site. Luciferase reporter constructs (named in the bottom) were generated by subcloning the fragment from the rat Ucp1 enhancer (2519/2283) into pGL3/basic containing the Ucp1 promoter (pGL3/cre4pro) or pGL3/SV40 (Promega). Nucleotides TGCT from the NF-E2 site were mutated to GTAG as described under "Experimental Procedures." Each construct was transfected into HIB-1B cells with pRL/SV40 vector (Promega), and the cells were cultured under the medium containing 0.5 mM 8-bromo-cAMP (cAMP) or 1 µM troglitazone (Trog) another 16 h. Luciferase activity was measured from cell lysates using the Dual Luciferase assay system (Promega). Data is presented as the mean ± S.D. from two experiments. D, effect of Nfe2l2 overexpression on the luciferase reporter construct containing the NF-E2 site from the rat Ucp1. Luciferase reporter constructs (pGL/cre4pro/Rat221 and pGL/cre4pro/mRat221) were co-transfected with expression vector, which was empty (pCMV/tag1, Stratagene) or containing cDNA for NFE2l2 into HIB-1B cells. Cells were cultured in the medium with or without 0.5 mM 8-bromo-cAMP (cAMP) for an additional 16 h. Luciferase activity was measured in cell lysates using the Dual Luciferase assay system (Promega). Data is presented as the mean ± S.D. from two experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mutation analysis of the four CREs in the upstream enhancer of mouse Ucp1 suggested that CRE2 and CRE4 were essential for the transient expression of a reporter construct following treatment of BAT cells cultures with norepinephrine (19). Because a mutation to CRE3 only slightly reduced transient expression, it was not considered to be important for the regulation of Ucp1, and mutations to CRE1 were without any effects on transient expression. In this current study some of the selectivity of mutations to the CREs has been lost. We think that this is because of loss in the differentiated phenotype, principally evidenced by the reduction in Ucp1 expression, of the available brown adipocyte cell lines, including the HIB-1B cells. Nevertheless, the transient expression evidence in this work continues to support a role for CRE2 and CRE4 in regulating Ucp1 expression by norepinephrine. The EMSA evidence for CRE2-CREB interactions corroborates the expression data to lead us to conclude that CRE2 interacts with CREB to directly regulate Ucp1 expression. Because Pgc1 is highly induced in brown adipocytes and is required for coactivation of the PPAR it suggests that a second mechanism exists for the adrenergic regulation of Ucp1. Finally the evidence in this paper showing that NFE2l2 interactions with NF-E2 are activated by cold exposure and that coexpression of NFE2l2 stimulates Ucp1 reporter gene expression points to a third mechanism for the adrenergic activation of Ucp1. We need to determine whether these mechanisms for adrenergic regulation of Ucp1 are coordinately utilized under all situations of adrenergic stimulation. Alternatively, it is possible that different physiological circumstances, for example, cold and diet use different mechanisms to activate Ucp1 or that the activation of Ucp1 in interscapular brown fat utilizes a different mechanism than activation in traditional white fat tissues. Such questions regarding Ucp1 expression are pertinent to our understanding of the physiological of thyroid hormone, retinoids, and PPAR ligands in Ucp1 expression.

Stimulus-induced activation of CREB is mediated by phosphorylation of the serine 133 residue through cAMP-dependent protein kinase A (48, 49). Phosphorylation of CREB can activate transcription of target genes by: (a) dimerization through endogenous leucine zipper (50) with CREB or ATF1/CREM (51, 52); (b) regulation of subcellular localization (53); or (c) interaction with CBP/p300 (54). As shown in Fig. 3A, norepinephrine treatment in the HIB-1B cell markedly increases the phosphorylation of both CREB and ATF1 within 5 min, followed by increased binding of dimerized CREB or CREB/ATF1 heterodimers to ³²P-end-labeled CRE2 probe in EMSA (Fig. 3B). Similarly, EMSA data showing an increase in complex formation with nuclear extracts from cells incubated with 1 µM NE suggests that phosphorylation of CREB and ATF1 increase their binding affinity to CRE in the Ucp1 gene. Alternatively, or in addition, CREB may form heterodimers with ATF1. Complex formation may also occur with CBP, a transcriptional adaptor that has intrinsic histone acetyltransferase activity (55, 56) and interacts with RNA polymerase II (57, 58). In addition to CREB interactions, CBP is reported to form nuclear partners with retinoic acid X receptor, c-Jun, c-Myb, Sap-1a, c-Fos, MyoD, and YY1 (see Ref. 59 for review). To test nuclear complexes involving CBP or other CREB partners, a supershift assay was performed using nuclear extracts from BAT of a cold-exposed A/J mouse (Fig. 2A). Other than the strong interference in band formation found with CREB antibodies, there was no evidence that antibodies against CBP, c-Jun, or c-Fos interfered with the interaction of the CRE2 probe with nuclear proteins in EMSA. Consistent with the lack of effects of these antibodies, a yeast two-hybrid screening for cDNA (4 × 10⁷ transformants were screened) from BAT of the A/J mouse with CREB as a bait, failed to detect positive clones.³ We conclude from the evidence that CRE2 is a major site for the transcriptional activation of Ucp1 expression by direct interaction with homodimers of CREB.

Inspection of the sequence around CRE2 showed that it overlapped with a NF-E2 at its 5' end (Fig. 1). Additionally, we were able to find a putative NF-E2-binding site (TGCTGYCNCT) in both the human and rat Ucp1 (Fig. 8A) by sequence comparison. Because a NF-E2 site is also present in the mouse in a comparable location, it means that two NF-E2 sites are present in the mouse gene, but only one occurs in the human and rat. NF-E2 is a binding site in the -globin gene locus control region (46) where the hematopoietic specific NF-E2 p45 subunit and the ubiquitously expressed small Maf protein, an important regulator of cell differentiation in various systems, form heterodimers (60-62). We found that a EMSA probe designed from the sequence in the mouse Ucp1 enhancer formed specific bands; however, supershift assays with antibodies showed that the proteins binding to the probe were not against the NF-E2 p45 subunit, but rather against NFE2l2, another member of the NF-E2 family. The binding of Nfe2l2 probes is increased in brown fat cells isolated by following treatment with cold or norepinephrine, and coexpression studies of Nfe2l2 vectors with the Ucp1 enhancer constructs suggests that transcription of Ucp1 is mediated by NFE2l2. Similar binding activities were found for the NF-E2 elements in rat and human. In addition, binding activity is higher in nuclear extracts from newborn mice than from cold-exposed adult mice. The potential relevance of NFE2l2 to Ucp1 regulation has been heightened by the fact that in situ hybridization with the sections from 15.5-day-old embryos has shown that the Nfe2l2 gene is highly expressed in brown fat (63). Our EMSA data in Fig. 5B corroborates this data by showing an intense NF-E2 binding activity in nuclear extracts prepared from fetal brown fat. This is a time that coincides with expression of Ucp1 in the embryo (64), and therefore suggests that Nfe2l2 is involved in Ucp1 expression during brown fat development in the embryo.

NFE2l2 is a member of cap"n"collar-basic leucine zipper (bZIP) superfamily (46). Cap"n"collar is a homeotic gene involved in the development of the head and neck structure in Drosophila (65, 66). The nuclear DNA-binding protein NF-E2 regulates expression of globin genes during the developing erythroid cells. Two additional members of the cap"n"collar-bZIP family, Nfe2l1 and Nfe2l2, have been cloned; they are expressed ubiquitously in tissues, but with variable expression among these different tissues (63). The binding sequence of the NF-E2 family is remarkably similar to the antioxidant responsive element consensus sequence (RGCNNN(C/G)TCA) (43). In this system Nfe2l1 and Nfe2l2 can form heterodimers with small Maf proteins and bind to the antioxidant responsive element-binding complex to activate downstream gene expression in response to reactive oxygen species or oxidative stress (43, 67-69). It has been demonstrated that NFE2l2 is retained in the cytoplasm through association with Keap1, then translocates into the nucleus by electrophilic agents as well as antioxidants and phorbol esters (67, 70). Previous studies also showed that mitogen-activated protein kinase pathways are involved in antioxidant responsive element-mediated transcription (71). In this study, we demonstrated that the binding activity of Nfe2l2 was increased by norepinephrine treatment (in vitro) and cold exposure (in vivo), furthermore, NFE2l2 overexpression induced the Ucp1 promoter activity only with norepinephrine and a cAMP analog in HIB-1B cell. That no induction occurred with the PPAR ligand, troglitazone, suggests that NFE2l2 activation is mediated by the protein kinase A signaling pathway, but is independent of PPAR. Further studies will be needed to address complexities of overlapping kinase pathways and interactions between transcription factors involved in transcriptional regulation in Ucp1 gene.

In summary, we have provided evidence for the direct involvement of CREB binding to CRE2 to regulate Ucp1 expression and have provided the first evidence that a member of the NF-E2 family of transcription factors is involved in the regulation of Ucp1. At present several transcription factors, including PPAR and , PGC-1, CREB, Nfe2l2, and RXR, have been identified that interact with regulatory motifs located in 30 bp of the Ucp1 distal enhancer. How these motifs and factors, together with others in the enhancer, regulate Ucp1 during variable physiological states that include fetal expression, conversion of white to brown adipocytes, and induction following cold exposure or cafeteria diets remain to be determined.

	ACKNOWLEDGEMENT

We thank Dr. Robert Koza for help in the review and preparation of the manuscript.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant HD08431.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.

To whom correspondence should be addressed. Tel.: 225-763-2771; Fax: 225-763-3030; E-mail: kozaklp@pbrc.edu.

Published, JBC Papers in Press, June 25, 2002, DOI 10.1074/jbc.M108866200

² J. S. Rim, B. Z. Xue, and L. P. Kozak, manuscript in preparation.

³ J. S. Rim and L. P. Kozak, unpublished data.

	ABBREVIATIONS

The abbreviations used are: BAT, brown adipose tissue; CRE, cAMP-responsive element; CREB, CRE-binding protein; PPAR, peroxisomal proliferator activator receptor; EMSA, electrophoretic mobility shift assay.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES