Original Article

HNF-4α Regulates Expression of Human Ornithin Carbamoyltransferase through Interaction with Two Positive Cis-Acting Regulatory Elements Located in the Proximal Promoter(ornithinecarbamoyltransferase/OTC/HNF-4α/transcription/regulation/promoter/DNaseIfootprinting/EMSA)

O. LUKŠAN1,L.DVOřáKOVá2, M. JIRSA1

1LaboratoryofExperimentalHepatology,InstituteforClinicalandExperimentalMedicineandInstituteofMedical Biochemistry and Laboratory Diagnostics, First Faculty of Medicine, Charles University in Prague, Czech Republic2Institute of Inherited Metabolic Disorders, First Faculty of Medicine, Charles University in Prague and General University Hospital in Prague, Czech Republic

Folia Biologica (Praha) 60, 133-143 (2014)

ReceivedOctober22,2013.AcceptedFebruary14,2014.

The study was supported by Charles University in Prague, Czech Republic, grantsGAUK-251168_99910 andSVV2013/266504,and by the Ministry of Health, Czech Republic – conceptual de-velopment of research organization („Institute for Clinical and Experimental Medicine – IKEM, IN 00023001“, and GeneralUniversityHospital–VFN,64165).

Correspondingauthor: MilanJirsa,InstituteforClinicalandEx-perimentalMedicine(IKEM),LaboratoryofExperimentalHepa-tology, Vídeňská 1958/9, 140 21 Prague 4, Czech Republic. E-mail:milan.jirsa@ikem.cz

Abbreviations:C/EBP–CCAAT-enhancer-bindingprotein,Cis+ – cis-acting regulatory element, COUP-TF – chicken ovalbumin upstream promoter-transcription factor, DPE – downstream pro-moter element, EMSA – electromobility shift assay, GATA1 – GATA-binding factor 1, HNF – hepatocyte nuclear factor, Inr – initiator element, OTC – ornithine carbamoyltransferase, OTCD –ornithinecarbamoyltransferasedeficiency,PCR–polymerasechain reaction, PIC – pre-initiation complex, Sp1 – specificityprotein1,TBP–TATA-boxbindingprotein,TFII–generaltran-scription factor for RNA-polymerase II, TIS – translation initia-tion site, TSS – transcription start site.

Abstract. OTC encodes ornithine carbamoyltrans-ferase, mitochondrial matrix enzyme involved in the synthesis of urea. The tissue-specific expression of OTC in the liver and intestine is dependent on the interaction of OTC promoter with an upstream en-hancer. HNF-4 and C/EBPβ are crucial for this inter-action in the rat and mouse. In the present study we focused on characterization of elements involved in the regulation of OTC transcription in human. Using a set of 5’-deleted promoter mutants in a reporter assay we identified two positive cis-acting regulatory elements located at c.-105 and c.-136 within the hu-

man OTC promoter. Both are essential for the tran-scriptional activity of the promoter itself and for the interaction with the enhancer. Protein binding at the corresponding sites was confirmed by DNase I foot-printing. Electro mobility shift assay with a specific competitor and anti-HNF-4α antibody identified the DNA-protein binding sites as HNF-4α recognition motifs. A third HNF-4α binding site has been found at the position c.-187. All three HNF-4α binding sites are located within 35 bp upstream of the transcrip-tion start sites at positions c.-95, c.-119 (major) and c.-169 (minor). A series of C/EBPβ recognition motifs was identified within the enhancer. Involvement of C/EBPβ and HNF-4α in the promoter-enhancer in-teraction is further supported by a massive DNA-protein interaction observed in the footprinting and EMSA assays. Since the OTC promoter lacks general core promoter elements such as TATA-box or initia-tors in standard positions, HNF-4α most likely plays an essential role in the initiation of OTC transcrip-tion in human.

IntroductionOrnithine carbamoyltransferase (OTC, EC 2.1.3.3;

MIM#300461)isamitochondrialmatrixenzymecata-lysing the synthesis of citrulline – the second step of urea biosynthesis. The human OTC gene, located on chromosome Xp2l.1, comprises 10 coding exons(Lindgrenetal.,1984;Hataetal.,1988).Ornithinecar-bamoyltransferasedeficiency(OTCD;OMIM#311250)is the most common inherited urea cycle disorder (BrusilowandHorwich,1995).Althoughitsincidenceis not high, the severity of clinical manifestations re-quires an early diagnosis, especially in males. Variations

134 Vol.60O. Lukšan et al.

in the OTC coding regions were thoroughly investigated (Tuchman et al., 2002;Yamaguchi et al., 2006; Stor-kanovaetal.,2013);however,moleculardiagnosisstillfailsin20–25%ofcases.Molecularbasisofthesemi-dominant disease in some of these cases might be rela-ted to genetic defects in the regulatory elements. Re-cently, we have reported a first patient manifesting amild form of OTCD caused by a point mutation in the promoter(Luksanetal.,2010).

The human OTC transcription start site (TSS) wasfirstassignedtothepositionc.-135(GenBankK02100;Horwich et al., 1984).Later publisheddata suggestedthat transcription of human OTC does not initiate at a particularsitebutwithina70bpregion,mostlyatpre-ferredpositionsc.-95,c.-120,c.-150,c.-161andc.-166(BrusilowandHorwich, 1995).Three alternativeTSSlocated at the positions c.-95, c.-119 and c.-169 up-stream of the initiation codon were identified usingrapid amplification of cDNA ends in total liver RNA(Luksanetal.,2010).

OTCisexpressedpredominantlyinhepatocytesandepithelial cells of the intestinal mucosa (Ryall et al., 1985;Hamanoetal.,1988).Themechanismofthetis-sue-specificpatternofOtcexpressionwasfirststudiedinthemouse.Intransgenicexperimentsusingsparsefurmice, Jonesetal. (1990)showed that the ratOtc pro-moterwascapableofdirectliver-andintestine-specifictranscription,buttheexpressionleveloftheintroducedgene was much lower than in the case of wild-type mouse with endogenous Otc expression.Inourpreviousstudy(Luksanetal.,2010)weinvestigatedthe793bphuman OTC promoterlocalized46bpupstreamofthetranslationinitiationsite(TIS).InlinewiththefindingsbyJonesetal.(1990)weobservedasignificantincreaseof the reporter luciferaseactivity inHep-G2cells,butnotinHEK293cells,achievedbysubcloninga465bpsequenceofthedistalenhancerlocalized9kbupstreamof TIS. This demonstrates the importance of the promot-er-enhancer interaction for the regulation of human OTCexpression.

Two positive cis-acting regulatory elements located aroundthepositions-187and-104upstreamofTISwereidentifiedintheratandmousepromoter.Bothcanbindan identical set of two members of the orphan receptor superfamily:hepatocytenuclear factor4 (HNF-4)andchicken ovalbumin upstream promoter-transcription fac-tor(COUP-TF).Inco-transfectionexperiments,HNF-4activated transcription of a reporter gene from the Otc promoter, whereas competition with COUP-TF had a repressoreffect(Kimuraetal.,1993).Moreover,HNF-4seemstocontributetothetissue-specificexpressionofOtc as its mRNA is present only in a limited number of tissue types such as the liver, kidney and intestine (Drewesetal.,1996).Ontheotherhand,COUP-TFex-hibitsratheraubiquitousexpressionpattern(Miyajimaetal.,1988),andthereforecompetitionforthebindingsite can result in repression of Otc in other tissues. HNF-4αiscriticalforureahomeostasisinthemouse

(Inoue et al., 2002) since liver-specific HNF4α-null

miceexhibitedelevatedserumammoniaanddecreasedserumurea,whileonlyasignificantdecreaseinproteinexpressionwithin theurea cycle enzymeswas that ofOTC.Similarlyas in therat, twoHNF-4αrecognitionmotifswereidentifiedwithinthemouse5’flankingre-gionatpositions-111bpand-191bpupstreamoftheTIS,andHNF-4αbindingwasconfirmedusingspecificantibody. Moreover, co-transfection of reporter constructs carrying the mouse OtcpromoterwitharatHNF-4αex-pression plasmid increased the reporter gene activity in CV-1cells(Inoueetal.,2002).Nishiyorietal.(1994)demonstratedthattheratOtc

enhancer is activated by co-operation of two liver-selec-tivetranscriptionfactors:HNF-4andCCAAT-enhancer-bindingprotein (C/EBP).Furthermore,co-transfectionexperimentsshowedthatbothHNF-4andC/EBPβarerequiredandneitherofthemaloneissufficientforacti-vation of the reconstituted enhancer in non-hepatic cells. Combinatorial operation of these two liver-enriched (butnotstrictlyliver-specific)transcriptionfactorsleadsto more restricted liver-specific transcription of Otc (Nishiyorietal.,1994).In the present studywe identified key elements re-

sponsible for the regulation of OTC transcription in hu-man. We suggest that initiation of OTC transcription in the absence of the general core promoter elements such asTATA-boxorinitiatorsinstandardpositionsismedi-ateddirectlybyHNF-4α.

Material and Methods

Reporter plasmid constructs

Fourteen reporter constructs were generated from pGL3-basicvector(Promega,Mannheim,Germany)bysubcloningthe465bpOTCenhancer(9constructs)andpromotervariants (12constructs) inSalI and XhoI re-striction site, respectively. Six promoter variants 685,458,223,173,133and81bpinlengthwerepreparedbyaseriesofdeletionsfromthe5’-endofthefull793bppromoter(Table1).Allsequencesandnucleotideposi-tions used in the study are in accordance with GenBank NG_008471.1andNM_000531.5.

Transient transfections and reporter gene assaysHumanhepatoblastomaHep-G2(ECACC85011430)

cells were cultivated in minimal essential Eagle’s me-dium supplementedwithL-glutamine, 10% foetal bo-vine serum, 100 units/ml penicillin, and 100 mg/mlstreptomycin, all purchased from Sigma-Aldrich, St. Louis,MO.Theculturesweremaintainedinhumidifiedatmosphereat37°Cand5%CO2. Transfection was per-formed 48 h after seeding using FuGeneHD reagent(Promega)inthereagenttoDNAratio5 : 1accordingtothemanufacturer’sprotocol.Thecellscultivatedin96-wellplateswereco-transfectedwithpGL3-derivedex-perimentalplasmids(0.1μgperwell)andaRenilla lu-ciferase-containingreporterplasmidpRL-TK(0.01μgperwell). Cell lysis and determination of the luciferase

Vol.60 135ExpressionofHumanOTCIsRegulatedbyHNF-4α

activity was performed according to the manufacturer’s instructions using the Dual Luciferase Reporter Assay System(Promega).Allmeasurementswereperformedin a Synergy 2 automatic microplate reader (BioTek,Winooski,VT).

Signal of the Photinus luciferase was normalized by Renilla luciferase luminescence in all measurements. Alltransfectionsweredoneinsixreplicatesandthere-sultswerecalculatedfromatleastsixseparateexperi-ments. Because of the high inter-assay variation, the transcriptional activity of each promoter variant is pre-sentedrelativetotheactivityofpGL3_793/Enh(posi-tive control, 100%). Statistical evaluationswere per-formed after logarithmic transformation using the SigmaPlotsoftware(SystatSoftware,Chicago,IL).Allgroups passed the test for normality and analysis of var-iance using one-way ANOVA. Multiple comparisons were performed using the Bonferroni method.

Nuclear extract preparationNuclearandcytoplasmicextractswerepreparedfrom

Hep-G2cellsusingtheNE-PERnuclearextractionkit(Thermo Scientific, Waltham, MA) according to themanufacturer’sinstructions.Extractsweredialysedfor4 h against 50 volumes of buffer containing 20 mMHEPES (pH 7.9), 20%glycerol, 0.1MKCl, 0.2mMEDTA, 0.5 mM PMSF, and 0.5 mM DTT. Protein con-centrations were determined using the Pierce BCA ProteinAssayKit(ThermoScientific).

DNase I footprintingThe DNase I footprinting method based on the sepa-

rationoffluorescentlylabelledproductsusingcapillaryelectrophoresis was derived from the literature (Wilson et al., 2001; Zianni et al., 2006).A fluorescent probespanning the 223bpOTC proximal promoter, 5‘- un-

translatedregionandapartofexon1waspreparedbypolymerasechainreactions(PCR)using6-carboxyfluo-rescein (6-FAM) labelled reverse primer 5’-(6-FAM)-TTTCGAACCATGAAGTTGTGACCA-3’andforwardprimer 5’-AAATGAGGAGGCCAGGCAAT-3’. The200μlreactionmixturescontaining100μloftheproteinbindingbuffer(2×),2μgofpoly(dI-dT),133μg(sam-ple) or 0 μg (control) of Hep-G2 nuclear extract andBSAupto400μgoftotalproteinwereincubatedwith600fmolesoffluorescentlylabelledprobefor20minonice.DNAcleavagewasperformedwith10μlofDNaseIat the concentration of 4 U/ml (Fermentas – ThermoScientific, Pittsburgh, PA) after addition of 20 μl ofMg2+/Ca2+ solution (10×).The cleavagewas inhibitedafter 5 min by addition of the DNase stop solution and supplemented with GeneScan™ 500 ROX™ (Applied Biosystems,FosterCity,CA)servingasasizestandardandassaycontrol.TheDNAfragmentswerepurifiedbyphenol-chloroformextractionfollowedbyethanolpre-cipitation, dissolved in 15 μl of Hi-Di formamide(AppliedBiosystems)andseparatedbycapillaryelec-trophoresis in the Applied Biosystems 3130 GeneticAnalyzer. The composition of the buffers and solutions used was as follows: protein binding buffer (2×) –50mM Tris-HCl, pH 8.0, 100 mM KCl, 12.5 mMMgCl2,1mMEDTA,1mMDTT,20%(v/v)glycerol;Mg2+/Ca2+ solution (10×) – 50 mM MgCl2, 25 mMCaCl2;DNase stop solution–200mMNaCl, 50mMEDTA,1%(v/v)SDS.

Electromobility shift assaysDouble-stranded biotin-labelled probes were pre-

pared by standard PCR or by hybridization of comple-mentary oligonucleotides. The binding reaction was car-riedoutin20μlofamixturecontaining 10 mM Tris-HCl (pH8.0),50mMNaCl,1mMMgCl2, 0.5 mM EDTA,

Table 1. Reporter constructs used in the dual luciferase reporter assay

Construct Promoter variant EnhancerFragment length Position

pGL3_793/Enh 793bp c.-838..c.-46 yespGL3_685/Enh 685bp c.-730..c.-46 yespGL3_458/Enh 458bp c.-503..c.-46 yespGL3_223/Enh 223bp c.-268..c.-46 yespGL3_173/Enh 173bp c.-218..c.-46 yespGL3_133/Enh 133bp c.-178..c.-46 yespGL3_81/Enh 81bp c.-126..c.-46 yespGL3_0/Enh - - yespGL3-793/0 793bp c.-838..c.-46 -pGL3_223/0 223bp c.-268..c.-46 -pGL3_173/0 173bp c.-218..c.-46 -pGL3_133/0 133bp c.-178..c.-46 -pGL3_81/0 81bp c.-126..c.-46 -pGL3_793(HNF4m)/Enh 793bp c.-838..c.-46 yespGL3_0/0 - - -

136 Vol.60

0.5mMDTT,4%glycerol1μgofpoly(dI-dT),25μgofnuclearextractand300fmolesofbiotin-labelledprobe.Inthecaseofcompetitionassays,0,1,4and12pmolesofthenon-labelledprobe[-98..-126]ora200-foldex-cess of a double-stranded probe AGTGAGGGTCAA-AGTTCATATCAC,derivedfromtheconsensusHNF-4αrecognitionsequence,wasusedasaspecificcompetitor.The sampleswere incubated for 20min at room tem-perature, separated using non-denaturing polyacryla-mide gel electrophoresis (6% polyacrylamide gel in0.5×TBE,100V,50min)andelectro-blottedat300mA,for 30min on theAmershamHybond-N nylonmem-brane (GE Healthcare Bio-Sciences AB, Uppsala, Sweden). TheDNA-protein complexeswere fixed bybakingthemembranefor2hinadryovenat80°C.Themembrane was processed using the Chemiluminescent Nucleic Acid Detection Module (Thermo Scientific,Rockford,IL)accordingtothemanufacturer’sinstruc-tions and analysed with the FUJI LAS 3000 system(Fujifilm,Tokyo,Japan).Insupershiftassays,goatanti-HNF-4αpolyclonalantibody(SantaCruz,Dallas,TX)waspre-incubatedwith thenuclearextract for20minprior to addition of labelled probes. Detection of the anti-HNF-4α antibody was performed using HRP-conjugated polyclonal rabbit anti-goat antibody (Thermo Scientific).

Prediction of transcription factor binding sitesPrediction of transcription factor binding sites was

performed with MatInspector (Genomatix Software,München,Germany)andMatch(Biobase,Wolfenbuettel,Germany) bioinformatic tools using the MatBase 9.0andTRANSFAC7.0databaseversions.

Results

In silico prediction of the main regulatory domains

The sequences of rat Otc and human OTC promoter reach about 65% similarity. Higher homology is pre-servedintheproximalpartofthepromoterupto450bpupstream of the TIS. DNase I footprinting and function-al characterization of deletion mutants generated from the rat Otc 5’-flanking region uncovered four regionsprotected against DNase I cleavage and twoHNF-4αbindingmotifslocated104and187bpupstreamofTIS(Murakamietal.,1990;Kimuraetal.,1993) (Fig.1).Homologous motifs were found in the mouse by Inoue etal.(2002).Similarbindingsiteslocatedatc.-105andc.-187 were predicted in the human sequence usingMatInspector (MI)andMatch (M) (Fig.1).While theHNF-4αbindingmotiflocatedatc.-105wasrecognizedwith both tools (MI score 0.884,M score 0.883), theotherlocatedatc.-187waspredictedonlywithMatch(score 0.910).MatInspector recognized a binding sitefor another member of the nuclear receptor superfamily atthesameposition–peroxisomeproliferator-activatedreceptor g (PPARg,MIscore0.833).BothHNF-4and

PPARg recognize similar sequence motifs. Transfection experiments revealed that preferential binding is par-tially determined by the sequence of the core motif, but it is much more dependent on the promoter context(Nakshatri and Bhat-Nakshatri, 1998). Therefore, thesequencemotif can be considered a HNF-4α bindingsite as well.TwoadditionalpotentialHNF-4αbindingsiteslocat-

ed136bp(MI/Mscores0.923/0.746)and588bp(MI/Mscores0.822/0.835)upstreamofTISandonesiterecog-nizedbyC/EBPβ located221bpupstreamof theTIS(Mscore0.970)werepredictedwithinthehumanOTC promoter. None of these binding sites was recognized either in the rat or in the mouse promoter because of variationsinthematrixcoremotifs.

As mentioned above, human OTC lacks a particular site of transcription initiation, indicating the absence of a strong core promoter element such as TATA-box.Accordingly,inourpreviousstudy(Luksanetal.,2010)we found three alternative TSSs – two major ones start-ingatpositionsc.-95,c.-119andaminoronestartingatc.-169.Ofthesethreeonlythec.-169transcriptvariantimplies a potential TATA-like element CATAAA located 25–30bpupstreamoftheTSS;however,itsregulatoryrole seems to be low because cutting off the correspond-ing region did not lead to a remarkable decrease in the promoteractivity(Fig.2,constructs173/Enhand133/Enh). Moreover, this fully conserved motif acts as aknownHNF-4αbindingsitelocated187bpupstreamofTIS in rodents.

Identification of cis-acting regulatory elements and delimitation of the core promoter

Thetranscriptionalactivityofsix5’-deletedpromotermutantsranging685,458,223,173,133and81bp(allwith3’-ends46bpupstreamofTIS–seeFig.1)wascomparedwiththefull793bphumanpromoterinpGL3constructs containing the human enhancer sequence. The empty plasmid pGL3_0/Enh containing only theenhancer was used as a negative control to eliminate the influenceoftheenhanceralone.Theresultshaveshownthata5’-deletionof108bphasnoinfluenceonthere-porter activity (both mean and median varied around 116 % of the activity of pGL3_793/Enh), while the5’-deletionof335bpleadstoadecreasedluciferasesig-nal(medianandmeanreached57%and54%,respec-tively) (Fig.2). Incontrast, increasedreporteractivitywasobservedwiththe223bppromotervariant(median146%,mean152%).Theresponseremainedconstantthrough the173and133bpvariants (medians137%and156%,means150%and167%respectively),whiletheshortesttestedvariantof81bpshowedslightlylow-er transcriptional activity when compared with the full promoter(median81%,mean80%)(Fig.2).Thede-creasewasstatisticallysignificantwhencomparedwiththe133,173and223bppromotervariants,butnotsig-nificantwhencomparedwiththe458and685bpvari-ants(Fig.2).Thereportergeneactivityobservedwith

O. Lukšan et al.

Vol.60 137

thepGL3vectorcarryingtheenhanceralonewaslowerthan20%ofthepGL3_793/Enhandsignificantlylowerthan with all other constructs. Based on these resultswe defined the humanOTC

corepromoterasthe223bpregionlocated-46..-268bpupstream of TIS containing three transcription initiation sites and two positive cis-actingelements:Cis1+ locat-edwithinthefirst81base-pairs,reaching83%ofthefull promoter activity, and Cis2+ located within the up-stream52bpsequence.Inthefollowingexperimentsweevaluatedthecontri-

bution of Cis1+ and Cis2+ to the transcriptional activity ofthepromoteralone(Fig.2,rightplot).Thecorepro-moterexhibitedtwicehigherreporteractivity thanthefull promoter and the level of luciferase signal remained stablewithin all tested constructs (81/0, 133/0, 173/0and223/0).Finally,thereporteractivityobservedwithallpromotervariantswassignificantlyhigherthanwith

thenegativecontrolpGL3_0/0.TheinfluenceofCis2+ was not observed in this case, indicating two possible explanations: first,Cis2+ regulatory element contrib-utes only to the interaction with the enhancer whereas Cis1+ plays a role in both interaction with the enhancer and activation of the promoter. Second, Cis1+ and Cis2+ operate simultaneously but they contribute to for-mation of different mRNA variants originating from various TSSs. In such case, absence of one element re-sponsible for formation of one transcript variant could be compensated by the action of the other element re-sponsible for production of the other transcript originat-ing from a different TSS. Such compensatory effect could maintain the overall transcriptional activity of the promoter. However, the compensatory capacity may reach its limit when the transcriptional activity is accel-erated by the promoter-enhancer interaction, and there-fore the promoter variant containing only the Cis1+ ele-

ExpressionofHumanOTCIsRegulatedbyHNF-4α

Fig. 1. Sequence alignment of the rat Otc(lowercase)andhumanOTC (uppercase)5’flankingregion. The regions pro-tected against DNase I cleavage in the rat are underlined. The 5’ ends of the rat promoter fragments studied by Murakami et al. (1990) and their relative activity in chloramphenicol acetyltransferase reporter assays aremarkedwith arrows.Transcriptionfactorbindingsitesidentifiedascis-acting regulatory elements in the rat are boldfaced. Transcription factor binding sites predicted by Match and MatInspector in human OTC promoter are boldfaced and highlighted in grey. Black arrows indicate human OTCpromotervariantstestedinreporterassays.ThreealternativeTSSsidentifiedwithinthehu-man liver mRNA are marked with black triangles.

138 Vol.60

mentaloneexhibitsloweractivitywhencomparedwiththevariantcarryingbothregulatoryunits(Fig.2,vari-ants81and133).

Identification of DNA-protein interaction sites by DNase I footprinting

Inaseriesofexperimentswe identifiedfiveDNaseI-protectedregionslocalized92–102(A),105–114(B),135–145(C),150–172(D)and180–192(E)bpupstreamofTIS(Fig.3A).Thesequencemotifsobservedintheprotected regions B and E are very close to the protect-ed regions in the rat promoter that partially overlap withtheHNF-4 andCOUP-TF binding sites (Fig. 1).AnotherHNF-4α recognition site predictedwith highmatrixsimilarityscoreislocatedclosetotheprotectedregion C.

No conventional transcription factor-binding sites were recognized within the protected region A that coin-

cides with the c.-95 bp TSS and overlaps with a se-quence motif related to initiator element Inr (MI score 0.967).ThesamesequencewasrecognizedwithMatchasaCapsitewithmatrixsimilarityscore0.995.Agen-eral core promoter motif Inr is known to moderate tran-scription by direct interaction with the transcription fac-torIID(TFIID)complexinbothTATA-containingandTATA-less promoters. Unfortunately, the role of the ini-tiator element in the OTC promoter remains unclear since the strictly unidirectional Inr is located in the tem-plate strand in the antisense orientation.

Structural motifs recognized by two different tran-scriptionfactorswithhighmatrixsimilarityscorewerepredicted in silicowithintheprotectedregionD:GATA--bindingfactor1(GATA1),located155bpupstreamofTIS(MIscore0.962,Mscore0.970)andspecificitypro-tein 1 (Sp1) at position -169..-177 (M score 0.962).Whereas GATA1 is essential for erythroid and mega-karyocytedevelopment,itsexpressionpatternislimited

O. Lukšan et al.

Fig. 2. Transcriptionalactivityofthemutantpromotervariantsincombinationwiththeenhancer(leftpanel)andwithouttheenhancer(rightpanel).Thedataarebasedonatleastsixindependentexperiments.Theboxrepresentstherangebe-tweenthe25thand75thpercentiles,horizontallinewithintheboxshowsthemedian.Whiskersindicatetheminimalandmaximalvalues.Diamondsareusedformeans.MultiplecomparisonsofvarianceusingtheBonferronimethodarepre-sentedinthetablebelowtheplots.Statisticalsignificanceisrepresentedbyasterisks:(***)indicatesthedifferenceswithP≤0.005and(**)isusedforstatisticalsignificancewithP≤0.01.

Vol.60 139ExpressionofHumanOTCIsRegulatedbyHNF-4α

Fig. 3. Analysis of the DNA-protein interactions within the human OTC core promoter. A – DNase I footprinting. Peaks representthefragmentsobtainedafterDNaseIcleavage(inblue)andinternalstandard(inred).Regionsprotectedagainstthe endonuclease cleavage are marked by red horizontal bars. B–Electromobilityshiftassays(EMSA).StandardEMSA(upperleftpanel),competitionassayswithaspecificcompetitorforHNF-4α(upperrightpanel)andspecificconfirmationofHNF-4αbindingtothepredictedsitesusingcompetitorandsupershiftassays(lowerpanels).Detectionoftheanti-HNF-4αantibodywasperformedbyprobingthesamemembranewithasecondaryantibody(rightelectrophoreogramsinlowerpanels).NX–incubatedwithnuclearextract.C – Schematic summary of results. Horizontal bars represent pro-moterfragmentsusedinthedualluciferasereporterassay(darkblue),protectedregionsobservedwiththeDNaseIfoot-printing(red)andEMSAprobes(lightblue).Theupperpartofthefigurerepresentstranscriptionfactorsandtheirbindingsites(underlinedsequence)predictedwithhighscoreusingMatInspectorandMatchprograms.

140 Vol.60

to the haematopoietic cell lineage and its involvement in the regulation of OTC transcription is thus unlikely. Sp1 is rather a ubiquitous transcription factor capable of regulating basal expression in genes with TATA-lesspromoters(Blacketal.,2001).TheSp1recognitionsiteoverlapsthealternativeTSSlocated169bpupstreamofthe translation origin. Since no other general core pro-moter has been found in a standard position related to TSSexceptfortheTATA-likedomaindiscussedearlier,the contribution of Sp1 to the initiation of transcription fromthepositionc.-169seemslikely.

Analysis of the DNA-protein complexes by EMSAThe DNA-protein interactions observed by DNase I

footprinting were verified in a set of electromobilityshift assays (EMSA). Six double-stranded biotin-la-belled probes were designed to cover the regions con-taining cis-acting regulatory elements, DNase I-pro-tected regions or in silico predicted transcription factor-bindingsites (seeFig.3C). Interactionwith theHepG2 nuclear extract led to an obvious shift: threeshifted bands were observed with the probes covering theregionslocatedatpositions-46..-126and-98..-178,twobandswithprobesspanningtheregions-168..-218and-204..-268andonezonewasdetectedusingaprobecorresponding to the sequence localized -98..-126 up-stream of the TIS. Theprobe[-98..-126]spanning29bpwasdesignedto

cover theproximalHNF-4αbindingsite.Themotif isconserved in human and rodents and overlaps the region B protected against the DNase I cleavage. The site is also contained in the [-46..-126] probe carrying theCis1+regulatoryelementandinthe[-98..-178]probe.Bothprobesexhibitedadifferentbandshiftpatternthanthe[-98..-126]probe,indicatingformationofadditionalDNA-proteincomplexes(Fig.3B).Oneof thesecom-plexes formedwith theprobe [-46..-126]may involvethe intermediates formed by the factors binding to the Inrelement.Thepatternobtainedwiththeprobe[-98..-178]likelyinvolvescomplexeswithotherHNF-4αandSp1 predicted to bind the sequences of protected regions C and D.Theprobe [-168..-218]wasdesigned tooverlap the

conservedHNF-4α binding site, found also in the rataround theposition -187bpupstreamofTIS, and theDNase I-protected region E. The probe [-204..-268]covered the distal part of the core promoter region, whereasequencemotifrecognizedbyC/EBPβwaspre-dicted.Specific interaction of HNF-4α with probes [-46..-

126]and [-98..-178]wasexamined incompetitionas-saysusinganon-labelledprobe [-98..-126].The reac-tion mixture containing both biotin-labelled andnon-labelledprobe[-98..-126]wasusedasacontrolin-dicating the specific manner of the competition. Re-pression of nuclear factor binding was observed with all three probes. A gradual decrease of two differentially shiftedbandswasobservedwith theprobe[-98..-178]carryingtwomotifsrecognizedbyHNF-4α(Fig.3B).

ConfirmationofspecificinteractionbetweenHNF-4αandthepredictedbindingsiteslocated105bpand136bpupstreamofTISwasachievedusingprobes [-98..-126]and[-132..-156](Fig.3C).Additionofaspecificcompetitor (a non-labelled probe corresponding to the consensus sequence of HNF-4α recognition motif) in200-foldexcessledtoasignificantdecreaseintheshift-ed band intensity. Supershifted bands of the DNA-pro-tein-antibodycomplexeswereobservedafterincubationwith the anti-HNF-4α antibody. The presence of theanti-HNF-4α antibody in these complexes was con-firmedusingHRP-conjugatedsecondaryantibody: theimmunoreactive zones co-localized with the supershift-edbands(Fig.3B).

Mutation in Cis1+ element abolished HNF-4α binding

ToconfirmtheroleoftheCis1+ element as a HNF-4αbindingsitewegeneratedamutant793(HNF4mut)inwhichthematrixsequenceCAAAGGwasreplacedby CAccaG and we then compared the transcriptional activity of both variants combined with the enhancer in a set of reporter assays (Fig. 4A).The activity of thepGL3_793(HNF4mut)/Enh construct was more thantwo times lower than in the case of normal promoter and thedecreasewashighly significant inStudent’s t-test. When compared with the reporter gene activity obtained withothervariantconstructs,thepGL3_793(HNF4mut)/Enh reached a similar level as the core promoter vari-ants without enhancer and it was about two times lower than pGL3_81/Enh (mean and median reached 41 %and34%ofpGL3_793/Enh,respectively).Interestingly,inspiteofthedisruptionoftheHNF-4αbindingsite,theGL3_793(HNF4mut)/Enh activity was partially pre-served,indicatingtheabilityoftheupstreamelement(s)to compensate for the function of the disrupted Cis1+.

DiscussionWe have shown that the minimal promoter responsi-

ble for both the promoter activity and activation of the enhancer is locatedwithin theregion-46..-126bpup-stream of TIS. As the shortest reported transcription variantstartsatthepositionc.-95,theminimalpromotercontains one strong positive cis-acting regulatory ele-ment Cis1+locatedwithin30bpupstreamoftheTIS.The motif is highly conserved in vertebrates and the af-finityofHNF-4tothisregionhasbeenclearlydemon-stratedintherat(Kimuraetal.,1993)andmouse(Inoueet al., 2002). Accordingly, we found that the Cis1+ spanning region is protected against DNase I cleavage, and experimentally confirmed binding of HNF-4α atthissite(Fig.3B).

The Cis2+elementwasidentifiedwithintheregionbetween-127and-178bpupstreamoftheTISoverlap-ping with the protected regions C and D. Cis2+ is in-volved in the interaction with the distal enhancer, but it doesnotdirectly influencethepromoteractivityalone

O. Lukšan et al.

Vol.60 141ExpressionofHumanOTCIsRegulatedbyHNF-4α

(Fig.2).SequencemotifsrecognizedbynuclearfactorsHNF-4αandSp1werepredictedin silico in this region (Fig. 3C). Binding of HNF-4α was confirmed in theband shift competition assay where two bands obtained withtheprobe[-98..-178]andonebandobtainedwiththe probe [-132..-156] decreased proportionally to theincreasedamountofspecificcompetitor(Fig.3B).Thepresence of two differentially shifted bands with the probe[-98..-178]canbeexplainedbyformationoftwodifferent complexes, probe-HNF4 and HNF4-probe--HNF4,astheprobecarriestwoHNF-4αbindingsitesandtheratioofbothcomplexesremainedconstantevenwhenrepressingtheirformationbyaspecificcompeti-tor.AdirectproofofHNF-4αbindingat this sitewasobtained by detection of supershifted complex DNA--protein-antibody(Fig.3B).BindingofSp1couldalsoexplaintranscriptionfrom

the alternative OTCTSS at the position c.-169, but itcannotexplaintheroleofCis2+ in the interaction with the enhancer. Unlike the two HNF-4α binding sites discussed

above,themotiflocated187bpupstreamofthetransla-tionorigindidn’texhibit significant influenceon tran-scriptional activation of OTC (Fig.2).Nevertheless,thematrixsequenceishighlyconservedanditdiffersfromthe rodent sequence only in one nucleotide located out-sideof thematrixcore.Theregionpartiallycorrelateswith theprotected regionE and interactionofHNF-4withthatsitehasbeenconfirmedinrodents(Kimuraetal.,1993;Inoueetal.,2002).Themotifoverlapswiththe probe [-168..-218], which competed for HNF-4αwiththeHNF-4α-specificcompetitorinEMSA.There-fore,itisobviousthattheHNF-4αbindingsiteisactive,despite that its contribution to the overall promoter

function is negligible when compared with the Cis1+ and Cis2+ elements. The low contribution is further supported by observations in the mouse (Inoue et al., 2002).Fangetal. (2012) recently reportedagenome-wide

studyevaluatingmatrixsequencemotifsinalargegroupof HNF4-specific genes using high-throughput ChIP-seq.Motifsrecognizedby1371specificsequenceswerestatistically evaluated based on their binding affinity.Matricespreferredby top10% (strongest), top33%(strong)andmiddle33%(middle)bindersoriginatingfrom the Fang’s study were compared with all four in silicopredictedHNF-4αbindingsitesandtheirratandmousehomologues (Fig. 4B).Thematrix sequence ishighly conserved and practically all compared regions couldbeconsideredasatleastmiddle-affinitybinderstoHNF-4α.

Activation of the enhancer in the rat is associated withtheinteractionofpromoter-boundHNF-4withtheenhancer-boundC/EBPβ(Nishiyorietal.,1994).Asim-ilar mechanism involving synergistic co-operation of HNF-4andC/EBPα,bothbound to thepromoter, hasbeen suggested for the human apolipoprotein B gene (Metzgeretal.,1993).Indeed,wepredictedabindingsiteforC/EBPβintheOTC core promoter between po-sitionsc.-221andc.-234,andtheresultsofEMSAre-vealed two shifted bands indicating formation of DNA--protein complexes within this region. In contrast,cutting off the region located -219 bp to -268 bp up-stream of the TIS did not change the level of the re-porter activi ty in luciferase assays.

In silicoanalysisoftheproximal200bppartoftheenhancerregionrevealedasetofsixrepeatedsequen-cemotifs corresponding to the C/EBPβ binding site.

Fig. 4.A–Functionalanalysisofthewild-type(793)andmutated(793(HNF4mut))OTC promoter, carrying a mutation intheproximalbindingsiteforHNF-4α,bothsubclonedtopGL3vectorcontainingtheenhancer.Thedataarepresentedasmean+SD.B–Sequencecomparisonofthein silicopredictedhumanHNF-4αbindingsitesandtheirratandmousehomologueswiththematrixsequencemotifspublishedbyFangetal.(2012).Greyboxesshowthecoresequencemotif.“n”indicatesequalincidenceofallfournucleotides.Lettersbelowthesequencesinthe“Matrix”sectionrepresentcom-mon(uppercase)orprobablebutnotcommon(lowercase)alternatives.

142 Vol.60O. Lukšan et al.

Accordingly,theDNaseIcleavageoftheproximalpartof the enhancer was decreased and a shifted band was observed in the gel retardation assay after incubation of the proximal enhancer fragment (data not shown).InteractionoftheliverenrichedHNF-4factorsboundtothe OTC core promoter with the set of periodically oc-curringliver-specificC/EBP-likeproteinsthusmaycon-tributetothetissuespecificityofOTCexpressionintheliver, as was seen in the rat model by Nishiyori et al. (1994).ThedistributionofHNF-4αrecognitionsiteswithin

the OTCcorepromoterregionexhibitsanimportantfea-ture:allthreeHNF-4αbindingmotifsarelocatedwithin35basesupstreamoftheTSSs(Fig.1).Moreover,thesequenceof223bpproximal promoter lacks any gen-eralcorepromoterelementssuchasTATA-box,Inr,thedownstreampromoterelement(DPE),theTFIIBrecog-nition element (BRE), the motif ten element (MTE),downstreamcoreelement (DCE),and theXcorepro-moterelement(XCPE)locatedatstandardpositionsup-stream of the TSSs, thereby suggesting the importance of HNF-4α in the initiation of OTC transcription. Initiation of transcription mediated by DNA-dependent RNA-poly merase II is dependent on the assembly of functional pre-initiation complex (PIC) consisting ofgeneral trans cription initiation factors TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH (Myer and Young, 1998;Hahn,2004).

The transcriptional machinery and the PIC formation has been well described in promoters equipped with general core promoter elements, but it has been poorly studied in promoters lacking these elements. However, fewindicessuggestingtheroleofHNF-4αinthetran-scription from TATA-less and Inr-less promoters can be foundintheliterature.HNF-4αwasshowntoactivatetranscription in vitro by facilitating assembly of the pre-initiation complex through direct physical interactionwith TFIIB. The interaction of TFIIB with DNA-bound HNF-4α led to TFIIB-mediated recruitment of TBP.Furthermore,HNF-4α participated in the assembly ofdownstream basal factors into the PIC (Malik and Karathanasis,1996).A similarmechanismcomprisingboththestablecomplexactivator-TFIIB-TBPassemblyand involvement in the association of downstream basal transcription factors TFIIE and TFIIF was observed lat-er in oestrogen receptor as another member of the nu-clear hormone receptor superfamily (Sabbah et al., 1998).Inarecentstudy,systematiccharacterizationoftheHNF4α-TFIIDlinkrevealedthattheHNF4αDNA-binding domain binds directly to the TBP and, through this interaction, can target TBP or TFIID to promoters containing HNF4α-binding sites (Takahashi et al.,2009).

In conclusion, our data strongly suggest the essential roleofHNF-4αintheregulationofOTC transcription in human. The regulatory mechanism may involve physi-cal interaction with TFIIB, recruitment of TBP and sta-bilization of TFIID to a particular site in the promoter followedbytheassemblyofthepre-initiationcomplex.

ReferencesBlack,A.R.,Black,J.D.,Azizkhan-Clifford,J. (2001)Sp1

and kruppel-like factor family of transcription factors in cell growth regulation and cancer. J. Cell. Physiol. 188, 143-160.

Brusilow,S.W.,Horwich,A.L.(1995)Ureacycleenzymes.In The Molecular and Metabolic Bases of Inherited Disease, eds. R. Scriver, A. L. Beaudet, W. S. Sly, D. Valle, pp.1187-1232,McGraw-Hill,NewYork.

Drewes, T., Senkel, S., Holewa, B., Ryffel, G. U. (1996)Humanhepatocytenuclerfactor4isoformsareencodedbydistinctanddifferentiallyexpressedgenes.Mol. Cell. Biol. 16,925-31.

Fang, B., Mane-Padros, D., Bolotin, E., Jiang, T., Sladek, F. M. (2012) Identification of a binding motif specific toHNF4bycomparativeanalysisofmultiplenuclearrecep-tors. Nucleic Acids Res. 40,5343-5356.

Hahn,S.(2004)StructureandmechanismoftheRNApoly-merase II transcription machinery. Nat. Struct. Mol. Biol. 11,394-403.

Hamano, Y., Kodama, H., Yanagisawa, M., Haraguchi, Y., Mori,M.,Yokota,S.(1988)Immunocytochemicallocali-zation of ornithine transcarbamylase in rat intestinal mu-cosa. Light and electron microscopic study. J. Histochem. Cytochem. 36,29-35.

Hata, A., Tsuzuki, T., Shimada, K., Takiguchi, M., Mori, M., Matsuda,I.(1988)Structureofthehumanornithinetran-scarbamylase gene. J. Biochem. 103,302-308.

Horwich, A. L., Fenton, W. A., Williams, K. R., Kalousek, F., Kraus, J. P., Doolittle, R. F., Konigsberg, W., Rosenberg, L. E. (1984) Structure and expression of a complementaryDNA for the nuclear coded precursor of human mitochon-drial ornithine transcarbamylase. Science 224,1068-1074.

Inoue, Y., Hayhurst, G. P., Inoue, J., Mori, M., Gonzalez, F. J. (2002)Defectiveureagenesisinmicecarryingaliver-spe-cificdisruptionofhepatocytenuclearfactor4α(HNF4α).J. Biol. Chem. 277,25257-25265.

Jones, S. N., Grompe, M., Munir, M. I., Veres, G., Craigen, W. J., Caskey, C. T. (1990) Ectopic correction of ornithinetranscarbamylase deficiency in sparse fur mice. J. Biol. Chem. 265,14684-14690.

Kimura, A., Nishiyori, A., Murakami, T., Tsukamoto, T., Hata, S.,Osumi,T.,Okamura,R.,Mori,M.,Takiguchi,M.(1993)Chicken ovalbumin upstream promoter-transcription fac-tor(Coup-Tf)repressestranscriptionfromthepromoterofthe gene for ornithine transcarbamylase in a manner an-tagonistictohepatocytenuclearfactor-IV(Hnf-4).J. Biol. Chem. 268,11125-11133.

Lindgren, V., de Martinville, B., Horwich, A. L., Rosenberg, L.E.,Francke,U.(1984)Humanornithinetranscarbamyl-aselocusmappedtobandXp21.1neartheDuchennemus-cular dystrophy locus. Science 226,698-700.

Luksan, O., Jirsa, M., Eberova, J., Minks, J., Treslova, H., Bouckova, M., Storkanova, G., Vlaskova, H., Hrebicek, M., Dvorakova,L.(2010)DisruptionofOTCpromoter-enhanc-er interaction in a patient with symptoms of ornithine carba-moyltransferasedeficiency.Hum. Mutat. 31,E1294-1303.

Malik, S., Karathanasis, S. K. (1996) TFIIB-directed tran-scriptional activation by the orphan nuclear receptor hepat-ocytenuclearfactor4.Mol. Cell. Biol. 16,1824-1831.

Vol.60 143ExpressionofHumanOTCIsRegulatedbyHNF-4α

Metzger,S.,Halaas,J.L.,Breslow,J.L.,Sladek,F.M.(1993)OrphanreceptorHnf-4andBzipproteinC/Ebp-αbindtooverlapping regions of the apolipoprotein-B gene promoter and synergistically activate transcription. J. Biol. Chem. 268,16831-16838.

Miyajima, N., Kadowaki, Y., Fukushige, S., Shimizu, S., Semba, K., Yamanashi, Y., Matsubara, K., Toyoshima, K., Yamamoto,T.(1988)IdentificationoftwonovelmembersoferbAsuperfamilybymolecularcloning:thegeneprod-ucts of the two are highly related to each other. Nucleic Acids Res. 16,11057-11074.

Murakami,T.,Nishiyori,A.,Takiguchi,M.,Mori,M.(1990)Promoter and 11-kilobase upstream enhancer elements re-sponsibleforhepatomacell-specificexpressionof theratornithine transcarbamylase gene. Mol. Cell. Biol. 10,1180-1191.

Myer,V.E.,Young,R.A.(1998)RNApolymeraseIIholoen-zymes and subcomplexes. J. Biol. Chem. 273, 27757-27760.

Nakshatri,H.,Bhat-Nakshatri,P.(1998)Multipleparametersdeterminethespecificityoftranscriptionalresponsebynu-clear receptors HNF-4, ARP-1, PPAR, RAR and RXRthrough common response elements. Nucleic Acids Res. 26,2491-2499.

Nishiyori, A., Tashiro, H., Kimura, A., Akagi, K., Yamamura, K.,Mori,M.,Takiguchi,M.(1994)Determinationoftissuespecificityof theenhancerbycombinatorialoperationoftissue-enriched transcription factors.BothHNF-4andC/EBPßarerequiredforliver-specificactivityoftheornith-ine transcarbamylase enhancer. J. Biol. Chem. 269, 1323-1331

Ryall, J., Nguyen,M., Bendayan,M., Shore, G. C. (1985)Expressionofnucleargenesencoding theureacycle en-

zymes, carbamoyl-phosphate synthetase I and ornithine carbamoyl transferase, in rat liver and intestinal mucosa. Eur. J. Biochem. 152, 287-292

Sabbah, M., Kang, K. I., Tora, L., Redeuilh, G. (1998)Oestrogen receptor facilitates the formation of preinitiation complexassembly: involvementof thegeneral transcrip-tion factor TFIIB. Biochem. J. 336,639-646.

Storkanova, G., Vlaskova, H., Chuzhanova, N., Zeman, J., Stranecky, V., Majer, F., Peskova, K., Luksan, O., Jirsa, M., Hrebicek,M.,Dvorakova,L.(2013)Ornithinecarbamoyl-transferase deficiency: molecular characterization of 29families. Clin. Genet. 84,552-559.

Takahashi, H., Martin-Brown, S., Washburn, M. P., Florens, L.,Conaway,J.W.,Conaway,R.C.(2009)Proteomicsre-veals a physical and functional link between hepatocyte nuclear factor 4α and transcription factor IID. J. Biol. Chem. 284,32405-32412.

Tuchman, M., Jaleel, N., Morizono, H., Sheehy, L., Lynch, M. G.(2002)Mutationsandpolymorphismsinthehumanor-nithine transcarbamylase gene. Hum. Mutat. 19,93-107.

Wilson,D.O.,Johnson,P.,McCord,B.R.(2001)Nonradio-chemical DNase I footprinting by capillary electrophore-sis. Electrophoresis 22,1979-1986.

Yamaguchi, S., Brailey, L. L., Morizono, H., Bale, A. E., Tuchman,M.(2006)Mutationsandpolymorphismsinthehuman ornithine transcarbamylase (OTC) gene. Hum. Mutat. 27,626-632.

Zianni, M., Tessanne, K., Merighi, M., Laguna, R., Tabita, F. R. (2006) Identification of theDNAbases of aDNase Ifootprint by the use of dye primer sequencing on an auto-mated capillary DNA analysis instrument. J. Biomol. Tech. 17,103-113.