A nuclear factor I-like activity and a liver-specific repressor govern estrogen-regulated in vitro transcription from the Xenopus laevis vitellogenin B1 promoter

Please download to get full document.

View again

of 15
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
A hormone-controlled in vitro transcription system derived from Xenopus liver nuclear extracts was exploited to identify novel cis-acting elements within the vitellogenin gene B1 promoter region. In addition to the already well-documented
Document Share
Document Transcript
  Vol. 9, No. 12 OLECULAR AND CELLULAR BIOLOGY, Dec. 1989, p. 5548-5562 0270-7306/89/125548-15 02.00/0 Copyright © 1989, American Society for Microbiology A Nuclear Factor I-Like Activity and a Liver-Specific Repressor Govern Estrogen-Regulated In Vitro Transcription from the Xenopus laevis Vitellogenin Bi Promoter BLAISE CORTHESY, JEAN-RENE CARDINAUX, FRANOIS-XAVIER CLARET, AND WALTER WAHLI* Institut de Biologie Animale, Universite de Lausanne, CH-1015 Lausanne, Switzerland Received 5 July 1989/Accepted 3 September 1989 A hormone-controlled in vitro transcription system derived fromXenopus liver nuclear extracts was exploited to identify novel cis-acting elements within thevitellogenin gene B1 promoter region. Inaddition to the already well-documented estrogen-responsive element (ERE), two elements were found within the 140 base pairs upstream of the transcription initiation site. One of them, a negativeregulatoryelement, is responsible for thelack of promoter activity in the absence of the hormone and, as demonstrated by DNA-binding assays, interacts with a liver-specific transcription factor. The second is required in association with the estrogen- responsive element to mediate hormonal induction and is recognized by the Xenopus liver homolog of nuclear factor I. Vitellogenin, fromwhich the yolk proteins are derived, is synthesized under estrogen control in oviparous vertebrates in the liver of the adult female. The vitellogenin genescan also be activated in males, in which they are not normallyexpressed, by a single injection of estradiol. The mecha- nisms regulatingthe activities of these genes act at both thetranscriptional and theposttranscriptional levels (44, 49,53, 55, 57). With respect to transcriptional regulation, it has been demonstrated in various genesystems that interactions between nuclear trans-actingregulatoryproteins and cis- acting promoter elements representa key step in thecontrol mechanism (29, 33,38, 43). Among the best-characterized trans-actingfactorsarethe steroid hormone receptors that, when associated with theirligand, stimulate transcription by binding to hormone-responsive promoterelements (9, 15,59). The comparison of frog andchicken vitellogenin genes first identified the13-base-pair (bp) promoter element 5'- GGTCANNNTGACC-3' as a possible candidate for an estrogen-responsive element (ERE) (56). Thiselementhas now been extensively tested and presents thecharacteristics of an estradiol-inducible enhancer element (5, 15, 22-24, 30, 42). The vitellogenin gene Bi 5'-flanking region containsthree imperfect copies of the 13-bppalindromicelement at positions -555,-334,and -314 (56). Functionalanalysesof the vitellogenin Bi promoter regiontransfected into the human estrogen-responsive cell line MCF-7 has shown that both the -334 and-314 imperfect 13-bpelements are required to form a strong ERE. The third -555 element present further upstream within the 5'-flanking region of the gene cannotconfer hormone responsiveness by itself. Interactions of the estrogenreceptor (ER) with the ERE of thevitellogenin gene and other estrogen-responsive genes have been extensively studied (16, 20,26,27, 31, 34, 50-52). However, further understanding of the molecular mecha- nisms involved in the hepatocyte-specific expressionof the vitellogenin genes requiresthe identification and character-ization of additionaltrans-acting factors and their target DNA sequences.Analysisof liver-specific transcription can * Corresponding author. benefit from the preparationof tissue nuclear extracts (14). Using such an extract prepared from estrogen-induced Xe- nopus liver nuclei, we succeeded in establishing the first steroid hormone-dependent in vitro transcription system (7). In these extracts, estradiol was necessary and sufficient to induce an ERE-containing vitellogenin promoter, while its activation was not possible in extracts prepared from un- stimulated male livers, which normally do notsynthesize vitellogenin. We concluded that both these extracts provide avaluable toolfor the study of vitellogenin gene cis-acting elements and the interaction with their cognate trans-acting proteinregulatory factors. In this paper, we present results obtained with extracts prepared from Xenopus female andmale livers, from a Xenopus kidney cellline, andfrom HeLa cells to further investigate the female- and hepatocyte-specific expression of the vitellogenin gene. We demonstrate that in addition to the ERE, at least twonewly characterized sequenceelements play crucialroles in the regulation of the vitellogenin gene promoter. MATERIALS AND METHODS Templates for in vitro transcription. The preparation of plasmid pBl(-596/+8)CAT8+ has been described previ- ously (42). Constructionof the 5' deletion mutants (-344, -302,-113,-95,and-67) carrying the vitellogenin (vit) promoter(-41/+8) will be described elsewhere (E. Martinez and W. Wahli, unpublished results). To generate the mutant truncated at position -138, we digested the AluI-Hinfl fragment (-138/+ 24) inserted in the SmaI site of pUC-9 with BamHI-BglII and subcloned the resulting -138/-42 digest into the linearized B1(-41/+8)CAT8+ vector resulting from thetreatment of pB+(-596/+8)CAT8+ with BamHI (site -596) and BglII (site -41). All mutants werechecked by dideoxynucleotide sequencing. For anestrogen-independent internal control, we used pSV2CAT (13), whose transcripts are primed by the same chloramphenicol acetyltransferase (CAT) primer as that which primes the specific vit-CAT transcripts. Plasmids were prepared by the alkaline sodiumdodecyl sulfate extraction procedure (1). DNA was further purified on one CsCl-ethidium bromide gradient and by gel filtration through a Bio-Gel A-50 (Bio-Rad Laboratories) 5548  STEROID-CONTROLLED IN VITRO TRANSCRIPTION 5549 column (18 by 1 cm) equilibrated and eluted in 10 mM Tris hydrochloride (pH 7.5)-150 mM NaCl-1 mM EDTA. Frac- tions(800 ,ul) containing the plasmid DNA were monitored by ethidium bromide fluorescence, pooled, and precipitated with 2 volumes of ethanol. The DNA was suspended in distilled water at aconcentration of 1 ,ug/,ul (measured by UV photometry at 260nm).Analysis on a 1% agarose gel indicated that at least 90% of the DNA was in thesuper-coiled form.Preparation of nuclear extracts fromXenopus livers, Xeno-pusB3.2 kidney cells, and humanHeLa cells. Xenopus laevisadults were induced twice by injection of 1 mg of 17p- estradiol(10 mg/ml in propanediol) into thedorsal lymph sac, with an interval of 3 days between the two injections, or they were left unstimulated. Extracts from liver nuclei were prepared essentially as described by Gorski et al. (14), except for thefollowing modifications. After removal, the livers were perfusedwithaheparin-phosphate-buffered sa-line buffer (for 1 liter, 0.5 g of heparin,0.15 gof KH2PO4, 0.07 g of Na2HPO4, 0.075 g of MgCl2 - 6H20, 0.075 gof CaCl2, 0.15 gof KCl, and 6 g of NaCl), using a syringe needle. Homogenization buffer I contained 2.4 M sucrose; homogenization buffer II was made of 2 M sucrose and 10% glycerol.All solutions except the dialysisbuffer contained the protease inhibitors aprotinin (0.5 jig/ml; Bayer),benza- midine (50 ,uM; Sigma Chemical Co.), leupeptin (0.5 jig/ml; Sigma), and pepstatin A (1 ,ug/ml; Sigma). KCI (60 mM) was included in the dialysis bufferinstead of40 mM KCl. Possible residual hormone in the extracts was eliminated during the dialysis step. The addition of 1 p.M 17,-estradiol to thebuffers used in the preparation did not generate more active extracts. In contrast, inclusion of the whole set ofprotease inhibitors prevented the appearance of bands of higher mobility in the gel shift assays and increased the relative abundance of transcription factors. Although the Xenopus liver is not a major sourceof digestive enzymes, this observation suggests that protease activities released from intracellular compartments during the homogenization steps can degrade a significant portion of the protein factors in the extracts. The protein concentrations, determined by the method of Bradford (2) with bovineserumalbumin as standard, were usually 4 and 7 mg/ml for maleand female extracts, respectively. TheXenopus kidney cellline B3.2 was a kind gift from R. Weber (University of Bern). Standard culture medium con- sisted of 60% Leibowitz L15 medium (GIBCO Laboratories) supplemented with 8% fetal calf serum (Inotech) and 5% tryptose phosphate (GIBCO), as well as 200 IU of penicillin (Sigma) per mland 0.2 p.l of streptomycin sulfate (Sigma) per ml.Cells were routinely grown to 50% confluence on 175- cm2 plastic dishes (Nunclon) in a Forma Scientific model 3157 water-jacketed incubator at 25°C. Cells were collected by washing the monolayer culture with a solution made of50 mM Tris hydrochloride (pH 7.4), 150 mM NaCl, and 1 mM EDTA. After further incubation at 25°C for 15 min, the cells no longer adhere firmly to theculture dish and are easily removedby gentle pipetting. This method presents two main advantages:scraping is not required, resulting in a higher yield of living cells (determined by the trypan blue assay), and the use of trypsin is avoided. HeLa S3 cells were grown in suspension in minimum essential medium modified for suspension cultures (GIBCO) supplemented with 8% fetal calf serum (Inotech) and mini- mum essential medium-nonessential amino acids (GIBCO), as well as with antibiotics and fungicides to a density of approximately 5 x 105 cells per ml. The culture medium was buffered with20 mM HEPES (N-2-hydroxyethylpiperazine- N'-2-ethanesulfonic acid) (Calbiochem) (pH 7.5, 25°C). Cell nuclear extracts were prepared as described by Shapiro et al. (45), with the addition of the same set of protease inhibitors used for the obtainment of liver nuclear extracts. The final dialysis buffer contained 10% rather than 20% glycerol. The protein concentration was determined with the Bradford reagent as described aboveand ranged between 7 and 15 mg/ml. Preparation of CTFINF-I extracted from HeLa cells and further purified by sequence-specific DNA affinity chroma- tography was kindly provided by N. Mermod and R. Tjian. In vitro transcription and transcript analysis. Transcription mixtures contained 33 mM HEPES (pH 7.9), 6 mM MgCl2, 60 mM KCI, 0.06 mM EDTA, 2.5 nM 17,B-estradiol (when not indicated otherwise), 0.6 mM dithiothreitol (DTT), 10% glycerol, 5 mM creatine phosphate, 1 p.1 of RNasin (40 U/,Il; Promega), and nucleoside triphosphates (at a concentrationof 0.6 mM for each) in a total volume of20 p.. For transcription with Xenopus liverextracts, each reaction contained 1.4 mg of protein per ml and 1,000 ngof the specific vit-CAT construct, as well as 250ng of the pSV2CAT template. With cell extracts, the final protein concentration was 2.0 mg/ml, and 750ng each of the vit-CATand pSV2CAT templates was added. The protein-to-DNA ratios correspond to optimal transcription conditions and were established for each extract. After preincubation on ice for 15 min, the reactions were initiated by theaddition of nucleoside triphosphates to the otherwise complete reaction and incubated at 30°C for 45 min. The reactions were terminated by the addition of180 p.1 of stop mix (1% sodium dodecyl sulfate, 200 mM NaCl, 25 mM Tris hydrochloride [pH 7.5], 5 mM EDTA, 1 p.1 of glycogen [20 mg/ml], 40 p.1 of proteinaseK), incubated at 37°C for 30 min, and extracted several times with chloropane (phenol-chloroform-isoamyl alcohol [50:48:2] saturated with 10 mM sodium acetate [pH 6.61-0.1 M NaCl-1 mM EDTA). After separation of the organic and aqueous phases, nucleic acids were precipitated by the addition of 20 p.1 of 3 M sodium acetate (pH 4.8) and 2.5 volumes of ethanol. The pellets were washed with 70% ethanol, dried, and suspended in 40 p.1 of distilled water. The RNAs synthesized in vitro were analyzed by primer extension. CAT primer (+35 to +65) was 5' labeled, using T4 polynucleotide kinase (Boehringer Mannheim Biochemi- cals) and [-y-32P]ATP (5,000 Ci/mmol; Amersham Corp.) according to the standard procedure (28). Primer (5 x 105 cpm) was then added to the RNA solutions, coprecipitated by ethanol, rinsed with 70% ethanol, dried under vacuum, taken up in 20 p.1 of a buffer consisting of 10 mM Tris hydrochloride (pH 7.5), 250 mM KCI, and 1 mM EDTA, and hybridized at 60°C for 60 to 90min. This was followed by the addition of 40 p.1 of a mixture containing 75 mM Tris hydrochloride (pH 7.5), 15 mM DTT, 12 mM MgCl2, 75 p.g of actinomycin D per ml, 0.75 mM deoxynucleoside triphos- phates, 12 U of RNasin (40 U/Iul; Promega), and 1 p.1 of mouse mammary leukemia virus reverse transcriptase (200 U/,Il; BethesdaResearch Laboratories, Inc.). Incubation was performed at 37°C for 60min. The reaction was stopped by the addition of 6.6 p.l of 3 M sodium acetate (pH 5.2) and 2.5 volumes of ethanol. After precipitation, the nucleic acids were dissolved in 10 p.l of formamide dye (99% deionized formamide, 0.25% xylene cyanol FF, 0.25% bromophenol blue, 1 mM EDTA), and 5-,u portions were analyzed on a 6% polyacrylamide (acrylamide-bisacrylamide [19:1])-7.5 M urea sequencing gel. Gels were exposed for 24 to 48h at -80°C with an VOL. 9, 1989  5550 CORTHtSY ET AL. intensifying screen. The correctly initiated vit-CAT tran- scripts were detected as products of 86 nucleotides, while transcriptssrcinating from pSV2CAT resulted in elongated molecules of 125 to 126 nucleotides. DNase I footprintassays. DNA-binding reactions were carried out in20-,ul volumes with approximately 3 ng of an end-labeled DNA fragment (1 x 104to 2 x 104 cpm),50 or 1,000 ng of thealternating copolymer poly(dI-dC)-poly(dI- dC) (Pharmacia) in 100 mM NaCl for reactions with liver nuclear extracts or cellline nuclear extracts, respectively, and variable amounts of extracts (as indicated in the figure legends) in a final buffer consisting of 45 mM HEPES (pH 7.9), 5 mM MgCl2, 0.08 mM EDTA, 0.08 mM DTT, 80 mM KCl, and 7.5% glycerol. The DNase I footprint assayswith purified CTF/NF-I were performed in 20-,ul volumes of amixture containing 12.5 mM HEPES (pH 7.9), 6.25 mM MgCl2, 5 ,uM ZnSO4, 0.5 mM DTT, 120 mM KCl, 10% glycerol, 0.05% Nonidet P-40, and 2% polyvinyl alcohol. The nuclear extract was added last, and the bindingmixtures were incubated for 30 min on ice. Freshly diluted DNase I (1 ,ul [50ng]; Boehringer Mannheim) in 10 mM HEPES (pH 7.9)-25 mM CaCl2 was added, and the digestion was allowed to proceed for 3 min at 0°C. The footprint reactions were terminated by theaddition of 80 ,ul ofa stop mix (20 mM Tris hydrochloride [pH 8.0], 250 mM NaCl, 20 mM EDTA, 0.5% sodium dodecyl sulfate, 50 ,ug of sonicated salmon sperm DNA per ml, and 125 ,ug of proteinase K per ml). The samples were incubated for 1 h at 45°Cand extracted twicewithchloropane, and thenucleicacids were precipitated by the addition of 2.5 volumes of ethanol. The DNA pellets were rinsed with 70% ethanol, dried under vacuum, sus- pended in 10 ptl of 99% formamide-1 mM EDTA with tracking dyes (0.25% xylenecyanol FF and 0.25% bro- mophenol blue), and denatured by heating at 95°C for 3 min. The footprints were analyzed on an 8% sequencing gel alongwith G and C+T chemicalsequencing reactions (32). The gel was then transferred to Whatman 3MM paper, dried, and autoradiographed at -80°C with a Fuji X-ray film, using an intensifying screen. Preparation ofthe probe for DNase I footprint assays. Plasmid pBl(-371/+24)-UC 9 was generated by digestion of a Xenopus vitellogenin gene B1 fragment (-596 to +24) with MaeI, filling in of the cohesive termini, separation on a 1.5% agarose gel, and insertion into the SmaI site of the pUC9 polylinker. The first step in the preparation of the DNA footprint probes was the digestion with a suitablerestriction enzyme of the cloned Bl vitellogenin promoter element. The linearized DNA plasmid was either 3' end labeled with the appropriate [a-32P]deoxynucleoside triphosphate (3,000Ci/ mmol) and the Klenow fragment of DNA polymerase I or treated with calf intestinal alkaline phosphatase and 5' end labeled with [y-32P]ATP (5,000 Ci/mmol) and T4 polynucle- otide kinasefollowing the standard procedures (28). A second restriction digestion yielded a labeled 413-bp frag- ment containing the Bi vitellogenin promoter. The generated 5'- or 3'-end-labeled DNA probes were purified on an 8% polyacrylamide preparative gel, isolated by electroelution, and extensively purified by severalextractions with phenol-chloroform. The DNA fragments were precipitated with ethanol, washed with 70% ethanol, dried, and finally stored in distilled H20 at -20°C (2 x 104cpm/l,l). Preparation of synthetic double-stranded oligonucleotides for band shift assays and competitions. Oligonucleotides wereused as probes forgel retardation assays and as competitors for in vitro transcription, DNase I footprints, and band shift experiments. They wereprepared as follows. Oligonucleo- tides ranging in length from 34 to 43 residues were chemi- cally synthesized on an automated DNA synthesizer (model 380B; AppliedBiosystems) and purified on a 20% polyacryl- amide-7.5 M urea gel(25 cm by 1.5 mm).The DNA was visualized by UV shadowing with polyethyleneimine-cellu- lose F precoated thin-layer chromatography plastic sheets (E. Merck AG, Darmstadt, Federal Republic of Germany) and electroeluted from the polyacrylamide gel at room temperature. The recovered aqueous phase was chloropane extracted once and ethanol precipitated. Complementary oligonucleotides were hybridized as described by Kadonaga and Tjian (21). For band shifts, the probeswere labeled by Klenow filling in with the appropriate radioactive and cold deoxynucleoside triphosphates. When theoligonucleotides were used as competitors for DNase I footprintinganalyses, 250ng of the selected oligonucleotide was incubatedwith the unspecific competitor [poly(dI-dC)-poly(dI-dC)], 2 to 3 ngof the end-labeled fragment, and the nuclear extract for 30 min at 0°C. For gel retardation assays, the oligonucleotide com- petitor, the nonspecific competitor [poly(dI-dC)-poly(dI-dC) or Escherichia coli DNA], and the proteins were mixed together and incubated on ice for 15 min before theaddition of the probe. Using thislatter mixing procedure, we deter- mined that approximately 10 times less specific competitor excess was necessary for detectionof the same effect. Continuation of the manipulation was as describedabove. For in vitro transcriptions, a 250-fold excess of oligonucle- otides was incubated with the nuclear extracts and the DNA plasmid templates on ice for 10 min before theaddition of the other constituents required for transcription. The competitor fragment, referred toas L, was prepared by digesting pUC19 plasmid DNA with EcoRl and HindIII and isolatingthe 50-bp polylinker from a 12% polyacrylamide nondenaturat- ing gel. Oligonucleotide concentrations were estimated by comparison with E. coli DNA standards, using the ethidium bromide dot method (28). Gel retardation assays. Protein-DNAcomplexes were formed as described above for the footprint experiments, except that0.5 to 1,ug of poly(dI-dC)-poly(dI-dC) or 0.5 to 8.0 ,ug of E. coli DNA was present in the incubation medium.The binding of CTF/NF-I was performedby the method ofJones et al. (19) in the following medium: 25 mM HEPES (pH 7.9)-50 mM NaCI-1 mM DTT-1 mM EDTA-10% glyc-erol-0.05 Nonidet P-40-250 ,ug of bovineserumalbumin per ml. After 30 min on ice, the samples were resolved on a low-ionic-strength native 8% polyacrylamide gel(15 by 0.15 cm; acrylamide-bisacrylamide weight ratio of 39:1). As a migration control, 2 ,ul of 25% Ficoll 400 (Pharmacia) and 0.25%bromophenol blue were added in the two lateral slots. Electrophoresis was performed at 150 V/20 mA in a cold room (4°C) in a buffer consisting of 6.7 mM Tris hydrochlo- ride (pH 7.5),3.3 mM sodium acetate, and 1 mM EDTA at the same voltage until bromophenol blue ran to two-thirdsof the gel (46). The buffer was recirculated at a speed sufficientto ensure that the pH remained constant in the two buffer compartments. The gel was then fixedfor 20 min in 50% methanol-10% aceticacid, transferred to Whatman 3MM paper, dried, and autoradiographed with an intensifying screen at -80°C. RESULTS In vitro transcription from the vitellogenin promoter in nuclear extracts from Xenopus liver, Xenopus kidney B3.2, and humanHeLa cells. In vitro transcription in nuclear extracts was used to identify and characterize cis-acting MOL. CELL. BIOL.  STEROID-CONTROLLED IN VITRO TRANSCRIPTION 5551 A ERE + 8 IF ~~ ___ pEMBL8B+ -596 CA T--- =a -344 -3 4 2 ---   -302 ----- 3 -- -1 13 -67 -41 -U-- -564 --t-- E nz _ -28 7- C I Stv N 0 nI sv o c - - r- _ ) COCO - - o I* vit   4*. I a   S L T: +H -H 1 00 1 10 <1 cl D SV Pisa   vt l FIG. 1. In vitro transcription of 5' deletion mutantsof the Bi vitellogenin promoter in various nuclear extracts. (A) Schematic representation of the vitellogenin sequences present in each of the mutants, pBl(A5'/-41/+8)CAT8+ (the CAT gene is not drawn to scale), and quantitative data from three independent transcription experimentswithfemale nuclear extracts and the A5' deletion constructs. The positions of the three 13-bp imperfect palindromic motifs are indicated by black inverted triangles. The two constructs used for experiments shown in Fig. 5 are also schematically drawn at the bottom:one of them lacks a portion (-287 to -42) of the vitellogenin Bi promoter region and has been called A3'/-41; the other was generated by cloning a 30-bpfragment of the promoter (the NRE) in front of the minimal (-41/+8) promoter. To allow quantification of these assays, each test vit-CATplasmid was transcribed in the presence of the control plasmid (pSV2CAT). The amount of RNA synthesized under thecontrol of the vit-CAT constructs was quantified and standardized according to the values obtained from the simian virus 40 promoter. The level of transcription (LT) of the pBl(-596/+8)CAT8+ template carried out in the presence (+H) or absence (-H) of hormone was arbitrarily fixed as 100%o. Data obtained from three independent experiments are given. (B) A series of constructs containing 5'-deleted portions of the vitellogenin Bi promoter (vit) was assayed for its ability to sustain transcription in vitro in nuclear extracts prepared from Xenopus female livers (Q). Foreach plasmid, the transcription was carried out either in the absence(-) or in the presence (+) of 2.5 nM 17,B-estradiol (H). (C) Utilization of the same set of 5' deletion mutants in extracts prepared fromuninduced Xenopus male livers (d) in the absence of estradiol. We demonstrated previously that these extracts do not respond to hormone (7). Lane C contains thetranscription product from the pSV2CAT DNA alone. (D) In vitro transcription of vit-CAT constructs in Xenopus B3.2 kidney cell nuclear extracts. (E) In vitro transcription of vit-CAT constructs in mammalian HeLa cell nuclear extracts. In lane C (control), the vit-CAT template was omitted and the pSV2CAT alone was tested;lane -41* contains the transcription products of the template pBl(-41/+8)CAT8+ alone. promoterelements. Towards this end, a promoter DNA fragment from thevitellogenin gene Bi was progressively truncated at its 5' endand placed in front of the CAT reportergene. The transcriptionalactivity of the mutants shown in Fig. 1A was tested in Xenopus liver extracts prepared from bothinduced females anduninduced males. The same constructs were also transcribed in nuclear ex- tracts from the Xenopus kidney cellline B3.2 and the human HeLa cell line. As an estrogen-independent internal control, we included the pSV2CAT construct. In the extract preparedfrom female livers that normally express the vitellogenin gene (Fig. 1B), the -596 and -344 ERE-containing constructs give maximal transcription in the presence of hormone but are silentin its absence. Thus, these results indicate that the ERE is also required for optimal transcription from the vitellogenin templates in B <1 <1 c1c1 10 10 9 9 77 4 4 A3'/-4 1 C/-41 Y E HeLa sv p4 vit 0 B3.2 I[ OC 0 C') U5 C' a n- 0-O1CO o __ me e m** *l df fttqt ft . VOL. 9, 1989  5552 CORTHESY ET AL. vitro. In addition, deletion of the imperfect element at position -555 does not result in a decreaseof activation, most likelyreflecting the factthat it is not essential to the function of the ERE. In contrast, templates fromwhich the ERE formed by the two -334and -314 imperfect palindro- mic elementshas been deleted behave differently and fall into two groups. The -302and-138 mutants are transcribed weakly or not at all, irrespective of the presence or absence of hormone. On the other hand, the -113,-95, -67, and -41 templates are constitutively expressed butwith decreas- ing efficiency as the vitellogenin 5'-end region becomes shorter. Furthermore, the rate of RNA synthesis is at least 10 times lower than that of the ERE-containing templates in the presence of hormone (Fig. 1B). Transcription of thecontrol template is unchanged in the different reactions, indicating that thefluctuations in transcriptional activity of the vitellogenin promoter mutants srcinate from selective activation or repression and are not due to competition between the two templates for transcription factors or RNA polymerase II. With the extract from the uninduced male liver not normally expressing the vitellogenin genes (7), the -596,-344,-302, and -138 constructs are not expressed. As in the case of the female liver nuclear extract, the -113, -95,-67, and -41 mutants are active (Fig. 1C). Together, these results suggest that, in these extracts, the vitellogenin promoter is under negative control. This repression can be released artificially by the deletion of the 25 bp between positions -113and -138 or can be overcome more naturally by the addition of hormone to the female liver extract. The apparently increased activity from the simian virus 40 (SV) promoter signal with deletions -113,-95, and -67 is due to a minor comigrating vit bandwhich was identified in an assay performed in the absence of the pSV2CAT control template (data not shown). It remains to bedetermined whether there is a direct relationship between these in vitro results and the behavior of the same constructs in cultured hepatocytes. Next, we examined whether the vitellogenin gene Bi is also repressed in nuclear extracts from cells of nonhepatic srcin. Since the preparation of sufficient amounts of nuclear extracts from other Xenopus organs is not feasible at the moment, we used the HeLa and Xenopus kidney B3.2 cell lines. In extracts from these two cell lines, all mutants tested are constitutively expressed (Fig. 1D and E). These results are consistent with the basal activity that we haveobserved after transfection of similar Bi constructs into cultured B3.2 cells (41). Since the B3.2 cells do not express their endoge- nous vitellogenin genes, thisresult suggests that the repres- sion mechanism in these cells must be different from that observed in the noninduced hepatocytes. In both extracts, and unlikethe liverextracts, the -41 deletion generates a second transcript revealed by a 130-nucleotide extension product (Fig. 1D, lane -41 and Fig. 1E, lanes -41 and -41*). The reason for the appearanceof this secondary start site was not further investigated. In addition, thesenuclear extracts and that obtained from the male liver appear to support basal transcription activity from the vitellogenin Bi promoter better than the female liver extract does We interpret this effect in terms of the relative abundance or nature of transcription factors present in the different nu- clear extracts. In summary, the data presented so far show that the 140 bp immediately upstreamfrom the transcription initiation site exhibit at least two different functions. The first 113 bp play a role in basal promoter activity in nuclear extracts derived from both hormone-treated anduninduced livers(Fig. 1B). Sequences between positions -113and -138 confer astrongrepressor effect on the promoter activity in vitro in the absence of hormone. With female liver nuclear extracts, release of this inhibition requires both the addition of estradiol to the reaction mixture and the presenceof the ERE on the template. Under these conditions, repression release is associated with a strong induction. Interaction of nuclear proteins from Xenopus liver, Xenopus kidney B3.2, and humanHeLa cells with promoter elements of the vitellogenin gene Bi. The functional identification in the liverextracts of two new regions, i.e., the one responsible for the repressed state in the absence of hormone (-138 to -113) and that involved in constitutive expression (-112 to +8), raises the possibility that trans-acting factors recogniz- ing specific sites withinthese regions are present inliver nuclear extracts. Since the 140 bp immediately upstream from the transcription startsite compose these two regions, we concentratedour attention on this promoter DNA frag- ment. Footprintanalyses (11) were carried out with a vitelloge- nin gene Bl fragment extending from positions -371 to +24. The four different nuclear extracts used in the transcription experiments described abovewere assayed for their abilitytointeract with the probe (Fig. 2). Several protected regions were observed at various locations around the startsite and within the upstreampromoter sequences, reflecting the presence of distinct DNA-binding activities. The left panelof Fig. 2A shows the DNase I protection pattern obtainedwith increasing amounts ofnuclear extract from induced female livers. In addition to the TATA box, three protected se- quences called A, B, and C are observed in the 140-bp region upstream of the transcription start site. That at least two different factors interact with sites B and C is suggested by the observation that the protection of the latter occurs at higherprotein concentrations and is confirmed (see below) by competitionexperiments. A closely related protectionpattern is generated withnuclear extracts from uninducedmale livers (Fig. 2A, center panel). With nuclear extracts from both the cell lines B3.2 and HeLa, careful analysis of the footprint profile reveals a slightly more extended shield- ing of region C. For region B, HeLa cells and liverextracts generate a similar pattern, in contrast to the B3.2 cell extract that confers only partial protection. The A region is pro- tected differently by the extracts from both cell lines when compared with the liver extracts, as revealed by the changes in the pattern of hypersensitive sites on the coding strand. These variations in the protection may reflect differences in the relative abundance or binding affinity of cell-specific factors ofvarious srcins. Alternatively, cell-specific post- translational modifications of pre-existing factors could also explain why similar binding activities affect the expression of the vitellogenin promoter in vitro differentially. A comparison of the nucleotide sequences of sites A, B and C protected by Xenopus nuclear liver factors with the recognition sites of known transcription factors revealeda strong homology between sequences of the B domain and the adenovirus type 2 nuclear factor I (NF-I) consensus recognition sequence [5'-TGG(A/C)N5GCCAA-3'] (19; Fig. 3B). A weak similarity was also found between a portion of the A region and the USF-binding site (39; Fig. 3B). In contrast, C is apparently unrelated to the recognition site of previously described transcription factors. Competition DNase I footprintreactions were carried outwith the sameBl promoter fragment as described above in the presence of anexcess ofdouble-stranded oligonucleotides correspond- ing to the A, B, and C binding sites (oligonucleotides A, B MOL. CELL. BIOL.
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks