Purification and biochemical characterization of the promoter-specific transcription factor, Sp1.

The biochemical analysis of cellular trans-activators involved in promoter recognition provides an important step toward understanding the mechanisms of gene expression in animal cells. The promoter selective transcription factor, Sp1, has been purified from human cells to more than 95 percent homogeneity by sequence-specific DNA affinity chromatography. Isolation and renaturation of proteins purified from sodium dodecyl sulfate polyacrylamide gels allowed the identification of two polypeptides (105 and 95 kilodaltons) as those responsible for recognizing and interacting specifically with the GC-box promoter elements characteristic of Sp1 binding sites.

Thiazolidinediones repress ob gene expression in rodents via activation of peroxisome proliferator-activated receptor gamma.

The ob gene product, leptin, is a signaling factor regulating body weight and energy balance. ob gene expression in rodents is increased in obesity and is regulated by feeding patterns and hormones, such as insulin and glucocorticoids. In humans with gross obesity, ob mRNA levels are higher, but other modulators of human ob expression are unknown. In view of the importance of peroxisome proliferator-activated receptor gamma (PPARgamma) in adipocyte differentiation, we analyzed whether ob gene expression is subject to regulation by factors activating PPARs. Treatment of rats with the PPARalpha activator fenofibrate did not change adipose tissue and body weight and had no significant effect on ob mRNA levels. However, administration of the thiazolidinedione BRL49653, a PPARgamma ligand, increased food intake and adipose tissue weight while reducing ob mRNA levels in rats in a dose-dependent manner. The inhibitory action of the thiazolidinedione BRL49653 on ob mRNA levels was also observed in vitro. Thiazolidinediones reduced the expression of the human ob promoter in primary adipocytes, however, in undifferentiated 3T3-L1 preadipocytes lacking endogenous PPARgamma, cotransfection of PPARgamma was required to observe the decrease. In conclusion, these data suggest that PPARgamma activators reduce ob mRNA levels through an effect of PPARgamma on the ob promoter.

Requirement of Sp1 and estrogen receptor alpha interaction in 17beta-estradiol-mediated transcriptional activation of the low density lipoprotein receptor gene expression.

Estrogen is one of the most important physiological regulators of low density lipoprotein receptor (LDLR) expression. Despite many studies conducted in animals and humans showing increased expressions of LDLR messenger RNA by hormone treatment, the molecular basis of the effect of estrogen on LDLR transcription has not been clearly elucidated. By using HepG2 cells that transiently express functional estrogen receptor alpha (ERalpha) and LDLR promoter constructs, we show that the specific interaction of ERalpha with the transcription factor Sp1 bound to the LDLR promoter is responsible for the activation of LDLR transcription by estrogen. We demonstrate that 1) mutations to abrogate the binding of Sp1 to its recognition sequences present in repeat 1 and repeat 3 elements of the LDLR promoter completely abolish the ERalpha-mediated activation of the LDLR promoter activity; 2) mutations that abolish the selective DNA-binding activity or inactivate the C-terminal transcription activation function (AF2) of ERalpha had no effect on the ability of ERalpha to activate LDLR transcription; however, transcriptional activation was completely lost by deletion of the N-terminal transcription activation region (AF1); 3) a subregion of AF1 (amino acids 67-139) was further identified to be important for ERalpha to activate the LDLR promoter; and 4) ERalpha enhanced the formation of Sp1-repeat 3 DNA complexes. We also show that mutation at the sterol-responsive element-1 site diminishes the activity of ERalpha on LDLR transcription, thereby suggesting that the sterol-responsive element-1-binding protein may interact with the Sp1-ERalpha complex to trans-activate LDLR gene transcription. This study for the first time provides a molecular basis for an understanding of the regulation of LDLR transcription by estrogens.

Androgen receptor-mediated antagonism of estrogen-dependent low density lipoprotein receptor transcription in cultured hepatocytes.

Postmenopausal women receiving hormone replacement therapy have a lower risk of coronary heart disease than women who do not receive hormone treatment. Multiple mechanisms are likely to underlie estrogen's cardioprotective action, including lowering of plasma low density lipoprotein (LDL) cholesterol. Using an in vitro system exhibiting normal regulation of LDL receptor (LDLR) gene transcription, we show that 17beta-estradiol activates the LDLR promoter in transiently transfected HepG2 cells. LDLR activation by estrogen in HepG2 cells is dependent on the presence of exogenous estrogen receptor, and the estrogen-responsive region of the LDLR promoter colocalizes with the sterol response element previously identified. The estrogen response is concentration dependent, saturable, and sensitive to antagonism by estrogen receptor antagonists. Further, we show that compounds with androgen receptor agonist activity attenuate the estrogen-induced up-regulation of LDLR in our model system. Progestins with androgen receptor agonist activity, such as medroxyprogesterone acetate, also suppress estrogen's effects on LDLR expression through their androgenic properties. Characterization of the interplay between these hormone receptors on the LDLR in vitro system may allow a better understanding of the actions of sex steroids on LDLR gene expression and their roles in cardiovascular disease.

Nuclear protein that binds sterol regulatory element of low density lipoprotein receptor promoter. I. Identification of the protein and delineation of its target nucleotide sequence.

The current paper reports the identification of a protein in rat liver nuclei that binds to the sterol regulatory element (SRE-1) in the promoter of the gene for the low density lipoprotein receptor. The 10-base pair SRE-1 is embedded within a 16-base pair sequence designated Repeat 2 located immediately upstream of a related sequence designated Repeat 3. To confirm that DNA recognition by the SRE-1 binding protein (SREBP) correlates with sterol-regulated transcription, we synthesized an artificial promoter that contains two copies of wild-type or mutant Repeat 2 + 3 sequences immediately upstream of a TATA box from adenovirus. The synthetic promoters were inserted upstream of a reporter gene and tested for transcriptional activity in the absence and presence of sterols after transient transfection into monkey CV-1 cells. The reporter gene with two copies of the wild-type Repeat 2 + 3 sequence was transcribed actively in sterol-deprived cells and was repressed by more than 80% when sterols were present. Binding of SREBP to the SRE-1 sequence, assessed by gel mobility shift assays, correlated precisely on a nucleotide-by-nucleotide basis with the transcriptional activity of each of 16 synthetic promoters with point mutations in Repeat 2. The SREBP bound to the nine mutant promoters that were positive for sterol-regulated transcription, and it did not bind to any of the nine point mutants that abolished transcription. We conclude that SREBP is a DNA binding protein that mediates sterol-regulated transcription of the low density lipoprotein receptor gene.

Identification of a novel sterol-independent regulatory element in the human low density lipoprotein receptor promoter.

The cytokine oncostatin M (OM) activates human low density lipoprotein receptor (LDLR) gene transcription through a sterol-independent mechanism. Previous studies conducted in our laboratory have narrowed the OM-responsive element to promoter region -52 to +13, which contains the repeat 3 and two TATA-like sequences. We now identify LDLR promoter region -17 to -1 as a sterol-independent regulatory element (SIRE) that is critically involved in OM-, transcription factor CCAAT/enhancer-binding protein (C/EBP)-, and second messenger cAMP-mediated activation of LDLR transcription. The SIRE sequence overlaps the previously described TATA-like element and consists of an active C/EBP-binding site (-17 to -9) and a functional cAMP-responsive element (CRE) (-8 to -1). We demonstrate that (a) mutations within either the C/EBP or CRE site have no impact on basal or cholesterol-mediated repression of LDLR transcription, but they completely abolish OM-mediated activation of LDLR transcription; (b) replacing the repeat 3 sequence that contains the Sp1-binding site with a yeast transcription factor GAL4-binding site in the LDLR promoter construct does not affect OM inducibility, thereby demonstrating that OM induction is mediated through the SIRE sequence in conjunction with a strong activator bound to the repeat 3 sequence; (c) electrophoretic mobility shift and supershift assays confirm the specific binding of transcription factors C/EBP and cAMP-responsive element-binding protein to the SIRE; (d) cotransfection of a human C/EBPbeta expression vector (pEF-NFIL6) with the LDLR promoter construct pLDLR234 increases LDLR promoter activity; and (e) OM and dibutyryl cAMP synergistically activate LDLR transcription through this regulatory element. This study identifies, for the first time, a cis-acting regulatory element in the LDLR promoter that is responsible for sterol-independent regulation of LDLR transcription.

DNA binding and transcription activation specificity of hepatocyte nuclear factor 4.

Hepatocyte nuclear factor 4 (HNF-4) is an orphan intracellular receptor that appears to be a key factor in the regulation of energy metabolism. In order to gain greater understanding of the binding and activation requirements of HNF-4, we performed genetic analysis of the apoCIII promoter, a promoter that has previously been shown to be highly sensitive to HNF-4-induced transcription. We identified two elements within the apoCIII promoter that bind HNF-4, either of which are sufficient to confer promoter induction in response to HNF-4. These two elements are both direct repeat-like in nature, but they differ significantly in motif sequence and the repeats are separated by either 1 or 2 nt. Therefore, to better define the DNA sequence recognition requirements of HNF-4, we utilized PCR-based binding site selection. HNF-4 selected for direct repeat-like elements with either 1 or 2 nt between the repeats. Surprisingly, the strongest selection was for a core motif that included the nucleotides between the repeats. Mutation of the nucleotide between the repeats resulted in a 6-fold reduction in affinity, indicating that the interaction between HNF-4 and the intervening nucleotide(s) is critical for high affinity binding.

Nuclear protein that binds sterol regulatory element of low density lipoprotein receptor promoter. II. Purification and characterization.

This paper describes the purification and characterization of a sterol regulatory element binding protein (SREBP) that recognizes the SRE-1 sequence in the 5' flanking region of the gene for the low density lipoprotein (LDL) receptor. The protein was purified more than 38,000-fold from nuclear extracts of human HeLa cells by ion exchange, gel filtration, and DNA-affinity chromatography. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the purified preparation revealed a cluster of bands at 59-68 kDa, each of which bound to the SRE-1 element as revealed by cross-linking experiments. Binding of SREBP correlated perfectly with transcriptional activity in a series of 16 sterol regulatory elements with point mutations. In the LDL receptor promoter, the 10-base pair SRE-1 is embedded in a 16-base pair sequence designated Repeat 2, which is adjacent to Repeat 3, a binding site for nuclear factor Sp1. Oligonucleotides containing Repeat 2 + 3 bound SREBP and Sp1 as revealed by mobility shift assays. SREBP produced a DNase I footprint over the SRE-1 sequence, which was immediately adjacent to the footprint produced by Sp1. The current data are consistent with the concept that SREBP acts in concert with Sp1 to achieve high level, sterol-suppressible transcription of the gene for the LDL receptor.

SREBP-2, a second basic-helix-loop-helix-leucine zipper protein that stimulates transcription by binding to a sterol regulatory element.

We report the cDNA cloning of SREBP-2, the second member of a family of basic-helix-loop-helix-leucine zipper (bHLH-Zip) transcription factors that recognize sterol regulatory element 1 (SRE-1). SRE-1, a conditional enhancer in the promoters for the low density lipoprotein receptor and 3-hydroxy-3-methylglutaryl-coenzyme A synthase genes, increases transcription in the absence of sterols and is inactivated when sterols accumulate. Human SREBP-2 contains 1141 amino acids and is 47% identical to human SREBP-1a, the first recognized member of this family. The resemblance includes an acidic NH2 terminus, a highly conserved bHLH-Zip motif (71% identical), and an unusually long extension of 740 amino acids on the COOH-terminal side of the bHLH-Zip region. SREBP-2 possesses one feature lacking in SREBP-1a--namely, a glutamine-rich region (27% glutamine over 121 residues). In vitro SREBP-2 bound SRE-1 with the same specificity as SREBP-1a. In vivo it mimicked SREBP-1a in activating transcription of reporter genes containing SRE-1. As with SREBP-1a, activation by SREBP-2 occurred in the absence and presence of sterols, abolishing regulation. Cotransfection of low amounts of pSREBP-1a and pSREBP-2 into human embryonic kidney 293 cells stimulated transcription of promoters containing SRE-1 in an additive fashion. At high levels transcription reached a maximum, and the effects were no longer additive. The reason for the existence of two SREBPs and the mechanism by which they are regulated by sterols remain to be determined.

Novel mechanism of transcriptional activation of hepatic LDL receptor by oncostatin M.

In this paper we describe a sterol-independent regulation of low density, lipoprotein receptor (LDLR) transcription by the cytokine oncostatin M (OM) in HepG2 cells. We show that OM-induced expression is independent of cholesterol regulation and occurs at the transcriptional level. To elucidate regulatory mechanism(s), we constructed a luciferase reporter system comprising either the native LDLR promoter including repeats 1, 2, and 3, or a synthetic promoter vector containing repeats 2+3 only, allowing us to directly examine OM effects on individual elements. Specific mutants in repeats 1, 2, and 3 were made to facilitate the mapping of the OM effect on the promoter. Wildtype and mutant constructs were assayed for cholesterol and OM regulation. The results show that mutation within the core SRE-1 element of repeat 2 totally abolished cholesterol regulation but had no effect on OM inducibility. More interesting, a mutation within repeat 1 reduced basal transcription activity to 10% of the native promoter, but OM induction was unaltered. However, the identical mutation engineered in repeat 3 significantly decreased OM induction of LDLR promoter activity. These results suggest a novel regulatory role for the repeat 3 element in LDLR transcription.

Activation of LDL receptor gene expression in HepG2 cells by hepatocyte growth factor.

The effect of recombinant human hepatocyte growth factor (HGF) on low density lipoprotein (LDL) receptor gene expression was studied in the human hepatoma cell line HepG2. HepG2 cells were incubated with serum-free media in the presence and absence of HGF for various times and 125I-labeled LDL specific binding at 4 degrees C, uptake at 37 degrees C, and the levels of LDL receptor mRNA were measured. Incubation with HGF produced time- and concentration-dependent increases in 125I-labeled LDL binding (2-fold), uptake (2.5-fold), and LDL receptor mRNA (6-fold). HGF increased the rate of LDL receptor gene transcription 4- to 5-fold relative to that of several "house-keeping" genes as measured by nuclear run-on transcription. The half-life of LDL receptor mRNA, measured with actinomycin D, was not increased in HGF-treated cells. The stimulation of LDL receptor expression occurred independently of changes in cellular cholesterol or DNA biosynthesis or total cell protein. HepG2 cells were transiently transfected with plasmids bearing either three copies of repeats 2 and 3 (pLDLR(23)3LUC) or one copy of the LDL receptor promoter from -556 to +53 (pLDLR600LUC) linked to firefly luciferase. Incubation of pLDLR(23)3LUC, or pLDLR600LUC-transfected cells with HGF for 4 or 24 h at 37 degrees C produced a concentration-dependent increase in luciferase activity. A maximal stimulation of 3 to 6-fold was achieved for each construct at an HGF concentration of 100 ng/ml. In contrast, HGF had little or no effect on reporter activity in HepG2 cells transfected with a luciferase reporter plasmid bearing the HMG-CoA reductase promoter extending from -325 to +22. Thus, when compared to the native LDL receptor promoter, multiple copies of repeats 2 and 3 of the LDL receptor promoter can fully support activation of the luciferase reporter gene by HGF, demonstrating that the effect of HGF is mediated through the SRE-1. The lack of HGF effects mediated through the HMG-CoA reductase sterol regulatory element suggests, however, that sterol depletion may not be responsible for the induction of the LDL receptor promoter by growth factors. The signalling pathways or effectors responsible for activation of the LDL receptor and HMG-CoA reductase genes thus differ in their response to HGF. These data suggest that the level of SREBP's reaching the nucleus may be determined by as yet unidentified second messengers as well as by sterols.

Utilization of recombinant adenovirus and dominant negative mutants to characterize hepatocyte nuclear factor 4-regulated apolipoprotein AI and CIII expression.

Using recombinant adenoviral vectors and a dominant negative mutant of HNF-4, we have examined the contribution of hepatocyte nuclear factor 4 (HNF-4) to endogenous apolipoprotein AI and CIII mRNA expression. Overexpression of HNF-4 leads to a 7.4-fold increase in apolipoprotein CIII expression, while infection with the dominant negative mutant of HNF-4 reduces the level of apolipoprotein CIII mRNA by 80%, demonstrating that endogenous HNF-4 is necessary for apolipoprotein CIII expression. Experiments using the hepatoma cell lines, HepG2 and Hep3B, indicate that HNF-4 is also involved in the regulation of apolipoprotein AI expression in these lines. However, the effect of HNF-4 on apolipoprotein AI expression is much more dramatic in cell lines derived from intestinal epithelium. Infection of the intestinal-derived cell line IEC-6 with the HNF-4 adenovirus resulted in a greater than 20-fold increase in the level of apolipoprotein AI mRNA. These results indicate that HNF-4 does regulate apolipoprotein AI and CIII mRNA expression and suggest that HNF-4 is critical for intestinal apolipoprotein AI expression.

Identification of DNA sequences required for mouse APRT gene expression.

The mouse aprt promoter contains four GC boxes, which bind transcription factor Spl in vitro, and lacks both TATA and CCAAT boxes. Removal of the two most distal GC boxes of this promoter had little effect on APRT enzyme levels produced in a transient expression assay. Deletion of the distal three GC boxes resulted in a 50% reduction, and deletion of all GC boxes resulted in essentially complete loss of APRT activity. There are two predominant transcription start sites which are located within the region containing the GC boxes. The promoter behaved as a relatively strong promoter when compared to the RSV LTR promoter in a transient CAT assay, and operated in one orientation only. No upstream anti-sense transcripts were detected in either mouse CAK or liver cells, confirming that the mouse aprt promoter, unlike some other GC-rich promoters appears not to support bidirectional transcription.

SREBP-1, a basic-helix-loop-helix-leucine zipper protein that controls transcription of the low density lipoprotein receptor gene.

Sterol regulatory element 1 (SRE-1), a decamer (5'-ATC-ACCCCAC-3') flanking the low density lipoprotein (LDL) receptor gene, activates transcription in sterol-depleted cells and is silenced by sterols. We report the cDNA cloning of human SREBP-1, a protein that binds SRE-1, activates transcription, and thereby mediates the final regulatory step in LDL metabolism. SREBP-1 contains a basic-helix-loop-helix-leucine zipper (bHLH-ZIP) motif, but it differs from other bHLH-ZIP proteins in its larger size (1147 amino acids) and target sequence. Instead of an inverted repeat (CANNTG), the target for all known bHLH-ZIP proteins, SRE-1 contains a direct repeat of CAC. Overexpression of SREBP-1 activates transcription of reporter genes containing SRE-1 in the absence (15-fold) and presence (90-fold) of sterols, abolishing sterol regulation. We suggest that SREBP-1 is regulated by an unknown factor that is overwhelmed when SREBP-1 is overexpressed. Understanding the regulation of SREBP-1 may be crucial for understanding the control of plasma cholesterol in humans.

Regulation of lipoprotein metabolism by thiazolidinediones occurs through a distinct but complementary mechanism relative to fibrates.

Thiazolidinediones are antidiabetic agents, which not only improve glucose metabolism but also reduce blood triglyceride concentrations. These compounds are synthetic ligands for PPAR gamma, a transcription factor belonging to the nuclear receptor subfamily of PPARs, which are important transcriptional regulators of lipid and lipoprotein metabolism. The goal of this study was to evaluate the influence of a potent thiazolidinedione, BRL49653, on serum lipoproteins and to determine whether its lipid-lowering effects are mediated by changes in the expression of key genes implicated in lipoprotein metabolism. Treatment of normal rats for 7 days with BRL49653 decreased serum triglycerides in a dose-dependent fashion without affecting serum total and HDL cholesterol and apolipoprotein (apo) A-I and apo A-II concentrations. The decrease in triglyceride concentrations after BRL49653 was mainly due to a reduction of the amount of VLDL particles of unchanged lipid and apo composition. BRL49653 treatment did not change triglyceride production in vivo as analyzed by injection of Triton WR-1339, indicating a primary action on triglyceride catabolism. Analysis of the influence of BRL49653 on the expression of LPL and apo C-III, two key players in triglyceride catabolism, showed a dose-dependent increase in mRNA levels and activity of LPL in epididymal adipose tissue, whereas liver apo C-III mRNA levels remained constant. Furthermore, addition of BRL49653 to primary cultures of differentiated adipocytes increased LPL mRNA levels, indicating a direct action of the drug on the adipocyte. Simultaneous administration of BRL49653 and fenofibrate, a hypolipidemic drug that acts primarily on liver through activation of PPAR alpha both decreased liver apo C-III and increased adipose tissue LPL mRNA levels, resulting in a more pronounced lowering of serum triglycerides than each drug alone. In conclusion, both fibrates and thiazolidinediones exert a hypotriglyceridemic effect. While fibrates act primarily on the liver by decreasing apo C-III production, BRL49653 acts primarily on adipose tissue by increasing lipolysis through the induction of LPL expression. Drugs combining both PPAR alpha and gamma activation potential should therefore display a more efficient hypotriglyceridemic activity than either compound alone and may provide a rationale for improved therapy for elevated triglycerides.

The adipocyte specific transcription factor C/EBPalpha modulates human ob gene expression.

The ob gene product, leptin, apparently exclusively expressed in adipose tissue, is a signaling factor regulating body weight homeostasis and energy balance. ob gene expression is increased in obese rodents and regulated by feeding, insulin, and glucocorticoids, which supports the concept that ob gene expression is under hormonal control, which is expected for a key factor controlling body weight homeostasis and energy balance. In humans, ob mRNA expression is increased in gross obesity; however, the effects of the above factors on human ob expression are unknown. We describe the structure of the human ob gene and initial functional analysis of its promoter. The human ob gene's three exons cover approximately 15 kb of genomic DNA. The entire coding region is contained in exons 2 and 3, which are separated by a 2-kb intron. The first small 30-bp untranslated exon is located >10.5 kb upstream of the initiator ATG codon. Three kilobases of DNA upstream of the transcription start site has been cloned and characterized. Only 217 bp of 5' sequence are required for basal adipose tissue-specific expression of the ob gene as well as enhanced expression by C/EBPalpha. Mutation of the single C/EBPalpha site in this region abolished inducibility of the promoter by C/EBPalpha in cotransfection assays. The gene structure will facilitate our analysis of ob mutations in human obesity, whereas knowledge of sequence elements and factors regulating ob gene expression should be of major importance in the prevention and treatment of obesity.

The organization, promoter analysis, and expression of the human PPARgamma gene.

PPARgamma is a member of the PPAR subfamily of nuclear receptors. In this work, the structure of the human PPARgamma cDNA and gene was determined, and its promoters and tissue-specific expression were functionally characterized. Similar to the mouse, two PPAR isoforms, PPARgamma1 and PPARgamma2, were detected in man. The relative expression of human PPARgamma was studied by a newly developed and sensitive reverse transcriptase-competitive polymerase chain reaction method, which allowed us to distinguish between PPARgamma1 and gamma2 mRNA. In all tissues analyzed, PPARgamma2 was much less abundant than PPARgamma1. Adipose tissue and large intestine have the highest levels of PPARgamma mRNA; kidney, liver, and small intestine have intermediate levels; whereas PPARgamma is barely detectable in muscle. This high level expression of PPARgamma in colon warrants further study in view of the well established role of fatty acid and arachidonic acid derivatives in colonic disease. Similarly as mouse PPARgammas, the human PPARgammas are activated by thiazolidinediones and prostaglandin J and bind with high affinity to a PPRE. The human PPARgamma gene has nine exons and extends over more than 100 kilobases of genomic DNA. Alternate transcription start sites and alternate splicing generate the PPARgamma1 and PPARgamma2 mRNAs, which differ at their 5'-ends. PPARgamma1 is encoded by eight exons, and PPARgamma2 is encoded by seven exons. The 5'-untranslated sequence of PPARgamma1 is comprised of exons A1 and A2, whereas that of PPARgamma2 plus the additional PPARgamma2-specific N-terminal amino acids are encoded by exon B, located between exons A2 and A1. The remaining six exons, termed 1 to 6, are common to the PPARgamma1 and gamma2. Knowledge of the gene structure will allow screening for PPARgamma mutations in humans with metabolic disorders, whereas knowledge of its expression pattern and factors regulating its expression could be of major importance in understanding its biology.