Cloning of the human estrogen receptor cDNA - Europe PMC

6 downloads 0 Views 1MB Size Report
Cloning of the human estrogen receptor cDNA. (Agtll expression screening/oligonucleotide screening/in vitotranslation/hybrid-selection/monoclonal antibody).
Proc. Natd. Acad. Sci. USA Vol. 82, pp. 7889-7893, December 1985 Biochemistry

Cloning of the human estrogen receptor cDNA (Agtll expression screening/oligonucleotide screening/in vito translation/hybrid-selection/monoclonal antibody)

PHILIPPE WALTER*, STEPHEN GREEN*, GEOFFREY GREENEt, ANDRtE KRUST*, JEAN-MARC BORNERT*, JEAN-MARC JELTSCH*, ADRIEN STAUB*, ELWOOD JENSENt, GEOFFREY SCRACE§, MIKE WATERFIELD§, AND PIERRE CHAMBON*¶ *Laboratoire de Gdndtique Moleculaire des Eucaryotes du Centre National de la Recherche Scientifique, Unite 184 de Biologie Moldculaire et de Genie Gdndtique de l'Institut National de la Sante et de la Recherche Medicale, Facult6 de M6decine, 67085 Strasbourg, France; tBen May Laboratory for Cancer Research, The University of Chicago, 5841 Maryland Ave., Chicago, IL 60637; MLudwig Institute for Cancer Research, Stadelhoferstrasse 22, CH 8001 Zurich, Switzerland; and §Protein Chemistry Laboratory, Imperial Cancer Research Fund, Lincoln's Inn Fields, London WC2A 3PX, United Kingdom

Contributed by Pierre Chambon, July 30, 1985

ABSTRACT Poly(A)' RNA isolated from the human breast cancer cell line MCF-7 was fractionated by sucrose gradient centrifugation and fractions enriched in estrogen receptor (ER) mRNA were used to prepare randomly primed cDNA libraries in the AgtlO and Agtll vectors. Clones corresponding to ER sequence were isolated from both libraries after screening with either ER monoclonal antibodies (Xgtll) or synthetic oligonucleotide probes designed from two peptide sequences of purified ER (XgtlO). Five cDNA clones were isolated by antibody screening and five were isolated after screening with synthetic oligonucleotides. The two largest ER cDNA clones, XOR3 (1.3 kilobase pairs) and XOR8 (2.1 kilobase pairs), isolated by using antibodies and oligonudeotides, respectively, were able to enrich selectively for ER mRNA by hybrid-selection. Furthermore, AOR8 contains the DNA sequence expected from the two ER peptides and crosshybridizes with each of the other ER cDNA dones. These results demonstrate that the clones isolated correspond to the ER mRNA sequence. Use of XOR8 as a hybridization probe revealed a single poly(A)+ RNA band of %6.2 kilobase pairs in the ER-containing human breast cancer cell lines MCF-7 and T47D. In contrast, no hybridization was seen in the human ER-negative cell line HeLa. The same probe hybridizes to a chicken gene that is expressed in oviduct tissue as a 7.5kilobase-pair poly(A)+ RNA.

Estrogens, in common with other steroid hormones, regulate gene expression in target cells through their interaction with specific receptors (for review, see ref. 1). The presence of estrogen receptors (ER) can be determined either by their high affinity binding for [3H]estradiol (2) or by using specific monoclonal antibodies (3, 4). Recent studies have suggested that the estrogen-free receptor is localized predominantly in the nuclear compartment (5, 6), where it is loosely bound until its association with estradiol converts the receptor to an active form with the ability to bind tightly in the genome (2). The activated complex is believed to act directly at some, as yet, ill-defined chromatin site(s), resulting in specific changes in gene expression, although the molecular mechanism by which ER complexes are able to modify the expression of specific genes is so far unknown. Further understanding of this mechanism has been severely hampered due to the low level of ER expression. A high level of expression of ER cDNA, in both homologous as well as heterologous systems, should allow further insight into ER structure and function at the molecular level. Since expression of the ER gene is both tissue-specific and developmenThe publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

tally regulated, isolation of the ER gene should lead to the identification of the responsible sequence elements. ER are believed to play an important role in the growth and development of a subset of hormone-dependent human breast cancers. Approximately one-third of all breast cancer tumors contain significant amounts of ER and about twothirds of these are able to respond objectively to some form of anti-estrogen endocrine therapy (2, 7). Therefore, a better understanding either of the mechanism of estrogen action in these tumors or of the way in which their expression is regulated may lead to an improvement in the therapy of these cancers. We describe here the isolation of several cDNA clones corresponding to the mRNA sequence of the human ER.

MATERIALS AND METHODS Preparation of cDNA Libraries. Sucrose gradient fractions 21-23 (see legend to Fig. 1) that are enriched in receptor mRNA were pooled and used for cDNA synthesis and cloning into Xgt vectors essentially as described by Huynh et al. (8) and Young and Davis (9, 10). Briefly, 5 ,ug of enriched RNA was reverse transcribed in the presence of 35 pmol of a random primer consisting of a 13-mer synthetic oligonucleotide synthesized by using a mixture of each nucleotide at each position (11). After boiling for 90 sec, the second strand was synthesized with DNA polymerase I and the cDNA was treated with S1 nuclease at 25°C. Any internal EcoRI sites were protected by treatment with EcoRI methylase and the extremities of the cDNA were made flush by using T4 DNA polymerase. An equal mass of EcoRI linker (generally 1 ,ug of cDNA with 1 ,ug of the linker GGAATTCC) was ligated to the cDNA overnight at 16°C by using T4 DNA ligase. The cDNA was digested with EcoRI, separated from excess linker by chromatography on a Bio-Gel A-50m column, and ligated at a molar ratio of 2:1 (cDNA:X phage DNA) overnight at 4°C by using either EcoRI-digested XgtlO or EcoRIdigested, alkaline phosphatase-treated Xgtll (8, 9). After packaging the DNA in vitro (12) the phage were amplified on either Escherichia coli C600hf1 (XgtlO) or Y1088 (Xgtll). This technique yielded -5 x 106 recombinants per ,ug of RNA for XgtlO and -1 x 106 recombinants per jig of RNA for Xgtll. Between 85% and 95% of the Xgtll plaques contained inserts. Antibody Screening. The Xgtll cDNA library was plated onto E. coli Y1090 (10), with -25,000 phage per 8.5-cm plate. These were screened with the receptor monoclonal antibodies H222, H226, D75, and D547 (1 ,tg of each per ml) (3, 4) essentially as described by Young and Davis (10) except that proteins expressed by the plaques were transferred in duplicate onto nitrocellulose filters (soaked in 10 mM isopropyl Abbreviations: ER, estrogen receptor(s); kb, kilobase pair(s).

tTo whom all correspondence should be addressed. 7889

7890

Proc. Natl. Acad. Sci. USA 82 (1985)

Biochemistry: Walter et al.

,B-D-thiogalactopyranoside) for 2 hr each. Only phage plaques producing duplicate signals were studied further. Oligonucleotide Screening. Oligonucleotides were synthesized either semimanually (11) or automatically (Sam One, Biosearch, San Rafael, CA) by using the phosphotriester method. The XgtlO cDNA library was plated onto E. coli C600hfl at 3000-5000 plaques per 8.5-cm plate and duplicate filters were prepared, which were hybridized with oligonucleotide mixture A (probe A, Fig. 3) at 35°C and probes B and C (Fig. 3) at 50°C as described by Wallace et al. (13). Filters were washed at 0°C in 0.9 M NaCl/90 mM sodium citrate for probe A or at 0°C in 0.9 M NaCl/90 mM sodium citrate, followed by a 15-min wash at 50°C in 0.3 M NaCl/30 mM sodium citrate for probes B and C. Sequencing of selected clones was performed in M13 mp8 by using the dideoxy technique (14). RESULTS Enrichment in ER mRNA. Poly(A)+ RNA from the human breast cancer cell line MCF-7 (15) was fractionated on sucrose gradients containing methylmercury hydroxide, and RNA sedimenting between =23 S and =32 S was translated in vitro. The proteins from each translation reaction were immunoprecipitated by using a mixture of monoclonal antibodies prepared against the MCF-7 ER (3, 4) and displayed by polyacrylamide gel electrophoresis and fluorography (Fig. 1). Two proteins of =65 kDa and -46 kDa were observed (Fig. 1). The peak of the corresponding mRNA was found in fractions 21 and 22 (Fig. 1, lanes 5 and 6), representing an mRNA that sedimented faster than 28S rRNA. Purified ER labeled with 125I comigrated with the 65-kDa in vitro translation product under these electrophoresis conditions (not shown), supporting the conclusion that fractions 20-23 were enriched in ER mRNA (compare lanes 4-7 with lane 8). The abundance of ER mRNA in total cellular poly(A)+ RNA was kDa - 92.5

65 46

@

- 68 43

- 25.7

1 2 3 4 56 7 8 FIG. 1. Sucrose gradient fractionation and in vitro translation of MCF-7 cell poly(A)+ RNA. RNA was isolated from the human breast cancer cell line MCF-7 by using the LiCl/urea technique (16) and poly(A)+ RNA was purified by chromatography using oligo(dT)cellulose. One hundred micrograms of poly(A)+ RNA was fractionated on 5-20% sucrose gradients (11.2 ml) containing 50 mM Tris HCl (pH 7.5), 1 mM EDTA, and 5 mM methylmercury hydroxide in an SW41 rotor at 40,000 rpm for 6 hr and 28 fractions were collected. Aliquots of fractions 17-23 were translated in vitro in the presence of [35S]methionine (1000 Ci/mmol; 1 Ci = 37 GBq) by using a rabbit reticulocyte lysate following the manufacturer's conditions (NEN). Samples containing -2 x 106 acid-insoluble cpm were precipitated (17, 18) with a mixture of the four monoclonal antibodies [H222, H226, D75, D547 (5 ,g of each); see refs. 3 and 4]. The antibody-selected proteins were then separated on a 12.5% NaDodSO4/polyacrylamide gel. Lanes 1-7 represent fractions 17-23, respectively, of the sucrose gradient. The peak of 28S rRNA was found between fractions 20 and 21 (lanes 4 and 5). Lane 8 represents the in vitro translation (10 x 106 cpm) and immunoprecipitation of 4 ug of poly(A)+ MCF-7 RNA. The sizes of the markers (BRL, high molecular weight) are shown at the right (phosphorylase B, 92.5 kDa; bovine serum albumin, 68 kDa; ovalbumin, 43 kDa; a-chymotrypsin, 25.7 kDa).

-0.003% as estimated by comparing the amount of [35S]methionine present in the immunoprecipitated protein to the amount of methionine incorporated into total protein during in vitro translation. Fractionation of poly(A)I RNA on sucrose gradients resulted in a 10- to 15-fold purification of the receptor mRNA as determined by the same assay. ER Cloning by Expression. Randomly primed cDNA was prepared from selected sucrose gradient fractions enriched in ER mRNA and inserted into the Xgt1O vector and the expression vector Xgtll (Materials and Methods). Each of the four ER monoclonal antibodies H226, H222, D75, and D547 (3, 4) was tested separately for its ability to react with purified ER spotted onto nitrocellulose under the same conditions as those used when screening the Xgtll library. All of the monoclonal antibodies reacted with the receptor. However, antibodies H226 and D75 were each capable of detecting 100 pg of purified receptor, whereas, in the case of H222 and D547, approximately five times more receptor was required to obtain the same signal (not shown). A mixture of all four antibodies was used to screen p4-1.5 million phage plaques. Twenty-one positive plaques were obtained, of which 10 remained positive after two additional rounds of screening. On further analysis it appeared that these 10 clones represented 5 individual clones, which were named XORO to XOR4. Two clones, XOR2 and XOR3, contained two inserted cDNA fragments, but, in each case, only one of the fragments corresponded to ER sequence. Fig. 2 shows the reaction of each of the four monoclonal antibodies with each of the 5 Xgt1l clones. It is interesting to note that the antigen expressed by all 5 clones reacted only with the most "efficient" monoclonal antibodies (H226 and D75). ER Cloning by Using Synthetic Oligonudeotides. Synthetic oligonucleotide probes were used as an alternative approach to the cloning of the ER cDNA. Homogeneous preparations of the MCF-7 ER were cleaved with cyanogen bromide and some of the peptides were sequenced (unpublished data). Two peptide sequences, 99 and 80, were chosen as the most suitable for the design of oligonucleotide probes (Fig. 3). Of the 300,000 XgtlO cDNA clones screened with a mixture of 24 14-mers (probe A, Fig. 3), 35 were positive. To exclude false positives due to hybridization of the probe to closely homologous sequences, these 35 clones were screened with a 41-mer oligonucleotide (probe B) corresponding to the whole of peptide 99 (Fig. 3). Three clones, XOR5, XOR6, and XOR8, were also positive with probe B. Two additional clones, XOR9 and XOR10, were isolated by screening the original 300,000 clones with a 41-mer oligonucleotide (probe C) corresponding to peptide 80 (Fig. 3). Interestingly, 1 of the Receptor monoclonal antibody

II (D

CNJ

M

N e o

IL)

01

C'J CNJ N

XOR1 * XOR3

Xgt1 1 clones

xOR2

0

xOR4 XOR0

FIG. 2. Expression of ER antigens in the Xgtll clones. Each of the five Xgtll clones was spotted onto a plate of E. coli Y1090 and grown at 42°C for 4 hr. The plate was overlaid with a nitrocellulose filter soaked in 10 mM isopropyl f-D-thiogalactopyranoside and left for a further 2 hr at 37°C. The filter was cut into strips and each was incubated with one ofthe four ER monoclonal antibodies by using the standard assay as described for the screening.

Proc. Natl. Acad. Sci. USA 82 (1985)

Biochemistry: Walter et A ER peptide sequence 99

7891

(Met)Ala Leu Leu His Gin lie Gin Giy Asn Glu Leu Glu Pro Leu

3'GTG

Oligonucleotide A

GTi TAG GTi CC 5' T

3'TAC CGG GAC GAC GTG GTC TAG GTC CCG TTG CTC GAC CTC GG 5'

Oligonucleotide B

x x x FGCA TCT GGG ATG GCC CTA CTG CAT CAG ATC CAA GGG AAC GAG CTG GAG CCC CTG AAC LAla Ser Giy Met Ala Leu Leu His Gin lie Gin Giy Asn Glu Leu Glu Pro Leu Asn

ER peptide sequence 80 (Met)Lys Cys Lys Asn Val Val Pro Leu Tyr Asp Leu Leu Gin Oligonucleotide C 3'TAC TTT ACG TTT TTG CAC CAC GGG GAC ATG CTG GAC GAC GT 5' -

x

x

x

x

x

rTAC AGC ATG MAG TGC MG AAC GTG GTG CCC CTC TAT GAC CTG CTG CTG GAG ATG LTyr Ser Met Lys Cys Lys Asn Val Val Pro Leu Tyr Asp Leu Leu Leu Glu Met

O

FIG. 3. Amino acid sequence of ER peptides, nucleotide sequence of the oligonucleotide probes, and amino acid sequences deduced from the corresponding cDNA sequence of the XOR8 clone. Probe A, a mixture of 24 14-mer oligonucleotides, was designed from peptide 99 by using all of the possible codons of the least ambiguous region (His-Gln-Ile-Gln-Gly). Probes B and C were chosen by using the most common codon for each amino acid after taking into consideration the frequency of human codon usage (19).

clones hybridizing with probes A and B, XOR8, also hybridized with probe C. That probes B and C were derived from two independent peptide sequences suggested strongly that XOR8 contains cDNA sequences corresponding to ER mRNA. The regions of XOR8 that hybridized to probes B and C were subcloned into vector M13 mp8 and sequenced (Fig. 3). The two corresponding amino acid sequences match almost perfectly those of peptides 80 and 99, confirming that XOR8 contains an ER cDNA insert. The last amino acid of peptide 80 is glutamine instead of the leucine predicted from the nucleotide sequence. This single discrepancy is most likely related to a decrease in the accuracy of the amino acid sequence after several steps of microsequencing. The cDNA insert from XOR8 was used as a probe for cross-hybridization to each of the XOR clones. As shown in Fig. 4, XOR8 cross-hybridized to all of them, irrespective of the technique used for their selection, indicating that they all

xgt11 clones

A

--o a:cc

0

CY

a: a: o-0 o~ 0 0

4

o

A

contain cDNA inserts corresponding to the ER mRNA sequence. Hybrid-Selection and in Vitro Translation of ER mRNA. The largest clone isolated by antibody screening, XOR3 (1.3 kb), and the largest clone obtained by oligonucleotide screening, XOR8 (2.1 kb), were used to select ER mRNA from sucrose gradient fractions. The mRNA selected by hybridization with either XOR3 or XOR8 was then translated in a rabbit reticulocyte lysate system and the products were electrophoresed on a NaDodSO4/polyacrylamide gel before or after immunoprecipitation with the four monoclonal antibodies (Fig. 5). Even without immunoprecipitation a strong band of =65 kDa was apparent in both cases, and after immunoprecipitation

Agt1O clones 1Wcc x cc = Ecc CC U~o cM

0

00 o

A444

0

oo

4

VO Oa

VV@ )O

00

4

OQ i: ocD

Z

0

CLO G

0

kb -2.1

kb

kDa - 92.5

1.3-1.1

0.7-

0.45-

-0.5

0.3-

O....

68

-0.3

0.15-

B ti g _

500 bp

0O R4 -IA.OR2

)\OR3 i

XOR1!

ORR8

I

~

)ORO

I XORIO

XOR51 AOR6 1-I

~-25. 7

1 2 3 4 M 5 6 7

I-I

XOR9

FIG. 4. Cross-hybridization of the XOR clones with the insert of XOR8. (A) Each of the isolated XOR clones was digested.with EcoRI and electrophoresed on a 1% agarose gel. After transfer to nitrocellulose, the filter was hybridized with the nick-translated insert of XOR8 by using standard techniques (12). Clones XORO to XOR4 were isolated from the Xgtll cDNA library by using monoclonal antibodies, whereas XOR5, XOR6, XOR8, XOR9, and XOR10 were isolated from a XgtlO library by using oligonucleotide probes. The size of the insert ofeach clone is indicated in kilobase pairs (kb). (B) The relative positions of each of the cDNA clones. bp, Base pairs.

FIG. 5. In vitro translation of hybrid-selected MCF-7 ER mRNA. The inserts from XOR3 (1.3 kb) and XOR8 (2.1 kb) were subcloned into pBR322 at the EcoRI site and used to hybrid-select ER mRNA (20, 21) by using sucrose gradient fractions of MCF-7 poly(A)+ RNA-enriched in ER mRNA (fractions 20-23, Fig. 1). The selected RNA was translated in vitro in the presence of [35S]methionine (1000 Ci/mmol) and the proteins were resolved on a 12.5% NaDodSO4/polyacrylamide gel before and after immunoprecipitation by using the four ER monoclonal antibodies. Lane 1, no RNA; lane 2, total translated proteins by using pOR3-selected RNA; lane 3, as lane 2, but by using pOR8; lane 4, proteins translated by using poly(A)+ RNA from gradient fractions 20-23; lanes 5-7, same as lanes 2-4, respectively, but after immunoprecipitation; lane M, markers (see legend to Fig. 1).

7892

Biochemistry: Walter et al.

Proc. Natl. Acad. Sci. USA 82 (1985)

only two bands of 65 kDa and 46 kDa were seen. These two bands corresponded to those seen when total poly(A)+ mRNA or sucrose gradient-enriched mRNA fractions were translated in vitro and immunoprecipitated (Fig. 1). These results support further the conclusion that the XOR clones contain ER cDNA inserts. Sequence Homology Between Human ER cDNA and Chicken RNA and DNA Revealed by RNA Transfer and Southern Analyses. The same amount at total poly(A)+ isolated from laying hen oviduct, HeLa cells, and the human breast cancer cell lines MCF-7 and T47D (22) was electrophoresed on an agarose gel and transferred to diazobenzyloxymethyl-paper. When hybridized with the insert from XOR8 a single band corresponding to an RNA of -6.2 kb was observed for both the T47D and MCF-7 cell lines (Fig. 6A). The T47D cell line contains less ER than the MCF-7 cell line when determined by using a hormone-binding assay (22). The RNA transfer blot suggests that this is also true at the mRNA level (Fig. 6, lanes 3 and 4). No hybridization was observed with the human ER-negative cell line HeLa. Interestingly, the human El probe was capable of hybridizing to a chicken oviduct poly(A)+ RNA of -7.5 kb under high-stringency hybridization conditions. In this respect, we note that H222 and H226 monoclonal antibodies react with chicken oviduct ER (4). Further evidence of a homology between the human ER cDNA sequence and the chicken gene was obtained from Southern analysis (Fig. 6B). Human EcoRI-digested genomic DNA hybridized with the nick-translated cDNA insert of XOR8 indicates that the organization of the human ER gene is complex (Fig. 6B). At present we are not certain of the exact copy number of the ER gene; however, hybridizations performed by using fragments of XOR8 cDNA give only one or two bands, suggesting that the gene may be present as a single copy (data not shown). Therefore, it would appear that -40 kb (the sum of the EcoRI fragments) of genomic sequence is required to code for 2.1 kb of cDNA sequence, which suggests the existence of long intronic sequences. In ram

II 00 :r

0

0a 't

B

kb

kb

8.5

9.3-

27.6

,,_

6.1-

-

-5.Q _

W

-3.6

-3.1

-2.7

2.0%O

4

1

-1.7

2

FIG. 6. (A) RNA transfer analysis. Poly(A)+ RNA (30 ,ug) (from laying hen oviduct, HeLa cells, MCF-7 cells, and T47D cells, in lanes 1-4, respectively) was electrophoresed on 1% agarose gels containing 10 mM methylmercury hydroxide as described (23). The RNA was transferred to diazobenzyloxymethyl-paper and hybridized with the nick-translated insert of XOR8 (24). The arrowheads indicating 7.5 kb and 6.2 kb represent the sizes of the bands seen in the chicken and human RNA samples, respectively (using DNA fragments as size markers). (B) Southern analysis. Genomic DNA (30 fig), digested to completion with EcoRI, was electrophoresed on a 1.5% agarose gel and transferred to diazobenzyloxymethyl-paper (24). The filters were hybridized with the nick-translated cDNA insert of XOR8. Lanes 1 and 2, chicken and human (MCF-7) DNA, respectively. The sizes of the chicken and human EcoRI fragments are given in kb on the left and right, respectively.

contrast, the genomic organization of the chicken gene appears to be less complex, but this may simply reflect the lack of hybridization of some portion of the human cDNA with the chicken gene when hybridized under high-stringency conditions. Hybridization of the cDNA of XOR8 with genomic blots of human-mouse hybrid cell lines has localized the human ER gene to human chromosome 6 (J. L. Mandel, personal communication).

DISCUSSION The use of well-characterized monoclonal antibodies against the human ER purified from MCF-7 cells in combination with a cDNA expression library in Xgtll and of synthetic oligonucleotide probes, derived from amino acid sequences of the purified protein, with a XgtlO cDNA library, has allowed us to isolate cDNA clones corresponding to this receptor. One of these clones, XOR8, was found to contain sequences corresponding to those expected from two peptide sequences of the purified receptor. Two of these clones, XOR3, isolated by antibody screening, and XOR8, isolated by using oligonucleotide probes, were able to hybrid-select an MCF-7 cell mRNA that could be translated in vitro to yield predominantly a protein that has the same size as the ER (-65 kDa) and is selectively immunopurified with the ER monoclonal antibodies. Thus, the cDNA clones presented here correspond to human ER mRNA. Whenever the in vitro, translation products of the MCF-7 poly(A)+ RNA were examined, either before or after hybridselection, a weaker additional band of -46 kDa was observed. The nature of the smaller protein is at present unknown. However, it may correspond to an in vitro degradation product of the larger protein or to a premature termination of translation, since it was only observed in those denaturing sucrose gradient fractions that yielded the 65-kDa protein (see Fig. 1). The 46-kDa component is not observed after immunoprecipitation of iodinated purified MCF-7 ER (unpublished results). The amount of information required to code for a protein of 65 kDa corresponds to an RNA of x2 kb. Since the size of the MCF-7 cell ER mRNA appears to be =6.2 kb, a large fraction of the mRNA should be untranslated. In the vast majority of eukaryotic mRNAs the most 5' AUG is used to initiate translation (25). Therefore, the human mRNA is likely to contain a short 5' and a very long 3' untranslated region. This is not unique to the ER mRNA as several other receptor mRNAs appear to have a similar structure (26-28). One of the most interesting questions related to steroid hormone research is how does the binding of a hormone to the receptor activate gene expression at the transcriptional level? It is likely that the ER behaves in a way similar to those of the progesterone and glucocorticoid receptors and that specific promoter elements are in some way responsible for specific gene activation, possibly by directly binding the hormone-receptor complex (refs. 29-33; see also references in ref. 34). Therefore, the ER protein may consist of at least two functional domains, the hormone and DNA binding sites, as shown previously (4). We have isolated the chicken ER cDNA by using XOR8 as a probe (unpublished results). Sequence comparison between the human and chicken ER cDNAs, together with in vitro genetics, should permit location of these functional domains, whose tertiary structure could subsequently be studied. Expression of ER cDNA in transgenic mice under the control of a variety of promoters may help to evaluate the role of ER in regulation of gene expression during development and in terminally differentiated cells. This could be further examined by transfection of an ER cDNA expression vector into cells together with an estrogen-regulated gene, such as pS2 (35). Similar studies using the ER gene should allow those regions involved in the

Biochemistry: Walter et al. tissue-specific and developmental regulation ofthe gene to be localized. Finally, loss of estrogen dependence in some human breast cancers is often associated with the appearance of more malignant tumors (7). The expression of anti-sense ER mRNA preventing the expression of the endogenous ER in established breast cancer cell lines will offer an opportunity to directly examine the effect of the ER on the growth of these cells. This may lead to a better understanding of the role of the ER in human breast cancer. We thank T. Huynh, R. Young, and R. Davis of Stanford University for the kind gifts of XgtlO, Xgtll, associated E. coli strains, and protocols. We are also grateful to Dr. L. S. Miller of Abbott for providing the H226 and H222 monoclonal antibodies. MCF-7 cells were kindly provided by the Michigan Cancer Foundation. We thank also Mrs. B. Heller and L. Heydler for growing the cells and the secretariat for their help in preparing this manuscript. This research was supported by the Institut National de la Sante et de la Recherche Medicale (Grant CNAMTS), the Centre National de la Recherche Scientifique (Grant ATP 0184), and the Association pour le Developpement de la Recherche sur le Cancer in France and by the National Cancer Institute (Grant CA07897) and the American Cancer Society (Grant BC86) in the United States. S.G. is a recipient of a Royal Society European Exchange fellowship. 1. Anderson, J. N. (1984) in Biological Regulation and Development, eds. Goldberger, R. F. & Yamamoto, K. R. (Plenum, New York), Vol. 3B, pp. 169-212. 2. Jensen, E. V., Greene, G. L., Closs, L. E., DeSombre, E. R. & Nadji, M. (1982) in Recent Progress in Hormone Research, ed. Greep, R. 0. (Academic, New York), Vol. 38, pp. 1-34. 3. Greene, G. L., Nolan, C., Engler, J. P. & Jensen, E. V. (1980) Proc. Natl. Acad. Sci. USA 77, 5115-5119. 4. Greene, G. L., Sobel, N. B., King, W. J. & Jensen, E. V. (1984) J. Steroid Biochem. 20, 51-56. 5. King, W. J. & Greene, G. L. (1984) Nature (London) 307, 745-747. 6. Welshons, W. V., Lieberman, M. E. & Gorski, J. (1984) Nature (London) 307, 747-749. 7. Henderson, I. C. & Canellos, G. P. (1980) New Engl. J. Med. 302, 78-90. 8. Huynh, T. V., Young, R. A. & Davis, R. W. (1985) in DNA Cloning: A Practical Approach, ed. Glover, D. M. (IRL, Oxford), Vol. 1, pp. 98-121. 9. Young, R. A. & Davis, R. W. (1983) Proc. Natl. Acad. Sci. USA 80, 1194-1198. 10. Young, R. A. & Davis, R. W. (1983) Science 222, 778-782. 11. Matthes, H. W. D., Zenke, W. M., Grundstrom, T., Staub, A., Wintzerith, M. & Chambon, P. (1984) EMBO J. 3, 801-805.

Proc. Natl. Acad. Sci. USA 82 (1985)

7893

12. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). 13. Wallace, R. B., Johnson, M. J., Miyake, T., Kawashima, E. H. & Itakura, K. (1981) Nucleic Acids Res. 9, 879-894. 14. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. 15. Brooks, B. C., Locke, E. R. & Soule, H. D. (1973) J. Biol. Chem. 248, 6251-6253. 16. Auffray, C. & Rougeon, F. (1980) Eur. J. Biochem. 107, 303-314. 17. Kraus, J. P. & Rosenberg, L. E. (1982) Proc. Natl. Acad. Sci. USA 79, 4015-4019. 18. Long, E. O., Gross, N., Wake, C. R., Mach, J. P., Carrel, S., Accolla, R. & Mach, B. (1982) EMBO J. 1, 649-654. 19. Lathe, R. (1985) J. Mol. Biol. 183, 1-12. 20. LeMeur, M., Glanville, N., Mandel, J. L., Gerlinger, P., Palmiter, R. & Chambon, P. (1981) Cell 23, 561-571. 21. Hoeijsmakers, J. H., Borst, P., Van den Burg, J., Weissman, C. & Cross, G. A. M. (1980) Gene 8, 391-417. 22. Horwitz, K. B. (1981) J. Steroid Biochem. 15, 209-217. 23. Bailey, J. M. & Davidson, N. (1976) Anal. Biochem. 70, 75-85. 24. Alwine, J. C., Kemp, D. J., Parker, B. A., Reiser, J., Remart, J., Stark, G. R. & Wahl, G. M. (1979) Methods Enzymol. 68, 220-242. 25. Kozak, M. (1984) Nucleic Acids Res. 12, 857-872. 26. Miesfeld, R., Okret, S., Wilkstrom, A. C., Wrange, O., Gustafsson, J. A. & Yamamoto, K. R. (1984) Nature (London) 312, 779-781. 27. Yamamoto, T., Davis, C. G., Brown, M. S., Schneider, W. J., Casey, M. L., Goldstein, J. L. & Russell, D. W. (1984) Cell 39, 27-38. 28. McClelland, A., Kuhn, L. C., Ruddle, F. H. (1984) Cell 39, 267-274. 29. Mulvihill, E., LePennec, J. P. & Chambon, P. (1982) Cell 28, 621-632. 30. Payvar, F., Wrange, O., Carlstedt-Duke, J., Okret, S., Gustafsson, J. A. & Yamamoto, K. R. (1981) Proc. Natl. Acad. Sci. USA 78, 6628-6632. 31. Govindan, M. V., Spiess, E. & Majors, J. E. (1982) Proc. NatI. Acad. Sci. USA 79, 5157-5161. 32. Dean, D., Gope, R., Knoll, B. J., Riser, M. E. & O'Malley, B. W. (1984) J. Biol. Chem. 259, 9967-9970. 33. Von der Ahe, D., Janich, S., Scheidereit, C. & Beato, M. (1985) Nature (London) 313, 706-709. 34. Chambon, P., Dierich, A., Gaub, M. P., Jakowlev, S., Jongstra, J., Krust, A., LePennec, J. P., Oudet, P. & Reudelhuber, T. L. (1984) Recent Prog. Horm. Res. 40, 1-42. 35. Masiakowski, P., Breathnach, R., Block, J., Gannon, F., Krust, A. & Chambon, P. (1982) Nucleic Acids Res. 10, 7895-7903.