Vol.
186,
July
31, 1992
No.
BIOCHEMICAL
2, 1992
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS Pages
PURIFIED
ESTROGEN
RECEPTOR TRANSCRIPTION
ENHANCES
803-810
IN VITRO
Vincenzo Nigro, Anna Maria Molinari, Ignazio Armetta, Antonietta de Falco, Ciro Abbondanza, Nicola Medici, and Giovanni Alfred0 Puca Istituto di Patologia Generale e Oncologia I Facolta di Medicina e Chirurgia S. Andrea delle Dame, 2. 80138 Naples, Italy Received
June
10,
1992
An in vitro transcription system was developed to investigate the mechanismsof gene regulation by the estrogenreceptor (ER). ER purified from calf uterus was highly active in enhancingRNA transcription from a template DNA containing estrogenresponseelements (EREs) upstreamfrom a minimal promoter. Under the conditions employed, no addition of tissuespecific factors was required and both estrogenor antiestrogenswere ineffective. The stimulation of transcription correlated with the copy number of EREs in the template. The addition of competitor ERE oligonucleotides specifically inhibited the ER-induced transcription. We suggestthat the ER may be involved in the formation of the stableinitiation complex. 0 1992Academrc PTfSS,1°C.
Eukaryotic classII promotersgenerally contain a TATA box and regulatory sequences.RNA polymeraseII forms a transcription complex with the generalinitiation factors (TFIIA, B, D, E, F), at the promoter start site (1, 2). Complex assemblybeginswith the stable binding of TFIID to the TATA box. This promotesthe subsequentordered recruitment of the general initiation factors and RNA polymeraseII resultingin a template-committedcomplex (3). The regulatory DNA sequencesare bound by specific transcription activating proteins that respond to regulatory cues. These activators generally contain distinct domains for DNA binding and transcriptionalactivation (4). Steroid receptorsbelong to a large superfamily of ligand-inducible transcription factors with an additional regulatory hormonebinding domain.Transfectionexperimentswith chimeric or truncated steroid receptorshasprovided muchinformation on their function and structure (for review 5, 6). Multiple transcription activation domains were found (7); hence, steroid receptors may stimulate different stagesin the transcription initiation and/or interact with different components of the transcriptional machinery. Understanding the mechanismby which steroid receptors regulate gene expression requires the development of cell-free transcription systemsthat respond to exogenousaddition of receptors. The availability of functional purified moleculesenablethe analysisof the direct effect of the receptoraddition on the transcription machinery. Purified steroid receptors can be modified, components of Abbreviations: AdML, adenovirusmajor late; EMSA, electrophoreticmobility shift assay; ER, estrogenreceptor; ERE, estrogenresponseelement; E2, estradiol-178
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCtI
COMMUNICATIONS
nuclear extracts and/or tissue specific factors can be isolated and their roles and interactions defined. However, the purification of large amount of steroid receptors from target tissue suitable for in vitro transcription studies may present cumbersome difficulties. Conventional procedures damage the molecules and/or dilute the samples, which are then unable to induce in vitro transcription. Furthermore, the receptor elution from ligand-affinity chromatography by steroid analogues hampers the evaluation of the hormone effect. Until now, only the progesterone receptor from chicken oviduct (8) or from rabbit uterus (9) was successful purified as functional molecules. The group of 0’ Malley demonstrated that the stimulation of cell-free transcription occurred through the enhancement of the formation of a stable preinitiation complex. Interestingly, partially purified hormone-free progesterone receptor activated by salt treatment induced transcription in a hormone-independent manner (8). In the attempt to overcome the purification problems for the other members of the steroid receptors family some Authors overexpressed the cDNA in heterologous systems (10). Bacterial or yeast systems yielded poorly functioning molecules, while recombinant human glucocorticoid (11) and mouse estrogen receptors (12) from baculovirus were shown to stimulate correct RNA synthesis in vitro in a way identical to the purified progesterone receptor and from an equivalent promoter construct. However, dissimilar posttranslational modifications have to be taken into account (13). Here we report the successful development of an irz vitro transcription system greatly stimulated by the ER. This ER is a hormone-free molecule highly purified from calf uterus by immunoaffinity that retains full functional properties of ERE binding and high affinity estradiol binding. Our findings suggest that there is a direct ER involvement in the enhancement of the transcription initiation from a template DNA of minimal structure with no tissue specific factor requirement. The stimulation of transcription is completely hormone independent, confirming previous observations with recombinant molecules.
MATERIALS
AND METHODS
Plasmids. pML(CzAT)19 was a generous gift of Dr. R. Cl. Roeder (14). pMLllO(C2AT) was obtained by exonuclease III deletion from the BarnHi site, after Ban&I and PstI digestion of pML(C2ATj19 and contains the AdML promoter linked to a shortened 1 IO-bp Gfree cassette. pMLTATA:O was constructed by recovering the Mspl-Mspl fragment from pML(C2AT) 19, which contains the AdML TATA box and the complete G-free cassette. The fragment was cloned into the AccI site of GEM 3 @omega), oriented with the EcoRI site upstream from the TATA box. pMLTATA: lE, pMLTATA:2E, and pMLTATA:3E derived from pMLTATA:O with the insertion of 1, 2, or 3 copies, respectively, of the estrogen response element of the Xenopus laevis A2 vitellogenin gene. Purification of ER. The hormone binding subunit of ER was purified as previously described (15) except that : i) the immunoaftinity chromatography was performed with AER 311 monoclonal antibody prepared in our laboratory (to be described elsewhere); ii) the immunoaffinity chromatography eluates were buffered with 50 mM Tris-Cl (pH 7.9), containing 40mM NaCl, 3mM dithiothreitol (DTT), 5% (v/v) glycerol and lOOpg/ml bovine serum albumin @omega). The purity of the ER was assessed each time by SDSPAGE and silver staining. The hormone binding capacity of each purified batch of ER was assayed. A control purification was performed as described (15). The purified ER was seen as a main 67 kDa and smaller silver stained bands, that were also affinity-labeled with 3Htamoxifen-aziridine (15). Western-blot analysis of the purified ER showed that all the silver stained SDS-PAGE bands were recognised by the monoclonal antibody AER 310 (not shown). The electrophoretic mobility shift assay (EMSA) showed that the ER was the only ERE binding protein, since the monoclonal antibody JS34/32 further retarded all the proteinoligonucleotide complexes (16, 17) Experimental design. A HeLa cell nuclear extract was employed as the source of basal transcription factors and RNA polymerase II. This cell line does not contain ER and its extracts are relatively nuclease-free. The test template contained the TATA box of Adenovirus 804
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
major late promoter (AdML) and the G-free cassette of 377 bp as reporter (14). AdML promoter provides a useful model system for in vitro analysis of specific transcription by RNA polymerase II. This promoter has a well known structure, with only two essential DNA elements: a TATA box and a single upstream element (USF binding site). We replaced the upstream element with sequences containing ERE(s). Two copies of the Xenopus luevis A2 vitellogenin ERE were inserted at about 90 bp and 140 bp, upstream from the transcription initiation point. A plasmid containing the AdML promoter linked to a shortened (110 bp) Gfree cassette was included in the assay as an internal control. Correct initiation at the test promoter yields transcripts of 387 nucleotides in length, while eventual read-through transcripts are longer (392 nt). In vitro transcription assays. HeLa S3 nuclear extracts were prepared according to Shapiro D.J. ef al. (18). Typical in vitro transcription reactions contained (30 ~1volume): 20 mM HEPES-NaOH (pH 7.9), 5% (v/v) glycerol, 35 mM NaCl, 8 mM MgC12, 2.5 mM DTT, 10 mM creatine phosphate,250 uM 3’-O-methyl GTP, 0.5 mM ATP, 0.5 mM CTP, 10 PM UTP, 4 l&i of [alphaJ*P] UTP, 12 U of RNase Tl, 20 U RNasin (Promega), 6 pg of sonicated salmon spermDNA, 330 ng of pML1 lO(C2AT) template and 225 ng of test template @MLTATA:2E, unlessotherwisespecified),3-5~1 (50-70 pg) of HeLa extract and, where indicated, 0.7 pmoles of purified ER, preincubated at 4°C for 45 min with 5 0 nM 17l3-estradiolprior to addition in the transcription assay.Salmon spermDNA carrier in the preincubation mix was shownto improve the transcriptrecovery and to repressmainly basal transcription. Preincubation step was performed in the absence of MgC12 and nucleotidesfor 30 min at 28°C. Transcriptionstartedby the addition of theselast components and the reaction continued for 30 min at 3O’C. Reactionswere terminatedwith the addition of 70 1.11 of 25 mM Tris-Cl pH7.5, 0.5% SDS, 5 mM EDTA, 100ug/ml of Ecoli transfer RNA, 400 pg/ml proteinaseK and incubatedat 45°C for 45 min. Then, 200 yl 8 M urea in 10 mM Tris-Cl pH 7.5, 5 mM EDTA was addedand the sampleswere extracted once with phenol-chloroform and ethanol precipitated. The pellet was suspendedin 80% formamide, denaturedat 86°C for 5 min and applied on a 8 M urea-6% polyacrylamide gel, run at 45 W (constant power) and visualised by autoradiography.Typical exposurewas 16hours with an intensifying screen.
RESULTS
Purified
ER stimulates in vitro transcription. Figure 1 shows the stimulation of accurate RNA transcription by purified ER (upper arrow). Little if any longer read-through transcripts from the test templatewere observed,as alsoconfirmed by higher-resolution gel electrophoresis(not shown). As expected, ER did not significantly alter the synthesisof the shorter 120 nt RNA from the control templatepML1 lO(C?AT) (lower arrow). The addition of 0.5 pg/ml a-amanitin completely abolishedthe transcription from both templates(lanes7 and 8), thus demonstratingthat the transcriptswere entirely due to polymeraseII. Under the conditions employed, the presenceof estradiol (lanes5 and 6) or the estrogenantagonists, ICI 164,384or 4-hydroxytamoxifen (data not shown),did not modify the fold stimulation by ER. The addition of an increasing amount of ER resulted in the enhancementof expected transcript synthesis, but estradiol had no effect even with lower ER concentrations (not shown). ERE is required for ER transcription enhancement. To verify the requirement of EREs for ER induction, we usedreporter plasmid containing 0, 1, 2, or 3 copies of EREs. As shown in Fig. 2, the addition of ER was totally unable to increaseRNA synthesisfrom a reporter plasmid lacking an ERE (lanes l-4). A dramatic increasewas already seenwith 1 ERE in the reporter plasmid (pMLTATA: 1E, lanes5-8). Two copiesof EREs provided the best result of ER-dependent induction, since the basallevel of transcription was reduced. Unexpectedly, RNA transcription from the test template with 3 ERE was heavily reduced (lanes 13-16). However, longer autoradiographic exposureof the gel displayed a high fold stimulation by ER from this template, comparableto that observedwith pMLTATA:2E. The reasonfor the lower basalactivity of the template with multiple EREs remainsunknown. It 80.5
Vol.
186,
No.
ER E2
2, 1992
-
BIOCHEMICAL
+ -
+ +
AND
BIOPHYSICAL
cy aman. - + +
0 ERE
RESEARCH
1 ERE
COMMUNICATIONS
2 ERE
3 ERE
ER’-T.
1 2 3 4 5 6 7 8 9 10111213141516
Figure 1. Effect of the ER on in vitro transcription. Elution buffer (lanes 1, 2) or 1 pmol of highly-purified ER, in the absence (lane 3, 4) or presence of 50 nM estradiol (lanes 5-6). were added in the assay. Samples in lanes 7-8 were identical to 5-6, except that a-amanitin was included. In vifro transcription was performed as described under Materials and Methods. Grey and black arrows indicate the correcdy initiated transcripts from the pML1 lO(C2AT) and pMLTATA:ZE, respectively. Fiwre 2. Dependence of ER-induced transcription from the copy number of EREs in the template. In vitro transcriptionwasperformedunderthe standardconditions with 225 ng of templatescontaining no copies (pMLTATA:O, lane l-4), one copy (pMLTATA:lE, lane 5-8), two copies(pMLTATA:2E, lanes9-12) and three copies @MLTATA:3E, lanes13-16)of ERE.
may be that repeatedpalindromicsequences weakenthe templatetranscriptioneither through a modified DNA conformation, or by drawing HeLa proteins that interfere with the process. EMSA of ERE by HeLa extract did in fact showmultiple retardedbands(data not shown). If the effect of ER is dependenton its binding with the ERE in the test template, the addition of ERE oligonucleotide excess should squelch the ER, preventing its stimulatory effect. Competitor ERE oligonucleotideswere thus addedto the transcription assay.Starting from a 10 to 20-fold molar excess of ERE, a decreaseof the receptor-stimulated transcripts was detected (Fig. 3, lanes 5-6). At 40-fold molar excessthe ER-dependent transcription was reducedto the basallevel. At higher competitor concentrations(lane 8 and data not shown), a consistentinhibition of both the testtemplateand the internal control transcription occurred. ER functions in the preinitiation step. The recruitment of preinitiation complexes iir vitro is a slow step preceding the transcription initiation. Many Authors have described the 806
Vol.
186,
No.
2, 1992
BIOCHEMICAL
(molar
ERE ratio) ER
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
0 0 0 0 10 204080 - -++++++
12345678
Figure 3. Specific competition of ER-induced transcription by ERE oligonucleotides. Increasing amounts of oligonucleotides containing the ERE from the Xenopus vitellogenin A2 gene were added to the assay prior to the addition of the ER and HeLa extract. The molar ratios of competitor ERE vs template ERE are indicated. Transcriptionwasperformedunderthe standardconditions,exceptthat pMLTATA:lE as inducible template DNA was employed.
importance of such a step to maximise the RNA synthesis, which promptly starts from preformed complexes,when nucleotidetriphosphatesare added(19,20). To elucidate whether ER is necessaryfor template-committedpreinitiation complexes, we preincubatedthe test template with the HeLa extract and the ER (Fig.4, lanes l-4). A parallel preincubation of the test template with HeLa extract alone wascarried out, while the ER was addedlater with the nucleotides(lanes5-8). It appearsevident that, when the ER was absent during the preincubation step, the stimulatory effect wasmuch weaker (Fig. 4, comparelanes l-4 with 5-8). When the preincubation step was omitted, a general decrease of RNA transcripts was noticed (Fig. 4, lanes 9-12) without disappearance of ER-dependent transcription. Finally, an experiment presentedwas performed to define the requirement of the general transcription factor in the preincubationstep.If comparedto the standardconditions, the ERinduced transcription was lessefficient when the preincubation step was carried out in the absenceof HeLa extract and the HeLa extract addedtogether with nucleotides(not shown). This suggeststhat the ER is required in the early step, not only to bind to its target sequence, but also to recruit the factors involved in the formation of the stableinitiation complex. In this way, both ER and HeLa cell extract are essentialto obtain maximal activity.
DISCUSSION In the present paper, we characterise an in vitro transcription system, stimulated by the highly-purified estrogen receptor. The receptor induced faithful RNA transcription 807
Vol.
186, No. 2, 1992
BIOCHEMICAL
A
r ER
AND BIOPHYSICAL
-
B +
-
RESEARCH COMMUNICATIONS
C +
-
I +
1 2 3 4 5 6 7 8 9101112
DENRA
A
B
NTPs, MgCll 4
HeLa I.
preincubation
transcription
b
ER NTPs, MgClz 4
DNA HeLa ,
preincubation
c
’
transcription
HeLa NTPs, MgCI, I transcription
-
*
Figure 4. ER-dependent RNA synthesis with or without preincubation. Lanes l-4 representthe standardassay,performedasin scheme A. In lanes5-8 theER wasomitted from the preincubationmix and addedtogetherwith NTPs and MgC12(B). Lanes9-12 indicateassays performedwithoutpreincubation (C).
specifically from ERE(s)-containing template DNA without the requirement for tissue- or species-specific factors and/or promoters. Similarly to what is known about other activating proteins (21-23, it is likely that ER enhancesthe transcription interacting with the basal factors involved in the preinitiation complex. The ER was quite ineffective when added after the preincubation step and a decreasein fold stimulation was observedwhen the HeLa extract was added subsequently. Hence, the ER may be required during the preincubation step. It may be involved in the recruitment or modification of a basal factor essential for transcription initiation. This hypothesisalsoagreeswith by a detailedreport on the preinitiation complex enhancementby purified progesteronereceptor (8). 808
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
In our system, the ER-induced transcription was totally unresponsive to the presence of estradiol at all conditions tested. Likewise, antihormones such as OH-tamoxifen or ICI 164,384 did not modify the activity of ER. Previously, it had been established that the highlypurifiedER was able to form specific complexes with ERE oligonucleotides in vitro, independently of the presence of estradiol or other proteins (18, 24). These results are in contrast with transfection experiments, where the hormone was absolutely required for DNA binding and gene regulation (7, 25, 26). One might speculate that in vivo receptors may be complexed with presumptive inhibitory proteins, such as hsp90, which dissociate upon the hormone binding. Extraction or purification procedures would reproduce the dissociation in vitro, making the hormone unnecessary. Since in our ER preparation the hsp90 or other associated proteins had been washed away from the immunoaffinity column, this explanation seemed correct. Various results, however, suggest that estradiol may be involved in the receptor action after the DNA binding. In fact, estradiol-receptor as well as antihormonereceptor complexes specifically bind DNA (27), while estradiol, but not antiestrogen is required for the transcription activation function associated with the hormone binding domain of the ER (28). Furthermore, the ER deprived of its hormone binding region, binds DNA but activates the transcription poorly (29). One cannot exclude that the observed hormoneindependence may be due to the simple structure of the constructs employed for in vitro transcription system. A more complete explanation about the hormone role could arise from the study of DNA organisation. In eukaryotes the DNA is packaged into nucleosomes and higher-order chromatin structures, that are now regarded as universal repressors of h-anscriptional initiation (1, 30-32). In vitro studies have shown that the initiation of transcription is inhibited by nucleosomes assembled on the promoter region (33). The question is how activator proteins that bind to specific DNA sequences gain access to their binding sites. The activators could act altering the chromatin conformation in order to allow the binding of initiation or other transcription factors (34, 35). It is worthwhile to note that steroid receptors have been involved in a modification of the chromatin structure in vivo and in vitro (36, 37). More pertinent to this point is the very recent description of a reversible disruption of the nucleosome by glucocorticoid hormone over an enhancer region at the rat tyrosine aminotransferase promoter (38). It would be suggestive if the estradiol-dependent hydrolytic activity that we recently described may be involved at this step (15). This calls for in vitro transcription systems that tend to accurately reproduce the steroid action on reconstituted chromatin. With our system we plan to analyse the effect of the natural ER in a chromatin context as well. Hence, the mechanism of ER action could be further elucidated.
ACKNOWLEDGMENTS We thank Robert G. Roeder for providing us with the pML(CzAT)19 plasmid, and Dario Acampora for helpful suggestions. This work was supported by the Italian Association for Cancer Research (AIRC). V.N. was recipient of an AIRC fellowship.
REFERENCES 1. Roeder, R.G. (1991) Trends Biochem. Sci. 16, 402-408. 2. Sawadogo, M. & Sentenac, A. (1990) Annu. Rev. Biochem. 59, 71 l-754. 3. Buratowski, S., Hahn S., Guarente, L., & Sharp, P.A. (1989) Ceil 56, 549-561. 4. Mitchell, P.J. & Tijan, R. (1989) Science 245, 371-378. 5. Evans, R.M. (1988) Science 242, 889-895. 6. Beato, M. (1989) Ceil 56, 335-344.
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
7. Tora, L., White, J., Brou, C., Tasset, D., Webster, N., Scheer, E., & Chambon, P. (1989) Cell 59. 477-487. 8. Klein-HitpaB, L, Tsai, S.Y., Weigel, N-L., Allan, G.F., Riley, D., Rodriguez R., Schrader, W.T., Tsai, M.-J., & O’Malley, B.W. (1990) Cell 60, 247-257. 9. Kalff, M., Gross B., & Beato (1990) Nature 344. 360-362. 10. Freedman, L.P., Yoshinaga,‘ S.K.‘, Vanderbilt; J.N., & Yamamoto, K.R. (1989) Science 245. 298-30 1. 11. Tsai, S.Y., Srinivasan, G., Allan, G.F., Thompson, E.B., O’Malley, B., & Tsai, M.-J. (1990) J. Biol. C/tern. 265, 17055-17061. 12. Elliston, J. F., Fawell, S.E., Klein-HitpaB, L, Tsai, S.Y., Tsai, M.-J., Parker, M.G., & O’Malley, B.W. (1990) Mol. Cell Biol. 10, 6607-6612. 13. Lewin, B. (1990) Cell 61, 1161-l 164. 14. Sawadogo, M. & Roeder, R.G. (1985) Proc. Natl. Acad. Sci. USA 82, 4394-4398. 15. Molinari, A.M., Abbondanza, C., Armetta, I., Medici, N., Minucci, S., Moncharmont, B., Nigro, V., & Puca, G.A. (1991) Proc. Natl. Acad. Sci. USA 88, 4463-4467. 16. Moncharmont, B., Su, J.L., & Pa&h, I. (1982) Biochemistry 21, 6916-6921. 18. Shapiro, D.J., Sharp, P.A., Wahli, W.W. & Keller, M.J.(1988)DNA 7, 47-55. 17. Medici, N., Nigro, V., Abbondanza, C., Moncharmont, B., Molinari, A.M., & Puca, G.A. (1991) Mol. Endo. 5, 555-563. 19. Fire, A., Samuels, M., & Sharp, P.A. (1984) J. Biol. Chem. 259, 2509-2516. 20. Hawley, D.K., & Roeder, R.G. (1985) J. Biol. Chem. 260, 8163-8172. 21. Sawadogo, M., & Roeder, R.G. (1985) Cell 43, 165-175. 22. Horikoshi, M., Hai, T., Lin, Y.-S., Green, M.R., Roeder, M.G. (1988) Cell 54, 1033-1042. 23. Lin, Y.-S., & Green, M.R. (1991) Cell 64, 971-981. 24. Curtis, S.W. & Korach K.S. (1990) Mol. Endo. 4, 267-286. 25. Tasset, D., Tora, L., Fromental, C., Scheer, E., & Chambon, P. (1990) Cell 62, 11771187. 26. Webster, N.J.R., Green, S., Jin, J.R., & Chambon, P. (1988) Cell 54, 199-207. 27. Kumar, V., & Chambon, P. (1988) Cell 55, 145-156. 28. Pham, T.A., Elliston, J. F., Nawaz, Z., McDonnell, D.P., Tsai, M.-J., & O’Malley, B.W. (1991) Proc. Natl. Acad. Sci. USA 88, 3125-3129. 29. Kumar, V., Green , S., Stack, G., Berry, M., Jin, J.R., & Chambon, P. (1987) Cell 51, 941-951. 30. Grunstein, M. (1990) Trends Genet. 6, 395-400. 31. Laybourn, P.J., & Kadonaga, J.T. (1991) Science 254, 239-245. 32. Kornberg, R.D., Larch, Y. (1991) Cell 67, 833-836. 33. Felsenfeld, G. (1992) Nature 355, 219-224. 34. Workman, J.L., Roeder, R.G., & Kingston, R.E. (1990) EMBO J. 9, 1299-1308. 35. Beato, M. (1991) FASEB J. 5, 2044-2051. 36. Richard-Foy, H. & Hager, G.L. (1987) EMBO J. 6, 2321-2328. 37. Perlmann, T. & Wrange, 0. (1988) EMBO J. 7, 3073-3079. 38. Reik, A., Schtitz, G., and Stewart A.F. (1991) EMBO J. 10, 2569-2576.
810
186,
July
31, 1992
No.
BIOCHEMICAL
2, 1992
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS Pages
PURIFIED
ESTROGEN
RECEPTOR TRANSCRIPTION
ENHANCES
803-810
IN VITRO
Vincenzo Nigro, Anna Maria Molinari, Ignazio Armetta, Antonietta de Falco, Ciro Abbondanza, Nicola Medici, and Giovanni Alfred0 Puca Istituto di Patologia Generale e Oncologia I Facolta di Medicina e Chirurgia S. Andrea delle Dame, 2. 80138 Naples, Italy Received
June
10,
1992
An in vitro transcription system was developed to investigate the mechanismsof gene regulation by the estrogenreceptor (ER). ER purified from calf uterus was highly active in enhancingRNA transcription from a template DNA containing estrogenresponseelements (EREs) upstreamfrom a minimal promoter. Under the conditions employed, no addition of tissuespecific factors was required and both estrogenor antiestrogenswere ineffective. The stimulation of transcription correlated with the copy number of EREs in the template. The addition of competitor ERE oligonucleotides specifically inhibited the ER-induced transcription. We suggestthat the ER may be involved in the formation of the stableinitiation complex. 0 1992Academrc PTfSS,1°C.
Eukaryotic classII promotersgenerally contain a TATA box and regulatory sequences.RNA polymeraseII forms a transcription complex with the generalinitiation factors (TFIIA, B, D, E, F), at the promoter start site (1, 2). Complex assemblybeginswith the stable binding of TFIID to the TATA box. This promotesthe subsequentordered recruitment of the general initiation factors and RNA polymeraseII resultingin a template-committedcomplex (3). The regulatory DNA sequencesare bound by specific transcription activating proteins that respond to regulatory cues. These activators generally contain distinct domains for DNA binding and transcriptionalactivation (4). Steroid receptorsbelong to a large superfamily of ligand-inducible transcription factors with an additional regulatory hormonebinding domain.Transfectionexperimentswith chimeric or truncated steroid receptorshasprovided muchinformation on their function and structure (for review 5, 6). Multiple transcription activation domains were found (7); hence, steroid receptors may stimulate different stagesin the transcription initiation and/or interact with different components of the transcriptional machinery. Understanding the mechanismby which steroid receptors regulate gene expression requires the development of cell-free transcription systemsthat respond to exogenousaddition of receptors. The availability of functional purified moleculesenablethe analysisof the direct effect of the receptoraddition on the transcription machinery. Purified steroid receptors can be modified, components of Abbreviations: AdML, adenovirusmajor late; EMSA, electrophoreticmobility shift assay; ER, estrogenreceptor; ERE, estrogenresponseelement; E2, estradiol-178
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCtI
COMMUNICATIONS
nuclear extracts and/or tissue specific factors can be isolated and their roles and interactions defined. However, the purification of large amount of steroid receptors from target tissue suitable for in vitro transcription studies may present cumbersome difficulties. Conventional procedures damage the molecules and/or dilute the samples, which are then unable to induce in vitro transcription. Furthermore, the receptor elution from ligand-affinity chromatography by steroid analogues hampers the evaluation of the hormone effect. Until now, only the progesterone receptor from chicken oviduct (8) or from rabbit uterus (9) was successful purified as functional molecules. The group of 0’ Malley demonstrated that the stimulation of cell-free transcription occurred through the enhancement of the formation of a stable preinitiation complex. Interestingly, partially purified hormone-free progesterone receptor activated by salt treatment induced transcription in a hormone-independent manner (8). In the attempt to overcome the purification problems for the other members of the steroid receptors family some Authors overexpressed the cDNA in heterologous systems (10). Bacterial or yeast systems yielded poorly functioning molecules, while recombinant human glucocorticoid (11) and mouse estrogen receptors (12) from baculovirus were shown to stimulate correct RNA synthesis in vitro in a way identical to the purified progesterone receptor and from an equivalent promoter construct. However, dissimilar posttranslational modifications have to be taken into account (13). Here we report the successful development of an irz vitro transcription system greatly stimulated by the ER. This ER is a hormone-free molecule highly purified from calf uterus by immunoaffinity that retains full functional properties of ERE binding and high affinity estradiol binding. Our findings suggest that there is a direct ER involvement in the enhancement of the transcription initiation from a template DNA of minimal structure with no tissue specific factor requirement. The stimulation of transcription is completely hormone independent, confirming previous observations with recombinant molecules.
MATERIALS
AND METHODS
Plasmids. pML(CzAT)19 was a generous gift of Dr. R. Cl. Roeder (14). pMLllO(C2AT) was obtained by exonuclease III deletion from the BarnHi site, after Ban&I and PstI digestion of pML(C2ATj19 and contains the AdML promoter linked to a shortened 1 IO-bp Gfree cassette. pMLTATA:O was constructed by recovering the Mspl-Mspl fragment from pML(C2AT) 19, which contains the AdML TATA box and the complete G-free cassette. The fragment was cloned into the AccI site of GEM 3 @omega), oriented with the EcoRI site upstream from the TATA box. pMLTATA: lE, pMLTATA:2E, and pMLTATA:3E derived from pMLTATA:O with the insertion of 1, 2, or 3 copies, respectively, of the estrogen response element of the Xenopus laevis A2 vitellogenin gene. Purification of ER. The hormone binding subunit of ER was purified as previously described (15) except that : i) the immunoaftinity chromatography was performed with AER 311 monoclonal antibody prepared in our laboratory (to be described elsewhere); ii) the immunoaffinity chromatography eluates were buffered with 50 mM Tris-Cl (pH 7.9), containing 40mM NaCl, 3mM dithiothreitol (DTT), 5% (v/v) glycerol and lOOpg/ml bovine serum albumin @omega). The purity of the ER was assessed each time by SDSPAGE and silver staining. The hormone binding capacity of each purified batch of ER was assayed. A control purification was performed as described (15). The purified ER was seen as a main 67 kDa and smaller silver stained bands, that were also affinity-labeled with 3Htamoxifen-aziridine (15). Western-blot analysis of the purified ER showed that all the silver stained SDS-PAGE bands were recognised by the monoclonal antibody AER 310 (not shown). The electrophoretic mobility shift assay (EMSA) showed that the ER was the only ERE binding protein, since the monoclonal antibody JS34/32 further retarded all the proteinoligonucleotide complexes (16, 17) Experimental design. A HeLa cell nuclear extract was employed as the source of basal transcription factors and RNA polymerase II. This cell line does not contain ER and its extracts are relatively nuclease-free. The test template contained the TATA box of Adenovirus 804
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
major late promoter (AdML) and the G-free cassette of 377 bp as reporter (14). AdML promoter provides a useful model system for in vitro analysis of specific transcription by RNA polymerase II. This promoter has a well known structure, with only two essential DNA elements: a TATA box and a single upstream element (USF binding site). We replaced the upstream element with sequences containing ERE(s). Two copies of the Xenopus luevis A2 vitellogenin ERE were inserted at about 90 bp and 140 bp, upstream from the transcription initiation point. A plasmid containing the AdML promoter linked to a shortened (110 bp) Gfree cassette was included in the assay as an internal control. Correct initiation at the test promoter yields transcripts of 387 nucleotides in length, while eventual read-through transcripts are longer (392 nt). In vitro transcription assays. HeLa S3 nuclear extracts were prepared according to Shapiro D.J. ef al. (18). Typical in vitro transcription reactions contained (30 ~1volume): 20 mM HEPES-NaOH (pH 7.9), 5% (v/v) glycerol, 35 mM NaCl, 8 mM MgC12, 2.5 mM DTT, 10 mM creatine phosphate,250 uM 3’-O-methyl GTP, 0.5 mM ATP, 0.5 mM CTP, 10 PM UTP, 4 l&i of [alphaJ*P] UTP, 12 U of RNase Tl, 20 U RNasin (Promega), 6 pg of sonicated salmon spermDNA, 330 ng of pML1 lO(C2AT) template and 225 ng of test template @MLTATA:2E, unlessotherwisespecified),3-5~1 (50-70 pg) of HeLa extract and, where indicated, 0.7 pmoles of purified ER, preincubated at 4°C for 45 min with 5 0 nM 17l3-estradiolprior to addition in the transcription assay.Salmon spermDNA carrier in the preincubation mix was shownto improve the transcriptrecovery and to repressmainly basal transcription. Preincubation step was performed in the absence of MgC12 and nucleotidesfor 30 min at 28°C. Transcriptionstartedby the addition of theselast components and the reaction continued for 30 min at 3O’C. Reactionswere terminatedwith the addition of 70 1.11 of 25 mM Tris-Cl pH7.5, 0.5% SDS, 5 mM EDTA, 100ug/ml of Ecoli transfer RNA, 400 pg/ml proteinaseK and incubatedat 45°C for 45 min. Then, 200 yl 8 M urea in 10 mM Tris-Cl pH 7.5, 5 mM EDTA was addedand the sampleswere extracted once with phenol-chloroform and ethanol precipitated. The pellet was suspendedin 80% formamide, denaturedat 86°C for 5 min and applied on a 8 M urea-6% polyacrylamide gel, run at 45 W (constant power) and visualised by autoradiography.Typical exposurewas 16hours with an intensifying screen.
RESULTS
Purified
ER stimulates in vitro transcription. Figure 1 shows the stimulation of accurate RNA transcription by purified ER (upper arrow). Little if any longer read-through transcripts from the test templatewere observed,as alsoconfirmed by higher-resolution gel electrophoresis(not shown). As expected, ER did not significantly alter the synthesisof the shorter 120 nt RNA from the control templatepML1 lO(C?AT) (lower arrow). The addition of 0.5 pg/ml a-amanitin completely abolishedthe transcription from both templates(lanes7 and 8), thus demonstratingthat the transcriptswere entirely due to polymeraseII. Under the conditions employed, the presenceof estradiol (lanes5 and 6) or the estrogenantagonists, ICI 164,384or 4-hydroxytamoxifen (data not shown),did not modify the fold stimulation by ER. The addition of an increasing amount of ER resulted in the enhancementof expected transcript synthesis, but estradiol had no effect even with lower ER concentrations (not shown). ERE is required for ER transcription enhancement. To verify the requirement of EREs for ER induction, we usedreporter plasmid containing 0, 1, 2, or 3 copies of EREs. As shown in Fig. 2, the addition of ER was totally unable to increaseRNA synthesisfrom a reporter plasmid lacking an ERE (lanes l-4). A dramatic increasewas already seenwith 1 ERE in the reporter plasmid (pMLTATA: 1E, lanes5-8). Two copiesof EREs provided the best result of ER-dependent induction, since the basallevel of transcription was reduced. Unexpectedly, RNA transcription from the test template with 3 ERE was heavily reduced (lanes 13-16). However, longer autoradiographic exposureof the gel displayed a high fold stimulation by ER from this template, comparableto that observedwith pMLTATA:2E. The reasonfor the lower basalactivity of the template with multiple EREs remainsunknown. It 80.5
Vol.
186,
No.
ER E2
2, 1992
-
BIOCHEMICAL
+ -
+ +
AND
BIOPHYSICAL
cy aman. - + +
0 ERE
RESEARCH
1 ERE
COMMUNICATIONS
2 ERE
3 ERE
ER’-T.
1 2 3 4 5 6 7 8 9 10111213141516
Figure 1. Effect of the ER on in vitro transcription. Elution buffer (lanes 1, 2) or 1 pmol of highly-purified ER, in the absence (lane 3, 4) or presence of 50 nM estradiol (lanes 5-6). were added in the assay. Samples in lanes 7-8 were identical to 5-6, except that a-amanitin was included. In vifro transcription was performed as described under Materials and Methods. Grey and black arrows indicate the correcdy initiated transcripts from the pML1 lO(C2AT) and pMLTATA:ZE, respectively. Fiwre 2. Dependence of ER-induced transcription from the copy number of EREs in the template. In vitro transcriptionwasperformedunderthe standardconditions with 225 ng of templatescontaining no copies (pMLTATA:O, lane l-4), one copy (pMLTATA:lE, lane 5-8), two copies(pMLTATA:2E, lanes9-12) and three copies @MLTATA:3E, lanes13-16)of ERE.
may be that repeatedpalindromicsequences weakenthe templatetranscriptioneither through a modified DNA conformation, or by drawing HeLa proteins that interfere with the process. EMSA of ERE by HeLa extract did in fact showmultiple retardedbands(data not shown). If the effect of ER is dependenton its binding with the ERE in the test template, the addition of ERE oligonucleotide excess should squelch the ER, preventing its stimulatory effect. Competitor ERE oligonucleotideswere thus addedto the transcription assay.Starting from a 10 to 20-fold molar excess of ERE, a decreaseof the receptor-stimulated transcripts was detected (Fig. 3, lanes 5-6). At 40-fold molar excessthe ER-dependent transcription was reducedto the basallevel. At higher competitor concentrations(lane 8 and data not shown), a consistentinhibition of both the testtemplateand the internal control transcription occurred. ER functions in the preinitiation step. The recruitment of preinitiation complexes iir vitro is a slow step preceding the transcription initiation. Many Authors have described the 806
Vol.
186,
No.
2, 1992
BIOCHEMICAL
(molar
ERE ratio) ER
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
0 0 0 0 10 204080 - -++++++
12345678
Figure 3. Specific competition of ER-induced transcription by ERE oligonucleotides. Increasing amounts of oligonucleotides containing the ERE from the Xenopus vitellogenin A2 gene were added to the assay prior to the addition of the ER and HeLa extract. The molar ratios of competitor ERE vs template ERE are indicated. Transcriptionwasperformedunderthe standardconditions,exceptthat pMLTATA:lE as inducible template DNA was employed.
importance of such a step to maximise the RNA synthesis, which promptly starts from preformed complexes,when nucleotidetriphosphatesare added(19,20). To elucidate whether ER is necessaryfor template-committedpreinitiation complexes, we preincubatedthe test template with the HeLa extract and the ER (Fig.4, lanes l-4). A parallel preincubation of the test template with HeLa extract alone wascarried out, while the ER was addedlater with the nucleotides(lanes5-8). It appearsevident that, when the ER was absent during the preincubation step, the stimulatory effect wasmuch weaker (Fig. 4, comparelanes l-4 with 5-8). When the preincubation step was omitted, a general decrease of RNA transcripts was noticed (Fig. 4, lanes 9-12) without disappearance of ER-dependent transcription. Finally, an experiment presentedwas performed to define the requirement of the general transcription factor in the preincubationstep.If comparedto the standardconditions, the ERinduced transcription was lessefficient when the preincubation step was carried out in the absenceof HeLa extract and the HeLa extract addedtogether with nucleotides(not shown). This suggeststhat the ER is required in the early step, not only to bind to its target sequence, but also to recruit the factors involved in the formation of the stableinitiation complex. In this way, both ER and HeLa cell extract are essentialto obtain maximal activity.
DISCUSSION In the present paper, we characterise an in vitro transcription system, stimulated by the highly-purified estrogen receptor. The receptor induced faithful RNA transcription 807
Vol.
186, No. 2, 1992
BIOCHEMICAL
A
r ER
AND BIOPHYSICAL
-
B +
-
RESEARCH COMMUNICATIONS
C +
-
I +
1 2 3 4 5 6 7 8 9101112
DENRA
A
B
NTPs, MgCll 4
HeLa I.
preincubation
transcription
b
ER NTPs, MgClz 4
DNA HeLa ,
preincubation
c
’
transcription
HeLa NTPs, MgCI, I transcription
-
*
Figure 4. ER-dependent RNA synthesis with or without preincubation. Lanes l-4 representthe standardassay,performedasin scheme A. In lanes5-8 theER wasomitted from the preincubationmix and addedtogetherwith NTPs and MgC12(B). Lanes9-12 indicateassays performedwithoutpreincubation (C).
specifically from ERE(s)-containing template DNA without the requirement for tissue- or species-specific factors and/or promoters. Similarly to what is known about other activating proteins (21-23, it is likely that ER enhancesthe transcription interacting with the basal factors involved in the preinitiation complex. The ER was quite ineffective when added after the preincubation step and a decreasein fold stimulation was observedwhen the HeLa extract was added subsequently. Hence, the ER may be required during the preincubation step. It may be involved in the recruitment or modification of a basal factor essential for transcription initiation. This hypothesisalsoagreeswith by a detailedreport on the preinitiation complex enhancementby purified progesteronereceptor (8). 808
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
In our system, the ER-induced transcription was totally unresponsive to the presence of estradiol at all conditions tested. Likewise, antihormones such as OH-tamoxifen or ICI 164,384 did not modify the activity of ER. Previously, it had been established that the highlypurifiedER was able to form specific complexes with ERE oligonucleotides in vitro, independently of the presence of estradiol or other proteins (18, 24). These results are in contrast with transfection experiments, where the hormone was absolutely required for DNA binding and gene regulation (7, 25, 26). One might speculate that in vivo receptors may be complexed with presumptive inhibitory proteins, such as hsp90, which dissociate upon the hormone binding. Extraction or purification procedures would reproduce the dissociation in vitro, making the hormone unnecessary. Since in our ER preparation the hsp90 or other associated proteins had been washed away from the immunoaffinity column, this explanation seemed correct. Various results, however, suggest that estradiol may be involved in the receptor action after the DNA binding. In fact, estradiol-receptor as well as antihormonereceptor complexes specifically bind DNA (27), while estradiol, but not antiestrogen is required for the transcription activation function associated with the hormone binding domain of the ER (28). Furthermore, the ER deprived of its hormone binding region, binds DNA but activates the transcription poorly (29). One cannot exclude that the observed hormoneindependence may be due to the simple structure of the constructs employed for in vitro transcription system. A more complete explanation about the hormone role could arise from the study of DNA organisation. In eukaryotes the DNA is packaged into nucleosomes and higher-order chromatin structures, that are now regarded as universal repressors of h-anscriptional initiation (1, 30-32). In vitro studies have shown that the initiation of transcription is inhibited by nucleosomes assembled on the promoter region (33). The question is how activator proteins that bind to specific DNA sequences gain access to their binding sites. The activators could act altering the chromatin conformation in order to allow the binding of initiation or other transcription factors (34, 35). It is worthwhile to note that steroid receptors have been involved in a modification of the chromatin structure in vivo and in vitro (36, 37). More pertinent to this point is the very recent description of a reversible disruption of the nucleosome by glucocorticoid hormone over an enhancer region at the rat tyrosine aminotransferase promoter (38). It would be suggestive if the estradiol-dependent hydrolytic activity that we recently described may be involved at this step (15). This calls for in vitro transcription systems that tend to accurately reproduce the steroid action on reconstituted chromatin. With our system we plan to analyse the effect of the natural ER in a chromatin context as well. Hence, the mechanism of ER action could be further elucidated.
ACKNOWLEDGMENTS We thank Robert G. Roeder for providing us with the pML(CzAT)19 plasmid, and Dario Acampora for helpful suggestions. This work was supported by the Italian Association for Cancer Research (AIRC). V.N. was recipient of an AIRC fellowship.
REFERENCES 1. Roeder, R.G. (1991) Trends Biochem. Sci. 16, 402-408. 2. Sawadogo, M. & Sentenac, A. (1990) Annu. Rev. Biochem. 59, 71 l-754. 3. Buratowski, S., Hahn S., Guarente, L., & Sharp, P.A. (1989) Ceil 56, 549-561. 4. Mitchell, P.J. & Tijan, R. (1989) Science 245, 371-378. 5. Evans, R.M. (1988) Science 242, 889-895. 6. Beato, M. (1989) Ceil 56, 335-344.
Vol.
186,
No.
2, 1992
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
7. Tora, L., White, J., Brou, C., Tasset, D., Webster, N., Scheer, E., & Chambon, P. (1989) Cell 59. 477-487. 8. Klein-HitpaB, L, Tsai, S.Y., Weigel, N-L., Allan, G.F., Riley, D., Rodriguez R., Schrader, W.T., Tsai, M.-J., & O’Malley, B.W. (1990) Cell 60, 247-257. 9. Kalff, M., Gross B., & Beato (1990) Nature 344. 360-362. 10. Freedman, L.P., Yoshinaga,‘ S.K.‘, Vanderbilt; J.N., & Yamamoto, K.R. (1989) Science 245. 298-30 1. 11. Tsai, S.Y., Srinivasan, G., Allan, G.F., Thompson, E.B., O’Malley, B., & Tsai, M.-J. (1990) J. Biol. C/tern. 265, 17055-17061. 12. Elliston, J. F., Fawell, S.E., Klein-HitpaB, L, Tsai, S.Y., Tsai, M.-J., Parker, M.G., & O’Malley, B.W. (1990) Mol. Cell Biol. 10, 6607-6612. 13. Lewin, B. (1990) Cell 61, 1161-l 164. 14. Sawadogo, M. & Roeder, R.G. (1985) Proc. Natl. Acad. Sci. USA 82, 4394-4398. 15. Molinari, A.M., Abbondanza, C., Armetta, I., Medici, N., Minucci, S., Moncharmont, B., Nigro, V., & Puca, G.A. (1991) Proc. Natl. Acad. Sci. USA 88, 4463-4467. 16. Moncharmont, B., Su, J.L., & Pa&h, I. (1982) Biochemistry 21, 6916-6921. 18. Shapiro, D.J., Sharp, P.A., Wahli, W.W. & Keller, M.J.(1988)DNA 7, 47-55. 17. Medici, N., Nigro, V., Abbondanza, C., Moncharmont, B., Molinari, A.M., & Puca, G.A. (1991) Mol. Endo. 5, 555-563. 19. Fire, A., Samuels, M., & Sharp, P.A. (1984) J. Biol. Chem. 259, 2509-2516. 20. Hawley, D.K., & Roeder, R.G. (1985) J. Biol. Chem. 260, 8163-8172. 21. Sawadogo, M., & Roeder, R.G. (1985) Cell 43, 165-175. 22. Horikoshi, M., Hai, T., Lin, Y.-S., Green, M.R., Roeder, M.G. (1988) Cell 54, 1033-1042. 23. Lin, Y.-S., & Green, M.R. (1991) Cell 64, 971-981. 24. Curtis, S.W. & Korach K.S. (1990) Mol. Endo. 4, 267-286. 25. Tasset, D., Tora, L., Fromental, C., Scheer, E., & Chambon, P. (1990) Cell 62, 11771187. 26. Webster, N.J.R., Green, S., Jin, J.R., & Chambon, P. (1988) Cell 54, 199-207. 27. Kumar, V., & Chambon, P. (1988) Cell 55, 145-156. 28. Pham, T.A., Elliston, J. F., Nawaz, Z., McDonnell, D.P., Tsai, M.-J., & O’Malley, B.W. (1991) Proc. Natl. Acad. Sci. USA 88, 3125-3129. 29. Kumar, V., Green , S., Stack, G., Berry, M., Jin, J.R., & Chambon, P. (1987) Cell 51, 941-951. 30. Grunstein, M. (1990) Trends Genet. 6, 395-400. 31. Laybourn, P.J., & Kadonaga, J.T. (1991) Science 254, 239-245. 32. Kornberg, R.D., Larch, Y. (1991) Cell 67, 833-836. 33. Felsenfeld, G. (1992) Nature 355, 219-224. 34. Workman, J.L., Roeder, R.G., & Kingston, R.E. (1990) EMBO J. 9, 1299-1308. 35. Beato, M. (1991) FASEB J. 5, 2044-2051. 36. Richard-Foy, H. & Hager, G.L. (1987) EMBO J. 6, 2321-2328. 37. Perlmann, T. & Wrange, 0. (1988) EMBO J. 7, 3073-3079. 38. Reik, A., Schtitz, G., and Stewart A.F. (1991) EMBO J. 10, 2569-2576.
810
