Proc. Nati. Acad. Sci. USAVol. 88, pp. 10638-10641, December 1991Biochemistry

Activation of the erythropoietin receptor promoter by transcriptionfactor GATA-1

(erythroid cells/hematopoietic growth factor/DNA-binding proteins)

LEONARD I. ZON*, HAGOP YOUSSOUFIANtt, CHERI MATHER*, HARVEY F. LODISHt§,AND STUART H. ORKIN*¶*Division of Hematology-Oncology, The Children's Hospital, Dana-Farber Cancer Institute, Department of Pediatrics, Harvard Medical School, Boston, MA02115; tWhitehead Institute for Biomedical Research, Nine Cambridge Center, Cambridge, MA 02142; tHematology-Oncology Unit, Massachusetts GeneralHospital, Boston, MA 02114; §Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139; and lHoward Hughes MedicalInstitute, Boston, MA 02115

Contributed by Stuart H. Orkin, August 19, 1991

ABSTRACT Erythropoietin, a glycoprotein produced bythe kidneys in response to anemia and hypoxia, is a majorgrowth factor for cells of the erythroid lineage. Erythropoietininteracts with high-affinity cell surface receptors (EpoR) pres-ent on developing progenitors and is required for their survival.Previously'we characterized the gene for EpoR and demon-strated that its promoter acts in a cell-specific manner. Here weshow that the hematopoeitic-specific' transcription factorGATA-1 is necessary, and indeed is sufficient as the solecell-restricted regulator, for activation of the EpoR promoterin fibroblast transfection assays. Hence, GATA-1, which par-ticipates in transcriptional control of the majority of erythroid-expressed genes, also acts on the promoter of an essentiallineage-restricted receptor (EpoR). This central contribution ofGATA-1 to EpoR promoter function provides a mechanismwhereby a cell-restricted regulator may ensure the viability andsubsequent maturation of progenitor cells during hematopoi-etic differentiation.

The hormone erythropoietin (Epo), which interacts with ahigh-affinity cell surface receptor (EpoR) on developingerythroblasts, is critical for normal erythroid development(1). Intracellular signaling consequent to binding of the ligandto its receptor provides a proliferative response and ensuresdifferentiation of the erythroid lineage (2). The EpoR isexpressed in a cell-restricted fashion': the polypeptide (or itsmRNA) is present only in erythroid, megakaryocytic, andmast cells (refs. 3 and 4; L.I.Z., unpublished data). We haverecently demonstrated that a 458-base-pair (bp) fragment ofthe EpoR promoter is active upon transfection into erythroidcells but is inactive in nonerythroid cells (5).One clue to the basis for the cellular distribution of the

EpoR is the presence of a consensus motif (GATA) in theproximal promoter of both munrne and human EpoR genes(5-8). GATA motifs (9, 10) are found in the vast majority oferythroid-expressed genes and are recognized by a he-matopoietic-specific transcription factor GATA-1 (11), whichin turn is expressed in cells oferythroid, megakaryocytic, andmast lineages (12, 13). Targeted disruption of the GATA-1gene in murine ES cells blocks normal erythroid development(14). As regulation of EpoR expression through the action ofGATA-1 would provide a biologically meaningful link be-tween a nuclear regulator and a lineage-restricted receptor,we have assessed the role of the GATA element in EpoRpromoter function.

MATERIALS AND METHODSGel Mobility Shift Analysis. Preparation of nuclear extracts

of mouse erythroleukemia (MEL) cells and gel mobility shift

analysis were performed as previously described (11). Incu-bation mixtures contained 2 pug of poly(dI-dC), 5-10 x lo,dpm of end-labeled DNA fragment (approximately 3 x 105dpm/ng), and 2-6 tug of nuclear extract in a total volume of20 AL. DNA-protein complexes were resolved from unboundDNA by electrophoresis on 5% polyacrylamide gels in 0.5 xTBE (90 mM Trisborate, pH 8.2/2.5 mM EDTA) buffer.

Analysis of Promoter Activity. A 458-bp fragment corre-sponding to nucleotides -337 to +121 of the wild-typemurine EpoR gene (+ 1 is defined as transcriptional start site,and the major translational start site is at +150) was gener-ated by PCR and cloned in a growth hormone (GH) reporterconstruct between the Xba I and Hind III sites (15). Corre-sponding 458-bp fragments with mutations of the GATA andSpO elements (Fig. la) were similarly assembled by usingPCR mutagenesis (17). For transient expression in MELcells, 2 X 107 cells were subjected to electroporation using100 Ag of reporter plasmid as previously described (18). GHproduction was measured by radioimmunoassay (NicholsInstitute) at 72 hr. NIH 3T3 cells were transfected by themethod of Okayama (19) as previously described (18), using15 pug of GATA-1 cDNA expression plasmid and 2 ,ug ofEpoR promoter-reporter plasmid.Northern Blot Analysis. Total cellular RNA was loaded at

20 Ag per lane. Hybridization and washing at high stringencywere performed as before (12). Probes included a partialmurine GATA-2 cDNA probe (unpublished data) and themurine GATA-3 cDNA probe (20). As a control for RNAloading, filters were hybridized with a 83-actin probe.

RESULTSGATA-1 and Spl Specifically Bind the EpoR Promoter. The

proximal promoter of the murine'EpoR gene contains poten-tial GATA (in the inverted orientation) and Spl protein-binding sites (see Fig. la) (21). Initially we examined bindingof nuclear proteins derived from MEL cells to these potentialtargets. As revealed by gel retardation assay, the GATAmotifforms a specific complex with GATA-1 present in MELcell extract (Fig. lb, lanes 1-5). Mutation of the GATA coreimpairs protein binding (lanes 6-10). Binding of the wild-typesequence is inhibited by competitor DNA containing aGATAsequence (lanes 2 and 4) and not by unrelated DNA (lanes 3and 5). The Spl consensus binding site (-16 to -21) is alsospecifically recognized by a protein in MEL extract (Fig. ic;lanes 1-5). Binding is prevented by mutation ofthis site (lanes6-10). Purified Spl also binds to this site with formation ofthesame complex as that seen with MEL extract (not shown).From these data, we conclude that the GATA and Spl

Abbreviations: Epo, erythropoietin; EpoR, Epo receptor; GH,growth hormone.

10638

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

June

26,

202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 10639

a GATA spi-53 ~~~~~~~~~~~~~~~+1

Wild-Type GCr X 3C C(- I

GATA Mutant TAC ToA1CGGSP1 Mutant TALCXGTT C1

b 1 2 3 4 5 6 7 8 9 10

, i4.

C 2 3 4 5 6 7 8 9 10

in EpoR promoter function, we transfected promoter-reporter constructs into MEL cells (which express endoge-nous EpoR). A 458-bp fragment of the EpoR gene (-337 to+121) was subcloned in the GH reporter plasmid pOGH.Mutations corresponding to those tested for in vitro proteinbinding were introduced by PCR-directed mutagenesis (Fig.la) (16). Using a transient transfection assay, we demon-strated previously that the wild-type EpoR promoter-GHconstruct is active in MEL cells and inactive in fibroblasts(5). Control experiments showed that transcription is-initi-ated appropriately on the EpoR-GH construct in MEL cells(not shown). EpoR promoter activity, as measured by GHproduction by transfected MEL cells, is dependent on theintegrity of both GATA- and Spl-binding sites (Fig. 2a).Thus, the GATA element is necessary for activity of theEpoR promoter. Furthermore, these findings reveal potentialcooperation between GATA-1 and Spl transcription factorsin establishing full promoter activity.GATA-1 Directly Transactivates -the EpoR Promoter. We

next asked whether the EpoR promoter might be activated byGATA-1 expressed in a cell that does not normally express

E

C)

z0

0

00C:0C

--Oewe

10 -

0 I I I I --I

WILD GATA SPiTYPE MUTANT MUTANT

FIG. 1. Gel mobility shift analysis of MEL nuclear extract withprobes derived from the promoter of the EpoR. (a) Sequence of partof the wild-type and mutant EpoR promoters. Mutations are under-lined. (b) Gel-shift assay with labeled wild-type (lanes 1-5) andmutant (lanes 6-10) probes (nucleotides -61 to -28) encompassingthe region of the GATA site. Unlabeled competitor DNA added at200-fold excess: wild-type -61 to -28 (lanes 2 and 7); GATA mutant-61 to -28 (lanes 3 and 8); a-globin probe containing a GATA motif(lanes 4 and 9) (11); and heterologous oligonucleotide (lanes 5 and 10)(16). (c) Gel-shift assay with labeled wild-type (lanes 1-5) and mutant(lanes 6-10) probes (nucleotides -38 to -2) encompassing the regionofthe Spl-binding site. Unlabeled competitor DNA added at 200-foldexcess: wild-type -38 to -2 (lanes 2 and 7); Spi mutant -38 to -2(lanes 3 and 8); heterologous oligonucleotide (a-globin probe con-taining a GATA motif) (lanes 4 and 9); and an Spl consensusoligonucleotide (lanes 5 and 10).

consensus sites in the EpoR promoter interact in vitro withproteins present in MEL nuclear extract.The GATA- and Spl Sites Are Required for EpoR Promoter

Function in Erythroid Cells. To assess the role of these sites

c

z

0

0

D

0.C,

60

40-

20

0

4- -4-

GATA-1

FIG. 2. Functional analysis of the EpoR promoter. (a) Transienttransfection ofMEL cells. Wild-type EpoR, GATA mutant, and Splmutant EpoR promoter fragments were each inserted into a GHreporter plasmid and introduced into MEL cells by electroporation.(b) Transactivation of the EpoR promoter in fibroblasts. NIH 3T3cells were cotransfected with a mammalian expression plasmidcontaining GATA-1 cDNA (+) with the wild-type (WT) and GATAmutant (MUT) EpoR reporter constructs (see a). Control experi-ments (-) were performed using a truncated form of the GATA-1cDNA [designated "mini" by Martin et al. (18)] that lacks activationdomains.

a

b EPO-R PROMOTER

I rWT--1 T

_~

IMA.-

Iz"- "O"qw,

Biochemistry: Zon et al.

.90 A*l.

A* .. ..4IF ..A

Dow

nloa

ded

by g

uest

on

June

26,

202

0

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

the protein. Prior experiments have demonstrated thatGATA-1 serves as a potent transcriptional activator of re-porter constructs containing minimal promoters bearingTATA and GATA motifs when expressed in fibroblasts (18).Although expressed chicken GATA-1 activates the chickenaD-globin promoter through a proximal GATA site (22), otherhuman and murine erythroid-specific promoters in whichGATA elements are more distally positioned (such as thehuman 'y-globin and mouse a-globin promoters) are notactivated in heterologous cells by expression of mammalianGATA-1 alone (18).To examine GATA-1 action on the EpoR promoter, we

cotransfected EpoR-GH and GATA-1 cDNA expressionconstructs into murine fibroblasts and measured reporter(GH) expression. Wild-type EpoR-GH was transactivatedapproximately 50-fold by wild-type GATA-1 relative to acontrol in which a mutant form of GATA-1 cDNA lackingactivation domains was expressed (Fig. 2b). Transactivationof the mutant EpoR promoter was markedly reduced (Fig.2b). Residual activity of this mutant reporter may reflectbinding of GATA-1 to cryptic sites in the plasmid vector andtranscriptional enhancement at a distance (18, 22). In contrastto other erythroid-specific promoters, the EpoR promoter ishighly sensitive and directly responsive to GATA-1 expres-sion in transient assay in a fibroblast environment (18). Assuch, GATA-1 appears to be the only cell-restricted factornecessary for activation of the EpoR promoter in a cotrans-fection-transactivation assay.

Potential Involvement of Other GATA-Binding Proteins inErythroid Cells. Cis-regulatory elements are often recognizedby multiple proteins with similar (or identical) DNA-bindingspecificities that may coexist in cells (23, 24). GATA-1 wasinitially believed to be the sole protein in erythroid cellscapable of high-affinity interaction with the GATA motif.More recently, however, other GATA-binding proteins, re-lated to GATA-1 by virtue of their highly homologous zincfinger domains, have been identified in other species (25). Todate, two additional GATA-binding proteins (GATA-2 andGATA-3) have been characterized in vertebrate cells (20,25-27). Whereas GATA-2 is widely but variably expressed,GATA-3 is largely restricted to T-lymphoid cells and fetalbrain.To evaluate the potential contribution of these proteins to

EpoR promoter regulation, we have examined their expres-sion at the RNA level in MEL cells. GATA-2 transcripts,though present, are detectable only after prolonged exposure(Fig. 3 a and b). GATA-3 RNA is undetectable in MEL cells(Fig. 3c). Forced expression of GATA-2 mRNA in NIH 3T3cells, at levels far greater than that which exists in MEL cells,transactivates the EpoR promoter specifically through theGATA element (data not shown). By quantitation using aPhosphorimager (Molecular Dynamics, Sunnyvale, CA) ofNorthern blots hybridized with probes having comparablespecific activity, we estimate that MEL cells express 30-foldmore GATA-1 mRNA at steady state than GATA-2 mRNA.In accord with this finding, we observe that the vast majorityof GATA-binding activity in MEL cells can be specificallyretarded in gel-shift assays using a polyclonal antiserumspecific to GATA-1 (28). The abundance of GATA-1 inerythroid cells (estimated to be approximately 0.2% ofmRNA in MEL cells) argues strongly that GATA-1, ratherthan another GATA-binding protein, is indeed the factor thatinteracts with the GATA motif in the EpoR promoter in vivo.In support of this conclusion, we find that another erythro-leukemia cell line (GM979) expresses abundant GATA-1 andEpoR RNA, yet lacks GATA-2 mRNA (not shown).

DISCUSSIONWe have demonstrated that the hematopoietic-restrictedtranscription factor GATA-1 specifically recognizes a bind-

C0

GATA 2 4 ._4

-a CI r

III!

GATA3A 41

arti n

FIG. 3. Northern blot analysis of RNA encoding GATA-bindingproteins in various cell lines. Total cellular RNA was loaded at 20 .gper lane. Hybridization and washing at high stringency were per-formed as before (12). (a) Murine GATA-2 cDNA probe (unpub-lished data): lane 1, MEL; lane 2, NIH 3T3; lane 3, P815 (mast cell);exposure time, 14 hr. (b) Murine GATA-2 cDNA probe: lane 1,MEL; lane 2, NIH 3T3; lane 3, P815; exposure time, 72 hr. This isan independent experiment from that shown in a. (c) MurineGATA-3 cDNA probe: lane 1, MEL; lane 2, NIH 3T3; lane 3, EL4(T-lymphoid cell); exposure time, 14 hr. As a control for RNAloading filters were hybridized with a /8-actin probe.

ing site in the promoter of the EpoR gene. In the context ofa transient transfection assay this site is critical for thefunction of the promoter in erythroid cells. In addition,forced expression of GATA-1 in heterologous cells, such asfibroblasts, transactivates the EpoR promoter in cotransfec-tion experiments. This distinguishes the EpoR promoter as asensitive target of GATA-1 activation, as contrasted with thepromoters of many erythroid-expressed genes (e.g., globins)that are not subject to direct transactivation (18). Genes, suchas that encoding EpoR, whose promoters are highly respon-sive to GATA-1 transactivation may constitute a distinctclass of erythroid-expressed products that are required earlyin the program of erythroid development. Although othermembers of the GATA-binding family can bind the GATAsite of the EpoR promoter in vitro, this interaction is unlikelyto be significant in erythroid cells, given their low level ofexpression in vivo.Our findings provide the first evidence (to our knowledge)

for involvement of a cell-restricted transcription factor inregulation of a lineage-specific hematopoietic growth factorreceptor and are of particular relevance to erythroid devel-opment. Two pathways mediated by GATA-1 may serve toguarantee subsequent maturation of developing progenitorcells. First, control of the EpoR promoter through the actionof GATA-1 should ensure continued survival of erythroidprogenitors in the presence of Epo. Second, evidence for afunctional GATA site in the GATA-1 promoter provides abasis for a positive regulatory loop by which GATA-1 main-tains its own expression (17, 29). Hence, a cell-specifictranscription factor and a cell-surface receptor effectivelycooperate to promote survival and maturation of erythroidprogenitors. As revealed through gene disruption in ES cells,erythroid development in vivo requires GATA-1 (14). Thoughthe point at which the block to differentiation is manifestremains to be established, involvement of GATA-1 in controlof the EpoR promoter suggests a simple mechanism wherebythe absence of GATA-1 might impair expression of the EpoRin developing progenitors. As a consequence, EpoR-deficientprogenitors would undergo programmed cell death, despiteadequate circulating Epo (30).

This work was supported by National Heart, Lung and BloodInstitute Grant 5PO1HL33262 (to H.F.L. and S.H.O.), PhysicianScientist Award HL02347 (to L.I.Z.), Physician Scientist AwardHL02277 (to H.Y.), and Grant HL32259 (to S.H.O.), and by the

10640 Biochemistry: Zon et al.

Dow

nloa

ded

by g

uest

on

June

26,

202

0

Proc. Natl. Acad. Sci. USA 88 (1991) 10641

Charles H. Hood Foundation (L.I.Z.). S.H.O. is an Investigator ofthe Howard Hughes Medical Institute.

Note Added in Proof. Subsequent to completion of this work, Chibaet al. (31) have also demonstrated that GATA-1 transactivates theerythropoietin receptor promoter.

1. Erslev, A. J. & Caro, J. (1986) Med. Oncol. Tumor Pharma-cother. 3, 159-164.

2. Sawyer, S. T. (1990) Prog. Clin. Biol. Res. 352, 145-152.3. D'Andrea, A. D., Lodish, H. F. & Wong, G. G. (1989) Cell 57,

277-285.4. Fraser, J. K., Tan, A. S., Lin, F. K. & Berridge, M. V. (1989)

Exp. Hematol. 17, 10-16.5. Youssoufian, H., Zon, L. I., Orkin, S. H., D'Andrea, A. D. &

Lodish, H. F. (1990) Mol. Cell. Biol. 10, 3675-3682.6. Kuramochi, S., Ikawa, Y. & Todokoro, K. (1990) J. Mol. Biol.

216, 567-575.7. Winkelmann, J. C., Penny, L. A., Deaven, L. L. Forget, B. G.

& Jenkins, R. B. (1990) Blood 76, 24-30.8. Jones, S. S., D'Andrea, A. D., Haines, L. L. & Wong, G. G.

(1990) Blood 76, 31-35.9. Wall, L., deBoer, E. & Grosveld, F. (1988) Genes Dev. 2,

1089-1100.10. Evans, T., Reitman, M. & Felsenfeld, G. (1988) Proc. Natl.

Acad. Sci. USA 85, 5976-5980.11. Tsai, S. F., Martin, D. I., Zon, L. I., D'Andrea, A. D., Wong,

G. G. & Orkin, S. H. (1989) Nature (London) 339, 446-451.12. Martin, D. I., Zon, L. I., Mutter, G. & Orkin, S. H. (1990)

Nature (London) 344, 444-447.13. Romeo, P. H., Prandini, M. H., Joulin, V., Mignotte, V.,

Prenant, M., Vainchenker, W., Marguerie, G. & Uzan, G.(1990) Nature (London) 344, 447-449.

14. 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.

15. Selden, R. F., Howie, K. B., Rowe, M. E., Goodman, H. M.& Moore, D. D. (1986) Mol. Cell. Biol. 6, 3173-3179.

16. Tsai, S., Strauss, E. & Orkin, S. (1991) Genes Dev. 5, 919-931.17. Ho, S. N., Hunt, H. D., Horton, R. M., Pullen, J. K. & Pease,

L. R. (1989) Gene 77, 51-59.18. Martin, D. I. & Orkin, S. H. (1990) Genes Dev. 4, 1886-1898.19. Chen, C. & Okayama, H. (1987) Mol. Cell. Biol. 7, 2745-2752.20. Ko, L. J., Yamamoto, M., Leonard, M. W., George, K. M.,

Ting, P. & Engel, J. D. (1991) Mol. Cell. Biol. 11, 2778-2784.21. Briggs, M. R., Kadonaga, J. T., Bell, S. P. & Tjian, R. (1986)

Science 234, 47-52.22. Evans, T. & Felsenfeld, G. (1991) Mol. Cell. Biol. 11, 843-853.23. Sturm, R. A., Das, G. & Herr, W. (1988) Genes Dev. 12,

1582-1599.24. Staudt, L. M., Clerc, R. G., Singh, H., LeBowitz, J. H.,

Sharp, P. A., & Baltimore, D. (1988) Science 241, 577-580.25. Yamamoto, M., Ko, L. J., Leonard, M. W., Beug, H., Orkin,

S. H. & Engel, J. D. (1990) Genes Dev. 4, 1650-1662.26. Ho, I., Vorhees, P., Marin, N., Oakley, B. K., Tsai, S., Orkin,

S. H. & Leiden, J. M. (1991) EMBO J. 10, 1181-1192.27. 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.

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

29. Hannon, R., Evans, T., Felsenfeld, G. & Gould, H. (1991)Proc. Natl. Acad. Sci. USA 88, 3004-3008.

30. Koury, M. J. & Bondurant, M. C. (1990) Science 248, 378-381.31. Chiba, T., Ikawa, Y. & Todokoro, K. (1991) Nucleic Acids Res.

19, 3843-3848.

Biochemistry: Zon et al.

Dow

nloa

ded

by g

uest

on

June

26,

202

0