Proc. Nail. Acad. Sci. USAVol. 88, pp. 10642-10646, December 1991Developmental Biology

Expression of GATA-binding proteins during embryonicdevelopment in Xenopus laevis

(hematopoiesis/mesoderm/transcription factor/erythrocytes/zinc-finger domains)

LEONARD I. ZON*, CHERI MATHER*, SHAWN BURGESS*, MARGARET E. BOLCEt, RICHARD M. HARLANDt,AND STUART H. ORKIN*t*Division of Hematology-Oncology, The Children's Hospital, Dana-Farber Cancer Institute, Department of Pediatrics, Harvard Medical School, and tHowardHughes Medical Institute, Boston, MA 02115; and tDepartment of Molecular and Cell Biology, Division of Biochemistry and Molecular Biology, Berkeley,CA 94720

Contributed by Stuart H. Orkin, August 19, 1991

ABSTRACT Proteins that recognize the core sequenceGATA are important regulators of hematopoietic-specific genetranscription. We have characterized cDNAs encoding theXenopus laevis homologues of three related transcription fac-tors, designated GATA-1, -2, and -3. Comparative sequenceanalysis reveals strong conservation of the zinc-finger DNA-binding domain among all vertebrate GATA-binding proteins.GATA-2 and GATA-3 polypeptides are homologous through-out their entire sequences, whereas GATA-1 sequence is con-served only in the region responsible for DNA binding. InXenopus, RNAs encoding GATA-binding proteins are ex-pressed in both larval and adult erythroid cells. GATA-1, -2,and -3 RNAs are first detectable in early gastrula (Nieuwkoopdevelopmental stage 11). This is earlier than the appearance ofthe early larval aTl globin RNA (stage 15), fiT1 globin RNA(stage 26), or blood island formation (stage 30). The expressionofGATA-1, -2, and -3 in early development may signal an earlycommitment of mesoderm to form hematopoietic tissue.

The developmental program of hematopoiesis is conservedthroughout vertebrate evolution (1). The site of blood for-mation changes sequentially during development from yolksac in the early embryo, to the liver in the fetus, and finallyto bone marrow in the adult. During embryonic and adulthematopoiesis, stem cells proliferate and differentiate intomultipotential progenitors, which subsequently mature alongindividual lineages, producing erythroid, myeloid, and lym-phoid cells (2). Pluripotent stem cells derived from eitherembryonic (primitive) or adult (definitive) hematopoieticregions can reconstitute hematopoiesis in a lethally irradiatedanimal (3).The earliest events determining the hematopoietic program

in vertebrates have yet to be defined. Groudine and Wein-traub (4) studied the presumptive hemoglobin-forming area of20- to 23-hr chicken embryos plated in fibrin clot assays andestimated that hematopoietic commitment begins duringearly gastrulation, long before blood formation is histologi-cally evident. Hence, early events accompanying the differ-entiation ofmesoderm lead to the formation of hematopoietictissue.The inductive process of dorsal and ventral mesoderm

formation has been intensively examined in Xenopus laevis(5). Although considerable information exists regarding em-bryonic muscle commitment and differentiation (6), the in-duction of ventral mesoderm to form blood is less wellcharacterized. X. laevis is an excellent species in which tostudy the developmental expression of the hematopoieticprogram. Developmental stages have been well defined, andembryos can be analyzed prior to the appearance of recog-

nizable hematopoietic cells (7). Embryonic ventral bloodislands form about 36 hr after fertilization (stage 30, Nieuw-koop). At metamorphosis, the liver becomes the predominantsite of hematopoiesis. Amphibians are the first organism inevolution to use bone marrow as a site for adult hematopoie-sis, although the spleen and liver retain some hematopoieticfunctions. As a result of a genomic duplication event, whichoccurred about 80 million years ago, X. laevis has two globingene clusters on distinct chromosomes (8, 9). Each individualcluster contains linked a and ,B globin loci organized in anaL-aLbA-SABLa arrangement (L for larval and A foradult). Hemoglobin switching occurs from early larval to latelarval globins at hatching and from larval to adult globins atmetamorphosis (10).Formation of yolk sac blood elements from pluripotent

cells of the embryo, the earliest stages of embryonic he-matopoiesis, must be regulated, at least in part, by cell-specific transcription factors. In vertebrates, DNA-bindingproteins that recognize consensus motifs containing the coresequence GATA appear to be critical for normal erythroiddevelopment (11-13). These transcription factors form afamily of proteins related by virtue of their highly conservedzinc-finger DNA-binding domains. Three members (desig-nated GATA-1, -2, and -3) have been characterized frommammalian and avian species (13). GATA-1, the foundingmember of the family, is expressed only in hematopoieticcells and is further restricted to erythroid, megakaryocytic,and mast cell lineages (14-19). GATA-2 is expressed in theselineages and in several nonhematopoietic cell types (20, 46).GATA-3 is highly expressed in T-lymphoid cells but is alsopresent in definitive erythroid cells of the chicken and in fetalbrain (20-23). Functional GATA-binding motifs are presentin the promoters or enhancers of the vast majority of eryth-roid-expressed genes (15) and in the locus control regions ofthe a- and ,3-globin gene complexes in humans (13, 24). Genetargeting in murine embryo-derived stem cells has shown thatGATA-1 is absolutely required for erythroid development(25). In the developing mouse embryo, GATA-1 expressionprecedes and parallels the increased expression of globinmRNA (26). GATA-1 mRNA is evident during yolk sachematopoiesis; however, no information is available regard-ing the onset of expression of this transcription factor sinceearlier tissues have not been examined.To examine the role of GATA-binding proteins in embry-

onic hematopoiesis, we have initiated studies in X. laevis.Here we describe the cloning of Xenopus homologues ofGATA-1, -2, and -3§ and their RNA expression duringembryonic development.

Abbreviation: RT, reverse transcription.§The sequences reported in this paper have been deposited in theGenBank data base (accession nos. M76563, M76564, M76565, andM76566).

10642

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 10643

MATERIALS AND METHODSCells. Xenopus eggs and staged embryos were obtained.

Xenopus erythroid cell sources included: reticulocytes fromnormal adult animals, peripheral blood cells from adultanimals made anemic with phenylhydrazine, and peripheralblood cells from phenylhydrazine-treated tadpoles arrestedbefore metamorphosis with thiourea (27). The cell lines XTCand XL were the generous gift of Douglas Melton (HarvardUniversity).cDNA Cloning. A 468-base-pair (bp) probe encompassing

the zinc-finger region of murine GATA-1 cDNA was gener-ated by the polymerase chain reaction (PCR) using primersspanning nucleotides 483-510 and 933-950 (15). Filters bear-ing S x 105 plaques from an anemic adult Xenopus erythroidcell Agtl0 cDNA library were screened at reduced stringency(final wash with 0.15 M NaCl/15 mM sodium citrate at 420C).After plaque purification, cDNA inserts for XenopusGATA-1A and GATA-1B were subcloned into pUC-13 forsequencing. Of 10 GATA-1 cDNAs isolated, 9 cDNAs cor-responded to xGATA-1A and 1 cDNA corresponded to apartial clone of xGATA-1B. Rescreening of the libraryyielded 15 full-length xGATA-1B cDNAs. Xenopus GATA-2and GATA-3 cDNA clones were obtained by screening 5 X105 plaques from a Xenopus stage 17 embryo Agtl0 cDNAlibrary (kindly provided by Douglas Melton) with a 627-bpprobe including the zinc-finger region of xGATA-1A. Thisprobe was generated by PCR using primers spanning nucle-otides 597-617 and 1206-1224 of the xGATA-1A cDNA. Thefilters were subjected to a final wash at reduced stringency(0.15 M NaCl/15 mM sodium citrate at 55°C). cDNA insertswere subcloned into pUC-13 for sequencing. Of 5 x 105plaques originally screened, 180 individual clones hybridizedwith the zinc-finger region probe. Examination of DNAsequences of 12 clones revealed 9 clones corresponding toxGATA-2 and -3 clones to xGATA-3. Northern blot analysiswas performed by standard methods (28).

Developmental Expression Using Reverse Transcription(RT)-PCR Analysis. RNA was prepared from staged embryosas described (29). Total ovary was used as a source of oocyteRNA. PCR primers were designed to encompass an exonboundary based on the known structure of the murine andhuman GATA-1 genes (30). The following primers of theXenopus cDNA clones were utilized: GATA-1A (nucleotides597-617 and 1206-1224 for product length of 627 bp),GATA-2 (nucleotides 1387-1412 and 1632-1655 for a productlength of 268 bp), GATA-3 (nucleotides 1186-1205 and 1671-1693 for a product length of 507 bp), ,f1Tl globin (nucleotides81-101 and 408-429 for a product length of 348 bp) (31), andMyoD (nucleotides 164-186 and 484-507 for a product lengthof 343 bp) (32). RT-PCR was performed as described (33)using 1 ,g of total cellular RNA. The PCR program used ina Perkin-Elmer Thermocycler consisted of 1 min at 94°C, 1min at 55°C, and 3 min at 72°C for 30 cycles, and a final 7-minextension. One-half of the product was visualized by ethid-ium bromide staining after electrophoresis through a 5%polyacrylamide gel. Controls included a RT reaction withoutRNA and a reaction with HeLa cell total RNA.

RESULTS AND DISCUSSIONCloning of cDNAs Encoding Xenopus GATA-Binding Pro-

teins. By crosshybridization to a murine GATA-1 zinc-fingercDNA probe, we isolated clones for the Xenopus homologueofGATA-1 from an erythroid cell cDNA library. Two distinctcDNAs encoding the xGATA-1 (termed A and B) wereidentified (Fig. 1). Xenopus homologues of GATA-2 andGATA-3 were isolated from a stage 17 embryo cDNA library(Fig. 1). In one xGATA-2 clone, a 33-bp insertion, corre-sponding to the peptide sequence RNSVSSSFHLE, was

present between amino acid residues 314 and 315 of thealignment in Fig. 1, presumably the result of alternative RNAprocessing. This corresponds to an exon boundary in themurine GATA-1 gene. Similar alternative processing ofcGATA-2 has been observed (J. D. Engel, personal commu-nication). All clones could be aligned with the known verte-brate GATA-binding proteins. The deduced sizes ofGATA-1, -2, and -3 are 39 kDa, 49 kDa, and 47 kDa,respectively. Xenopus and chicken GATA-1 proteins aresmaller than GATA-2 and -3, whereas all other mammalianGATA-binding proteins are approximately the same size.

Comparative Sequences of GATA-Binding Proteins. Char-acterization of the Xenopus GATA-binding proteins nowpermits comparison ofdeduced amino acid sequences amongall vertebrate species in which hematopoietic developmenthas been studied at the molecular level (Fig. 1). Severalfeatures are notable and merit comment. (i) The zinc-fingerdomains, which are necessary for sequence-specific DNAbinding (34), reveal extraordinary conservation among allspecies (Fig. 1). (it) GATA-2 and GATA-3 are highly con-served throughout their entire polypeptide structure and areboth related more closely to each other than to GATA-1. (iii)Outside the finger domains, mammalian, avian, and amphib-ian GATA-1 protein sequences are strikingly divergent, par-ticularly in contrast to the strong maintenance of primarysequence of the other GATA proteins through evolution.Although the implications are unknown, such divergence isunexpected given the common features of erythroid generegulation noted in all vertebrates. The presence of twodistinct xGATA-1 sequences presumably reflects genomicduplication in Xenopus (8). Distinct forms of GATA-2 havealso been isolated (unpublished data). (iv) Despite extensiveconservation of the zinc-finger domains of all species, align-ment of the numerous members now highlights amino aciddifferences that distinguish GATA-1, -2, and -3 (Fig. 2). Themajor differences are situated near exon boundaries. This isparticularly evident in the region corresponding to the N-ter-minal segment of exon 5 (aligned to mGATA-1). Previousdata indicate that the N-terminal finger domain cooperateswith the C-terminal finger to maximize highly stable se-quence-specific DNA binding (34). Although all members ofthe GATA-binding family recognize highly similar consensustarget sites of the general form [(A/T)GATA(A/G)], it ispresently unknown whether they discriminate among indi-vidual sites. If the binding specificities of the various GATA-binding proteins are indeed subtly different, the amino aciddifferences shown here may provide a structural basis for finerecognition.RNA Expression of GATA-Binding Proteins. Consistent

with a presumptive role of GATA-1 in erythroid gene expres-sion, xGATA-1 mRNAs are abundant in both tadpole andadult erythroid cells and not detectable in oocytes, as shownby Northern blot analysis (Fig. 3A). Adult reticulocytes lackGATA-2 RNA (data not shown) but do contain GATA-3RNA (Fig. 3B). Although the egg and a fibroblast line (XL)lack xGATA-3 RNA, the embryonic cell line XTC, which isknown to produce mesodermal-inducing activity (activin-A)(35), expresses xGATA-3 at a high level (Fig. 3B). Whereashematopoietic cells in the bone marrow or spleen have beenshown to express activin (36), there is no evidence that theXTC cell line has any hematopoietic properties.A more detailed analysis of expression of xGATA-1, -2,

and -3 RNAs throughout Xenopus development suggestspotential roles for these transcription factors during embryo-genesis. As shown in Fig. 4A, xGATA-1, -2, and -3 areexpressed by Nieuwkoop stage 11 of development. The verylow level of xGATA-1 detected in ovary RNA is thought toresult from contaminating adult erythroid cells. xGATA-2 ispresent at a low level as a maternal transcript and abundantlyexpressed at stage 11 and thereafter. The RNA expression of

Developmental Biology: Zon et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

10644 Developmental Biology: Zon et al. Proc. Nati. Acad. Sci. USA 88 (1991)i ~~~~~~~~~~~~50100

xGATA-LA 1DY.........ITL TTIDPDP.....NYTESG LASTSEDSQ. FL..........Y GL(=ESP0. HY.CIOAVSSR AG.GFR. .. .HSPVJFQTF 1U-KPETSAGI P.....SN LTAYGRST..xGATA-1B liDY.........ITL TIOSP.....LONYTESA LASTSEDSE. FL..........Y GLOO=SPG. HY.L1OIVSSR AVG.G(TR... . HSPVLQTFP LHWP'ETSAGI P.....OS LTTYGRSP. .c-GATA-1 DIEF.........VAL G. .G.PDAGS .PTPFPD ...EPA FL..........GLOGnE.N........... .......... .. T...rnOATA-1 tDF.........PGL GAIDI7SEPLP .....FVDSA LVSSPSDSTG FFSSGPEGID AASSSTSPNA ATAAASALA. YY.....R ,A.AYR... .HSPVFVYTP LLNSME..GI P(XSPYASW. AYGK(TAL.hGATA-1 KfF.........GL GSLGISEP1P .....1FVPA LVSSTPESGV FFPSGPEGID AAASSTAPST ATAAAAAIA. YY.....R DAE.AL. . . HSPVFQVYP LIAI.IE. GI PGGSPYAGW. AYGKT`GL.xGATA-2 1.EVATDEPFN M. . .HHA\TL NGSQHPDSHHP GLAINYMEPT QLIPPDEVDV F.....NHLD S.G.. .PYY ANSAHARARV SYSOHARLT GCSa41PPHL .HPG....2 WLESGKTAIS A.AHHHPWI VSPF(aPLHcCATA-2 1MEVATDQPFU r....HHAVL NG2QHPESHHP GLAHNYI.EPA QLLPPDEVDV IFF.....NHID S.G. ..PYY ANSAHARARV SYSONHARLT .GSa4IRPHL IHPG.... IP WIDSSKAALS . .AHHHNPWr VNPPIKTPLHhGATA-2 WWEVAPE1 MA....HPAVL NAO-IPDSHHP GLAHNYMEPA HVIPPDEVDV IF... .NHID .G.. .PYY ANPAQ.RG.V SYSPAHARIT .Q242PHL .HPG....12 WLfXXKAAIS A.A.HHKWr VSPFSKTPLHxGATA-3 NEVSAEQPIM VSHPHHPALL NG2Q[MISHH9Q .VHNM4PS QYPPPED!4DV IF.....N. ID PQ24HVPSYY SNSVRTVP.. RYPPPH....H .GSLVCHIPSI LOS...... WMGKS1GG P. .HASAWi L.PPLYSIHcGATA-3 1.EVSTD.PR14 VS .HHHPAVL NK2QHPDSHHP TLG4TYI.IPT QYPLAEEVDV LI ....N. ID (2OGWVPPYY GNSVRATMD. RYTA....H .GSQVCRPPL LHGS ....12 WTIDGSK.ALS S.HHSASPWN LSPISKTSIHrmGATA-3 1.EVIADYQPFEt VS .HHHPAVL NGQHPDML1P GLGHSYMEA. QYPLTEEVDV LI ... N. ID QDSNHVPSYY GN4SVRATVQ. R.PTH....H .GSQVWPPPL LHGS ... .12 WLLXXI.AIG S.HHTASP1.N LSPFISKTSIHhGATA-3 MEIADQF8.J VS..HHHPAVL Nf3QIDI11HP GLSHSYMAA QYPLPEEVDV LI ....N. ID G2GNH\VPPYY GNSVRATM2. RYPTH....H .GSQVCIDPPL .HG. .. 12L WUXIKAIG S.HHTASPI4N LSPISKXTSIHCCNSENSUS t.EVA-DQ'FRW VS-H-.HPAVL Nf2aPDSHHP GLAHNYMEPA QY-PPEEVDV LF-SGPENLD -Q24N-VPPYY -NSVPAPA-V -YPPAHAR1k4 -GSQVU4P-L LHSPVLQ~-P WUDGSTAIS --HHASPMf LSPI-KTSLH

141 190 240xGATA-LA . I.....GLSFYPS. .AA SAIXIPIT. .S P.PLYSASS. FL.LGSAPPA ER~EGSP.. .............KIT.. .L.ET.LKTE RASPLTSDLL PE....PR SPSILQAIYI .G.CE...xGATA-1B.....GTL .PLYPS. .AA SALGLIT. .S P.PLYSATP. FL.LGSAPLA ERPCGP.. .............KF.. .L.ET.1KTE RTSPLTSDIL SL ....PR SPSLLQVGYL GXXX)215. .cGATA-1 ...EAGGL LAYP... PS GRV.......RVSLVP. WAUICLGTP .QWVPPAT..0.....4....5 PP.... HY.. .L.EL.LSQ'P PCOSPPHPSSG PLI......PISSP........rrGATA-1 ...... .P.....A SVPS.HE......D..... PSFWQA. LED..... J....C KSNN.T.F.. .L.DT.UKTE RLSP....DLL TLC.......TAIPASL PVIGSAYCLAhGATA-1 ........YP .... A SIVR.T.............REDSPPA. '/ED..... IX...LGKGST.S.F.. .L.ET.LKTE RLSP....DLL TLC.......PAIPSSL PVPNSAYCPxGATA-2 P.. .AARGGD. SLYPG. ¶TG SSAC~PSSS.....HSSP HL.FIPPIP PKDVSPDP.. .SP1ASPPSSS ....RIMEDSI0. .KY0Ci SISE.OIKNE GGSPIESSLA PtiG.TQCSTH HPIPTIYPSYV P .AAHDYSS.cGIATA-2 PSAALO.PGAI SVYPGSSTS STASVSS... LTP.ASHSGS HL.FISFPPTP PKEVSPEPNS TSAASPSSSA ....CAROSD KDSI. .KY2V SLSE.94KI4 SASPIESSLT SMGl.A<PSII HPIPTYPSYAV P.AAHDYSS.hGATA-2 PSAAIXI'GJI .SVYPGAGIS GXGSCSSVAS LTPTAAHSGS HL.FU`PPRH PKELSPCPST TGIAASPASSS AGGISSAR8ID KDGN. .KYOA SLTE.SMKM.E SGENPIRPGLA TIht.IQ'ATH HPIPTYPSYV PAAAHDYSS.xGATA-3 H....NS:GIL SVY....PA SSASLAIG.....HSSP HL.=PIFRP PKD0VSPEPSI STSGSTSSS.RH....Pi KE)IWSKYQV SIIIDT24KIE SLHP .RNSMS GIGG. .VSA~H HPITTYPDY.....YCA.cCATA-3 H....SSPGPL SVYP....PA SSSTISA .....HSSP HL.PIFPTP PKDVSPIP.. .SISTPGSTC .....STIRXE K8001. .KYQV' SLADT.MKLE .SSHSRSSMA SLOGATSSAH HPIM1PPYVT P... .EYSS.rrGATA-3 H....CSPCG~L SVYP. .. .PA SSSSLAA .....HSSP HL.PIFPTP PKDVSPCPSL STS9GSPGSA......RE KCLl. .KYO/ QLPDS.MKIE .TSHSI9ISMIf TLGGA.SSSAH HPIMFIPPYV P... .EYSS.hCATA-3 H.. .CSPGPL .SVYP. .. .PA 5555S50.....HASP HL.PIFPTP PKDVSPCPSL STPGSAGSA..... 1E KIIL. .KY(Y P1PDS.MKIE .SSHSRGISW~AIGGASS.ST14 HPITTYPPYV P.....EYSS.CCNSENSUS PSAAGSPG-L -SVYP--PA SSASPSS- -HP-YHSSP HL-=~PTIP PKDVSPCPSL ST-AS-CGSS- ----SAR--EE K1111--KY2' SL-T-MKIE SASPL2SSLL TLC-ASSS--H HPITO'PSYV PVAAHEYSS-

281 330 380xCATA-1A .......SF.........51.0SIEDRECVNC CATVI'PL R CMSGHYLCN AOGLYHK1q4Z ONRPLIPPKK PRLIVSKPACI QCSNCIHTSTr TLU4RFNASGD P\WNACGLYY KLHNvW~LT MKKEGIQTRNxGATA-1B ......HSLFQ..........STEDRE-CVNC GATVIPILWR RDLSGHYLCN ACGLYHKMNG ONRPLIRPKK RLIISKRAGT QISNC1HTSIT TLW4RFNAGLD P\C8UPLYY KGHNVNPPLT MKKE2GIQIRNcGATA-1 .................PPCF.ARECVNC GATATPILWR PZGTGHYJCN ACGLYHRLI9 0aPLIRPKK RLLVSKRACGT'\ESNC015ST TLW4RRSPMil P'.CNAOGLYY KLH18VNRPLT 14RKDGIQTRNrrGATA-1 IFPSPIFSPT CSPLSSAAYS SPKFHSIPL APCIARECVN9I GATATPIIR RERTGHYLCNI AIGLYHKMEW ONRPLIPPKK R AIVSKRAGTI QCTNCIQFF TLMF0NASG P\ZNWIYF KLHQVNRPLT MRKDGIQRNNhGATA-1 EOFSSTFFSPT GSPLNSAAYS SPKIR=ITL PPCEAIECVN9C GATATPIWA RLETGHYICN AOLYHKO4N ONRPLIPPIQ RLIVSKRAG CTN[CQOFI TL9EUNASGD PMOCG0~3LYY ELHQVNRPLT 14FKDCQTRNxGATA-2 ...CLI. .HP GSLGI.PASS FI¶PK0RSESR SCSEXRECVNC2 GATATPLWA RDGTIGHYLCN AEGLYHKMN11 O4RPLIKPKR RISAAPRAGT OZANCOISIT TLURFPNANGD P\CNACG00YY KDENVNRPLT M.KIEGIQTRNcGATA-2 ...S... HP GSII=PASS PIPKPRSKAR SCSEGEECAI GATAIPLWR RDGTSGNYLCN AOGLYHK149 ONEPLIKPKR RISAARRALII OIANCQFFF TLMiRUNANZ P\CAIWGOLYY KUWJNRPLT MKKEGIQIRNhGATA-2 ...G.. .HP GSIL0PASS PIPK0RSKTIR SCSEGFECVNC CATAIPkiRE RDXGTGHYLCN AOFYHKM4KG ONRPLIKPKR RLSAAPRRA OCANCIITF TLEWRFNAEC P\CWADILYY 1CAWNRPLT MKKEGIQIPNxGATA-3 .. .CLI. .C GSILIYISPTH ISSNARPKTR SSTEGRECVNC GATSTP~ku'R RDXIIOHYLQN ACX1YHK4f8 ONRPLIKPKR RISAARRAGI SCANCQFFF TLWPPNANCD P\C0NAEDYY KLHNINRPLT MKEGIQINNcGATA-3 .. .CLI..PP SSLLO&SPTG FGMKRPKAR SS¶1EGFECVNC GATSTPLA4R RDGIIONYLQN ACGLYHKM4Z ONRPLIKPKR RISAARRAGT SCANCQI'T TLWR~iNRI PVCNWACGLYY KTHNINPPLT MKKEGIQIRNnrGATA-3 .. .CLI..PP SSLILGGSPD` FGMKSPKAR SSTEaEICVNC GATSTPL84R RDGIGiYLCN AOGLYHKIM ONRPLIKPKR RISAAFRRAJ SCANCI2IIT TLWRFOWM1 P'/EN.gEDYY KLHNINPPLT WKKEGIQ¶IRNhGATA-3 .. .CLI..PP SSLLOGSPTG FGZKRPKAR SSTIEGFICVNC GATSTPIWR RDGTG(2YLCN ACILYHK.Rn ONRPLIKPKR RISAARRAC SCANCYFFI TLW4R~iANGD P\/ENAOLYY K1EINI8PLT WKKEllIQTRNCONSENSUS DISGLFFSPP GSLI0GSATS I-PK-RSKAR SSTIE2RCVNC GAJ¶ATPIW4 RDGTGIHYLrON ACGLLYHK~flG QNRPLTEKERLSAARRAIT -CANIDTIT TLWRRHANID EyZMNP XLZLKIHNMt~LTMKMZCIYIPN

421 470 520xGATA-1A RI(VSRSK<K K10......DNPI......PKAG VEEPSPYPFG P.LLIIAG24P PM.GHM. .IN PPHHILOSPR ISHSAPAVSY ROA.....AGIPxGATA-1B 1I(VSKSI(KQ10......ENPI..E ....PKT VEEPSPYQYG P.LLIHGQ.'IPTCM4.G..IS PPHHFLOSPR MSHSTPAVSY RYS.....ASGVTPPE1AcGATA-1 RW(SSKIKKR RPPG32PSA TAGOllA.M... XXI33P SENPPPPPPPA AAPPOSDALY .AIXIVWLSG .. .HF1PIG4 SGG1... F.1.1 (llRIDYIAP PGIP.....mG~ATA-1 RK(ASGKGKKK RGSNIAGAGA AEGPAGXIW VAGDSSSSGNC GEVTASGIAIG T. AGTA.HLY QGLa'VL.SG PVSHUIPIPG PLICSP¶0IS PTIGPAP¶ITSS TSVIAPLSS.hGATA-1 R%(SC~q2QK RGSSIGIIC AEGPAGGFW VAGSGSGNC GEVASGLTIL P.PGTA.HLY QGLCPVVLSG PVSHUMPFPG PLIC,%TOIGS P¶IGPI4PPTS TIWVAPLSS.xGATA-2 RII-ilNKSKKN K.....K GSFCFEEL......SRC MUMSS.PI'S A.AAIASHMA PM.GHLAPIS HSGHILQTPT PIHPSSSLISI GHPHHS.. SMVTA4I. .cGATA-2 RI1SNKSKKS 1.....K GSECFEEL......SEC E4jKSS.PFS A.AALASHIMA PM.GHIPPIS HSG-JIIPTPT PIHPSSSISI GHPHPS.. SMVrA4I...hGATA-2 RKI45NKSKKS K.....K GASED= .....SEC EMEEKSS .PFS A.AALAGHMA PM.GHIPPFS HSGHILPTP PIHPSSSISI GHPHPS.. SMVAtIi. .

xGATA-3 RKISSKSKKS K.....K EHDSLEDY.........PKSSSFS P.AAISRHMS SL.SHI.P15 HSSHML¶IIP PMHPPSSISF GPHHPS.. S1MVTAWE..cGATA-3 RKI.SSKSKKC K.....K VHDNIED3I.........PKSSSIN P.AALSRHMS SI.SHISPIS HSSHM4LTIP PMHPSSISI GPHHPS.. SMVrNI. .rrGATA-3 RKISESKSKKC K.....K VHDALEDI .........P1555S3 P.AALSRHMS SL.SHISPIS HSSHML¶II PMHPPSGLSI GPHHPS.. SM4VTAMG. .

hGATA-3 KINSISKSKC K.....K VHDSLED .........PKNSSIN P.AALSRHMS SL.SHISPIS HSSHML1TP PMHPPSSLSF GPHHPS.. SME/TAME..CONSENSUS PE 5SKSKK ---101-00 --OPPEDIN- -AG---KC MEEPSSSPI- P AALARI-E -L-GH4-SPIS HSSHML-TPT PIIHPSSSLSF GPHHPS-T-S T51MTrNnE--

FIG. 1. Comparison of the predicted amino acid sequences of vertebrate GATA-binding proteins. The zinc-finger DNA-binding domain isunderlined. The conserved amino acids outside the DNA-binding domain are marked (*). x, Xenopus; c, chicken; mn, mouse; h, human. (Forsequences, see refs. 14, 15, 18-23, and 46.)

the GATA-binding proteins temporally follows the expres- pressed differentially, restricted to particular cell types andsion of xMyoD, the muscle-restricted helix-loop-helix reg- developmental stages, and in some tissues, coexpressed withulator, which is initially detected at stage 9 (Fig. 4A). The other members. Two interesting conclusions emerge fromrelative levels of xGATA-2 and -3 mRNAs detected by this analysis. (i) The expression of these transcription factorsRT-PCR roughly approximate the ratio of cDNA clones after the midblastula transition suggests that they may serveobtained from the screening of the stage 17 cDNA library, important roles in early development, perhaps leading to aHence, as seen in birds and mammals (14, 15, 18-23, 46), determination of mesoderm to form blood. Alternatively,

members of the GATA-binding protein gene family are ex- GATA-binding proteins may have additional nonhematopoi-

314 409

IVS-3 IVS-4 IVS-5

E IXE OnCATXTPLWRRGTHYLCNAQ3iLYHK QNRLIXPKXRL IXXXXPAGTXCXNCQTSTTTLWRRNAXGDPVCNACGLYYKLHN IXNRPL

GATA-1 D V R K TICK QC C V

A A LV V T C

M

GATA-2 C A K R SAAR C A N I

GATA-3 G S K R SAAR S A N I

FIG. 2. Distinctive features of the finger region of the vertebrate GATA-binding proteins. Amino acids 314-409 of the finger domains ofmGATA-1 are shown at the top with positions ofdivergence noted by X. Amino acid residues characteristic of the various GATA family membersare shown below. Intervening sequence position is marked based on the structure of the murine GATA-1 gene (30).

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 10645

A w

OH

F J0

.X wz0

a.0J

A

/3T1 GLOBIN

cMCDLO

1 2 3 4 5 6 7 8 9 10 11 12131415

GATA-1

GATA-2

- 28S

.^

X -r

-

.:

GATA-3

MYOD-18S

B

aTl Globin

1 2 3 4 5 6 7 8 9 10 11 12

*10 *I n

B ,

w 0 w

2 cc x-J

-28 S

-18 S

FIG. 3. Expression of GATA-binding protein genes in Xenopuserythroid cells. Total cell RNAs (15 ,ug) were subjected to Northernblot analysis and hybridized with a xGATA-1B cDNA probe (A) ora xGATA-3 probe (B). Cells studied include Xenopus eggs or oocytes(egg), tadpole erythroid cells, adult erythroid cells, adult reticulo-cytes (retic), XTC cell line, XL cell line, murine erythroleukemiacells (MEL), and human erythroleukemia cells (K562).

etic roles during this stage ofdevelopment. GATA-3 has beenshown to be expressed in the embryonic brain (20, 21) and,consequently, may be anticipated to act during neurogenesis.Our preliminary RNA in situ analyses in Xenopus indicatethat GATA-1 and GATA-2 are expressed in the developingventral blood island region, and GATA-2 and GATA-3 aredetected in defined regions of the embryonic nervous system(unpublished results). (it) As shown in Fig. 4, the onset ofxGATA-1 RNA expression (stage 11) occurs earlier than theappearance of early larval aT1 globin RNA (stage 15, Fig.4B), 3T1 globin RNA (stage 26, see Fig. 4A), or blood islandformation (stage 30). Of note, the initial detection of aT1globin mRNA during the late neurula stage is earlier than hadbeen demonstrated (37-39) and precedes the developmentalstage when terminal erythroid differentiation was thought tobegin (stage 22) (40, 41). It is currently believed that expres-sion of GATA-1 is restricted to hematopoietic tissues invertebrates. As such, the considerable lag period between the

FIG. 4. Developmental expression of xGATA-binding proteinand larval globin genes. (A) Total cellular RNAs (1 /g) from stagedXenopus embryos were subjected to RT-PCR using specific primersfor ,T1 globin, xGATA-1, xGATA-2, xGATA-3, and xMyoD. Con-trols consisted of RT-PCR products from total RNA from Xenopusadult erythroid cells and HeLa cells and H20. Lanes: 1, oocyte; 2,4-cell embryo; 3, 32-cell embryo; 13, erythroid cells; 14, HeLa cells;15, H20. The Nieuwkoop stage is indicated in the following lanes.Lanes: 4, 7; 5, 9; 6, 11; 7, 12.5; 8, 13; 9, 18; 10, 24; 11, 26; 12, 30. (B)Total cellular RNAs from staged Xenopus embryos were subjectedto Northern blot analysis and hybridized with an aT1 globin cDNAprobe (37). Nieuwkoop stages are as follows. Lanes: 1, 10.5; 2, 12;3, 13; 4, 15; 5, 17; 6, 20; 7, 24; 8, 26; 9, 30; 10, 33; 11, 38; 12,erythroid-cell control.

first expression of xGATA-1 and the appearance of ventralblood islands is reminiscent of studies of hematopoieticcommitment in chicken embryos. Through study of thepresumptive hemoglobin forming area of 20- to 23-hr chickenembryos plated in fibrin clot assays, Groudine and Weintraub(4) estimated that hematopoietic commitment begins duringearly gastrulation. Perhaps early expression of xGATA-1 (orxGATA-2) allows for subsequent hematopoietic differentia-tion, which may be promoted further by specific mesodermalinducing factors in the developing embryo.Smith and coworkers (42, 43) have reported that multipo-

tential animal pole explants from Xenopus embryos formblood-like cells when cultured in the presence of mesoder-mal-inducing factors, such as basic fibroblast growth factor.These cells, however, are not mature erythroid cells sinceglobin was not detected. By transplantation of cytogeneti-cally distinct ventral blood island tissue, it has been demon-strated that a small ventral portion of the embryo containsstem cells that will ultimately yield erythrocytes, macro-phages, thymocytes, and B lymphocytes (41). Of great inter-est, the location of the embryo to which the stem cells aretransplanted influences the frequency of contribution tocommitted cell lineages (44). Lymphoid contribution is sig-nificantly increased when cells are transplanted to a periph-eral location in the embryo, whereas erythroid contributionis increased when cells are placed in a central region. Hence,environment apparently exerts a directive effect on stem cellsof the embryo to form particular hematopoietic tissues. Inthis context, the expression of xGATA-1 RNA at stage 11may reflect a potential commitment of mesoderm to hema-topoietic differentiation, analogous to the low-level wide-

Developmental Biology: Zon et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

10646 Developmental Biology: Zon et al.

spread expression of MyoD at midblastula transition (45).Preliminary experiments reveal that cultured animal-cap mul-tipotential cells express GATA-1 RNA after MyoD RNAlevels have declined (unpublished results), perhaps indicatinga temporal expression of mesodermal commitment. Furtherstudy of the mechanisms by which growth factors or cell-cellinteractions activate expression of xGATA-binding proteinsor, subsequently, localize these proteins to a particular regionof the embryo (e.g., ventral blood islands) may providevaluable insights into the inductive process of hematopoieticmesoderm formation.

We thank Gerry Thomsen forRNA from stagedXenopus embryos,Doug Melton for helpful discussions, and Lynda Schneider forproofreading nucleotide sequence data. This work was supported byNational Institutes of Health Physician Scientist Award HL02347(L.I.Z.), NIH Grant HL32259 (S.H.O.), the Charles H. HoodFoundation (L.I.Z.), and NIH Grant GM42341 (R.H.). S.H.O. is anInvestigator of the HHMI.

1. Rowley, A. F. & Ratcliffe, N. A. (1988) Vertebrate BloodCells, eds. Rowley, A. F. & Ratcliffe, N. A. (Cambridge Univ.Press, Cambridge, U.K.), pp. 1-19.

2. Till, J. E., McCulloch, E. A. & Sininovitch, L. (1964) Proc.Natl. Acad. Sci. USA 51, 29-36.

3. Moore, M. A. & Metcalf, D. (1970) Br. J. Haematol. 18,279-2%.

4. Groudine, M. & Weintraub, H. (1981) Cell 24, 393-401.5. Melton, D. A. (1991) Science 252, 234-241.6. Hopwood, N. D. & Gurdon, J. B. (1990) Nature (London) 347,

197-200.7. Turner, R. J. (1988) Vertebrate Blood Cells, eds. Rowley, A. F.

& Ratcliffe, N. A. (Cambridge Univ. Press, Cambridge, U.K.),pp. 129-210.

8. Hosbach, H. A., Wyler, T. & Weber, R. (1983) Cell 32, 45-53.9. Meyerhof, W., Stalder, J., Koster, M., Wirthmuller, U. &

Knochel, W. (1990) Mol. Biol. Rep. 14, 17-26.10. Banville, D. & Williams, J. G. (1985)J. Mol. Biol. 184, 611-620.11. Evans, T., Reitman, M. & Felsenfeld, G. (1988) Proc. Natl.

Acad. Sci. USA 85, 5976-5980.12. Wall, L., deBoer, E. & Grosveld, F. (1988) Genes Dev. 2,

1089-1100.13. Orkin, S. H. (1990) Cell 63, 665-672.14; Zon, L. I., Tsai, S. F., Burgess, S., Matsudaira, P., Bruns,

G. A. & Orkin, S. H. (1990) Proc. Natl. Acad. Sci. USA 87,668-672.

15. Tsai, S. F., Martin, D. I., Zon, L. I., DAndrea, A. D., Wong,G. G. & Orkin, S. H. (1989) Nature (London) 339, 446-451.-

16. Martin, D. I., Zon, L. I., Mutter, G. & Orkin, S. H. (1990)Nature (London) 344, 444-447.

17. Romeo, P. H., Prandini, M. H., Joulin, V., Mignotte, V.,Prenant, M., Vainchenker, W., Marguerie, G. & Uzan, G.(1990) Nature (London) 344, 447-449.

18. Evans, T. & Felsenfeld, G. (1989) Cell 58, 877-885.19. Trainor, C. D., Evans, T., Felsenfeld, G. & Boguski, M. S.

(1990) Nature (London) 343, 92-96.

Proc. Nati. Acad. Sci. USA 88 (1991)

20. Yamamoto, M., Ko, L. J., Leonard, M. W., Beug, H., Orkin,S. H. & Engel, J. D. (1990) Genes Dev. 4, 1650-1662.

21. Ko, L. J., Yamamoto, M., Leonard, M. W., George, K. M.,Ting, P. & Engel, J. D. (1991) Mol. Cell. Biol. 11, 2778-2784.

22. Joulin, V., Bories, D., Eleouet, J.-F., Labastie, M.-C., Chre-tien, S., Mattei, M.-G. & Romeo, P.-H. (1991) EMBO J. 10,1809-1816.

23. Ho, I. C., Vorhees, P., Marin, N., Oakley, B. K., Tsai, S. F.,Orkin, S. H. & Leiden, J. M. (1991) EMBO J. 10, 1187-1192.

24. Higgs, D. R., Wood, W. G., Jarman, A. P., Sharpe, J., Lida,J., Pretorius, I. M. & Ayyub, H. (1990) Genes Dev. 4, 1588-1601.

25. Pevny, L., Simon, M. C., Robertson, E., Klein, W. H., Tsai,S. F., DAgati, V., Orkin, S. H. & Costantini, F. (1991) Nature(London) 349, 257-260.

26. Whitelaw, E., Tsai, S. F., Hogben, P. & Orkin, S. H. (1990)Mol. Cell. Biol. 10, 6596-6606.

27. Widmer, H. J., Hosbach, H. A. & Weber, R. (1983) Dev. Biol.99, 50-60.

28. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D.,Smith, J. A., Seidman, J. G. & Struhl, K. (1987) CurrentProtocols in Molecular Biology (Wiley, New York).

29. Thomsen, G., Woolf, T., Whitman, M., Sokol, S., Vaughan, J.,Vale, W. & Melton, D. A. (1990) Cell 63, 485-493.

30. Tsai, S., Strauss, E. & Orkin, S. (1991) Genes Dev. 5, 919-931.31. Banville, D., Kay, R. M., Harris, R. & Williams, J. G. (1983)

J. Biol. Chem. 258, 7924-7927.32. Hopwood, N. D., Pluck, A. & Gurdon, J. B. (1989) EMBO J.

8, 3409-3417.33. Kawasaki, E. S. (1990) PCR Protocols, eds. Innis, M. A.,

Gelfand, D. H., Sninsky, J. J. & White, T. J. (Academic, SanDiego), pp. 21-27.

34. Martin, D. I. & Orkin, S. H. (1990) Genes Dev. 4, 1886-1898.35. Smith, J. C., Price, B. M., Van, N. K. & Huylebroeck, D.

(1990) Nature (London) 345, 729-731.36. Meunier, H., Rivier, C., Evans, R. M. & Vale, W. (1988) Proc.

Natl. Acad. Sci. USA 85, 247-251.37. Widmer, H. J., Andres, A. C., Niessing, J., Hosbach, H. A. &

Weber, R. (1981) Dev. Biol. 88, 325-332.38. Patient, R. K., Banville, D., Brewer, A. C., Elkington, J. A.,

Greaves, D. R., Lloyd, M. M. & Williams, J. G. (1982) NucleicAcids Res. 10, 7935-7945.

39. Banville, D. & Williams, J. G. (1985) Nucleic Acids Res. 13,5407-5421.

40. Nieuwkoop, P. D. & Faber, J. (1957) Normal Table ofXenopuslaevis (Daudin) (North-Holland, Amsterdam).

41. Smith, P. B., Flajnik, M. F. & Turpen, J. B. (1989) Dev. Biol.131, 302-312.

42. Smith, J. C., Cooke, J., Green, J. B., Howes, G. & Symes, K.(1989) Development 107, Suppl. 1, 149-159.

43. Green, J. B., Howes, G., Symes, K., Cooke, J. & Smith, J. C.(1990) Development 108, 173-183.

44. Turpen, J. B. & Smith, P. B. (1989) J. Immunol. 143, 3455-3460.

45. Rupp, R. A. W. & Weintraub, H. (1991) Cell 65, 927-937.46. Dorfman, D., Wilson, D. & Orkin, S. H. (1991) J. Biol. Chem.,

in press.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0