Proc. Nati. Acad. Sci. USAVol. 85, pp. 7119-7123, October 1988Biochemistry

Transient administration of estradiol-17.8 establishes anautoregulatory loop permanently inducing estrogen receptor mRNAMICHELLE C. BARTON AND DAVID J. SHAPIRODepartment of Biochemistry, University of Illinois, 1209 West California Street, Urbana, IL 61801

Communicated by Robert T. Schimke, June 13, 1988

ABSTRACT A single transient dose of estradiol-17fi issufficient to elicit the permanent induction of hepatic estrogenreceptor mRNA, which is induced 18-fold (from 0.13 to 2.4molecules per cell) and then remains fully induced for at least125 days. In primary liver cultures, extremely low concentra-tions of estradiol-17fi, which are below the Kd of the Xenopuslaevis estrogen receptor, maintain persistent induction ofestrogen receptor mRNA but not of estrogen-inducible vitel-logenin mRNA. These data and the ability of the antiestrogen,hydroxytamoxifen, to reverse persistent induction of estrogenreceptor mRNA, support a model in which transient doses ofestradiol-17fi induce the estrogen receptor and thereby estab-lish an autoregulatory loop. The low levels of estradiol-17Pnormally circulating in male X. laevis and the elevated level ofreceptor provide sufficient hormone-receptor complex to per-manently maintain the induced level of expression of theestrogen receptor gene. The permanent induction of the estro-gen receptor may be the regulatory switch that results in thepersistent expression of a recently identified class of proteinsthat exhibit long-term responses to estrogen.

A central problem in the molecular biology of development isthe question of how the expression of genes encoding generegulatory proteins is controlled. Variations in the levels ofthese regulatory proteins initiate a cascade of long-termchanges in the expression of genes subject to developmentalregulation and tissue-specific expression. The recent obser-vations that the retinoic acid receptor, which is likely to playa key role in vertebrate morphogenesis, and the transcriptionfactor SP1, are members of the superfamily of DNA-bindingproteins containing metal-binding fingers (1-4), reinforcesthe conclusion that this superfamily includes many eukary-otic gene regulatory proteins. Since other members of thissuperfamily mediate the effects of steroid hormones (5-15)and are oncogenes (14), they are currently generating intenseresearch interest. The steroid hormone receptors currentlyrepresent some of the best-understood models for nuclearreceptor proteins containing metal-binding fingers (reviewedin refs. 3, 5, 6, and 14). Although a great deal of progress hasbeen made in both the identification ofDNA sequences thatinteract with steroid hormone receptors and in the cloning ofhormone receptors, very little is known about control of theexpression of steroid hormone receptors and other nuclearreceptor genes (3, 5, 6, 14). We have exploited an experi-mentally tractable steroid hormone receptor system to studyone gene regulatory protein, the estrogen receptor. Ourstudies indicate that induction of the estrogen receptormRNA by its ligand estradiol-17,8 results in the establishmentof a stable, and probably permanent, regulatory loop. In-duced levels of the estrogen receptor mRNA and proteinmaintain this autoregulatory loop long after serum estradiollevels in male Xenopus laevis have returned to an extremelylow basal level. Our experiments show the very long-term

induction of estrogen receptor mRNA in vivo, reproduce thiseffect in primary Xenopus liver cultures maintained in levelsof estradiol that are below the Kd of the estrogen receptor,and use the antiestrogen hydroxytamoxifen to demonstratean in vivo requirement for residual estradiol-estrogen recep-tor complex in maintaining the autoinduction of estrogenreceptor mRNA. These studies, combined with detailedstudies showing the long-term induction of type II nuclearestrogen binding sites and apolipoproteins in human HepG2cells (16) and in avian liver (17), and of retinol-binding protein(RBP) in Xenopus liver (18), suggest that the administrationof estradiol results in long-term effects that persist far beyondthe period when the added hormone is present.

MATERIALS AND METHODSHormone Administration and Isolation of RNA Samples.

Male X. laevis (Xenopus I, Ann Arbor, MI) received injec-tions of 2 mg of estradiol-17P (dissolved in propylene glycol)on days 0, 2, 4, and 6 of the 14-day study of induction in vivo.Two animals were used for each time point. For the study ofthe long-term effects of estradiol in vivo, male X. laevisreceived a single 2-mg dose of estradiol-17P on day 0.Primary X. laevis liver fragment cultures were prepared

from untreated frogs essentially as described (19). Hormoneswere added to the culture medium in dimethyl sulfoxide. Atthe indicated times, total liver RNA was isolated byphenol/chloroform extraction (20). Poly(A)+ mRNA wasselected by batchwise adsorption to oligo(dT)-cellulose (Col-laborative Research, Waltham, MA) followed by elution (21).

Quantitation of X. laevis Estrogen Receptor (xER) mRNALevels. In quantitative RNA dot hybridizations, 5-1.l aliquotsof each sample containing 2 and 4 Ag of poly(A) RNA werespotted onto GeneScreen, and fixed to the membrane bycross-linking with ultraviolet light. The amount of poly(A)RNA in each spot was standardized by hybridization to a,f-actin cDNA clone. Actin mRNA levels are not regulated byestrogen (data not shown).[32P]DNA hybridization probes (specific radioactivity, 5-

10 x 108 cpm/4ug) were prepared by hexamer-primed syn-thesis (22) from the isolated insert of the xER cDNA clone,which encompasses the entire protein coding region of xERmRNA (13). Hybridizations were performed essentially asdescribed for genomic sequencing by Church and Gilbert (23)with the following modifications: the prehybridization periodwas 6-8 hr, the first wash contained 1% bovine serumalbumin, and a total of three washes were performed.The absolute amount of xER mRNA in each sample was

determined by comparison to an xER complementary RNA(cRNA) standard curve. A tritiated xER RNA standard wasprepared by T7 RNA polymerase transcription of the xERcDNA insert subcloned into the vector pGEM3 (PromegaBiotec, Madison, WI) (24). The tritiated cRNA standards andpoly(A) RNAs were spotted onto GeneScreen and hybridized

Abbreviations: xER, Xenopus laevis estrogen receptor; RBP, reti-nol-binding protein.

7119

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 18

, 202

1

7120 Biochemistry: Barton and Shapiro

to probes as described above. The spots were visualized byautoradiography. The autoradiograms were scanned by den-sitometry, and the spots were cut out and counted. Theresults obtained by scintillation counting of the spots and bydensitometry were repeated four or five times for each timepoint and were very similar. The absolute amount of xERmRNA in each sample was therefore calculated by averagingthe measurements. The number of molecules ofxER mRNAper cell was calculated from the absolute amount of xERmRNA per sample, as we have described for vitellogeninmRNA (20). Despite the extremely low levels ofxER mRNAin Xenopus liver, the accuracy and reproducibility of mea-surements made by this protocol are very high. For example,the level of xER mRNA in control liver was determined intwo independent hybridizations of two samples from differ-ent animals. One sample yielded values of 0.13 and 0.13molecules of xER mRNA per cell in the two hybridizations,and the other sample yielded values of 0.15 and 0.09 mole-cules per cell [mean, 0.125 ± 0.015 (SEM) molecules ofxERmRNA per cell].

RESULTS AND DISCUSSIONEstrogen Induction of xER mRNA in Vivo. In this study, we

used a xER cDNA clone (13) to investigate the regulation ofestrogen receptor gene expression. Several types of evidenceindicate that this clone encodes the xER. It encodes a proteinboth identical in size and exhibiting extremely high amino acidsequence similarity with the human estrogen receptor. In theDNA-binding domains ofthe two proteins, 82 of83 amino acidsare identical (13). The hormone-binding domains of the twoproteins contain 44 and 46 amino acid tracts that exhibit 100losequence identity (13). The Xenopus clone induces estrogen-dependent vitellogenin transcription upon transfection intoXenopus fibroblasts (T. C. Chang, D. Lew, and D.J.S., unpub-lished observations) and encodes a protein whose affinity forestradiol-17p is similar to that ofthe Xenopus estrogen receptor(data not shown; ref. 25). In RNA blot hybridizations ofXenopus poly(A) mRNA carried out under highly stringentconditions, both the intact cDNA clone and the hormone-binding domain (M. Hiken and D.J.S., unpublished observa-tions) ofthe xER hybridize to the same family offour estrogen-inducible mRNAs (ref. 13; data not shown). To provide aconvenient method of quantitating the levels of these fourextremely rare mRNAs, we carried out high stringency quan-titativeRNA dot hybridizations with the xERcDNA insert. Thebasal level ofxER mRNA in control Xenopus liver is only 0.13molecule per cell (1 molecule per 106 molecules ofmRNA). Thisis perhaps the lowest level ever measured for an mRNA in ahigher eukaryotic cell. In contrast to the transient down-regulation of the glucocorticoid (26) and progesterone (27)receptor mRNAs by their hormone ligands, estradiol-17,B elicitsan 18-fold induction of xER mRNA, which increases over a2-week period from -1 molecule for every 6 or 7 hepatocytesto >2 molecules per cell (Fig. 1). The time course of inductionof xER mRNA is similar to the time course of estrogeninduction of vitellogenin mRNA (28). These data demonstratethat a steroid hormone induces the mRNA encoding thereceptor to which it binds. While this I20-fold induction ofxERmRNA represents a large change in an extremely rare messagepopulation, the absolute number ofxER molecules per hepaticcell remains low. Even after induction to 2.4 molecules per cell,xER mRNA comprises only 0.002% of cellular mRNA.

Induction ofxER mRNA in Primary Cultures. To determinewhether the activation of xER gene expression was a directeffect of the hormone within the liver or resulted fromextrahepatic production of other hormones or factors, weinvestigated the effects of estradiol-17,B in primary Xenopusliver cultures. Two separate studies of the effects of estrogenon these cultures illustrated that the estradiol-17,8 induction

3

mn20

-

-5

zE

w

0

25

.120

z15 0U

z

100-J0

5

0

.

0

H-t0 2 4 6 8 10 12

DAYS AFTER FIRST E2 INJECTION14

FIG. 1. Estrogen induction of hepatic estrogen receptor mRNA.Poly(A) + RNA was isolated from the livers of control and estrogen-treated male X. laevis on the indicated days after estrogen (E2)stimulation and analyzed by quantitative RNA dot hybridization.Fold induction represents the increase in xER mRNA contentrelative to the uninduced day 0 control sample.

of estrogen receptor mRNA is a direct response that isreproducible in culture (Fig. 2). The similarity of the induc-tion in vivo and in liver cultures suggests that estrogen exertsa direct effect on hepatic xER gene expression. These studiesalso illustrated a second major difference between the regu-lation of estrogen receptor mRNA levels and those ofglucocorticoid receptor. Down-regulation of glucocorticoidreceptor mRNA by dexamethasone is a transient response incultured cells. Glucocorticoid receptor mRNA levels returnto a basal level within 3 days after addition of dexamethasoneto the culture medium (26). The induction of xER mRNA byestradiol continues for 12-14 days both in vivo (Fig. 1) and inprimary cultures (Fig. 2). The extremely low levels of xERmRNA make it impossible for us to definitively state whetherxER mRNA accumulation is nearly linear or exhibits amodest increase in rate late in induction when the level ofxER is increased.

zU

U)

00E

z

E

w'C

3 6 9TIME IN ESTROGEN (days)

z0

(.)

0z

0-J0UL.

FIG. 2. Estrogen induction of xER mRNA in primary livercultures. Xenopus liver cube cultures, prepared as described (19),were continuously maintained in medium containing 1 ,uM estradiol-17,8. At the indicated times, RNA was isolated and xER mRNAlevels were determined. The data are from two separate experiments.

Proc. Natl. Acad. Sci. USA 85 (1988)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 18

, 202

1

Proc. Natl. Acad. Sci. USA 85 (1988) 7121

Long-Term Stimulation of xER mRNA in Vivo. To examinelong-term effects of estrogen on xER mRNA levels, we ana-lyzed the RNA isolated from male animals that had received asingle injection ofestrogen up to 125 days earlier. The surprisingfinding was that hepatic estrogen receptor mRNA remains at afully induced level for at least 4 months after a single injectionof estrogen (Fig. 3). The persistent induction ofxER mRNA isnot due to residual estradiol from the injection, since estrogenlevels in serum return to basal levels within 1 day after injection(29). Several years ago, we used hormone-binding studies todemonstrate that estradiol-17,( evoked a long-term increase inhigh-affinity estrogen binding sites (25). The significance ofthose data was obscure, as long-term effects of steroid hor-mones had not been proposed.We propose that a transient dose of estradiol, which elicits

the long-term, and probably permanent, induction of estro-gen receptor mRNA and protein (25), establishes an auto-regulatory loop resulting in the essentially permanent induc-tion of estrogen receptor mRNA. Since the estrogen receptorprotein induces the synthesis of its cognate mRNA, anelevated level of xER mRNA could be maintained indefi-nitely by the residual level of circulating estradiol in the maleanimal. The observation that the circulating estradiol level incontrol male Xenopus serum (0.2 nM) is sufficient to load=25% of the receptor is consistent with this model (29).Although only a fraction of the estrogen receptor will beloaded with hormone, the elevated receptor level providesenough estradiol-estrogen receptor complex to maintain theinduced rate of estrogen receptor gene expression. Theobservation that only 10-20% of cellular estrogen receptormay be required for some hormone responses is supported bystudies that showed estrogen induction of uterine growthrequired only a small fraction of the cell's complement ofestrogen receptor (30).

This model predicts that a low level of residual hormone-receptor complex plays a central role in maintaining persis-tent activation ofestrogen receptor gene expression. To more

3.0

2.5

z

It:E

Lrx

cf

To

2.*H

.0-

0.51-

I

//

0 14 40 46 60 125 60 4dHT 60 7dHT

Days After Estradiol Injection

FIG. 3. Persistent estrogen induction ofxERmRNA is dependenton hormone-receptor complex. Hepatic estrogen receptor mRNAlevels were determined at the indicated times after male X. laevisreceived a single 2-mg dose of estradiol-17,8 on day 0. The quantityof xER mRNA was determined as described. Hepatic xER mRNAlevels in animals that also received injections of the antiestrogenhydroxytamoxifen are indicated by cross-hatched bars. One milli-gram of hydroxytamoxifen (HT) (in dimethyl sulfoxide) was admin-istered either once at day 60 and the mRNA levels were determined4 days later (60:4d HT) or twice 7 days and 4 days before sacrifice(60:7d HT). Hydroxytamoxifen in dimethyl sulfoxide is rapidly lostfrom the organism. Two injections are therefore required to obtainhigh levels of hydroxytamoxifen in vivo.

directly evaluate the role of the estrogen receptor in long-term induction, we used the antiestrogen, hydroxytamoxifen,a competitive inhibitor of the binding of estradiol-17P3 to theestrogen receptor (31-33). Hydroxytamoxifen was adminis-tered during the final week of a 60-day "withdrawal" periodto whole animals given the single dose of estradiol used toestablish persistent induction. Hydroxytamoxifen reducesthe level of xER mRNA (Fig. 3, 60:7d HT sample) in thepresence of inhibitor-saturated receptor protein. We havealso recently shown that hydroxytamoxifen blocks the per-sistent estrogen induction of serum RBP mRNA in Xenopusliver (18). These data establish a requirement for estradiol-estrogen receptor complex in the maintenance of persistentinduction and provide direct support for a key prediction ofour model.

Subsaturating Concentrations of Estradiol Maintain FullyInduced Levels ofxER mRNA. The estradiol required to formhormone-receptor complexes during withdrawal must bederived from the extremely low concentration of estradiol-17, circulating in the serum ofmale X. laevis. To directly testthis idea, we used our primary culture system to establish adose-response curve for persistent estrogen induction ofxER mRNA (Fig. 4). Vitellogenin mRNA induction providesan appropriate control for these experiments, as detailedstudies have unambiguously shown that induction of vitello-genin gene transcription and its mRNA requires the contin-uous presence of exogenous estradiol (19). An initial inducedlevel of xER and vitellogenin mRNAs was established bymaintaining the cultures in medium containing a saturatinglevel of estradiol (1 AM) for 1 week. Separate cultures werethen maintained for an additional 5 days in medium contain-ing various concentrations of hormone including 0.5 nM, theKd of the estrogen receptor, and 0.2 nM, the concentration ofserum estradiol in control male Xenopus. All levels of estra-diol added to the medium, even 0.2 nM, continued to inducexER mRNA to a level similar to that observed for thesaturating 1 ,uM level. When no estradiol was added to themedium after the initial induction, the continued induction ofxER mRNA was halted but not reversed. We observed nodecline to the basal level during this 5-day period even in theabsence of exogenous estrogen.Comparison of the levels of xER and vitellogenin mRNA

in these samples illustrates the dramatic differences that thevitellogenin and estrogen receptor genes display upon expo-sure to the same levels of hormone. No level of hormonebelow the full 1 AM saturating level is capable of continuingthe induction of vitellogenin mRNA. The level of vitellogeninmRNA declines in a dose-dependent manner when lowerconcentrations of hormone are present (Fig. 4). Since vitel-logenin transcription stops soon after estrogen is removedfrom the culture medium (19), the dose-dependent decline invitellogenin mRNA levels is likely to result from dose-dependent differences in the time course with which vitello-genin mRNA shifts from the estrogen-stabilized (ty2, 500 hr)to the destabilized (ty2, 16 hr) state (19). Clearly, there are vastdifferences in the threshold levels of hormone needed forregulation of these two estrogen-responsive genes. Maintain-ing vitellogenin gene expression requires a concentration ofestradiol-17f3 >1000 times higher than is required for estro-gen receptor gene expression.Other Genes Exhibit Long-Term Responses to Estradiol.

Our in vivo studies on the estrogen-inducible RBP mRNA(18) indicate that this gene also displays a persistent inductionby estradiol. Run-on transcription studies demonstrate thatthe persistent estrogen induction of RBP mRNA in Xenopusliver is due to a specific long-term increase in the rate ofRBPgene transcription (18). The extremely low level of xERmRNA, which is present in liver cells at a level lower than thelevel of RBP (21) and very low density apolipoprotein IImRNAs by a factor of >1000 (17), precludes direct measure-

Biochemistry: Barton and Shapiro

1-.0-

Dow

nloa

ded

by g

uest

on

Nov

embe

r 18

, 202

1

7122 Biochemistry: Barton and Shapiro

z0

to

N

U)0

.0

E

zcrE'Irw'C

6

4

2

o [--Days 0

Estradiol (nM) 0

Proc. Natl. Acad. Sci. USA 85 (1988)

'060 _

0

0CD

0

-40

zE

* 20 e=,CP20- ~~0'

1000 1000 1.0 0.5 0.2 0

FIG. 4. Trace concentrations ofestradiol maintain full induction ofxER raRNA. The levels ofxER mRNA (solid bars) and vitellogenin mRNA(hatched bars) were determined in cultures of liver cubes maintained for 1 week in medium containing 1 ,uM estradiol-17f3 (days 0-*7). Separateportions ofthe cultures were then maintained in medium containing the indicated concentrations of estradiol for 5 days (days 7-12). Vitellogeninand xER mRNA concentrations were then determined as described. The induced level of vitellogenin mRNA (-50,000 molecules per cell) issimilar to the maximum level we reported previously (34).

ments of the contributions of gene transcription and mRNAstability to long-term induction by currently available meth-ods. However, the observation that persistent estrogeninduction of RBP mRNA in the same cells is caused by aspecific increase in the rate of RBP gene transcriptionsuggests a similar mechanism is responsible for long-termeffects on estrogen receptor gene expression.The phenomenon of long-term induction of proteins by

estrogen is not restricted to Xenopus liver. In a series ofstudies, Deeley and coworkers have established that estrogenelicits a long-term induction of apolipoproteins in culturedhuman cells and in avian liver (16, 17). In contrast to Xenopusliver, they find that the avian liver high-affinity type I nuclearestrogen receptor is only transiently induced by estrogen(17). Since they observe the long-term estrogen induction ofa lower affinity type II estrogen-binding protein, they pro-pose that this protein mediates the persistent effects ofestrogen by serving as a constitutive estrogen-independentactivator of gene transcription (17). Three lines of evidenceleads us to conclude that the persistent effects of estradiol ongene expression in our system are directly mediated by thewell-characterized high-affinity estrogen receptor, and not bya secondary effector. (i) We find that protein synthesis is notrequired for the initial induction of vitellogenin (35) and RBPgene transcription (21). (ii) Our earlier hormone bindingstudies established that the induced estradiol-17,8 bindingprotein, present 70 days after estrogen stimulation, was thehigh-affinity estrogen receptor with the same Kd for estradiolas the estrogen receptor in control male liver cells (25). (iii)Each member of the family of xER mRNAs is estrogeninducible and hybridizes at high stringency to our cDNAclone of the "classical" high-affinity estrogen receptor (ref.13; data not shown).Two Classes of Estrogen-Regulated Genes Exist. The strik-

ing differences between the effects of estrogen on genes suchas vitellogenin and on the estrogen receptor and RBP geneslead us to propose the existence of two classes of hormone-regulated genes and the model shown in Fig. 5. The mostwidely studied class of hormone-regulated genes, typified bythe vitellogenin genes, requires the presence of exogenoushormone and a high level of loaded hormone-receptorcomplex to induce and maintain their transcription. A second

class of target genes exemplified by RBP and xER exhibitslong-term induction by estradiol. The seminal event thatallows these genes to exhibit persistent estrogen induction isthe establishment of an autoregulatory loop that maintainsxERmRNA levels via a combination ofhighlevels ofinducedestrogen receptor protein and residual estrogen circulatingeven in the serum of male animals (25, 29). This additionalclass of hormone-regulated genes presumably exhibits alower threshold response to loaded hormone-receptor com-plex. The very different threshold levels exhibited for estro-gen-regulated transcription by these two classes of genessuggests their promoters may contain different hormone-responsive elements. Further studies are necessary to deter-mine whether the estrogen receptor gene contains a singleclass of estrogen response element in its 5' flanking region ormultiple copies ofthe same estrogen response elements foundin the vitellogenin genes.A detailed understanding of the apparent differences in

threshold levels of hormone receptor complex required fortranscription of the two classes of genes requires additionalexperimentation. Our simplified schematic model does notconsider the potential contribution of negative regulation tocontrol of the short-term response class of genes such asvitellogenin. The possibility that unloaded native estrogenreceptor binds to the estrogen response element without acti-vating transcription is raised by the recent estrogen receptordeletion studies of Kumar et al. (36). This unoccupied receptorcould act as a negative regulator, competing with the estradiol-estrogen receptor complex for binding to the response elementsof short-term regulated genes, and making the ratio of unoccu-pied to occupied receptor critical for the activation of thesegenes. The possibility that activating the transcription of somegenes requires concentration-dependent association of recep-tors to form estrogen receptor dimers or multimers also must beconsidered in evaluating threshold levels of receptors requiredfor induction.The demonstration of the long-term, and probably perma-

nent, induction of a gene regulatory protein provides aninteresting and experimentally accessible model system forthe study of irreversible switch systems in development. Oneswitch system in which the autoregulation of estrogen recep-tor gene expression might play an important role is related to

Dow

nloa

ded

by g

uest

on

Nov

embe

r 18

, 202

1

Proc. Natl. Acad. Sci. USA 85 (1988) 7123

A. SHORT TERM INDUCTION--HIGH E21HIGH XER

x} X mRT[NA \

- f"I

IMJ iT1M171NA

ViI

ONOI 1Id-1 INP uf;l ION IOWt , 11( Xlx

XIxl I51NA

xL Pk

V

R P

V II

FIG. 5. Schematic representation of the autoregulation of estro-gen receptor gene expression and the permanent induction of xERand RBP gene transcription. (A) High level of exogenous estradiol-17(8 (v) and the fully loaded receptor induce transcription ofthe xER,vitellogenin (VIT), and RBP genes. (B) Persistent induction. Theinduced level of xER and the residual level of estradiol provideenough hormone-receptor complex to permanently maintain xERgene transcription and an elevated level ofxER protein. This level ofhormone-receptor complex is sufficient for long-term induction ofRBP gene transcription but is below the threshold level required forvitellogenin gene transcription.

sex determination in Xenopus. When Xenopus tadpoles areincubated in water containing estradiol-17P3 during a briefdevelopmental window, complete conversion of males tofemales is observed, and 100o of the resulting adults arefertile females (37, 38). The recent demonstrations in anotherdevelopmental system that the fushi tarazu gene, whichfunctions as a modulator of Drosophila melanogaster devel-opment, regulates its own expression (39) and that theDrosophila engrailed gene product binds to its own 5'flanking region (40) suggest that autoregulatory loops mayprovide a common mechanism by which cells control theexpression of key regulatory protein genes and therebyprovide long-term modulation of the expression of large setsof target genes.

We thank B. S. Katzenellenbogen for helpful comments on themanuscript and W. Soo Hoo for artwork. This work was supportedby National Institutes of Health Grant HD 16720. M.C.B. is aNational Institutes of Health predoctoral trainee.

1. Petkovich, M., Brand, N. J., Krust, A. & Chambon, P. (1987)Nature (London) 330, 444-450.

2. Giguere, V., Ong, E. S., Segui, P. & Evans, R. M. (1988)Nature (London) 331, 91-94.

3. Robertson, M. (1987) Nature (London) 330, 420-421.

4. Kadonaga, J. T., Carner, K. R., Masiarz, F. R. & Tjian, R.(1987) Cell 51, 1079-1090.

5. Yamamoto, K. R. (1985) Annu. Rev. Genet. 19, 209-252.6. Ringold, G. M. (1985) Annu. Rev. Pharmacol. Toxicol. 25, 529-

566.7. Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G.,

Oro, A., Lebo, R., Thompson, E. B., Rosenfeld, M. G. &Evans, R. M. (1985) Nature (London) 318, 635-641.

8. Miesfeld, R., Okret, S., Wikstrom, A.-C., Wrange, O., Gus-tafsson, J.-A. & Yamamoto, K. R. (1984) Nature (London) 312,779-781.

9. Greene, G. L., Gilna, P., Waterfield, M., Baker, A., Hort, U.& Shine, J. (1986) Science 231, 1150-1154.

10. Green, S., Walter, P., Kumar, V., Krust, A., Bornert, J. M.,Argos, P. & Chambon, P. (1986) Nature (London) 320,134-139.

11. Connelly, 0. M., Sullivan, W. P., Toft, D. O., Birnbaumer,M., Cook, R. G., Maxwell, B. L., Zarucki-Schulz, T., Greene,G. L., Schrader, W. T. & O'Malley, B. W. (1986) Science 233,693-716.

12. Danielsen, M., Northrop, J. P. & Ringold, G. M. (1986) EMBOJ. 5, 2513-2522.

13. Weiler, I. J., Lew, D. & Shapiro, D. J. (1987) Mol. Endocrinol.1, 355-362.

14. Green, S. & Chambon, P. (1986) Nature (London) 324, 615-617.

15. Klein-Hitpass, L., Schorpp, M., Wagner, U. & Ryffel, G. U.(1986) Cell 46, 1053-1061.

16. Tam, S.-P., Hache, R. J. G. & Deeley, R. G. (1986) Science234, 1234-1237.

17. Hache, R. J. G., Tam, S.-P., Cochrane, A., Nesheim, M. &Deeley, R. G. (1987) Mol. Cell. Biol. 7, 3538-3547.

18. McKearin, D. M. & Shapiro, D. J. (1988) J. Biol. Chem. 263,3261-3265.

19. Brock, M. L. & Shapiro, D. J. (1983) Cell 34, 207-214.20. Baker, H. J. & Shapiro, D. J. (1978) J. Biol. Chem. 253, 4521-

4524.21. McKearin, D. M., Barton, M. C., Keller, M. J. & Shapiro,

D. J. (1987) J. Biol. Chem. 262, 4939-4942.22. Feinberg, A. P. & Vogelstein, B. (1984) Anal. Biochem. 137,

266-267.23. Church, G. & Gilbert, W. (1984) Proc. NatI. Acad. Sci. USA 81,

1991-1995.24. Melton, D. A., Kreig, P. A., Rebagliati, M. R., Maniatis, T.,

Zinn, K. & Green, M. R. (1984) Nucleic Acids Res. 12, 7035-7056.

25. Hayward, M. A., Mitchell, T. A. & Shapiro, D. J. (1980) J.Biol. Chem. 255, 11308-11312.

26. Okret, S., Poellinger, L., Dong, Y. & Gustafsson, J.-A. (1986)Proc. Natl. Acad. Sci. USA 83, 5899-5903.

27. Read, L. D., Snider, C. E., Miller, J. S. & Greene, G. L. (1988)Mol. Endocrinol. 3, 263-271.

28. Shapiro, D. J., Baker, H. J. & Stitt, D. T. (1976) J. Biol. Chem.251, 3105-3111.

29. Wright, C. V. E., Wright, S. C. & Knowland, J. (1983) EMBOJ. 2, 973-977.

30. Clark, J. H. & Peck, E. J., Jr. (1976) Nature (London) 260,635-637.

31. Katzenellenbogen, B. S., Norman, J. J., Eckert, R. L., Peltz,S. W. & Mangel, W. F. (1984) Cancer Res. 44, 112-119.

32. Jordan, V. C., Collins, M. M., Rowsby, L. & Prestwich, G.(1977) J. Endocrinol. 75, 305-316.

33. Riegel, A. T., Aitken, S. C., Martin, M. B. & Schoenberg,D. R. (1987) Endocrinology 120, 1283-1290.

34. Brock, M. L. & Shapiro, D. J. (1983) J. Biol. Chem. 258,5449-5455.

35. Hayward, M. A., Brock, M. L. & Shapiro, D. J. (1982) NucleicAcids Res. 10, 8273-8284.

36. Kumar, V., Green, S., Stack, G., Meera, B., Jin, J.-J. &Chambon, P. (1987) Cell 51, 941-951.

37. Chang, C. Y. & Witschi, E. (1956) Proc. Soc. Exp. Biol. Med.93, 140-144.

38. Gallien, L. (1948) C.R. Hebd. Seances Acad. Sci. Paris 241,998-1000.

39. Hiromi, Y. & Gehring, W. J. (1987) Cell 50, %3-974.40. Desplan, C., Theis, J. H. & O'Farrell, P. H. (1985) Nature

(London) 318, 630-635.

VI mRNA

Biochemistry: Barton and Shapiro

Dow

nloa

ded

by g

uest

on

Nov

embe

r 18

, 202

1