Proc. Natl. Acad. Sci. USAVol. 90, pp. 1097-1101, February 1993Biochemistry

SRY recognizes conserved DNA sites in sex-specific promoters(Mullerian inhibiting substance/P450 aromatase/transcription factors/DNA-binding proteins/sexual dimorphism)

CHRISTOPHER M. HAQQ*t, CHIH-YEN KINGt, PATRICIA K. DONAHOE*, AND MICHAEL A. WEISSt*§tDepartment of Biological Chemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA 02115; and Departments of *Pediatric Surgery and*Medicine, Massachusetts General Hospital, Boston, MA 02114

Communicated by Charles C. Richardson, November 2, 1992

ABSTRACT Formation of male-specific structures andregression of female primordia are regulated in early maleembryogenesis by SRY, a singe-copy gene on the Y chromo-some. Assignment of SRY as the testis-determining factor ineutherian mammals is supported by molecular analysis ofcytogenetic sex reversal (i.e., XX males and XY females) andby complementary studies of transgenic murine models. Herewe characterize the putative DNA-binding domain of SRY,which contains a conserved sequence motif shared by high-mobility group nuclear proteins and a newly recognized classof transcription factors. The SRY DNA-binding domain spe-cifically recognizes with nanomolar affinity proximal upstreamelements (designated SRYe) in the promoters of the sex-specificgenes encoding P450 aromatase and Muflerian inhibiting sub-stance (MIS). P450 aromatase catalyzes the conversion oftestosterone to estradiol, and in the male embryo its expressionis down-regulated. Conversely, MIS is expressed in the maleembryo to induce testicular differentiation and regression offemale reproductive ducts. SRYe-binding activity is observed innuclear extracts obtained from embryonic urogenital ridgeimmediately preceding morphologic testicular differentiation.Our results support the hypothesis that SRY directly controlsmale development through sequence-specific regulation of tar-get genes.

Morphogenesis and pattern formation are regulated by de-velopmental programs of selective and sequential gene ex-pression (1). Sexual dimorphism in the embryogenesis ofeutherian mammals provides a model for a genetic switchbetween alternative programs (2-4). In humans the malephenotype is determined by a gene or set ofgenes on the shortarm of the Y chromosome (5). SRY, a single-copy gene ininterval lA1 (6), has recently been identified as encoding thetestis-determining factor, the master switch that initiates acascade of events leading to testicular differentiation andmale development (7-9). This genetic program involves bothpositive and negative regulation: induction of male-specifictissues and regression of female primordia (Fig. 1). Criticalelements of the downstream pathway are regulated by auto-somal loci, including those encoding Mullerian inhibitingsubstance (MIS) (10, 11) and P450 aromatase. MIS, a memberof the transforming growth factor ( (TGF-(3) family (12), isresponsible for regression of female reproductive ducts inmale embryos; P450 aromatase (13, 14) catalyzes the con-version of testosterone to estradiol during female gonadalspecification and is down-regulated in the male embryo (15,16). The molecular mechanisms involved in these and otherpathways (17) of gene expression in early vertebrate devel-opment represent problems of fundamental importance.SRY contains a motif [designated the HMG box, for

high-mobility group (6, 7)] that is shared by a newly recog-nized family of transcription factors (18). This homology

suggests that SRY likewise encodes a sequence-specificDNA-binding protein. Indeed, weak sequence-specific DNAbinding by SRY has recently been observed to DNA targetsites of other HMG-box proteins: an insulin-responsive ele-ment [IRE-A (19)] and enhancer elements of the T-cellreceptor gene [TCF-la-binding sites (20)]. The significance ofsuch binding is supported by analysis ofmutant SRY domainsidentified among human XY females: loss of DNA-bindingactivity is associated with incomplete male development (19,20). However, the low affinities of the IRE-A/TCF-la DNAsites and their unclear physiologic relationship to male de-velopment suggest the existence of a distinct family of SRYresponse elements. Accordingly, we have investigated theDNA-binding properties of the SRY HMG box in relation togenes involved in gonadal differentiation. Because of thepresumed importance of P450 aromatase and MIS in sexualdimorphism (2, 3, 10, 11, 15, 16), we have focused onpotential SRY target sites in the promoter regions of thesegenes.

MATERIALS AND METHODSSRY Expression. The SRY HMG box (85 residues) was

expressed in Escherichia coli as a thrombin-cleavable fusionprotein (pGEX-2T vector; Pharmacia) using primers 5' ATGGGATCC CATATG CAG GATAGA GTGAAG CGA CCCand 5' AAGGGA TCC TTAATT CTTCGG CAG CAT CTTCGC [bp 579 to 835 as identified (6)]. Four nonnative residues(Gly-Ser-His-Met) were expressed at the N terminus of thecleaved peptide (designated SRY-p).

Protein Purification. The fusion protein was expressed in E.coli strain DH1. The cell lysate (pH 8.0, 100 mM NaCl) wasmixed with 0.6 vol of glutathione agarose; SRY-p was re-moved from the agarose-carrier protein complex by thrombincleavage. Final purification (>98% by SDS/PAGE) wasachieved by using FPLC chromatography (Pharmacia) withan 8-ml phenyl-Superose column (gradient conditions: 1.5-0.0M (NH4)2SO4 atpH 7.0; SRY-p was eluted at 0.75 M). Theprimary structure ofthe peptide was confirmed by N-terminalsequencing and amino acid composition.DNA Probes. The human MIS and rat P450 aromatase

promoters were obtained by PCR of pP3 and pArom-CAT(10, 13). Synthetic oligonucleotides were obtained from Oli-gos, Etc. (Wilsonville, OR).

Gel Retardation Conditions. Duplex 32P-labeled oligodeoxy-nucleotide and protein were combined and incubated for 30min at 40C in 10 mM Tris-HCl, pH 7.4/50 mM NaCl/2 mMMgCl2/0.5 mM dithiothreitol/4% (vol/vol) glycerol contain-ing poly(dI-dC) at 40 .g/ml.DNase Protection Conditions. Reactions were in 12% (vol/

vol) glycerol/10 mM Tris HCl, pH 7.4/50 mM NaCl/2 mMMgCl2 in 20 1.d. After 30 min at 40C, 20 p.l of 10 mM MgCl2/5

Abbreviations: MIS, Mullerian inhibiting substance; TGF-,f, trans-forming growth factor 3; HMG, high-mobility group; IRE-A, insulin-responsive element A.§To whom reprint requests should be addressed at t.

1097

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.

Proc. Natl. Acad. Sci. USA 90 (1993)

IBI POTENT IAL EMBRO

I

I MIS GENE EXPRESSIONI

|REGRESSION OF MUL

I IMALE STEROIfDOGENES IS1

[INHIBITION OF AROMATASE +LERIAN DUCT| | IANDROGENS

DECREASED ESTROGENSI WOLFFIAN DUCT PROLIFERATION

ABSENT FEMALE SEX STRUCTURES |

FIG. 1. Proposed pathway of sexual dimorphism in mammalian development. At day 12 of male rat embryogenesis (days 35-40 in human)expression ofSR Yin the indifferent urogenital ridge initiates testicular differentiation (Wolffian duct; right-hand branch) and regression offemaleprimordia (Mullerian duct; left-hand branch). Possible roles of downstream factors Mullerian inhibiting substance (MIS) and P450 aromataseare indicated. Shaded arrow indicates proposed indirect role of aromatase down-regulation in male differentiation.

mM CaC12 was added, reaction mixtures were warmed toroom temperature for 1 min and treated with DNase I (-0. 1ng) for 1 min, and reactions were terminated with 5 Al of 3 Msodium acetate, pH 5.2/0.25 M EDTA containing tRNA at0.2 mg/ml and 50 A.l of phenol/chloroform.

Polyclonal Antisera. To obtain polyclonal antisera toSRY-p, FPLC-purified SRY-p (100 ,ug in complete Freund'sadjuvant) was injected into the popliteal lymph nodes ofNewZealand White rabbits. Booster injections of 25 A&g in incom-plete Freund's adjuvant were given 1 and 2 months later.Serum was collected at serial intervals and purified by pre-cipitation with (NH4)2SO4 and chromatography on staphylo-coccal protein-A-Sepharose.

Nuclear Extracts. Nuclear extracts (21) were obtained fromstaged microdissected embryonic and postnatal tissues. Em-bryonic days 11.5-13.5 of rat gestation were chosen tocorrespond to embryonic days 10.5-12.5 in the mouse (22).Sprague-Dawley rats (Holtzmann, Madison, WI) were usedin accord with the protocol for animal use approved by theinstitutional review board of the Massachusetts GeneralHospital (accession no. 88-1186).

Spectroscopy. Circular dichroism spectra were obtained atpeptide concentrations of 10 and 100 ,M and pathlength 0.1cm in an Aviv spectropolarimeter equipped with a tempera-ture-control unit. 'H-NMR spectra were obtained at 500 MHzat the Harvard Medical School NMR Facility.

RESULTS AND DISCUSSIONPeptide Characterization. SRY-p is monomeric at a peptide

concentration of 100 ,uM as determined by FPLC size-exclusion chromatography. Circular dichroism (Fig. 2A)indicates a helix content of40-50%, which does not vary withprotein concentration. Thermal unfolding of SRY-p is coop-erative and reversible (Fig. 2A Inset). The 1H NMR spectrumof SRY exhibits a broad range of secondary shifts charac-teristic of a folded domain (Fig. 2B). These results demon-

strate that the SRY HMG box encodes an autonomousdomain.Promoter Footprints. The proximal promoter regions of the

mammalian MIS (23) and P450 aromatase (13, 14) genes

A A5 -20

-40E 0 --60

CM -80

~ 5 -100.(D 0 20 40 609) ~~~~~~~~temp(or)CD

20,

-20200 210 220 230 240 250

X (nm)

B H20

. if ~

11 10 9 8 7 6 5 4 3 2 1 0 -1 ppm

FIG. 2. (A) Far-ultraviolet CD spectrum of SRY-p in phosphate-buffered saline (pH 7.4) demonstrates the presence of orderedsecondary structure; the content of a-helix is estimated to be40-50%. (Inset) Cooperativity of protein folding is indicated bythermal unfolding studies. The monitoring wavelength was 222 nm,and data are shown as a percentage of initial ellipticity ([G]0). (B) The500-MHz 'H-NMR spectrum of SRY-p at 5°C and pH 3.7 (0.1%deuterated acetic acid) demonstrates dispersion of chemical shifts.

I. *i .

1098 Biochemistry: Haqq et al.

Biochemistry: Haqq et al. Proc. Natl. Acad. Sci. USA 90 (1993) 1099

A MIS promoters

M3

.*.<ll.:A.±-AA<. r;'f;-CT TTGO.7< 7c+a:: :5:::: 2 ;~;.C.:V AA CT TTTG |7.

SRYerGgGGtCTGTCCTG CACAAACACCCC ttTAAg

0! GC cGcCCCTGcacCACAAACAgCtC1 TATAAa.'a"2ETGTCC cLCcCAAgCACtt 'TAT

B P-450 Aromatase promoters

SRlYeGG-T T GGTTAGTTGAGTTTGCI

'TTG' GTTTTGGTTAGC AATTTGCL 7..' .'..

CTTT. ..

. TATAAAjIATAUAA

[SRY], nMc Ah

.. ;, i .1. -

.5 3,

I

[SRY]. nMD

-r

: s-

Fin. 3. (Ai lro\1n1 1i promnotert 1e-mions of the tihuian utop) box I'imd-

dIe). and IIIitb ttom .II/S Icit1.o\ker_;.c:c leUCtt'IINIIi.iiinu(itIctdcthalt aiT nt thut' Nn1n11W ±1 thr)eeqLeneCe. Shokx n In 'holdI ac A't hehigh- aMnd loi0 -thfiit SRY-pr bindin,itOCS (l .ilICISR?) a d A'1 resrec-tivelR and T\ATA\h\bo lB o'mfi1Iironi(merre1l-nsr tof hti IFit to<a1n!d-1]Lchicken bhottom) 134 )4O-tom sttec eine-i13. 14). %, Ith SR% c-indl TAA'I\ ii-ments showin in bhakll'fc. [N i.scfootprinlt Iof hLI man11 .S1/IS [)p-oteIll1 too-Lra -

nlent 4 to i. Ilr1)i UNoc oo'00tr liiVf ra t aromataLise pIi mo r L iI IL29 to - 1.811. In (I-tlmd 1) SR\Y pr10ot1.III

cOn1cenII tratIon i"iFe fflluli. ti~d iloi)\each lanc: 11le reizioltls (it DNaii pro.-tection and their DN\ 1qii',, e rc,1nitiilcated

contain conserved DNA sequences (Fig. 3 A and B), which cific binding of SRY-p to these regions is demonstrated byare of interest as possible cis-acting control elements. Spe- DNase protection (Fig. 3 C and D); the apparent dissociation

5G \.

G '`rE \.,

c7

0C

0CGc

-Ci

cAAA

C'Ac /

c

fa iC:3

mS. I.

L-.. ,

Proc. NatL. Acad. Sci. USA 90 (1993)

ASPY

9SRY PDNA-

Free DNA--

I _9lle prche.C..no t

P,11

.,.. ,, "' fr-1 ." :'l-l',;.,_,t--r-,Xtr. E-'2 E

'Sv-,N-

Free,SRYDDNA--..,

FIG. 4. (A) Binding of SRY-p protein to a low-affinityCTCTTTGAG (-143 to -134 in the human MIS promoter;3A) and analogs is demonstrated by using the gel-retardatioThe analogs were chosen to resemble the high-affinity site SJM3 analog containing two transversions (CTGTTTGTG; -4;in the human MIS promoter) partially regains high-affinit)binding. Lane 1, wild-type M3 DNA in the absence of SRK2, wild-type M3 DNA in the presence of25 ,uM SRY-p; lane 3M3 DNA CT[C -* G]TTTGAG in the absence of SRY-p;mutant M3 DNA CT[C -- G]TTTGAG in the presence olSRY-p; lane 5, mutant M3 DNA CTCTTTG[A - T]G in theof SRY; lane 6, mutant M3 DNA CTCTTTG[A -- T]Cpresence of 25 ,uM SRY; lane 7, double mutant M3 DNA G(-. G]TTTG[A -- TIG in the absence of SRY; and lane 8mutant M3 DNA GGGCT[C -- G]TTTG[A -- T]G in the Iof 25 IzM SRY. Relative amounts of protein-DNA complexlane (normalized for total radioactivity) were determined tphorlmager analysis (Molecular Dynamics, Sunnyvale, C.half as much radioactivity was loaded in lane 8 as inaccounting for the apparent difference in the intensity of thbands. The autoradiogram was overexposed to reveal thedifference between low- (lane 2) and high-affinity (lanes 6and(B) Stage-specific expression of SR Ye-binding activity inextracts obtained from embryonic day 12 (E12) urogenil(unsexed) in the absence (lane 2) and presence (lane 3) of arcompetitor antiserum; no gel shift is observed in the ab;

constants (Kapp) are <4 nM (aromatase) and 8 nM (MIS).Discrete regions ofprotection (DNase footprints) indicate thepresence of one or more binding sites in each promoterfragment. These regions contain similar sequence elements,which we designate SR Ye. No DNase protection was ob-

9 served in control studies of promoter fragments from thesimian virus 40 early region and the a-gonadotropin gene. Inanalogous DNase protection studies an apparent dissociationconstant of 6 ,LM was observed between an SRY fusionprotein and a control element in the promoter of the insulin-responsive gene for glyceraldehyde-3-phosphate dehydroge-nase (19).

Oligonucleotide Binding and Mutational Analysis. To inves-tigate the SR Ye site in more detail, a series of oligonucleo-I tides were synthesized, and their SRY-binding propertieswere analyzed by a gel-retardation assay (24). Two specificbinding sites were observed in a 30-mer duplex oligonucle-otide (containing human MIS base pairs -69 to -45). Addi-

:- tional gel-retardation experiments demonstrate that a 15-bpoligonucleotide containing this site (-63 to -49; designatedSR Ye,) forms a single specific complex with SRY-p (Fig. 4A).The sequence of SR Ye, (5'-GGGGTGjTTITGCAG and

complement, 5'-CTGCACAAACACCCC) is similar but notidentical to the sequences of IRE-A and TCF-la-relatedenhancer elements (5'-CAGCCTTTGAAAGAA and 5'-AAGCCCTTTGAAAA, respectively). In control studiesthese sites exhibited only weak binding to SRY-p (Kapp 100,uM), in accord with previous studies (19, 20). In fact, asequence closely related to the IRE-A/TCF-la elementsoccurs further upstream in the MIS promoter (5'-GGGC-TCTTTGAGAAG; base pairs -145 to -131; designated M3in Fig. 3A). This region does not exhibit DNase protectionunder conditions in which SRYe is fully occupied; corre-sponding oligonucleotides exhibit weak SRY-p binding (com-parable to that of IRE-A and TCF-la-related oligonucleo-tides). A single A -* T transversion in the MIS upstreamsequence confers 5-fold-enhanced SRY-p binding (Fig. 4A;lanes 6 and 8); an additional 2-fold enhancement is conferredby a C -- G transversion:

5'-GGGCT[C -* G]TTTG[A -- T]G-145 -134

Conversely, T -* A transversion of the central thymidine(asterisk) reduces SRY-p binding by at least 100-fold (data notshown), consistent with the absence of A in this position in

site M3: previously characterized binding sites for TCF-la (26).see Fig. These results demonstrate that SRY and homologous HMGin assay. transcription factors have related but distinct DNA-bindingRYe. An specificities. Similar relationships among target sites are8 to -56 found among other DNA-binding motifs [e.g., helix-turn-y SRY-p helix, helix-loop-helix, and bZIP motifs (27)]. Among HMGr-p; lane family members, such variation may provide a protein-DNA

lIane 4 recognition code to specify and resolve distinct regulatoryfa25 4M pathways.absence Ontogeny of SRYe-Binding Activity. Sequential and over-i in the lapping expression of SRY and MIS mRNA in the differen-jGCT[C tiating testis, as previously reported (7, 22, 28), motivated,, double study of the developmental regulation of SRYe-binding ac-presence tivity in testicular nuclear extracts. To demonstrate a tem-in each poral correlation between SRYe-binding activity and MIS~Y Phos- expression, nuclear extracts were obtained from rat urogen-Al-nn.lane 6,e shiftedmarked

18) sites.nuclear

tal ridgeiti-HMGsence of

nuclear extract (lane 1; control). The antiserum recognizes with equalavidity the homologous HMG domain of IRE-A-binding protein (19)and with 1/5th the avidity the HMG domain of mtTF1 (25) asdetermined by ELISA. The total protein concentration in each lanewas normalized by using the Bradford assay and confirmed byCoomassie-stained SDS/PAGE. No specific SRYe-binding activitywas observed in any nuclear extract after a freeze-thaw cycle.

1100 Biochemistry: Haqq et al.

L---

A i . - j-.m _O

f,

P4 .: -1.

Proc. NatL. Acad. Sci. USA 90 (1993) 1101

ital ridge microdissected at successive stages of embryogen-esis, and the temporal pattern of SRYe-binding activity wasdefined by gel-retardation assays (Fig. 4B). Extracts obtainedat embryonic day 12 from pooled urogenital ridges exhibit aprominent gel shift (lane 2). Similar specific DNA-bindingactivities were noted until day 16. This activity could beblocked by a rabbit polyclonal antibody raised against SRY-p(lane 3) but not by the nonspecific antibody MGH-1 (29). Anadditional SR Ye-binding activity of decreased mobility wasobserved in nuclear extracts obtained from postnatal day 2testes. SR Ye-binding activity was not detectable in nuclearextracts obtained from embryonic male limb (presumably anon-sexually dimorphic tissue) or in nuclear extracts ob-tained from female gonadal ridge or limb on embryonic day14 (data not shown).

CONCLUSIONSThe SRYHMG box encodes an autonomous a -helical motifthat exhibits sequence-specific DNA binding at nanomolarprotein concentrations. High-affinity binding sites in thepromoters of sex-specific genes define a candidate SRY-response element that is distinct from the target sites ofhomologous HMG transcription factors. These results pro-vide a foundation for the further dissection of the role ofSRYin regulating downstream genes involved in gonadal differ-entiation.

Note Added in Proof. SRY-p/SRYe peptide-DNA dissociation con-stants obtained by gel-retardation experiments in the absence ofnonspecific competitorDNA [poly(dI1dC)] are in quantitative accordwith those obtained by DNase protection under similar conditions.

We thank S. Fitzpatrick and J. S. Richards for generously pro-viding the rat aromatose promoter; R. Cate for encouragement in theearly phases of the work; M. Alexander-Bridges for communicationof results prior to publication and for the gift of IRE-BP protein; D.Clayton for the gift of mtTF1 protein; R. Ragin and G. Deegan forassistance with rabbit injections; M. Kenneally for protein-A chro-matography; T. F. Manganaro and S. Hirobe for instruction inembryonic surgery; L. Asmundson and D. T. MacLaughlin forperforming ELISAs to determine antiserum avidity; the BostonBiomedical Research Institute for use ofaCD spectropolarimeter; U.Hansen for plasmid pSVS; L. Madison and J. L. Jameson forproviding the humana -gonadotropin gene; M. C. Botfield and S. E.Shoelson for advice; and D. C. Page for the SRY cDNA and helpfuladvice. This work was supported in part by grants from the NationalInstitutes of Health (CA-17393 to P.K.D. and HD-26465 to M.A.W.);the Lucille Markey Charitable Trust and an American CancerSociety Junior Faculty Research Award (to M.A.W.); and theRowland Foundation for Biomedical Research in the Harvard-Massachusetts Institute of Technology Program in Health Sciencesand Technology and a National Institutes of Health Genetics ofCancer Training Grant (T32CA-09595) at Harvard Medical School(C.M.H.).

1. St. Johnston, D. & Nusslein-Volhard, C. (1992) Cell 68, 201-219.

2. Jost, A. (1947) Arch. Anat. Microsc. Morphol. Exp. 36, 271-315.

3. Jost, A. (1972) Arch. Anat. Microsc. Morphol. Exp. 61, 415-438.

4. de la Chappelle, A. (1972) Am. J. Hum. Genet. 24, 71-105.5. Page, D. C., Mosher, R., Simpson, E., Fisher, E. M. C.,

Mardon, G., Pollack, J., McGillivray, B., de la Chapelle, A. &Brown, L. G. (1987) Cell 51, 1091-1104.

6. Sinclair, A. H., Berta, P., Palmer, M. S., Ross Hawins, J.,Griffiths, B. L., Smith, M. J., Foster, J. W., Frischauf, A.-M.,Lovell-Badge, R. & Goodfellow, P. N. (1990) Nature (London)346, 240-244.

7. Gubbay, J., Collignon, J., Koopman, P., Capel, B., Economou,A., Munsterberg, A., Vivian, N., Goodfellow, P. & Lovell-Badge, R. (1990) Nature (London) 346, 245-250.

8. Berta, P., Hawkins, J. R., Sinclair, A. H., Taylor, A., Grif-fiths, B. L., Goodfellow, P. N. & Fellous, M. (1990) Nature(London) 348, 448-450.

9. JAger, R. J., Anvret, M., Hall, K. & Scherer, G. (1990) Nature(London) 348, 452-454.

10. Cate, R. L., Mattaliano, R. J., Hession, C., Tizard, R., Farber,N. M., Cheung, A., Ninfa, E. G., Frey, A. Z., Gash, D. J.,Cow, E. P., Fisher, R. A., Bertonis, J. M., Torres, G., Wall-ner, B. P., Ramachandran, K. L., Ragin, R. C., Manganaro,T. F., MacLaughlin, D. T. & Donahoe, P. K. (1986) Cell 45,685-698.

11. Picard, J.-Y., Benarous, R., Guerrier, D., Josso, N. & Kahn,A. (1986) Proc. Natl. Acad. Sci. USA 83, 5464-5468.

12. Lyons, K. M., Jones, C. M. & Hogan, B. L. (1991) TrendsGenet. 7, 408-412.

13. Hickey, G. J., Krasnow, J. S., Beattie, W. G. & Richards,J. S. (1990) Mol. Endocrinol. 4, 3-12.

14. Matsumine, H., Herbst, M. A., Ignatius Ou, S.-H., Wilson,J. D. & McPhaul, M. J. (1991) J. Biol. Chem. 266,19900-19907.

15. George, F. W., Milewhich, L. & Wilson, J. D. (1978) Nature(London) 274, 172-173.

16. Vigier, B., Forest, M. G., Eychenne, B., Bezard, J., Garrigou,O., Robel, P. & Josso, N. (1989) Proc. Natl. Acad. Sci. USA86, 3684-3688.

17. Jessell, T. M. & Melton, D. A. (1992) Cell 68, 257-270.18. Ner, S. S. (1992) Curr. Biol. 2, 208-210.19. Nasrin, N., Buggs, C., Kong, X. F., Carnazza, J., Goebl, M.

& Alexander-Bridges, M. (1991) Nature (London) 354, 317-320.20. Harley, V. R., Jackson, D. I., Hextall, P. J., Hawkins, J. R.,

Berkovitz, G. D., Sockanathan, S., Lovell-Badge, R. & Good-fellow, P. N. (1992) Science 255, 453-456.

21. Dignam, J. D., Lebovitz, R. M. & Roeder, R. (1983) NucleicAcids Res. 11, 1475-1489.

22. Kuroda, T., Lee, M. M., Haqq, C. M., Powell, D. M., Man-ganaro, T. F. & Donahoe, P. K. (1990) Endocrinology 127,1825-1832.

23. Haqq, C. M., Lee, M. M., Tizard, R., Wysk, M., DeMarinis,J., Donahoe, P. K. & Cate, R. L. (1992) Genomics 12, 665-669.

24. Fried, M. & Crothers, D. M. (1981) Nucleic Acids Res. 9,6505-6525.

25. Parisi, M. A. & Clayton, D. A. (1991) Science 252, 965-969.26. Waterman, M. L. & Jones, K. A. (1990) New Biol. 2, 621-636.27. Harrison, S. C. (1991) Nature (London) 353, 715-719.28. Koopman, P., Munsterberg, A., Capel, B., Vivian, N. &

Lovell-Badge, R. (1990) Nature (London) 348, 450-452.29. Ragin, R. C., Donahoe, P. K., Kenneally, M. K., Ahmad,

M. F. & MacLaughlin, D. T. (1992) Protein Expression Purif.3, 236-245.

Biochemistry: Haqq et al.