3~7?
/V /
MOLECULAR MECHANISM OF ACTION OF STEROID HORMONE RECEPTORS
DISSERTATION
Presented to the Graduate Council of the
University of North Texas in Partial
Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
By
Zafar Nawaz, B.S., M.S.
Denton, Texas
May, 1992 Nawaz, Zafar, Molecular Mechanism of Action of Steroid Hormone Receptors.
Doctor of Philosophy (Biology), May, 1992, 136 pp., V tables, 46 illustrations, bibliography, 204 titles.
A novel bacterial expression system that is capable of producing high levels of soluble, stable, biologically active human vitamin D3 and estrogen receptors has been developed. The method utilizes ubiquitin fusion technology and a low temperature nalidixic acid induction of the lambda PL promoter. This system can produce large quantities of receptor antigen, but only a small fraction displays wild-type DNA and hormone binding properties. Therefore, the use of this system to overproduce receptors for crystallization studies is not practical. To overcome these problems, a 2 um based ubiquitin fusion system which allows regulated expression of the estrogen receptor in yeast (Saccharomyces cerevisiae) was developed. This system produces the estrogen receptor to a level of 0.2% of the total soluble protein. Moreover, this protein is undegradable, soluble, and biologically active. To test the transcriptional activity of the estrogen receptor produced in yeast, a cis-trans transcription assay was developed. This assay revealed that the transcriptional activity of the human estrogen receptor expressed in yeast was similar to that observed in transfected mammalian cells. This reconstituted estrogen transcription unit in
Saccharomyces cerevisiae was utilized to examine the regulation of estrogen receptor functions by antiestrogens, to develop a random and rapid approach for identifying novel estrogen response elements, to characterize estrogen receptor variants cloned from human breast tumors, and to examine the effect of estrogen receptor on the regulation of osteocalcin gene. PREFACE
It has been my hope that during these years of research work, I could have help to
clarify the "mystery" that surrounded the many aspects of the steroid receptor gene
superfamily. I have the hope that the results of this work can serve either as the starting
point for further detailed studies on the same topic, or as applications in other related or
unrelated research studies. I also hope that a reader will find this thesis a useful reference.
My research work has gone on for the last four years benefiting from the help of so
many people that it is impossible to thank everyone concerned. But let me make a start with
apologies to those unmentioned others whose involvement is also greatly appreciated.
I owe a special debt of gratitude to Donald P. McDonnell who provided not only
valuable guidance and the opportunity to work independently, but who also gave me
financial support throughout the course of this research. I would not have embarked on and completed this endeavor had it not been for his warm support and good counsel.
For most of the research work period, I am also deeply grateful to Bert W.
O'Malley for access to the facilities of his laboratories at Baylor College of Medicine in
Houston, Texas. I thank him also for precious scientific advice and for his assistance in
part of the work.
I thank Tauseef R. Butt for the valuable time spent in his laboratory at Smith,
Klein, and Beecham Laboratories in Philadelphia, and for kindly providing the ubiquitin
fusion technology.
I am also deeply grateful to Don W. Smith for his understanding, kindness, and for having encouraged me throughout my graduate studies.
As some of the work presented in this study was done with the support of other members of the laboratory in Houston, I would also like to express my thanks to them:
iii West Pike for kindly providing vitamin D3 antibodies; Ahmed Usman for helping with the bandshift experiments using E. coli expressed estrogen receptor; Geoff Green for kindly providing the estrogen receptor cDNA and estrogen receptor antibodies; Charles Densmore for helping with the Scatchard analysis and calf thymus DNA-cellulose chromatography of yeast expressed estrogen receptor; Nancy Weigel for assisting with immunoprecipitation analysis of estrogen receptor; John Elliston and Salman Haider for helping with transfections and CAT assays; Tony Pham for helping with in vivo footprinting; Suzanne
Fuqua for kindly providing the variant receptors cloned from human breast tumors; and Paula Howard for skilled technical assistance.
I thank the members of my thesis committee for their advice and support during the course of this study: Tauseef R. Butt, John Knesek, Donald P. McDonnell, Gerard A.
O'Donovan, Mark S. Shanley, Don W. Smith, and G. R. Vela.
All this immense help from those mentioned and from unmentioned others is gratefully acknowledged.
Houston, Texas
May, 1992
iv TABLE OF CONTENTS
Page
LIST OF TABLES ...... x
LIST OF FIGURES ...... xi
Chapter
I. INTRODUCTION ...... 1
General Aspects of the Steroid Receptors ...... 1
Structure and Functions of Steroid Receptors ...... 3
DNA-binding domain ...... 5
Steroid receptor DNA-binding sites ...... 9
Steroid-binding domain ...... 11
Transcriptional activation domains ...... 13
Steroid Signal Transduction ...... 15
Effect of hormone on DNA binding ...... 16
Effect of hormone on transactivation ...... 17
Recent Developments ...... 17
II. MATERIALS AND METHODS ...... 19
Materials ...... 19
Bacterial Strains ...... 20
Yeast Strains ...... 20
Cell Lines and Human Tumor Specimens ...... 20
DNA Constructions ...... 21
Bacterial expression vectors ...... 21
Yeast expression vectors ...... 22
V Yeast reporter vectors ...... 24 Mammalian expression and reporter
vectors...... 25
Induction of Recombinant Proteins in Bacteria ...... 26 Western Immunoblotting of Proteins Expressed in Bacteria ...... 27 Hormone Binding Determinations of Bacterial
Expressed Proteins ...... 27 Hormone-binding assays of
vitamin D receptor ...... 0.....6...... 27 Hormone-binding of estrogen
receptor...... 28 DNA-binding Properties of Bacterial Expressed
Proteins ...... 28 DNA-binding properties of vitamin D receptor...... 28 DNA-binding properties of estrogen receptor...... 0.4.0.0.0.0.29 Induction of Estrogen Receptor in Yeast and
Preparation of Yeast Extracts ...... 29. ....* Preparation of MCF-7 Estrogen Receptor ...... 30 Hormone-binding Analysis of Estrogen Receptor from Yeast...... 30
DNA-binding Analysis of Estrogen Receptor from Yeast...... 31
vi DNA-cellulose chromatography...... 31
Band shift assays...... 31
Western Blot Analysis of Estrogen Receptor
from Yeast...... 31
Antibody Immunoprecipitation of Estrogen
Receptor ...... 32
Immunodot Analysis of Estrogen Receptor ...... 32
Assay of Estrogen Receptor Function in Yeast ...... 32
Construction and Propagation of (SRE)n
Library and Development of the Screen for
Identification of SREs ...... 33
Transient Transfection Assays ...... 33
III. RESULTS ...... 34
Development of Bacterial Expression System ...... 34
Expression and Biochemical Characterization
of Vitamin D3 Receptor ...... 34
Expression and Biochemical Characterization
of Estrogen Receptor ...... 40
Development of Saccharomyces cerevisiae as
a Host System for Steroid Hormone Action
Studies...... 45
Production of Mammalian Steroid Hormone
Receptors in Yeast* ...... 46
Biochemical Characterization of Yeast
Expressed Estrogen Receptor ...... 50
v11 Reconstitution of Steroid Hormone Responsive
Transcription Units in Saccharomyces
cerevisiae ...... 59
Applications of Saccharomyces cerevisiae
System...... 67
Effect of antiestrogens on the
regulation of estrogen receptor ...... 67
Identification of novel estrogen response
elements...... 75
Study of variant receptors cloned from
breast tumors ...... 84
Regulation of osteocalcin gene by
estrogen receptor ...... 99
IV. DISCUSSION...... 107
Bacterial Expression System...... 107
Development of Saccharomyces cerevisiae as
an Expression System ...... 110
Development of Estrogen Responsive
Transcriptional Unit in Saccharomyces
cerevisiae...... 113
Application of Saccharomyces cerevisiae
System...... 114
Effect of antiestrogens on the regulation of
human estrogen receptor ...... 114
Identification of novel estrogen response
v111 elements...... 116 Study of variant receptors cloned from
breast tumors ...... 118
Regulation of the osteocalcin gene by
estrogen receptor ...... 120
REFERENCES ...... 122
ix LIST OF TABLES
Table Page I.Sequence of Degenerate Oligonucleotides and
Comparison with Other SREs ...... 78 ... II.Sequence and Transcriptional Activities of Hormone
Dependent Response Elements ...... 79 III.Comparison of Synthetic EREs Sequences with
Mammalian Gene Bank ...... 83 IV.Transcriptional Activities of Wild-Type and Variant
Estrogen Receptors Cloned from ER-/PR+ Tumors ...... 87 V.Transcriptional Activities of Wild-Type and Variant
Estrogen Receptors Cloned from ER+/PR- Tumors ...... 9 1
x LIST OF FIGURES
Figure Page
1.The chemical structure of the major steroid
hormone receptor ligands ...... 1
2.Homology within the steroid receptor gene
superfamily ...... 4
3.Amino-acid comparison of the functional domains of
the nuclear receptor superfamily ...... 6
4.Sequence alignment of the DNA-binding domain, C1,
of the different steroid receptors ...... 8
5.Steroid response elements of different steroid
hormone receptors...... 10
6.Location of the dimerization motif of the
human estrogen receptor...... 12
7.Steroid receptor transactivation domains ...... 14
8.Bacterial expression vectors ...... 35
9.Expression of UB-VDR in E. coli strain AR58 and
its analysis by Western blot ...... 37
10.Western immunoblot analysis of the soluble form
of UB-hVDR produced in E. coli ...... 38
S1.Analysis of functional properties of
recombinant receptor...... 39
12.Expression of UB-ER in various E. coli strains and
its analysis by Western blots ...... 41
xi 13.Ligand titration curve and Scatchard blot analysis of ER
expressed in E. coli, induced by nalidixic acid ...... 43
14.Specific binding of E. coli expressed hER to the
estrogen response elements ...... 44
15.Schematic depiction of the strategy used to produce
biologically active steroid receptors in yeast ...... 47
16.Western immunoblot analysis of recombinant
estrogen receptor produced in yeast ...... 49
17.Scatchard plot analysis of estrogen receptor
expressed in yeast ...... 51
18.Hormone-binding and Western immunoblot analyses
of recombinant estrogen receptor ...... 52
19.Lmmunoprecipitation of labeled human estrogen
receptor isolated from yeast and mammalian
sources...... 54
20. Expression of human estrogen receptor in yeast
from leu 2d plasmid ...... 56
21.The DNA-binding properties of the yeast
expressed estrogen receptor ...... 57
22.Specific binding of yeast expressed estrogen
receptor to estrogen response elements ...... 58
23.Assay of transcriptional activity of steroid
receptor in yeast ...... 60
24.Yeast expression vector (YEPE10) and
reporter plasmid (YRPE2) ...... 61
Xii 25.Transcriptional activity of yeast
produced estrogen receptor ...... 62 26.Ligand induced activity of YEPKB 1 and YEPE1O...... 64
27.Western immunoblot analysis and transcriptional
activity of estrogen receptor expressed from an
integrated expression unit ...... 65
28.Induction of ER-VP16 chimeras by both estrogen and
antiestrogen in mammalian cells ...... 69
29.Hormone and antihormone regulation of
estrogen receptor function in yeast ...... 71
30.Effect of level of estrogen receptor expression
on transcriptional regulation by estradiol
and nafoxidine ...... 72
3 1.Alteration in chromatin structure induced by
estrogen and antiestrogen...... 74
32. Identification of estrogen response elements
using a yeast based screen ...... 76
33.Transcriptional activity of human estrogen receptor
from reporter plasmids containing receptor
binding sites as direct or indirect repeats ...... 81
34.Dideoxy sequence analysis of the variant
estrogen receptor ...... 85
35.Yeast expression vector expressing variant
estrogen receptor...... 86
36.Western blot analysis of the wild-type and
xiii variant estrogen receptors...... 88
37.Schematic diagram of the sequencing results for
the 6/7 and 7/8 exon bounderies ...... 90 38.Transcriptional activity of the wild-type estrogen
receptor in the presence of the variant YEPE13 ...... 93 39.Transcriptional activity of the wild-type estrogen
receptor in the presence of the variant YEPE15 ...... 94 40.Transcriptional activity of the wild-type estrogen
receptor in the presence of the YEPAR ...... 95 41.Western blot analysis of the wild-type and
variant estrogen receptors ...... 96 42.GAL4 transcriptional interference assay ...... 98 43.Estradiol dependent and independent activation of
transcription of human osteocalcin gene ...... 100 44.Hormone and receptor dependent and independent
activation of transcription of human
osteocalcin gene ...... 101 45.Regulation of osteocalcin gene by 1,25(OH)2D3
and estradiol ...... 103 46.Deletion analysis of osteocalcin gene promoter...... 105
xiv CHAPTER 1
INTRODUCTION
GENERAL ASPECTS OF THE STEROID RECEPTORS
Steroids are derivatives of cholesterol which are produced in primary tissues and act on the target tissues by activating their intracellular receptors. These intracellular receptors, upon hormone binding, acquire the ability to regulate transcription of genes linked to steroid-responsive DNA elements (SREs) located in the promoters of target genes (1).
Activated steroid receptors coordinate the expression of these genes by either promoting the formation of an active transcription complex, or, in certain cases, by preventing it (2-3). The steroid receptors are a subgroup of a much larger superfamily of transcription factors, including those that are regulated by known ligands, ranging from thyroid hormone to vitamin D, and others for which ligands have not been identified (4-5). Members of this superfamily have similar structural organization and perhaps similar mechanisms of regulation.
The ligands for the steroid receptor superfamily can be broadly divided into three groups based on their function: adrenal steroids, reproductive steroids, and metabolic or developmental hormones (Figure 1). Thyroid hormone and retinoic acid are also included in Figure 1. Their receptors are known to belong to the same family of ligand activated transcription factors as the classical steroid receptors.
The adrenal steroids regulate several processes associated with the stress response, neuroendocrine control, homeostasis, including glycogen and mineral metabolism, and the growth and differentiation of cells (6). The reproductive steroids are responsible for the
1 2 normal development and determination of the reproductive organs, sexual behavior, and the development of secondary sexual characteristics (7). Vitamin D plays an important role in normal bone development, cell differentiation, and regulation of calcium metabolism.
Thyroid hormone is required for the normal development and regulation of the central nervous system, skeleton, regulation of skin temperature and oxygen metabolism, and the maintenance of proper gonadal function (8). Retinoic acid also functions as a morphogen in the process of intervertebrate limb regeneration (9), and in the anteroposterior development of the central nervous system (10).
REPRODUCTIVE STEROIDS ADRENAL STEROIDS I METABOLIC HORMONES
CH,
C - 0 CHOH OH
C = 0 O~O
HO ,0
HO HOH PROGESTERONE 1, 25 OJHYOROXYCHOLE-CALCIFE ROL
OH CORTISOL COOH
O4 H 0 CHOH RETINOIC ACIC OIHYDROTESTOSTERONE HO OH NH,
0AisN HO CH?- - COOH ALOOSTERONE HO H
ESTRADOOL TRIODOTHYRONINE
Figure 1. The chemical structure of the major steroid hormone receptor ligands. 3
Structure and Functions of Steroid Receptors
A great deal is now known about steroid receptor structure since many of these
receptors have been cloned: progesterone receptor, PR (11-13), estrogen receptor, ER (14
15), androgen receptor, AR (16-17), glucocorticoid receptor, GR (18-20),
mineralocorticoid receptor, MR (21-22), thyroid receptor, TR (23-24), retinoic acid
receptor, RAR (25-28), and vitamin D receptor, VDR ( 29-30, 4).
Homology between members of the steroid receptor gene superfamily was revealed
by sequence comparisons of the cloned receptors. The three regions of consensus homology, C1, C2, and C3 are shown in Figure 2. The first region of homology, C1, a cysteine-rich DNA-binding domain, is the most highly conserved. The other two regions,
C2 and C3, located within the C terminal are less highly conserved, yet they still have significant homology. The C2 and C3 regions are considered to be associated with ligand
binding, transcriptional activation, and potential protein-protein interactions with other steroid receptors as well as with potential inhibitory factors. These structural observations suggest that the steroid receptor superfamily represents an old family of regulatable transcription factors. It seems that the early forms of these receptors were regulated by
intracellular metabolic ligands, in an intracrine fashion (31). During evolution, some of the receptors may have lost the ligand-binding domain, and became constitutive transcription factors, such as V-erbA (32). Other receptors may have acquired the ligand specificity for steroids, thyroid hormones, retinoic acid. An abundance of related genes (33) have been identified by low stringency Southern analysis using a DNA sequence corresponding to the
DNA-binding domain of GR. Some of these genes have been cloned and show to be closely related to the steroid hormone receptors. However, their ligands are still unknown 4
(33-35). These proteins, called "orphan receptors," may represent the earliest forms of the steroid receptor gene superfamily, and may not in fact require a ligand for function.
421 6 53 7M I ** hGR
.~.::c~x:rn$$:.~.:*....~8 2 I hMR
gig hAR
%% %Q M'34 hPR
7 S~t.7x.. > .36:.. : n ::::.x.x.':.> hER
x. .\a.*.. 43 $*.~29 k... hl ERR 1
hEFRR 2
40? hRR
466 TR 1
Q27 41 s...>. .. . *.>=03m x*$ 5 hVDR Domain Length I 66aa I 42aa 44GM..~.4.. .~. s*\:CM:*.k.~&:~4f III 22aa COUP
Figure 2. Homology within the steroid receptor gene superfamily. There are three regions of homology, referred to as Cl, the DNA-binding domain, C2, and C3, for which the functional significance of the homology remains to be determined. 5
Molecular cloning and mutational analysis have revealed that the steroids receptors
within the steroid receptor superfamily are organized into functional domains (36-37, 12).
Proteolytic cleavage analysis first revealed receptor fragments which separated DNA
binding activity from steroid binding (38-39). Deletion studies of cloned receptors allowed
these domains to be more clearly defined (40-45). These functional domains can be
classified into a hypervariable N-terminal domain, a basic cysteine-rich DNA-binding
domain, and a C-terminal hormone-binding domain (Figure 3). Also, embedded within
these major domains, there are regions important for transactivation, dimerization, nuclear
localization, and interaction with heat-shock proteins.
DNA-binding domain
Sequence comparisons of the earliest cloned steroid receptors suggested that the
most conserved 66-68 amino acid region among the different steroid receptors, C1, coded
for the DNA-binding domain (46-48). Proof that C1 region contains determinants for DNA
binding specificity was obtained by use of chimeric receptors in which the DNA-binding
domain of one receptor was swapped with that of another (49, 26, 36). Thus, for example,
an ER which contains a 66-amino acid DNA-binding domain from the human
glucocorticoid receptor (hGR) activates glucocorticoid responsive target reporter genes in
the presence of estradiol (49).
Initial analysis of sequence data suggested that the DNA-binding domain includes
nine perfectly conserved cysteines. This domain forms two zinc fingers analogous to the
structures formed by transcription factor IIIA (50). Sequences of C1 from several human
receptors and the proposed finger structures are shown in Figure 4. Positive identification of zinc in the finger structures coordinated by cysteines was reported for the glucocorticoid receptor using EXAFS (extended X-ray absorption fine structure) and visible light 6
Maximum activity DNA Hormone 421 486 528 777 Hypervariable GlucoGR
603 568 634 984 <15 MR 194 1 1 57
567 633 680 934 <15 1901 1 55p R 1 4
1 185 250 311 551 5995 <15 1521 /1 301 1E R
178 241 295 521 <15 156 1 1 28 1 ERIR1
1 103 168 212 433 <151 56L 1 28 ER *R2
1 24 59 192 4277 42 <15 VDR
1 162 169 232 456 <15 47 17 T 3 R,
1 53 120 183 410 <151 4 1717 T3 Rc
291 358 421 638 GAG 47 1 1 17 V-ert A
1 88 153198 462 <15 4I 1 15 RAkR
81 146 193 448 <15 44 16 HA
245 311 374 624 1 237 <15 44 <15 E75
Figure 3. Amino acid comparison of the functional domains of the nuclear receptor superfamily. Primary amino acid sequences have been aligned on the basis of regions of maximum amino acid similarity, with the percentage amino acid identity indicated for each region in the relation to the hGR. Domains shown are: a domain at the NH2-tenninal end, required for maximum activity; the 66-68-amino acid DNA-binding core, and the 250-amino acid hormone-binding domain. The amino acid position of each domain boundary is shown. 7
spectroscopy (51). The loop of the finger consists of 12-13 amino acids, and there is a linker region of 15-17 amino acids between the two fingers (52) as shown in Figure 4.
Mutagenesis experiments and physical-biochemical analyses have shown that the zinc-finger motifs of the receptors are structurally distinct from classical zinc fingers as typified by the DNA-binding transcription factor IIIA (53-54). The major structural difference is that unlike TFIIIA, which coordinates zinc by a pair of histidines, steroid receptors coordinate it only with cysteines. Substitution of cysteine pairs with histidines results in the inactivation of DNA binding (55), while nonpaired substitutions of cysteines with histidines only partially inactivates DNA binding (56).
The ninth highly conserved cysteine appears not to be essential for DNA binding, thus eliminating the possibility of an alternative finger structure involving this cysteine.
Genomic clones of some of the steroid receptors revealed that the two zinc fingers are coded in separate exons (57-60); however, these fingers are sufficiently different that if they arose by duplication, they have since diverged considerably.
Once it was established that DNA binding occurred through the specific arrangement of two zinc fingers, experiments were carried out to elucidate the structural features of these fingers responsible for the determination of DNA binding specificity (53).
The first finger was largely responsible for DNA sequence specificity, as determined by swapping individual fingers of the estrogen and glucocorticoid receptors (61). The second finger stabilizes the binding of the receptor to its DNA response element.
Detailed point mutation analysis was used to identify amino acids within the two zinc fingers required for the determination of target gene specificity (62-63, 52). These experiments demonstrated that the amino acid residues that are important for DNA-binding specificity occur at the base of the first zinc finger (P-box), and in the region between the first and the second finger (D-box). These experiments reveal a class heterogeneity in the 8
The Zinc Finger Regions of Steroid Receptors
First Finger
GR KLCIVCS0#.AS C C ItITYC V 1.T C C S CK9 VbF VKU It A VVF. -QisUVYV. MR K1. C L V C C 1) K A S C C It V C V V T C C S C K V F V K K A V K C - - Q I UV1. AR 9 . c1.1 C CSn K.A S C CIIV C A. T C C S CK VV V KUA AKA f C - -Q Vit. PR K C. I C CDVASC CH is IC V 4. T C CS C K V VKU9 N A V K C - -Q isM V I. ER KsYVC A V CU9 OYA S C VII1 V C V W S CKV.C C KAPV F KKA SIQ 9C -If Is OTE10 RAR KP C V V C Q9OK9S S C -1IVVCVSACkCCK9 C FV AU S I Q K 01104 - -E VV T o.c cI K cC v vEE K r I QKcEv .Ps V S T 3 R0 CLCV VCC KATCVItYIC I T c s VOR I I C C v C C U K A K C V T I A N T CV C C K C V VAA S S K U - - K A N T
Second Finger
GR CAC KUOC IK I N UKAUCPACNVUNUC.Q ACEL MR CA CUNC i 1 0 K I XUU KU4 C P A CK.Q K KC. QA CUN . AR C ASUNUC I IK IK V CP5CKLNK CQA;NTI. PR CACKUC IVUI K N UKCPAC.VAKC1.QAC1V. ER CPATQC1)I IxKUXKSCQACUNI.UKCTVVCENK RAR Cit I)KKU I UKI W V TU*UN N C QWVCUN1. QK1(C VFWCEM1 S K T 3 R0 C KVCCCWV I UKVTUQ6CQKCU V K K C IVCAT VOR C P V UC OC K I T K O 0UfXIt C Q A CU0 L I KUCWV11IIC4N K
Gly His - Arg - Tyr Ala Gly Lys - Val Asp Cys Asp Gly Thr - Ala Cys Cycys Cys Val'Zn- _- ,' Arg_ - 0 Gly Zn L Cys Cys Cys- NNye - Lys Cy Cs -PhePhLyAr---Gly---Tyr -Gly Met
Figure 4. Sequence alignment of the DNA-binding domain, C1, of the different steroid receptors. The two zinc-fingers are separated by a linker region of 15-17 amino acids. 9
structure of steroid receptor DNA-binding fingers, as outlined in Figure 5. There are two
major classes of receptors which differ primarily in the three variant amino acids of the P box. The first class, capable of stimulating the same steroid response element, includes the androgen, progesterone, glucocorticoid, and mineralocorticoid receptors. The second class of receptors can be subdivided further into subclasses as these receptors have similar P boxes but different D-boxes. Thus, the estrogen receptor exists in a class related to the thyroid hormone receptors. There are subtle differences in the mode of interaction with the target even though certain receptors may recognize the same or similar response elements within a class grouping (64-66). Such differences emphasize the role of additional amino acids in determining the final mode and outcome of receptor-DNA interactions.
The DNA-binding domain contains also a dimerization function. For the estrogen receptor, a weak dimerization site has been mapped to the DNA-binding domain (67).
Steroid receptors DNA-binding sites
Steroid hormones regulate transcription of responsive genes by interacting with specific sequences in target promoters (68). These sequences known as steroid response elements, SRE, are cis-acting and enhancer-like since they function in a relatively orientation and position-independent manner (69-70). SREs are 15bp palindromes comprising two 6bp 'arms' separated by a variable 'spacer.' The thyroid hormone receptor binds to thyroid response element, TRE, with no 'spacer' nucleotides while the vitamin D receptor can bind to a response element with a 5bp 'spacer' (Figure 5). The dyad symmetry of these elements suggest that nuclear receptors bind to their respective SREs as dimers
(71). Single base changes within these SREs can alter receptor binding and destroy hormone response (72). Surprisingly, as little as two base changes in an SRE can convert a glucocorticoid responsive gene to an estrogen responsive gene by weakening the binding of 10
the glucocorticoid receptor while enhancing the affinity of the estrogen receptor for the element.
SREs for each of the known ligand-activated nuclear receptors have been described (Figure 5). The glucocorticoid response elements (GRE) of the MMTV long terminal repeat (73-75) and the GRE of the chick lysozyme gene (76) were among the earliest described SREs. As more GREs were discovered (72, 77-78), a common consensus sequence
CLASS RECEPTOR P BOX 0 BOX STEROID RESPONSE ELEMENT SOURCE
GR, PR, MR GSCKV AGRND Consen AR GSCKV ASRND GGTACA-N,- TGTTCT sus
i A ER EGCKA PATNO AGGTCA - N1,- TGACCT s n
B TR-alpha EGCKG KYDSC AGGTCA- N,- TGACCT sus TR-beta EGCKG KYEGK _____ GATCA-N.- TGACCrGH
RAR GATCA - N,- TGACC rGH If C (alpha, beta & gamma) EGCKG HRDKN GTTCAC - N,- GTTCAC RAR-beta
VDR EGCKG PFNGD hOsteo I D RGACTCA - N,- TGAACG calcin
Figure 5. Steroid response elements of different steroid hormone receptors. 11
emerged. The estrogen response element, ERE, (79-81) is similar to, but structurally
distinct from, the canonical GRE (82). The consensus ERE is identical with that of TRE
(81), except that the ERE contains a trinucleotide spacer. The same sequences that conferred steroid responsiveness to glucocorticoids could function as androgen response
elements, ARE (82-83), progesterone response elements, PRE (84-85), and as
mineralocorticoid response elements, MRE (21).
It has been shown that the retinoic acid receptor was able to activate gene expression through a thyroid response element (86). The physiological implications of this overlapping enhancer specificity is not clear. By coordinating the levels of thyroid hormone and retinoic acid, a larger range of steroid responsiveness may be achieved since an unliganded thyroid hormone receptor acts as an inhibitor of thyroid and retinoic acid responsive genes, while the unliganded retinoic acid receptor appears to lack this ability.
Steroid-binding domain
The steroid-binding domain has been roughly localized to the C-terminal portion of the receptors and is approximately 25 KDa. This domain contains approximately 250 amino acids. The large number of amino acids in this domain are hydrophobic (87-88). In the estrogen, progesterone, glucocorticoid, and vitamin D receptors, deletion of amino acids, insertional mutations, or point mutations in the C-terminus resulted in a loss of hormone binding activity (18, 12, 40, 47, 89-91).
The hormone-binding domain also contains regions responsive for repression, transactivation, and for dimerization. The repression function of the hormone-binding domain keeps the receptor in an inactive state in the absence of ligand. It has been proposed that interaction with heat shock protein 90, HSP90, is the mechanism by which the hormone-binding domain represses receptor function (5). When the hormone-binding 12 domain is deleted, the receptor becomes a constitutive activator, but activates transcription less efficiently due to the loss of the transactivation and dimerization functions (36, 92-93). The hormone-binding domain of the estrogen receptor has been shown to be important for dimerization (67). The dimerization appears to be hormone dependent, however a direct requirement of hormone for dimerization was never demonstrated. The hormone binding and dimerization are closely localized but separable (94). Moreover, the dimerization site localized to a conserved region of the hormone binding domain. Figure 6 shows that a 14 amino acid region (507-521) is essential for dimerization of the estrogen receptor.
Human Estrogen Receptor
C1 C2 C3 N C
H
0 100 200 Amino Acids
Figure 6. Location of the dimerization motif of the human estrogen receptor. 13
The hormone-binding domain of the glucocorticoid receptor contains nuclear translocation signals. The nuclear translocation of glucocorticoid is hormone dependent
(95). The trans-activation function of hormone binding domains of different receptors will be discussed under transcriptional activation domains.
Transcriptional activation domains
Early studies of receptor structure developed the concept that the highly divergent amino terminus was probably the principle region for protein-protein interactions leading to target gene activation. The nti glucocorticoid receptor (96), a naturally occurring mutant isolated from mouse lymphoma lines which lacks 1-403 amino acids, binds hormone and
DNA, but is unable to activate transcription (97).
Further studies have revealed that the steroid receptors do not posses a single isolated transactivation domain, like the yeast regulatory proteins. In yeast certain transcriptional regulatory proteins have a transactivation domain which is separable from their DNA-binding domain (98-100). But the steroid receptors contain multiple transcriptional activation domains which are located throughout the steroid receptor sequences. Mapping studies using insertional and deletional mutagenesis of the estrogen
(36, 101), progesterone (89, 102) and glucocorticoid receptors (37, 101, 103) have shown that there are transactivation domains within the N-terminal region, the DNA-binding domain, as well as within the hormone-binding domain (Figure 7). The most consistent finding among the three best studied receptors is that a deletion of all but the DNA-binding domain along with about 20 carboxy flanking amino acids results in constitutively active transactivating protein. An 88 amino acid fragment including the DNA-binding domain of hGR was sufficient to activate transcription, but with less than 10% of the efficiency of the full-length receptor (104-105). As Figure 7 shows, the carboxy flanking region has been 14
referred to as tau#2 for hGR (36) and is included in the regions referred to as DNA-tau
(104-105) and enh#1 (106). The tau#2 domain is a 30 amino acid region ranging from amino acid 526-556.
Steroid Receptor Transcriptional Activation Domains lf Glucocorticold N C Glcooticold Receptor
tau1 -- - H oau2 Human
" DNA-tau Human
enh2 F- F--enhl Rat
------~------Mouse
C1 C2 C3 N IC Estrogen Receptor ------TAF1 TAF2 Human ------Mouse
C1 C2 C3 Nr a CmProgesterone N-0 M Receptor ~------
H Human
0 00 200 Amno acds
Figure 7. Steroid receptor transactivation domains. 15
The location of the amino-terminal activation regions of the glucocorticoid receptor varies
slightly among human (tau#1), rat (enh#2), and mouse. The human tau#1 domain is 185
amino acids in length spanning from amino acid 77-262. In the rat glucocorticoid receptor
enh#2, an enhancer region was localized from amino acid 237-318. The transcriptional
activity of the N-terminal and the hormone-binding domain can be observed by fusing
individual domains to a heterologous DNA-binding domain, such as that of the yeast
activator GALA (101, 107). These experiments reveal that the N-terminal activating region
has no intrinsic activity of its own. However, optimal activity of the carboxy transactivation
region required the presence of the N-terminus. Similarly, the N-terminus required the
presence of C-terminal activation region.
The estrogen and progesterone receptors also possess multiple transactivation
domains (89, 108), and individual activation functions appear to cooperate to produce the
full activity of the wild-type receptor.
Steroid Signal Transduction
The binding of hormone to the steroid receptors induces conformational changes in the receptor structure that trigger the process of signal transduction leading to changes in gene activity (109-112). Although there is still a controversy regarding the mechanisms of steroid hormone action such as, the elucidation of receptor functions, interaction with hsp's, binding ligand and DNA, dimerizing, recognizing specific target enhancer elements and promoting the formation of an active transcription complex. There is now convincing evidence that the mechanism of hormone activation involves changes in DNA-binding activity of the receptors, as well as processes that occur subsequent to DNA binding.
Interestingly, the steroid hormone receptors can mediate signal transduction when introduced into yeast cells (113-118). This suggests that accessory components of the 16 signal transduction pathway are ubiquitious and are highly conserved among divergent species.
Effect of hormone on DNA binding
In vitro and in vivo data clearly indicate that there is a definite role of hormone in promoting the association of the receptor with its target enhancer element (119-120).
Hormone is not absolutely required for DNA-binding in vitro. In fact, all the receptors, when exposed to high salt concentration and/or elevated temperatures, can be induced to bind DNA in the absence of hormone. Results from this and other studies showed that cloned mutant hER required hormone for DNA binding, but only under certain conditions (67). However, the recently cloned wild-type hER could not be shown to require hormone for DNA binding in vitro. under conditions where the mutant ER requires hormone (121, D.P. McDonnell, pers. comm.).
In vivo. data suggest that hormone is required for DNA binding in the cell (114
115, 119). A widely accepted view is that DNA binding itself is not dependent on hormone, but that, in vivo., hormone is required for the dissociation of the receptor from inhibitory proteins (121).
However, Bagchi et al., (122) showed that hPR from T47D cells when free of hsp's still required hormone for binding DNA in vitro. Also, Picard et al., (92) demonstrated that the GR function in yeast strains deficient of hsp90 is still dependent on hormone. These experiments suggest that inhibitory proteins not yet identified may be responsible for repressing receptor functions. 17
Effect of hormone on transactivation
The prediction that hormone may be required for the receptor to interact with trans acting factors to promote the formation of an active transcription complex was true for the thyroid hormone receptor group. It has been suggested that these receptors actually repress target gene transcription in the absence of ligand, presumably by binding to the target DNA sequences and preventing the formation of an active transcriptional complex (123).
Recently, several groups have independently developed methods to assay steroid receptor mediated target gene transcription in vitro (124-126). Klein-Hitpass et al., (125) have shown that chicken PR functions in vitro by enhancing the formation of a stable preinitiation complex at the promoter. This activity is not dependent on hormone, as purified unliganded chicken PR can also activate. Similar results have been obtained with the estrogen receptor (127).
Since the receptors have multiple transactivation domains, it is possible that not all of these require hormone for activity. It is suggested that the N-terminal transactivation domain of ER (TAF1) is hormone independent (108). The presence of hormone independent transactivators could explain why hormone is not required for the receptor activity in vitro (125), especially if the function of hormone independent transactivators were artificially dominant under in vitro. transcription conditions.
These experiments do not demonstrate a direct role of hormone in mediating the interactions of the receptors with the transcriptional machinery.
Recent Developments
There are many exciting developments within the fields of steroid receptor regulation, structure and function. The tools to perform expression studies of various 18 receptor mutants, provided by the relatively recent cloning of the steroid receptors, have elucidated some of the structural and functional features of the different steroid receptors as well as the target sequences they regulate.
High-level expression of the steroid receptors has been achieved in bacteria (128 13 1), yeast (113, 116-118), baculovirus (132), and mammalian cells (133).
Overexpression of receptors in these systems allows the production of sufficient amounts of receptor proteins to perform detailed structural analyses such as X-ray crystallography in the absence and presence of DNA. These studies should provide new insight into our understanding of receptor structure and of its interactions with DNA.
In addition, the powerful genetic tools available in the yeast system, will also allow an alternative method of probing receptor structure-function relationships (92, 134) and of probing factors involved in the signal transduction pathways which are highly conserved among divergent species (113, 116, 135).
Also special attention will be directed toward exploiting the relatively recent development of several steroid receptor dependent cell free transcription systems (125-127, 136). CHAPTER 2
MATERIALS AND METHODS
Materials
DNA manipulating enzymes were obtained from Promega Biotec (Madison, WI), Boehringer Mannheim (Indianapolis, IN), or New England Biolabs (Bethesda, MD).
[12511 protein A, [35S] ATP and [3 2P] ATP were obtained from ICN (Indianapolis, IN). The immobilon-P (PVDF) transfer membranes (IPVM 30uRo) were purchased from
Millipore (Bradford, MA) and rabbit anti-rat antibody (IgG) was purchased from Zymed (San Francisco, CA). 1,25- Dihydroxyvitamin D3 and [3H]-1,25-(OH)2D3 were obtained from Duphar (Holland) and Amersham Corp. (Arlington Heights, IL), respectively.
Radioinert 17 beta-estradiol, tamoxifen, nafoxidine, and general reagents were purchased from Sigma (St. Louis, MO). [3 H]-estradiol was from New England Nuclear. ICI 164,384 was obtained from ICI Pharmaceuticals, and LY156 and LY117 were from Lilly. Calf thymus DNA was purchased from Worthington Biochemicals (Freehold, NJ) and used together with N1 Cellulose (BioRad, Richmond, CA) to prepare DNA-cellulose columns as described previously (137-138). Dithiothreitol (DTT) was from Boehringer Mannheim (Indianapolis, IN), while HPLC grade ammonium sulphate was obtained from BioRad (Richmond, CA). Yeast media components were obtained from Difco-BRL (Bethesda, MD). Oxalyticase was obtained from Enzogenetics (Corvallis, OR).
19 20
Bacterial Strains
E. coli strain N99 C1+ (139) was used for cloning purposes. Strains AR58 (a cryptic lambda lysogen derived from N99 that is galE::Tn10, A-8 (chlD-Pgl), A-Hi (cro chiA), N+ and C1857), AR68 (protease deficient strains, derived from AR58), and AR120
(a lambda cryptic N99 derivative that is CI+, A-gal, and ::TnlO; the left arm of lambda distal to the N gene is substituted by host DNA carrying Lac Z; the arm contains the T 11 lesion that inhibits cro expressions A-(c t1-UVrB) were employed for expression studies
(140).
Yeast Strains
The expression of recombinant proteins was carried out in the protease deficient strains BJ3505, (Mat a, pep4::His3, Prbl-A.GR, his3, lys2-208, trpl-A1O1, ura3-52) or
BJ2168, (Mat a, prcl-407, Prbl-1122, pep4-3, leu2, trp1, ura3-52). These were obtained from the yeast genetic stock center (Berkley, CA). All transformations into these strains were done using the lithium acetate transformation protocol (141).
Cell Lines and Human Tumor Specimens
The ER+MCF-7 breast cancer cell line and CV-1 cell line were grown in Eagle's minimal essential medium with 10% charcoal-stripped fetal bovine serum. Frozen human breast tumor specimens, stored at -700C, were obtained from the San Antonio Breast
Tumor Bank and consisted of tissue remaining after receptor assays performed in the San
Antonio Clinical Receptor Laboratory. 21
DNA Construction
Bacterial expression vectors
The expression vector PUB-VDR was constructed as follows:
A synthetic linker with the sequence AATTGCTTAAGACTAAGAGGTGGTGGAATTCGGTAC CGAATTCTGATTCTCCACCACCTTAAGC
which encoded the carboxyl six amino acids of ubiquitin and contained an internal EcoRi
site, was synthesized. This linker was cloned into the EcoRl-Kpn1 site of PGEM4. The resulting plasmid, PG42, was digested with EcoRi, and the 2.0 kbp EcoR1-EcoR1
fragment of PHVDR13 (142) was inserted into that site. In the correct orientation, this gave
plasmid BCPV1. After amplification, the AflII-Kpnl fragment was excised and subcloned
into the corresponding sites of an E. coli human ubiquitin expression vector (143). In the
final construct, the human ubiquitin-vitamin D3 receptor fusion (PUB-VDR) gene was
under the control of the lambda phage PL promoter.
The human estrogen receptor cDNA clone PGEM ER35 (144) was digested with
BamH 1 and TthIII1 to release the complete translation frame. The TthIII1 site is 30 bases
upstream of the ER initiator methionine codon. To facilitate the construction of ubiquitin
carboxy terminus fusion with ER, an AflII-TthIII oligonucloetide linker was synthesized
that contained six amino acids of the carboxy terminus of ubiquitin followed by a sequence
that reconstructed the TthIII1 site, thus, the amino terminus of the ER protein. The
oligonucleotide linker sequence is:
5' GATCCC TTAAGACTAAGAGGTGGT TCCATGGGGACA 3'
Afl II UB-ER Junction TthlII1
The resulting fusion has an authentic carboxy terminus of ubiquitin but contains an extra 14 amino acid sequence SMGTRSAPCPRSRT before the authentic initiator 22
methionine of the ER. The resulting plasmid was called BCPEL. A Kpn1 site was
engineered at the 3' end of the cDNA by linearizing BCPE1 with EcoRi and filling the site
with a Klenow DNA polymerase fragment. The plasmid was recircularized in the presence
of an oligonucleotide that contained a Kpn1 site. The resulting plasmid BCPE2 was
digested with AflII-Kpnl and the ER DNA was inserted into an AflIL-KpnI digested E.
coli human ubiquitin expression vector (143). In the final construct, the human ubiquitin
estrogen receptor fusion (PUB-ER) gene was under the control of the lambda phage PL
promoter.
The original cDNA clone obtained from MCF-7 cells contained a point mutation (a
cloning artifact) which resulted in the substitution of a valine for a glycine at the amino acid
position 400 (108). This mutation in the hormone-binding domain of the ER decreased its
affinity for estradiol binding at 250C but did not appear to alter it at 40C when compared
with that of the wild-type receptor (108). In this study the expression of the mutant ER was
examined in E.coli.
Yeast expression vectors
The yeast expression vector YEPE10 expressing human estrogen receptor was
constructed as follows: the human estrogen receptor plasmid (PGEM35hER, a gift from
Geoff Greene) was digested with BamHl and TthIII 1, and ligated to a linker which
encoded the last six amino acids of ubiquitin and contained also an AflI1 and a Ncol site.
The EcoRi site of the resulting plasmid was converted to a Kpn1 site. The AflII-Kpn 1 fragment of this vector was purified and cloned into the corresponding sites of the yeast expression vector YEPV1, yielding the plasmid YEPE2. This plasmid will produce receptor with 14 additional amino acids at the amino terminus. A subsequent publication demonstrated that the original cloned cDNA contained a critical point mutation in the 23
hormone binding domain (108). The mutated hormone-binding domain of the vector
YEPE2 was corrected by replacing the ASP718-Smal fragment of this vector with the
analogous fragment from the wild-type cDNA (a gift from Dr. Geoff Greene, University of
Chicago). The resulting plasmid was called YEPEIO.
The ER(wt) expression plasmid (YEPKB1) was constructed from YEPE10 by
using site-directed mutagenesis to remove the extra N-terminal 14 amino acids. YEPKB 1
was kindly provided by Gisela Nilsson (Karo Bio, Sweden).
An integrative version of YEPE10 plasmid, YIPE10, was constructed by cloning
the BamHI-MluI fragment of this plasmid into the cognate sites of the vector PYSK152 (a
gift from Dr. G. Gorman, Smith, Kline, and Beecham, Inc.). Integration into the yeast
genome was accomplished by linearizing the plasmid at a unique BstEl11 site within the
leucine2 sequence and selecting for tryptophan auxotrophs.
Another derivative of YEPE1O, YEPE15, was constructed by cloning the Hpal-StuI
fragment of PLQ6 (a gift from Dr. T. Butt, Smith, Kline, and Beecham, Inc.), which
contains a defective leucine2 allele, into the Csp45 end-filled YEPEIO.
To construct YEPE23, the BamHI-XhoI fragment of YEPE10 was replaced with a similarly digested fragment of PYSK153 (a kind gift from Dr. George Livi, Smith, Kline,
and Beecham, Inc.). This construct expresses wild-type ER under the control of a
constitutive yeast triose-phosphate dehydrogenase (TDH) promoter (145). To change the
selection marker from tryptophan to leucine, this construct was next digested with Csp45
and Clal, and the vector fragment repaired with Klenow, followed by ligation to an end
filled SalI fragment of PM853 ( a kind gift of Dr. G. Gorman, Smith, Kline, and Beecham,
Inc.), yielding YEPE23.
The plasmid YEPAR and the yeast ER-VP16 expression plasmid were gifts from
Tony Pham. 24
To construct the variant ER expression plasmid missing exon 5, a portion of BCPE2 was replaced with a HindIII-BglIl ER PCR-derived fragment from the ER-/PR+
tumor. An AflII-KpnI fragment from this intermediate vector was then inserted into the
similarly digested YEPE2 to generate the variant ER expression vector YEPE9.
To reconstruct the variant ER expression plasmids missing exon 7 in an inducible
yeast expression vector system, a portion of the wild-type ER yeast plasmid YEPE10 under
the control of metallothionein promoter was replaced with a HindIII-BglII ER PCR-derived
fragments from ER+/PR- tumor expressing the exon 7 ER deletion variants. This generated
the variant ER expression vectors YEPE13 and YEPE15. The variant expression plasmid
YEPE13 has a point mutation converting valine to alanine (at the amino acid position 422)
in the hormone binding domain, otherwise this variant ER is similar to the variant
expressed from the plasmid YEPE15. The sequences of YEPE13 and YEPE15 were
confirmed by dideoxysequence analysis (146).
Yeast reporter vectors
Reporter plasmids were constructed as follows:
The parent vector PC2 for these constructions has previously been described (116
117). The plasmid PC2 was modified by the insertion of an EcoRI site and a BgII site into
the unique XhoI site, thus creating the vector PC3 containing a mini-cloning cassette.
The yeast reporter plasmid YRPE2 was constructed from PLG67OZ (147) by inserting a 75-base-pair oligonucleotide (containing two estrogen responsive elements) into the XhoI site upstream of the CYC 1 (iso- 1-cytochrome c) promoter. To construct the reporter plasmid, YRPD2, in which the two receptor binding half sites are aligned as direct repeats, two copies of an oligonucleotide were inserted into the unique BgII site of the 25
plasmid PC3 (114). The oligonucleotide sequenceis as follows: GATCCTAGAGGTCACAGGGTCATACGA.
The copy number and the orientation of the inserted sequences were determined by
sequencing.
In order to yield GAL4ERE and GAL4VDRE reporters, the oligonucleotides were constructed with BamHI ends and inserted into the plasmid PC3 (114) as single and multiple copies. The number of oligonucleotides inserted was determined by restriction analysis, and the orientation was confirmed by sequencing. The sequences of the oligonucleotides used were as follows:
GAL4ERE: -GATCAGGAAGACTCTCCTCCGGTCAGTGACC
GAL4VDRE: -GATCCAGGAAGACTCTCCTCCGAATGAACGGGGGCA
The reporter plasmid YRPV4 carrying the osteocalcin gene promoter was constructed by inserting Xhol-Sall fragment of the human osteocalcin gene promoter into the unique XhoI site of PC2 (116-117). To generate 5' and 3' deletions in the human osteocalcin gene promoter, the 1-kbp BamHI fragment of YRPV4 was cloned into the
BamHI site of PGEM-4Z vector in both (+) and (-) orientations. Both 5' and 3' deletions were performed in the osteocalcin gene promoter by using the Promega erase-a-base system. All deleted fragments were excised from PGEM-4Z vectors by BamHI-HindII, and repaired by Klenow, then inserted into the end-filled XhoI site of the vector PC2 (116
117).
Mammalian expression and reporter vectors
The ER and ER-VP16 mammalian expression plasmids were constructed as described (148). The ERE-tk-CAT reporter plasmids were constructed by inserting oligonucleotides with BamHI-BglII ends into the BamHI site of the plasmid PBLCAT2 26
(80). The number of oligonucleotides inserted was determined by restriction analysis, and
the orientation was confirmed by sequencing. The sequences of the oligonucleotides used were as follows:
EREwt: -GATCCTAGAGGTCACAGTGACCTACGG
ERE4: -GATCTGTGGTTCGGCTCGTGCGAAACGA
ERE9: -GATCTCGGGGTCAGGCGGTCCCCTACGA
Induction of Recombinant Proteins in Bacteria
The transcription of the genes controlled by the lambda PL promoter was induced in
E.coli AR58 and AR68 cells by elevating the temperature of the culture media. E. coli
strains AR58 and AR68 are cryptic lambda lysogens derived from N99, that contains a
temperature sensitive mutation in the phage lambda CI gene (CI857). This allows the PL
directed transcription to be activated by temperature shift from 300 C to 420 C (139).
Reducing the temperature to 300C presumably renatures the CI repressor which terminates
transcription. As an alternative to the thermal inactivation of the CI857 repressor, nalidixic
acid was added to a cryptic lysogen (AR120) that contains the CI+ repressor at 300C or 0 37 C (140). For chemical reaction studies, E. coli AR120 cells were grown at 300C to
OD550= 1.0 and gene transcription was induced by the addition of nalidixic acid at a final
concentration of 40 ug/mI by incubation at 300C for 4-5 h. The induced cells were
harvested by centrifugation and washed twice with 5 ml of ice cold buffer (10 mM Tris
HCI pH 7.5,5 mM DTT). For PAGE analysis, the final pellet was suspended in 1% SDS
PAGE loading buffer. The cell suspension was mixed vigorously on a vortex to lyse the
cells, and then boiled for 10 min. The cellular debris was removed by centrifugation at 10,000xg for 10 min at 40C. The supernatant was recovered for protein determination and subsequently 5-25 ug of proteins were analysed on SDS PAGE. 27
Western Immunobloting of Proteins Expressed in Bacteria
5-25 ug of proteins were analyzed on SDS PAGE. The UB-VDR and UB-ER fusion proteins were detected by Western blots as described previously (149). Blots were probed with either antibodies directed against ubiquitin or with monoclonal antibodies
(9A7) directed against human VDR, and H222 or D75 directed against human ER (H222 and D75 antibodies were kindly provided by Dr. Geoff Green, University of Chicago).
Hormone Binding Determination of Bacterial Expressed Proteins
Hormone binding assays of vitamin D receptor
After induction, both the induced and uninduced cells were collected by centrifugation and washed in assay buffer (10mM Tris-HCl, 1.5 mM EDTA, 10% glycerol, 400 mM KC1 and 5mM DTT). The washed cells were broken by sonication 4 times, 30 seconds each time, using a Fisher Sonic Desimembrator Model 300, followed by centrifugation at 40,000 rpm for 30 min in a Beckman 70.1 Ti rotor. The supernatant was used as the source of recombinant receptor. Single point assays were done by incubating the supernatant with 1 nM radiolabeled 1,25(OH)2D3 with or without a 100-fold molar excess of radioinert hormone. After a 2-4 h incubation at 40C, specific binding was determined by hydroxyapatite binding assays as described previously (149). Saturation analysis was carried out analogously, using increasing concentration of labeled ligand and overnight incubation. 28
Hormone binding assays of estrogen receptor
After four hour induction period in AR120 cells, both the induced and uninduced cells were collected by centrifugation and washed in assay buffer (either 10 mM Tris-HCl pH 7.4, or potassium phosphate buffer pH 7.4 containing 1.5 mM EDTA, 10% glycerol, 400 mM KCI, 5 mM DTT, and 10 mM sodium molybdate). The washed cells were broken by sonication 4 times, 30 seconds each time, using a Fisher Sonic Dismembrator Model
300, followed by centrifugation at 40,000 rpm for 30 min in a Beckman 70.1 Ti rotor. The supernatant was used as the source of recombinant receptor. To evaluate the ligand binding characteristics, the supernatant was incubated 16-20 h at 40C with increasing concentration of radiolabeled estradiol in the presence and absence of radioinert estradiol at a 200-fold excess. 17 beta-PH] estradiol served as a specific radiolabeled ligand. Dextran-coated charcoal suspension was used to remove unbound estradiol (150).
DNA Binding Properties of Bacterial Expressed Proteins
DNA binding properties of vitamin D receptor
The protein from recombinant cells was prepared as described above and labeled for 2 h with I nM [3 H]-labeled 1,25(OH2)D3. The salt concentration was lowered by dilution, and the protein fraction was loaded onto a 5 ml calf thymus DNA-cellulose column. After extensive washing, the specifically bound proteins were eluted by using a linear gradient of
KCl. The protocol is described in more detail elsewhere (151). 29
DNA binding properties of estrogen receptor
Protein extracts from induced and uninduced cells were prepared as described
above. The estrogen response element, ERE, from Xenopus vitellogenin A2 gene was used
as a wild-type ERE (152) with the sequence shown:
5' GTCCAAAGTCAGGTCACAGTGACCTGATCAAAGTT 3' 3' CAGGTTTCAGTCCAGTGTCACTGGACTAGTTTCAA 5'
The oligonucleotides were end labeled with polynucleotide kinase. The binding reaction
contained 10 mM Tris-HCl pH 7.5,80 mM KCl, 5 mM DTT, 1 mM EDTA, 0.5 ng of 3 2 [ pj oligonucleotide (100,000 cpm), and 10% glycerol. To ensure that the binding to the
ERE of E. coli expressed UB-ER was specific, one-to-two thousand fold excess of
nonradioactive PhiX174, HaeIIl digested DNA was added to each tube as a competitor.
The reaction was initiated by adding 5 ug of the protein extract and the samples were
incubated for 15 min at room temperature. Nucleoprotein complexes were separated on 7%
polyacrylamide gels containing 89 mM Tris-HCl, pH 7.5,89 mM boric acid and 2 mM
EDTA buffer. Electrophoresis was performed at 30 mA until the bromophenol blue was 2
inches from the bottom of the gel. The vacuum dried gels were examined by autoradiography to identify the associated receptor species.
Induction of EstrogenReceptor in Yeast and Preparation of Yeast Extracts
Yeast cells carrying receptor expression plasmid were grown overnight in minimal media containing 2% glucose and essential amino acids. When the cells reached an optical density of 1.0 at 600 nm they were induced with 100 uM CuSO4 and allowed to continue growing for 4-8 h. When cells were grown in a 51 fermenter, they were induced at lower optical densities. Cells were harvested by centrifugation and washed twice with water. The 30
pellet was resuspended in 1.2 M sorbitol, 40 mM KPO4, 20 mM beta-mercaptoethanol (pH
7.4) and incubated at 300C for 30 min. The cells were again pelleted and redissolved in the
same buffer (5 mg/i). The enzyme oxalyticase was added at a concentration of 30 ug/ml and
incubated at 300C for 90 min. The mixture was washed twice with sorbitol buffer to
remove the enzyme and the cells were lysed via hypotonic shock by redissolving the pellet
in 10 mM Tris-HCl pH 7.4, 1 mM EDTA. The cytosol was then clarified by centrifugation.
Preparation of MCF-7 Estrogen Receptor
MCF-7 cells were grown in DMEM medium containing 10% stripped fetal bovine
serum without phenol red. The cells were harvested by scraping, washed twice with ice
cold PBS (Phosphate Buffered Saline) and resuspended in 2X packed cell volume of 10 mM HEPES, 1.5 mM MgCI2, 10 mM KCI, 2 mM DTT, pH 7.9 (Buffer A). After 10 min,
the cells were homogenized using a B pestle in a dounce homogenizer and 1/9 volume of 5
M NaCl in buffer A was added. The sample was incubated on ice for an hour with
occasional gentle vortexing. Salt extract was prepared by centrifugation at 15,000 rpm for
30 min in a TiSO rotor. The supernatant was labeled with 40 nM [3 H] estradiol overnight at
40 C and free steroid was removed by stripping with charcoal.
Hormone Binding Analysis of Estrogen Receptor from Yeast
To evaluate ligand binding characteristics, yeast extract was incubated 16-20 h at
0 4 C with increasing concentration of radiolabeled steroid in the presence and absence of radioinert steroid at a 100-fold excess. l7beta-[3H1 estradiol served as a specific radiolabeled steroid. Specific binding was determined by the hydroxyapatite binding assay (149). 31
DNA Binding Analysis of Estrogen Receptor from Yeast
DNA-cellulose chromatography
The extract from yeast cells carrying receptor expression plasmid was prepared as described above and labeled for 2 h with 1 nM 3 H-labeled 17beta-estradiol. The salt concentration was lowered by dilution, and the protein fraction was loaded onto a 5 ml calf thymus DNA-cellulose column. After extensive washing, the specifically bound proteins were eluted by using a linear gradient of KCl (151).
Band shift assays
Gel retardation assays were performed essentially as described above using the consensus vitellogenin A2 ERE oligonucleotide probe labeled with gamma-3 2P-ATP by means of T4 polynucleotide kinase. 20 ug of the protein extract was used with 1.7 ug of
Hinf1 digested PBR322 plasmid and the nucleoprotein complex was analysed on 4% polyacrylamide gels.
Western Blot Analysis of Estrogen Receptor from Yeast
The protein samples used for Western blot were obtained from the yeast cells transformed with receptor expression plasmids and induced with copper sulphate for 4 h as described above. The cells were then harvested and lysed by glass-bead homogenization in
Z buffer (117) with 0.3 M NaCl. Western blot analysis was performed as described (117) except that a human ER monoclonal antibody (D75 or H226, a gift from Dr. Geoff Greene,
University of Chicago) was used. 32
Antibody Immunoprecipitation of Estrogen Receptor
Increasing concentrations of OV245 antiserum were incubated for 30 min at 40 C with fixed volumes of either MCF-7 cytosol or yeast extract to a final volume of 200 ul. The samples were assayed using a Protein A Sepharose column assay as described (153).
Immunodot Analysis of Estrogen Receptor
Cytosols were prepared as described above and dotted onto nylon membranes. All samples and dilutions were adjusted to contain 5 ug of total protein. Bovine serum albumin was used as a carrier. The filter was then subjected to a solid-phase radioimmunoassay as described previously (117), using H226 or D75 as primary antibodies.
Assay of ER Function in Yeast
Cells containing the receptor expression plasmid and the reporter plasmid were grown overnight in selective medium. This culture was used to inoculate fresh medium at a density of 100,000 cells per ml. For experiments requiring activation of GAL4, galactose was used as a carbon source instead of glucose. Estrogen and nafoxidine were added as required at 2 X 10-7 and 2 X 10-6 M, respectively. When the cells reached an optical density of 1.0 at 600 nm, they were harvested and the level of b-galactosidase produced was measured as described previously (116-117) except that 300 mM KCl was added to allow the extraction of the receptor. 33
Construction and Propagation of (SRE)n Library and Development of the Screen for Identification of SREs
Single-stranded oligonucleotides corresponding to the sequences below were synthesized. Degeneracy was achieved by using a mix of all four bases at each random position (N). The oligonucleotide sequence is:
5' GCATGCATAGATCTNNNGNNCNNNNNGNNCNNNGGATCCAAGCTTCG 3' The degenerate oligonucleotide was converted to double stranded DNA (154-155)
using a 14-mer primer complementary to the 3' end of the degenerate oligonucleotides. A library of 40,000 clones was prepared in E. coli by cloning the double-stranded
oligonucleotides into the unique BglII site of the yeast reporter plasmid YRPC3 (114), yielding YRP(SRE)n. DNA was prepared from the E. coli library and then transformed into the yeast strain BJ3505 carrying the receptor expression plasmid. Transformants were identified by tryptophan and uracil auxotrophy on plates containing X-gal and 17beta estradiol.
Transient Transfection Assays
Transient transfections into CV- I and MCF-7 cells and chloramphenicol acetyltransferase (CAT) assays were performed exactly as described (148). CHAPTER 3
RESULTS
Development of a Bacterial Expression System for Steroid Receptors
To facilitate the expression of full length human vitamin D3 and estrogen receptors in E. coli, the carboxyl-terminal glycine of the ubiquitin gene was fused to the amino terminus of the receptor genes. The expression of recombinant proteins was placed under the control of the lambda PL promoter (Fig. 8). Induction of recombinant proteins in this system is accomplished by the inactivation of the lambda repressor and the subsequent activation of the PL promoter. Inactivation of the repressor can be mediated in either of two ways: by heat inactivation of a heat sensitive lambda repressor; or by using
DNA damaging agents that induce the cellular SOS response (a protease component of which degrades the repressor) (130).
In this study both methods of induction were analyzed to determine which one was the most suitable for the production of recombinant human vitamin D3 and estrogen receptors in E. coli. For those studies, using heat induction, the expression vectors were introduced into the E. coli strains AR58 and AR68. For chemical induction studies, the expression vectors were transformed into the E. coli strain AR120.
Expression and Biochemical Characterization of Vitamin D3 Receptor
The expression plasmid PUB-VDR was introduced into the E. coli strain AR58. 0 The cells were initially grown at 30 C until they reached an optical density of 1.0 at 550
34 35
nm. At this time, the temperature of the media was elevated to 420C. At regular intervals following induction, the cytosols were examined by Western immunoblotting for the
production of a ubiquitin-vitamin D3 receptor protein. The results of this analysis are shown in Figure 9. There is no detectable receptor produced in the absence of induction.
However, receptor is detectable 15 minutes after induction, and maximal induction is observed by 90 minutes.
ClIr.b.s. Cllr.b.s. Ndel PL. Ndel
AMil Afill
Ampr pUB-VDR) Amp' pUB-ER
Kpni KpnI
Figure 8. Bacterial expression vectors. The expression of ubiquitin-hVDR and ubiquitin-hER fusion genes were placed under the control of lambda PL promoter (PUB-VDR and PUB-ER). 36
Previously, the vitamin D3 receptor was produced in E. coli in an unfused form.
This receptor was incapable of binding hormone or DNA, and was totally insoluble
(McDonnell, Ph.D Thesis). To examine the solubility of the protein produced in this
system, the cytosol from the induced cells was clarified by centrifugation (10,000xg). A
Western analysis of the soluble and pelleted fractions demonstrated that about 30% of the
receptor was in the soluble form (Fig. 10A). However, as before (McDonnell, Ph.D.
Thesis), this material was unable to bind hormone.
As an alternative approach, the PUB-VDR expression vector was introduced into
the E. coli strain AR120. This strain has a wild type lambda repressor. The cells were
grown to the required optical density at 300C, and the induction of the PL promoter was
initiated by the addition of nalidixic acid. The kinetics of the nalidixic acid mediated
transcription of this promoter were slower than heat induction [data not shown, (130)].
Following induction, cytosols were prepared as described in "Materials and Methods". A
Western immunoblot analysis of receptors produced in this strain is shown in Fig. 10B.
As it was the case using AR58, there was no receptor synthesis in the absence of
induction. However, following induction, a distinct immunoreactive band is observed at
the expected molecular weight. As with the heat induction protocol, there is a significant amount of proteolytic degradation of the receptor. It was also noticed (data not shown) that the fraction of the receptor produced as soluble protein is unaffected by the induction protocol.
One significant difference between the receptors produced in the two systems was that the soluble material produced in AR120 was capable of specifically interacting with
1,25(OH)2D3. The measured binding affinity of this material was similar to that reported for the wild-type receptor (Fig. 1I1A). This analysis did not detect any forms of receptor with intermediate binding affinities, suggesting that the soluble receptor may be 37
quantitatively active. The total specific binding activity was disappointingly low (600-800
fmole of receptor per mg of protein). The DNA binding properties of the E. coli produced receptor were analyzed by calf thymus DNA-cellulose chromatography. The results shown
in Fig. 1 lB demonstrate that the peak of the hormone labeled receptor elutes at 0.22M KCl, which is consistent with what is observed for the wild-type receptor.
Western Blot Probed with Antibody Against UB
Mins. after u 5 10 15 30 45 60 90 120 induction
4W I - UB-VDR
& - UB
L Ol A.R. 58
Figure 9. Expression of UB-VDR in E. coli strain AR58 and its analysis by Western immunoblotting.
Receptor synthesis was induced by elevating the temperature from 300 C to 420 C. Aliquots of 25 ug of protein from each culture were analyzed on 15% gels using PAGE. Immunodetection by Western blot was performed using a monoclonal antibody against ubiquitin. 38
Western Blots Probed with A. Antibody Against VDR U I S. P. S. P.
-UB-VDR
LC.LaA.R. 58
B. U I I U I
- UB-VDR
VDR -
L CaL A.R. 120
Figure 10. Western immunoblot analysis of the soluble form of UB-hVDR produced in E. coli. (A) Crude
lysate from heat induced AR58 cells was clarified by centrifugation, and Western analysis of soluble and
particulate fractions was performed. (B) The PUB-VDR expression vector was transformed into AR 120
strain, and transcription 0 from lambda PL promoter was induced at 30 C by addition of nalidixic acid.
Aliquots of 25 ug of proteins were analyzed on 15% PAGE. Western blots were probed with a monoclonal antibody against hVDR. 39
0.3
0.25
0.2 Kd= 2.8 x1O''M
V 0.15
0 0.1
0.05
0 1 2 3 4 5 8 7 8 Bound (fmole)
20 'Q
0.3
-0.1
0 I *0 0 I
FPacula Nambsr
Figure 11. Analysis of functional properties of recombinant receptor. Protein extract containing the
recombinant receptor was analyzed by Scatchard method. (A) No specific binding of 1,25 (OH)2D3 was
observed in cell extracts from non-transformed cells. (B) The DNA binding properties of the receptor
analyzed by calf thymus DNA-cellulose chromatography. The bound proteins were eluted with a linear
gradient of KCL. 40
Expression and Biochemical Characterization of Estrogen Receptor
To express the estrogen receptor, the expression plasmid PUB-ER was introduced
into E. coli strains AR58 and AR68. Receptor synthesis was induced by elevating the
incubation temperature to 420C, and samples of total protein from cells either uninduced
(U) or induced for 10, 15,30, 60, and 90 minutes, were analyzed by SDS-PAGE using
gels with 15% cross-linking (Fig. 12).
The expression of the fused proteins was detected by Western immunoblot. The
analysis of the recombinant estrogen receptor synthesized in the AR58 strain indicate that
the expressed ubiquitin-ER was rapidly degraded and intact receptor was not present even
at early times of induction (Fig. 12A). In AR68, a protease-deficient strain, synthesis of the
receptor was observed shortly after induction although only small quantities of intact
receptor were seen in response to induction by the 420C temperature shift (Fig. 12A).
Analysis of the solubilized proteins from induced AR58 or AR68 cells by ligand
binding assays showed that neither strain exhibited specific estrogen binding (data not
shown). The PUB-ER expression vector was transformed into AR120 E. coli where the
lambda PL promoter transcription was induced at 300C by the addition of nalidixic acid.
The kinetics of the nalidixic acid-mediated transcription of the PL promoter was slower than the transcription induced at 420 C in AR58 and AR68 cells. The western blot analysis of an extract from AR120 cells showed that a considerable amount of intact ubiquitin-ER protein was produced from 120 to 240 min after the addition of inducer (Fig. 12B). No. further increase in the quantity of UB-ER was observed after 240 min of induction in
AR120 cells.
Only a small fraction of the receptor was degraded under these conditions (Fig.
12B). Western blots probed using human ER antibodies gave similar results (data not shown). Although the fusion of the ER gene to that of ubiquitin substantially increased the 41
Western Blots Probed with Antibody Against ER A.
Mins. after U 10 15 30 60 90 u 10 15 30 60 90 induction
- UB-ER
0
LCIL A.R. 58 L CJiLA.R. 68
B. Mins. after u 10 30 45 60 120 180 240 induction - UB-ER- o
E, cL A.R. 120
Figure 12. Expression of UB-ER in various E. coli strains and its analysis by Western immunobloting. (A)
Receptor synthesis was induced in AR58 and AR68 cells by elevating the temperature from 300 C to 420 C.
Aliquots of 25 ug of protein from each culture were analysed on 15% gels using PAGE. (B) Receptor synthesis in AR120 cells was induced by the addition of nalidixic acid, and 25 ug aliquots of protein were analysed on 15% gels using PAGE. Immunodetection of receptor by western blot was performed using a monoclonal antibody against ubiquitin. 42
synthesis of ER in AR58 and AR68 strains of E. coli, these results suggest that thermal
induction of UB-ER produced largely inactive UB-ER. Therefore, the characterization of
E. coli-produced ER was performed on AR120 cells induced for 4 h with nalidixic acid to
ascertain if various domains remained intact.
To evaluate the intactness of the ligand-binding domain of the ER, protein
expressed in E. coli was conducted using [ 3 H]estradiol. In each experiment, several
control samples of wild type receptor were assayed to ensure method performance. The
binding isotherm resulting from ligand saturation analysis with [3 H]estradiol (Fig. 13A)
clearly demonstrated wild-type characteristics. Scatchard analyses of these data indicated
only one type of binding site with Kd value of 3.2 X 10-10 M. This Kd value is identical to
that of native ER in human tissues (89, 150). A representative Scatchard plot is shown in
Fig. 13B. Specific ligand binding capacities of ER from extracts of induced E. coli usually
ranged from 400-1000 fmol/mg of extract protein. One of the factors contributing to the
variability in ER detected appears to be the efficiency of induction and the ability to extract
protein following sonication. Under the present extraction conditions, it was possible to solubilize 30% of the expressed UB-ER as analyzed by Western immunobloting.
To demonstrate that E. coli-produced UB-ER molecules retained the ability to associate specifically with ERE, receptor extracts were incubated with [3 2 P]-labeled synthetic Xenopus vitellogenin A2 gene ERE (152) in the presence of a thousand-fold excess of nonradioactive, HaeIII-digested PhiX174 DNA. The nucleoprotein complexes formed were separated from the unassociated oligonucleotide on polyacrylamide gels.
Results presented in Fig. 14 indicate that protein extracts from uninduced cells did not appreciably retard the migration of the labeled ERE (lane 2). When receptor synthesis was induced in E. coli, a significant increase in the band shift of [3 2 P]-labeled ERE was observed (lane 3). Increasing concentrations of nonradioactive ERE in the reaction mixture 43
effectively competed with the labeled ERE (lanes 4-7). In this experiment, the DNA probe was labeled with polynucleotide kinase and [3 2PATP; as a result of the endogenous 3 2 alkaline phosphatases present in the E. coli extract, some cleavage of the terminal P was observed (Fig. 14). When 2,000-4,000 ng of HaeIIL-digested PhiX174 DNA was used as
500 A
. 400
a. 1,300-4 E
~200* 00
z o 100 Cn 0 100 200 300 400 500 BOUND (fmo/mg protein) 0 0.0 0.5 1.0 1.5 2.0 2.5 [3 H-ESTRADIOL] (nM)
Figure 13. Ligand titration curve and Scatchard plot analysis of ER expressed in E. coli cells induced by
3 nalidixic acid. (A) ER in the extract was detected by titration with increasing concentrations of [ Hlestradiol in the presence and absence of unlabeled estradiol. Only specific binding is plotted showing saturation of the receptor at approximately 1.0 nM ligand. (B) Scatchard analysis of these data gave a straight line indicating one type of binding site. The specific binding capacity determined was 484 fmol/mg of soluble extract protein with a Kd = 3.2 X 10- 1 0 M. 44
Competitor DNA (ng)
ERE 0X174 DNA
Bound -+
Free -+55
1 2 3 4 5 6 7 8 9
Figure 14. Specific binding of E. col-expressed human estrogen receptor to estrogen response elements. E.
coli cell extracts were prepared and 5 ug of extracted protein was incubated with 100,000 cpm of [3 2P
labeled estrogen response elements. Free labeled ERE and the receptor associated ERE were separated on a
7% gel by PAGE and visualized by autoradiography. Lane 1, labeled ERE only. Lane 2, ERE incubated with 5 ug of protein extract from uninduced E. coli. Lane 3, ERE incubated with 5 ug of protein extract
from induced cells. Lanes 4,5,6, and 7, labeled ERE incubated with induced cell extract and increasing amounts (ng) of unlabeled ERE. Lanes 8 and 9, protein extracts from induced cells were incubated with labeled ERE in the presence of 2,000 and 4,000 ng of PhiX 174, HaeIII-digested DNA. 45
a nonspecific binder, no competition was observed with the labeled ERE (lanes 8-9). These
studies demonstrate that E. coli-expressed ubiquitin-ER retains its specific ERE binding
properties.
These studies demonstrate that the vitamin D3 and estrogen receptors can be
produced in a biologically active form in bacteria and indicate that all the processes required
for the production of the hormone binding species are present. These studies further
demonstrate that the utilization of the nalidixic acid induction protocol can substantially
improve the quality of the recombinant receptors. However, the use of this system to
overproduce receptor for crystallization studies is not practical due to the low recovery of receptor proteins and due to heterogeneity as a result of proteolysis. Furthermore, the
receptor proteins produced in bacteria are not post-translationally modified, thus, they cannot be used for transcriptional studies. These limitations demonstrated the need to develop a system where one can overexpress intact receptor proteins and study the regulation of eukaryotic genes by steroid hormone receptors. To overcome these problems an expression system in Saccharomyces cerevisiae was developed.
Development of Saccharomyces cerevisiae as a Host System for Steroid Hormone Action
Saccharomyces cerevisiae, also known as Baker's yeast, is the most extensively genetically characterized eukaryote. A genetic map will soon be correlated with its complete physical genomic map. With a total haploid genome of about 15 Mb (only 3.5 times greater than that of E. colt), it displays most of the features of higher eukaryotes. Recent demonstrations of apparent functional similarity of mammalian and yeast RNA polymerase
11 (156), and the functional similarity between yeast and mammalian TATA (157) and
CCAAT binding proteins (158) supports that many of the yeast cellular processes are mechanistically conserved among different eukaryotic species. These similarities, coupled 46
with the powerful genetic tools available to Saccharomyces cerevisiae, have quickly made
yeast the organism of choice for many problems in eukaryotic molecular biology.
Saccharomyces cerevisiae can be successfully used as a means of overexpressing
heterologous proteins (159). Most importantly, these products are authentically post
translationally modified (i.e. phosphorylation, acetylation, glycosylation, etc.) by the host
yeast cell (160-161). These modifications are usually necessary to confer full biological
activity or to ensure proper localization within the cell. This fact, coupled with the recent
demonstrations that yeast enhancers and yeast enhancer binding proteins function in
mammalian cells (162), and that steroid receptors are fully functional in yeast (113-117),
led to the engineering of reconstituted hormone responsive transcription units, which will
ultimately permit an efficient genetic microdissection of steroid receptor structure.
Additionally, maintenance of yeast stocks is cost-effective and decreases the need
for animal models, and/or the use of animals as the sole source of receptor protein.
Production of Mammalian Steroid Hormone Receptors in Yeast
Saccharomyces cerevisiae is one of the few eukaryotes to possess a natural non
viral plasmid, the 2 um circle, which is capable of attaining a copy number of 50-100 per
cell. Furthermore, the plasmids are stably inherited so that most cells in a culture contain
this high copy number plasmid. These properties make the 2 um circle an excellent basic
molecule for the construction of a highly tractable gene expression vector.
A 2 um based ubiquitin fusion system was developed, which allows regulated
expression of an active steroid receptor (Fig. 15). YEpSR is a high copy number vector that uses the copper responsive yeast metallothionein (CUP 1) promoter to drive synthesis of receptor mRNA. The steroid receptor cDNA is fused to the carboxyl terminus of a synthetic cassette adapted ubiquitin molecule. Therefore, synthesis is from the ubiquitin 47
AUG producing a fusion protein. The fused ubiquitin is removed soon after translation by
a yeast processing enzyme, leaving a natural receptor molecule (163). This system has been
Production of Mammalian Steroid Receptors in Saccharomyces cerevisiae
Ampr 2p3
YEpSR CUP1
Ub TRP1
Term Receptor cDNA
I Cu2 + Induction
UbV,Receptr7
Arg-Gly-Gly-Met-Thr
Processing
Ub Rc ept o
Figure 15. Schematic depiction of the strategy used to produce biologically active steroid receptors in yeast using a 2 urn based, ubiquitin fusion expression vector. Locations of the CUP1 promoter, ubiquitin
molecule, CYC terminator sequence, tryptophan selection marker, and the 2 um sequences are indicated. 48
shown to enhance the level of production of many proteins, presumably by enhancing
stability and solubility, and by affording protection from endogenous proteolysis (164).
In the current study, this expression system was utilized to produce the human
estrogen receptor. The strategy for the production of the recombinant receptor is indicated
in Fig. 15. A fusion cDNA was created between ubiquitin and hER. This cDNA was then
inserted into the expression vector YEPV 1, creating the vector YEPE10 (Fig. 24). In this
construct, the recombinant protein was produced under the control of the CUPI
(metallothionein) promoter. This plasmid was then introduced into the protease deficient
strain BJ3505 by standard techniques (141). Receptor production was induced by growing the cells to mid-log phase and adding CuSO4.
Receptor protein, which is initially produced as a fusion with ubiquitin, is rapidly
processed following folding to yield an unfused receptor. The results of such an
experiment are shown in Fig. 16. Cells containing the estrogen receptor were induced for
different times with 100 uM copper sulfate. The cells were harvested by centrifugation,
lysed, and the cytosol was analyzed by Western immunoblot. Using a mono-specific
antibody, D75, the production of a 65KDa protein that co-migrated with the authentic
estrogen receptor produced in mammalian CHO (Chinese Hamster Ovary) cells was
detected. This antibody does not interact with any endogenous yeast proteins (data not shown).
Using mammalian estrogen receptor standards that have been accurately quantitated by hormone binding, the amount of ER produced by yeast cells was quantified. These cells
produce 3-4 pmoles of receptor per mg of total protein. The receptor was homogeneous and no significant breakdown of the receptor was apparent. The slightly higher molecular weight of the recombinant receptor was due to 14 amino acids that remained on the receptor following removal of the ubiquitin (resulting from a cloning artifact). 49
Cupi Promoter Stds. kDa
93--b
I.
Dye- 0 1 2 4 8 3.5 7.0
Hours hER (ng)
Figure 16. Western immunoblot analysis of the recombinant estrogen receptor produced in Saccharomyces cerevisiae. The cells were grown overnight and induced with copper sulfate at the times indicated. The human estrogen receptor standards were isolated from CHO cells transformed with a mammalian expression vector. They were quantitated by immunoblot and hormone binding. 50
Biochemical Characterization of Recombinant yER
The hormone binding properties of the recombinant ER molecule are
indistinguishable from the wild-type receptor (Fig. 17). The affinity of the receptor was
found to be 0.24 nM when the binding affinity data were transformed by Scatchard's
method. These analyses, however, revealed a significant discrepancy between the levels of
expressed protein when estimated by ligand binding compared to Western immunoblot.
This phenomenon has previously been observed for the estrogen and progesterone
receptors expressed in yeast (113-116). It has been proposed that Saccharomyces
cerevisiae may not perform all of the necessary modifications required for ligand binding
(113). An alternative explanation is that the harsh homogenization methods using glass
beads lead to the denaturation of the receptor. To address these questions, a rigorous
examination of the quality of the receptor extracted by the glass bead disruption method and
that isolated by an enzymatic digestion protocol was performed.
Cells were grown as normal and then split into two batches. One aliquot of cells
was disrupted by vortexing in the presence of glass beads. The second aliquot was digested
for 1 hour with oxalyticase and the resulting spheroplasts were lysed osmotically. The
resulting supernatants were clarified by centrifugation. Cytosols were analyzed for hormone binding. A comparison of the results is shown in Fig. 18. The cytosol prepared by enzymatic digestion yielded 5.5 pmoles/mg of protein whereas that prepared by using glass beads yielded 700 fmoles/mg of protein. The binding affinity of the receptor isolated by both methods was identical (data not shown), but, clearly, the material produced by enzyme digestion was of superior quality as estimated by ormone binding.
A comparison of an equal amount of receptor as estimated by hormone binding from yeast and mammalian CHO cell extract by Western immunoblot showed that both cytosols contained approximately the same amount of immunoreactive material. These 51
experiments, using cytosols derived from different mammalian sources, have been repeated with analogous results. This is an indication that the specific activity of the receptor produced in yeast is identical to that produced in mammalian cells.
Scatchard Analysis of Binding Data for Wild Type ER in Yeast
0.6 I
0.5
0.4 -
Li. 0.3- 00
0)0n
0.2 aa
0.1 -
a \
4. 0.0. I . I ...... 0.15 0.20 0.25 0.30
Bound
Figure 17. Scatchard plot analysis of ER expressed in Saccharomyces cerevisiae. The cytosol was prepared from yeast cells containing receptor expression plasmid. The cytosol was incubated with increasing concentration of [3H] estradiol in the presence and absence of unlabeled estradiol. 52
6A. B.
c5
0 4
E --- 66kDa 3
w -2 E
OO Bead Enzymatic Homogenization Extraction Procedure
Figure 18. Hormone binding and Western immunoblot analysis of recombinant estrogen receptor indicating
that the mammalian and yeast produced receptor bind hormone equally well. (A) Comparison of the
hormone binding properties of receptor isolated from the same batch of cells by either a bead
homogenization or enzymatic digestion protocol. (B) Western immunoblot analysis of an equivalent amount of labeled receptor from yeast and CHO cells indicating that the hormone binding activity of both preparations is identical. 53
To confirm these observations, a receptor immunoprecipitation assay using an
antibody raised against the DNA binding region of the receptor was employed. The abilities
of this antibody to immunoprecipitate an equal amount of labeled receptor from MCF7
cells and from the yeast extracts containing recombinant receptor were then compared.
When antibody is limiting, less labeled receptor would be precipitated from the yeast
cytosol if it contained significant amounts of a non-binding species. The results of this
analysis are shown in Figure 19. As Figure 19 indicates, at very low concentrations of antibody, equal amount of labeled estrogen receptor were immunoprecipitated. As the concentration of the antibody was increased, more recombinant binding activity was immunoprecipitated. It can be inferred from these data that the estrogen receptor produced in yeast is undegraded and it is of a similar quality to that produced in mammalian cells.
The strategy outlined in Figure 15 indicates that several components of the system can be changed to increase expression levels of the receptor in yeast. Such parameters as physical changes of the promoter, the origin of replication, the selectable marker or changes in the protocol for induction of the receptor are all possible and were evaluated systematically.
The only significant increase in production level was observed following the inclusion of a defective leu2 allele into the plasmid, creating the expression vector YEPE15.
This strategy has previously been shown to increase plasmid copy number in strains defective in leucine biosynthesis, as more of the plasmid is required to complement the cellular mutation (165). The protocol has been used successfully to produce recombinant proteins (145). The plasmid YEPE15 was transformed into the strain BJ2168 which was grown under conditions that were auxotrophic for leucine. The cells were harvested, the cytosol was prepared and the receptor production was quantitated by Western immunoblot as shown in Figure 20. By comparison to known standards it was estimated that this 54
Immunoprecipitation of Labelled ER Isolated from Yeast and Mammalian Sources
10-a
3C 8 0 o^6
cc E .~4
,--, MCF-7 S- OYeast ER
1 2 3 4 5 6 p1 OV245
Figure 19. Immunoprecipitation of labeled hER isolated from yeast and mammalian sources. The indicated amounts of OV245 antibody were incubated with either yeast or with MCF7 estradiol labeled receptor and assayed as described. (o), yeast: (e), MCF7. 55
construct produced 5-fold more receptor up to a minimal level of 0.2% of the total soluble yeast protein.
The DNA-binding properties of the yeast produced ER were analyzed by calf thymus DNA-cellulose chromatography. The result shown in Fig. 21 demonstrates that the peak of hormone labeled receptor elutes at 0.35 M KCI, and it is consistent with what is observed for the wild-type receptor. To study the interaction of the yeast expressed estrogen receptor with the estrogen response DNA elements, gel retardation assays were performed as described in "Materials and Methods". Figure 22 suggests that the recently cloned wild-type hER in vitro binds to DNA in the absence of hormone, whereas, under similar conditions, the original cloned ER, which possess a point mutation that results in the presence of valine instead of a glycine at position 400 (108), requires hormone. The nonradioactive ERE in the reaction mixture effectively competed with the labeled ERE. These studies demonstrate that yeast expressed wild-type ER retains its specific ERE binding properties in the absence of hormone. 56
Expression of hER in Yeast (Cupi Promoter/Leu2d Marker)
Stds. kDa
-66
+-30
Dye 0 2 4 8 3.5 7 Hours hER (ng) Po st- Indu ct ion
Figure 20. Expression of hER in yeast using YEPE15. Cells were grown in minimal media and induced for
the time indicated. The level of receptor produced was quantitated by Western imnunoblot. The standards used for quantitation were isolated from CHO cells transformed with the human estrogen receptor cDNA. 57
Salt Gradient Elution Profile of High Salt-Activated Yeast hER from DNA-cellulose 100000 1
80000 0.8
60000 0.6 aLL 0 --- CPM/Fractibn 0 ----.--- [KCqf(M) 40000 2 - 0.4
40'
20000 -0.2
pr 04h 0T0 . 0 10 20 30 40 50
Fraction Number (2 ml/fraction)
Figure 21. The DNA-binding properties of the yeast expressed ER. The DNA-binding properties of the receptor were analyzed by calf thymus DNA-cellulose chromatography. The bound proteins were eluted with a linear gradient of KCl. 58
In Vitro Association of hER with DNA
0 AL. 6
lOpg Cytosol + + + + + + + + + + + + 100ng D75Ab- + - - + + - - + 1bOX ERE- - - - + + - + - -+ Hormone - - + + - - + + + hER(Gly4oo--Val) Wt hER
Figure 22. Specific binding of yeast expressed ER to estrogen response elements. Yeast cell extracts were
prepared and 10 ug of extracted protein was incubated with 20,000 cpm/reaction of [3 2P]-labeled estrogen response elements. Free labeled ERE and the receptor associated ERE were separated on a 4% PAGE and visualized by autoradiography. 59
Reconstitution of Steroid Hormone Responsive Transcription Units in Saccharomyces cerevisiae
A cis-trans transcription assay to test the transcription activity of the steroid receptors produced in yeast cells was developed (Fig. 23). This beta-galactosidase assay consists of cotransforming a receptor expression plasmid and a reporter plasmid into a yeast cell, and determining the ability of the receptor to activate transcription of the reporter.
The reporter plasmid contains the yeast iso- 1-cytochrome C proximal promoter elements fused to the E. coli beta-galactosidase gene. The behavior of this particular CYC-lacZ fusion has been well characterized (166). The CYCI upstream activating sequences have been removed and replaced with synthetic oligonucleotides containing the steroid responsive elements. The plasmid also contains all the sequences required for selection and propagation in yeast. Moreover, the transcriptional activity of the expressed receptor at different cellular concentrations can be measured by varying the amounts of copper added to induce the CUP1 promoter.
To test the transcription activity of yeast derived ER, the reporter plasmid YRPE2 was constructed as shown in Fig. 24. This plasmid contains the proximal promoter elements of the yeast iso-1-cytochrome C promoter fused to the structural gene for beta galactosidase (117). Two tandemly linked EREs derived from the consensus vitellogenin
ERE (152) were placed 250 bases upstream of the initiation site. This plasmid was co transformed into yeast containing the expression vector YEPE10 (Fig. 24). Transformants were selected for by uracil and tryptophan prototrophy. 60
Assay of Transcriptional Activity of Steroid Receptors in Saccharomyces Cerevisiae
2p3 URA3 / mpR 2p4
YRpSR YEpSR AmpR SRE TRP1 Cyci CUP1 TERM Ub LacZ cDNA
Tranform Yeast Select for TRP and URA Auxotrophs
URA 0 TRPO
-H -H +H +H -Cu +Cu -Cu +Cu
Figure 23 Beta-galactosidase assay used to measure transcriptional activity of steroid receptors produced in yeast. The reporter plasmids are created such that they contain the proximal promoter elements of the yeast iso-1-cytochrome C promoter fused to the structural gene for beta-galactosidase. A unique Xhol site at -250 allows the insertion of appropriate SREs. Receptor expression and reporter plasmids contain different
markers and can be cotransformed into host cells. 61
The transcriptional activity of ER in these transformants was determined by measuring the beta-galactosidase activities of transformants grown overnight in the absence of hormone, or in the presence of varying concentrations of estradiol or nafoxidine (Fig.
25). The yeast-produced estrogen receptor clearly activates transcription of the target gene in a hormone dependent manner. Half maximal activation of transcription occurred at about
2p URA3 AmpR2po
Y E P \YRPE2 E 1 0 TRP1 AmpR TRPI EREs Cyc 1 Ub TERM LacZ hER cDNA
Figure 24. Yeast reporter plasniid (YRPE2) and expression vector (YEPE1O). The reporter (YRPE2)
contains two copies of EREs inserted into the XhoI site of yeast reporter plasmid YRPC2. CYCI is the
yeast iso-I cytochrome C promoter fused to the structural gene of E. coli Lac Z. UJRA3 is the selective
marker for uracil. The expression vector (YEPELO) contains the wild type human (hER) fused to ubiquitin
(UB). CUP1 is the yeast metallothionein promoter. TRPI is the tryptophan selective marker. 2 micron is
the replicating DNA of yeast. 62
1 nM hormone. This is similar to that required for activation of the receptor in mammalian cells (148). The appropriate response of the receptor to nafoxidine, which is a partial receptor agonist, further supports the idea that this receptor is biologically active. Subsequently, it was found out that, due to a subclonal artifact, ER expressed from the YEPE1O construct carries 14 extra amino acids at the N-terminus not found in the wild-type ER. These 14 extra amino acids in YEPE10 were removed and the plasmid was referred to
Yeast ER is Transcriptionally Active
3000 I
-7'1----- U 2500
2000
U 01500
N 0 1000
500
0 5 10 600 0 0.5 1.0 1.5 2.0 2.5 Estradiol (nM) Nafoxidine (pM)
Figure 25. Transcriptional activity of yeast produced ER on the reporter YRPE2 in the absence or the presence of increasing concentrations of (a) estradiol or (b) nafoxidine. The beta-galactosidase activity is expressed in Miller units/mg of protein (117). 63 as YEPKB 1 [ER(Wt)] to distinguish it from YEPE10, the original construct. The corrected receptor was also highly inducible by hormone, implying that the extra amino acids did not appear to compromise signal transduction.
Fig. 26 shows the activities of YEPE10 and YEPKB I in the absence of hormone and in the presence of estradiol or the antiestrogen nafoxidine. The YEPKB I displays about 3-fold higher activity than YEPE10 upon the estrogen induction. Nafoxidine has partially agonistic activity with both plasmids, 15% and 30% relative activity, respectively, compared to estradiol. The lower agonistic activity of nafoxidine is not due to a lower degree of binding to the receptor, as the concentration of nafoxidine used (2 uM) is sufficient to saturate the system. The apparent dissociation constant for nafoxidine in this system is approximately 0.2 uM for both YEPKB I and YEPE10.
These results suggest that this transcriptional system reconstituted in yeast is hormone dependent. In the current expression system, transactivation is maximal even at the lowest receptor concentrations. The possible explanation could be that ER is present in excessive amounts in cells, even in uninducible conditions. To test this hypothesis, a yeast strain was constructed, in which a single copy of the ER expression unit was integrated into the genome. Expression of ER in this strain is approximately 10 to 20-fold lower than for cells carrying multicopy episomal expression units, and is inducible by copper (Fig.
27B. Note that 5-fold more protein was loaded in lanes 3 and 4 compared to lanes 1 and 2). As shown in Fig. 27A, the transcriptional activity of ER expressed in this strain is dependent on receptor concentration (copper concentration). These results suggest a parallel increase in receptor concentration and in transactivation. 64
8000
0
6000 -
4000 -
N w
2000
0
0 -- Hormone -N E - N-E ER(YEPE 10) ER(wt)
Figure 26. Ligand-induced activity of YEPEIO (ER) and YEPKB 1 [ER (wt)]. Activity of plasmids on
YRPE2 in the absence of hormone (bars), in the presence of nafoxidine (bars N), or in the presence of
estradiol (bars E). Yeast cells were grown in the absence of CuSO4, and in the appropriate samples, induced with hormone for 4 hrs. LacZ activities are expressed in Miller units/mg of protein. 65
+H Miller Units -H U Miller Units 800
600
0
400 -
200 -
o 4- 0 5 10 20 50
CuSO4(uM)
from an Figure 27. Transactivation by ER in yeast. (A) Activation of transcription by ER expressed the leu2 locus of strain integrated single-copy expression unit. The ER expression unit was integrated into
BJ3505. These cells were then transformed with the reporter construct YRPE2. LacZ expression was assayed under different concentrations of CuSO4, and with hormonal induction. 66
040
- O4Aw A
1 2 3 4
(B) Western blot analysis of ER expressed from episomal and integrated vectors. Cells were grown with
(lanes 2, 4) or without (lanes 1, 3) 50 uM CuSO4. 50 ug (lanes 3, 4) or 10 ug (lanes 1, 2) of protein were loaded onto an SDS-PAGE, blotted onto immobilon, and probed with an hER monoclonal antibody (D-75). 67
Application of Saccharomyces cerevisiae System
The hormone-responsive transcription unit developed in these studies can be used to dissect the functional domains of steroid receptors as well as other transcriptional factors which participate in the steroid receptor transductional pathway. The system also will be useful to enhance our understanding of hormone and antihormone actions. The yeast system has been applied in this study for the following purposes :
(1) To examine the in vivo regulation of ER functions by antiestrogens;
(2) To identify novel estrogen response elements;
(3) To study variant receptors cloned from breast tumors;
(4) To address whether the estrogen receptor can stimulate the osteocalcin gene, and
to identify the estrogen response element in the osteocalcin gene promoter.
Effect of antiestrogens on the regulation of ER
It is thought that ER possesses two independent nonacidic transactivation domains, one in the N-terminal region and another within the hormone binding site (167). N-terminal domain may modulate transactivation and other receptor functions, whereas the hormone binding domain alone is able to confer ligand responsiveness (101, 168).
If ligand functioned solely by regulating the endogenous transactivation regions of the ER, it follows that the addition of a strong constitutive transactivation function onto the
ER may create a receptor molecule having transcriptional activity even when bound to an estrogen antagonist. To evaluate this possibility, a chimeric receptor was used, in which the acidic activator of the herpes simplex virus protein VP16 was inserted into the N-terminal region of the ER in mammalian cells (148). This chimera has been shown to be 10-fold more potent than ER in activating transcription (148). 68
The activity of native and chimeric ER in response to hormone was assayed by
cotransfection with a CAT reporter into mammalian CV- 1 cells. Fig. 28 shows that both
ER and ER-VP16 require hormone to activate transcription. The target promoter consists of
EREs linked to a thymidine kinase promoter. However, ER-VP16 can be induced by
nafoxidine to a similar extent as with estradiol, whereas the ER is not induced by the
triphenylethylene antiestrogen nafoxidine. The dose-response curve for nafoxidine is
shifted compared to that for estradiol most likely because the ER has a 20-50 times lower
affinity for nafoxidine (169). Since ER-VP16 has a powerful transactivation region that
functions independent of ligand, it is likely that, in the absence of hormone, ER-VP16 is
inactive because it is prevented from reaching its target element. These data imply that both
estradiol and nafoxidine are able to stimulate DNA binding. However, only the agonist can
promote transactivation by native ER. All of the antiestrogens examined could induce the
transcription of ER-VP16, as shown in Fig. 28, suggesting that their functions are
mechanistically similar.
The original cloned hER, used in constructing the ER-VP16 chimera, was found to
possess a point mutation in the hormone binding domain, which reduces the affinity of the
receptor for hormone, but appears to have negligible effects on other aspects of receptor
function (108). When the activity of the wild-type hER clone was examined in transient
transfection, it was found to have a high level of constitutive activity. Apparently, the
greater hormone binding affinity of the wild-type ER causes it to be activated by remaining
traces of estrogenic substances in phenol-red-free charcoal-stripped mammalian cell culture
medium (108). Therefore, a simpler eukaryotic system, Saccharomyces cerevisiae, in which ER function had been successfully reconstituted (113-115), was used to study further the ligand-dependent activity of the ER. The advantage of yeast is that it can be 69
grown in defined synthetic medium, and thus the activity of ER is fully dependent on the added hormone (108).
Estrogen- and Antiestrogen-Dependent Induction of CAT Activity
-3A c
ER:+ + + + + + + ERoVP16: + + + + E2 - TAM NAF ICI LY156 LY117
Figure 28. Induction of CAT expression by ER and ER-VP16 in the absence (-) or presence (+) of 10 nM
estradiol (E2) or an antiestrogen at 0. luM: tamoxifen (TAM), nafoxidine (NAF), ICI 164,384 (ICI),
LY156758 (LY 156), and LY117018 (LY117). The same amount of protein (20 ug) was assayed in all lines.
Ac, acetyl group.
The wild-type human ER expressed in yeast can stimulate an estrogen responsive reporter gene. The antiestrogen nafoxidine is a competitive inhibitor of estradiol activity
(Fig. 29A). At 2 uM, a saturating concentration, the activity of nafoxidine is about 15% of 70
that obtained with estradiol, as nafoxidine itself has weak agonistic activity in yeast. This is
in general agreement with the results obtained in animals, in which nafoxidine has been
shown to be a partial agonist (10% activity) (170-171). When 2 uM nafoxidine is used, at
low levels of estradiol (from I to 5 nM), the inhibition of transcription is as great as 70%.
When the estradiol concentration is increased, transcription is essentially restored. The
reason that much higher levels of nafoxidine are required to inhibit estradiol is primarily a
result of the known lower affinity of the ER for nafoxidine compared to estradiol (169).
Figure 29B shows that, in yeast, the ER-VP16 chimera is equally inducible by both
estradiol and nafoxidine, supporting the mammalian transfection results described earlier.
The ER-VP16 chimera is actually much more active than ER, since the steady-state
concentration of ER-VP16 is approximately 100 times lower than that of ER, even if basal
level expression of ER-VP16 results in the same level of transcription of lacZ compared to
that obtained with ER.
Testing the hypothesis that the observed antagonistic function of antiestrogens is
due to the interference with the DNA binding of the ER (172), resulting in only a small
percentage of the target sites being bound by ER, was possible, as the yeast system is a
regulatable ER expression system. If this hypothesis were correct, it should then be
possible to increase the saturation of EREs in vivo by increasing the total amount of ER in
the cell, thereby leading to a higher level of transcriptional activity. In the CUP1 promoter
ER expression system, there is a substantial basal level of expression. Moreover, ER
expression can be increased by adding copper to the growth medium. In this way, ER
levels can be increased approximately 6 to 10-fold from the basal level, as detected in a
Western immunoblot (Fig. 30A). The induced ER is active in DNA binding, as determined
by a mobility-shift assay. However, even when ER expression is increased at least 6-fold, the amount of lacZ transcription in the presence of either estradiol or nafoxidine remains 71
Regulation of ER Function in Saccharomyces cerevisiae
3000
+'
22000 |F
E N
Z 1000 E+N N
0.y 5 10 600nM Estradiol Concentration (nM)
Regulation of an ER/VP16 Chimera by Ligand
3000
6.
0 2000
1000
N
..J CL 0 - xx + xqx ER ER/VP16
Figure 29. Hormone and antihormone regulation of ER function in S. cerevisiae. (A) Induction of lacZ expression at various estradiol concentrations with estradiol (E) alone or with estradiol plus a constant concentration (2 uM) of nafoxidine (N). (B) Comparison of the transcriptional activity of ER and ER-VP16 in the presence of 0.6 uM estradiol (bars +E), 2 uM nafoxidine (bars +N), or without added hormone (bars
H). 72
hER Expression is Copper Inducible
k D a -9 7'.4 -- 66.2 -42.7
CuSO 4 0 1 2 5 10 (p M)
Effect of Receptor Levels on Transcriptional Activation
) I
E 0
'2000
1000 N
.j 0 0 2 4 6 8 10 1I2 CuSO 4 Concentration (pM)
Figure 30. Effect of the level of ER expression on transcriptional regulation by estradiol and nafoxidine. (A)
Western blot analysis showing the level of hER expressed as a function of CuSO4 concentration. (B) Effect of CuSO4 on the expression of lacZ in the presence of 0.6 uM estradiol (E) or 2 uM nafoxidine (N). 73
unchanged (Fig. 30B). These results suggest that, even under basal level expression of
ER, the target sites are maximally occupied in the presence of either estradiol or nafoxidine.
It can be said that the lower amount of transcription in the presence of nafoxidine is likely
to be due to the intrinsically lower transactivation potential of the bound ER.
The DNase I sensitivity of the chromatin template at target responsive elements
could be altered since antiestrogens can induce DNA binding of ER in vivo. It is known
that the action of agonist ligands of steroid receptors is correlated with specific DNase I
hypersensitive sites in chromatin (173-175). And an antiglucocorticoid, dexamethasone 21
mesylate, was found to suppress the hormone-induced DNase I hypersensitive site over the
mouse mammary tumor virus promoter (176). However, whether less active or inactive,
antihormone-receptor complexes that interact with responsive elements can induce similar
alterations in the chromatin structure is not known.
Direct examination of alterations in DNase I sensitivity over the ERE-CYC1-lacZ
reporter construct was done by using indirect end labeling (Fig. 31B). As shown in Fig.
3 1A, both estradiol and nafoxidine can induce alterations in chromatin structure. Moreover,
the promoter region becomes hypersensitive to DNase I. The hypersensitive sites are highly
reproducible and are not the result of artifacts in DNase I digestion. The same DNase I
hypersensitive sites are induced by estradiol and nafoxidine. These sites occur adjacent to
the EREs. The actual ER binding sites are not hypersensitive. It is believed that the interaction of the ER with the ERE alters the chromatin structure, resulting in hypersensitivity to DNase I , while the ERE itself is protected from DNase I by the bound
ER. A hypersensitive region without internal protection should be observed if nafoxidine bound ER were to alter chromatin structure without occupying the ERE. The magnitude of the hypersensitivity induced by estradiol or nafoxidine is very similar, suggesting that chromatin structure alterations occur on the same percentage of the minichromosomes 74
-820 -240 900 Cyc 1 LacZ
I EREs Probe Kx B C
Estrogen and Antiestrogen Induce Alterations in Chromatin Structure
Naked M +E2 +N -Hormone -ER DNA DNase I: a b c a b a b c a bd f C
EREs D 2 TATAC B-
LacZ
Figure 31. Alterations in the chromatin structure induced by estrogen and antiestrogen. (A) Diagram of
ERE-CYCI-lacZ reporter used in the chromatin structure analysis. C, ClaI; X, XhoL; B, Bami. DNase I
hypersensitive sites (DHS) induced by hormone are shown. (B) DNase I sensitivity of chromatin analyzed
by indirect end labeling. The hypersensitive sites are bracketed and labeled I and 2. Shown are the DNase I
sensitivity of the chromatin from cells not treated with hormone (lanes -H) and cells treated with nafoxidine
(lanes -N) or estradiol (lanes E2). Samples from cells that were not transformed with the expression plasmid
(lanes -ER) are also shown. DNase I digestions were performed at 600 units/ml for 6 min (lanes a), 300
units/ml for 6 min (lane b), and 300 units/ml for 3 min (lane c). The marker lane shows the position of the
XhoI site (upper band) and the BamHI site (lower band). 75
inside the cell. This result suggests that antihormone does not interfere with ER binding to
its cognate ERE in vivo. However, the actual pattern of the hypersensitivity is slightly
different and it is perhaps a reflection of the differences in the level of transcription. For
reasons that are not clear, the transcribed region becomes less DNase I sensitive with active
transcription. The ER has no effect on the chromatin structure in the absence of hormone as
shown in Fig. 31B. This is consistent with the fact that hormone is required for DNA binding.
Identification of novel estrogen response elements
The existing methodology (72, 166, 177-180) for identification of steroid response
elements is laborious and limited to comparative analysis of sequences from a few steroid
responsive genes. To overcome these problems, an alternative method was developed
using an estrogen responsive transcription unit that was reconstituted in yeast. An outline
of this method is shown in Fig. 32.
A degenerative oligonucleotide was designed following a comparison of the known
steroid response elements (Table I). Nucleotides common to all response elements are
boxed and were left constant in the otherwise random oligonucleotides (72, 181). An
evident pattern from this comparison of different response elements was the variable nature
of the spacer region between the two receptor binding half sites. For the estrogen receptor,
a spacer of 3 oligonucleotides, all of which were randomized, was chosen. The
oligonucleotides were converted to double-stranded DNA by using a specific primer
complementary to the 3' end of the degenerate oligonucleotide (154-155). The double stranded oligonucleotides were cloned in the BglII site of the YRPC3 vector (114) and
propagated in E. coli, generating a library of approximately 40,000 clones. 76
Identification of Estrogen Response Elements Using a Yeast Based Screen
Create a Library of Degenerate Oligonucleotides
N Clone into Reporter Plasmid Cyc
YRP(SRE)N Amp
.?4f
Transform into 8J3505/hER
0* 0 Select for TRP+URA+ o0 0 on X-Gal+Estradiol Plates
Pick Blue Colonies
SCharacterize Transformants -H +H
Rescue Plasmid and Sequence
Figure 32. Identification of estrogen response elements using a yeast based screen. A library of degenerate oligonucleotides YRP(SRE)n was created by cloning the double-stranded oligonucleotides into the unique
BglII site of the yeast reporter plasmid YRPC3. The library was introduced into the yeast cells carrying the estrogen receptor expression plasmid (YEPE 10) and the cells were selected for tryptophan and uracil prototrophy on plates containing X-gal and 17 beta-estradiol. 77
To assay for functional steroid response elements, the library was introduced into
yeast cells carrying an estrogen receptor expression plasmid and the transformants were
selected for tryptophan and uracil prototrophy on selective plates containing estradiol and
X-gal. The plates were incubated at 300C for several days. A total of 157 blue colonies
were picked, purified, and reassayed in the presence and absence of estradiol.
96% of the blue colonies exhibited hormone independent transcriptional activity.
This is attributed to the generation of cryptic transcriptional factor binding sites in the
library. However, 4% of the blue colonies exhibited hormone dependent transcriptional activity. Reporter plasmids were rescued from hormone dependent transformants and the
DNA inserts were sequenced. The sequence and transcriptional activities of hormone dependent transformants are shown in Table II.
The transcriptional activity of YRPE2, the reference reporter carrying the vitellogenin A2 ERE and transformant #12 were strictly hormone dependent.
Transformants #1, #4, #9, #13, #39, and #45 exhibited a basal activity which was inducible by hormone. The basal activity was approximately 4-28% of the inducible activity. The induction varied 7-23 fold from the basal level. To demonstrate that the transcriptional activity exhibited by these transformants was receptor dependent, the reporter plasmids containing synthetic EREs were transformed into a yeast strain that did not contain estrogen receptor and transformants were selected by uracil auxotrophy. These transformants were grown in the presence or the absence of estradiol and copper sulfate.
The cytosols were prepared and analyzed for beta-galactosidase activity. As expected neither reference reporter, YRPE2, nor the reporter plasmids containing synthetic EREs stimulated induction of beta-galactosidase activity under any conditions (data not shown). 78
Sequence of Degenerate Oligonucleotide and Comparison with Other SREs
Sequence
G C - (N) 0 - 5 Consensus 0 C N N N G N N C N (N) 3 N G N N N N N Oligolnucleotide G C G T C A (N) 3 T G A C ERE T Ct - - G G T A C A (N) T G T T T GRE/IPRE T 3 C C G G A C G (N) T G A C TRE T 0 C --G A C r C A (N) 5 T G A A G V DR E
Table I. Sequence of degenerate oligonucleotide and comparison with other SREs. The sequence of the degenerate oligonucleotide is indicated and aligned to the response elements for estrogen, glucocorticoid/progesterone, thyroxine, and vitamin D3. The sequences found in all response elements are boxed and were left constant in oligonucleotide. 79
Sequence and Transcriptional Activities of Hormone Dependent Response Elements 1-Galactosidase Activity Construct Miller Units Xho I --H +H 47 43 PC3 -25CyclacZ
-- YRPE2 C A G G TC C A G G A C C T G - - LacZ 0 3200 # 1 -G C G G G G C G C A G T G T G C T G T - LacZI 240 846
#4 -G T G G T T C G G C T C G T G C G A A - LacZ |I 1,40 2313 #9 -C G G G G T C G G C G G T C C C T - LacZ| 120 1173
# 12 -A C A G C A C A A C G G G A C C C G G 4 LacZ 0 1280 #13 -C C C G G A C A C A A G G C C C A G G - LacZI 1106 1286 -- #39 A C G G T A A C A G T C G T T C G - -- LacZ I 306 3540
#45 -T C G G T G C T G G T G T G C C T C -- |LacZ 120 2780 F I
Table II. Sequence and transcriptional activities of hormone dependent response elements. The sequence and transcriptional activities of hormone dependent transformants were aligned to the reference reporter
(YRPE2), carrying the vitellogenin A2 ERE, and YRPC3. 80
Sequence analysis of hormone dependent transformants suggests that the sequences
of all the transformants were different from the consensus sequence of the estrogen
response element. In the latter response element, the two half sites are organized as inverted
repeats. Surprisingly, the two half sites of clones #9 and #45 consisted of direct repeats. In
the clone #45, the direct repeats are misaligned by one base. These results suggest that the
estrogen receptor can bind and activate transcription from a response element, where the
two half sites are present as either direct or indirect repeats. In the clone #39, the conserved
C in the first half is mutated to A, while in the second half both conserved G and C are
mutated to C and T, respectively. These changes in the conserved nucleotides might have
originated during oligonucleotide synthesis. This suggests that these bases are not critically
important for the functional interaction of the estrogen receptor with its response elements.
Using this system, several progesterone responsive elements have been also identified.
Some of the progesterone responsive elements are also present in direct repeats as well
(data not shown).
To confirm that the estrogen receptor can bind and transactivate transcription from a response element where the two receptor binding sites are aligned as direct repeats, an oligonucleotide was constructed containing the receptor binding sites as perfect direct repeats. This oligonucleotide was cloned into the BglII site of the plasmid PC3 (114),
YRPD2. This plasmid was transformed into the yeast cells containing the receptor expression plasmid. The transcriptional activity of the estrogen receptor through this response elements was compared to those of a perfect palindrome under varying copper and estradiol concentrations. Cytosols were prepared and induction of lacZ activity was measured (Fig. 33). These data suggest that in the absence of hormone, receptor can not stimulate induction of transcription from either response elements. Transcriptional activity was observed only in the presence of hormone. Both reporters demonstrated a similar 81
Transcriptional Activity of hER from Reporter Plasmids YRPE2 & YRPD2
5000 YRPE2 YRPD2
4000
aim 3000 +H13 Miller Units *.A -H d Miller Units +H IE Miller Units 2000
1000--
0 &YRPE2SLYRPD2 0 20 40 60 80 100 120 CuSO4 (uM)
Figure 33. Transcriptional activity of hER from repoter plasmids containing receptor binding sites as direct or indirect repeats. The yeast cells containing YEPE1O and YRPD2 plasmids were grown in minimal media in the absence and the presence of estradiol with increasing concentrations of copper sulfate. Cytosols were prepared and induction is expressed as Miller units per mg protein. 82 transcriptional activity. These results confirmed that the estrogen receptor can transactivate transcription from yeast response elements where the receptor binding half sites are aligned in direct or indirect repeats.
The EREs identified in this study were compared to the mammalian gene bank by computer scanning analysis (Table III) to see if novel estrogen responsive genes could be identified. This analysis revealed that the homologous sequences were present in the promoters of several mammalian genes. A review of the literature indicated that most of these genes had been reported to be under estrogen regulation, but these reports did not identify the sequences responsible for this effect. The ERE in transformant #4 is homologous to the sequences found in the promoter region of the human phosphoglycerate kinase gene. The ERE sequences of transformant #12 are homologous to the sequences present in the promoter of the human serum retinol binding protein. The ERE of transformant #13 is similar to the sequences found in the promoter regions of the human elastin, muscle creatin kinase, complement C3 and HMG CoA reductase genes, as well as to those found in the promoter region of the mouse muscle creatin kinase and chicken apo VLDLII genes. 83
Table III
Sequences of Estrogen Response Genes Containing Homologous Elements Identified from Sequences in the Promoter Degenerate Oligonucleotide Libraries Regions and Found to ]Be
Regulated by Estrogen
IIIIIIEillilli IIIM
#4 GGTTCGNNNCGTGC Human phosphoglycerate kinase gene
#12 GCACANNNGGACC Human serum retinol binding protein
#13 GGACANNNGGCCC Human elastin gene
Human muscle creatin kinase gene
Human complement C3 gene
Human HMG CoA reductase gene
Mouse muscle creatin kinase gene
Chicken apo VLDL II gene
Table III. Comparison of synthetic EREs sequences with mammalian gene bank. The estrogen response elements identified from large libraries of degenerate oligonucleotides were compared to the mammalian gene bank. The genes that contain homologous sequences and are regulated by estrogen were identified. 84
Study of variant receptors cloned from breast tumors
Estrogen and progesterone receptors belong to a superfamily of ligand-modulated
transcription factors that play a role in the progression of breast cancer. Based on hormone
binding assays, breast tumors are classified into four groups:
ER-/PR- ER-/PR+ ER+/PR- ER+/PR+
20% 10% 10% 60%
Since progesterone receptor (PR) is normally induced by estrogen, breast cancer
lacking estrogen receptor (ER) would also be expected to lack progesterone receptor.
However, 10% of breast cancers are estrogen receptor negative, yet progesterone receptor
positive. The ER-/PR+ phenotype could result from a variant ER which is unable to bind
hormone, thus it appears negative in a ligand-binding assay but is still functional in
stimulating the estrogen-responsive PR. The possibility of a transcriptionally active ER
variant might also be a potential mechanism for the escapeof a tumor from hormonal
control. Such an altered ER gene might thus function as a type of oncogene. To test this
hypothesis, simplified polymerse chain reaction (PCR) amplification of cDNA was used to
detect and clone variant ER transcripts from individual breast cancers. A variant ER RNA
was amplified from ER-/PR+ tumors and sequence analysis revealed that the variant ER
was completely missing exon five within the hormone binding domain of the receptor (Fig.
34), suggesting aberrant splicing.
A yeast expression system was used to determine how the exon 5 deletion affected the ability of the ER product to activate the estrogen-responsive genes (116). An expression vector for the variant ER was constructed by replacing the HindIII-BglII fragment of the hER cDNA (YEPE2) with the equivalent fragment of the ER-/PR+ tumor cDNA which contains the deleted exon 5 region (Fig. 35). This variant ER construct (YEPE9) and a 85
A. T T T T c c C MCF-7 Tumor T c (Wild Type) (Variont) C c GA TC C A A ,,err, JO A c MCF-7 Tumor c C (Wild G 40 C Type) (Variant) cA EXON G 3 EGO G A T CGT G 4 T TEXON EXON G f l "" G .. ,.CC C 4 5 CT c e,4 EXON EXON T EXON EXON T A - T 6 .6 T T EXON A G G h ai T 6 A G G G G A c C c c C c c C c C T
B. 4 65 161 MCF-7 161
ER-/PgR+
. 4/5 5/6
xon 4 sxon 5 exon 6 WildType .... GGTCCGAA...... TGTCCTTG.... Variant .... G G T C C T T G....
Figure 34. Dideoxy sequence analysis of the variant ER. (A) Wild-type ER fragments and variant ER
fragments were prepared from MCF-7 and the ER-/PR+ tumor, respectively, and cloned into PGEM 7zf+
(Promega), for dideoxy sequence analysis. (B) Schematic diagram of the sequencing results for the 4/5 and
5/6 exon boundaries. 86
Smo I E R
a. 8 YEPE 2
A mpR
Hind M
Bg L r
Kpn I EcoR I
Figure 35. Yeast ER expression vector. Expression vectors were prepared for the wild-type and the variant
ER using an intermediate ER plasmid containing ubiquitin sequences placed in frame with the ER-coding region from the pORB cDNA (144). This plasmid, termed YEPE2, is under the control of the CUP 1 metallothionein promoter. UB, ubiquitin; AmpR, ampicillin resistance.
similarly prepared wild-type ER construct (YEPEIO) were then transformed into a protease-deficient strain of yeast. An estrogen-responsive reporter, YRPE2 (116), which contains 2 copies of the vitellogenin A2 ERE within the proximal promoter elements of the yeast iso-I -cytochrome C promoter (promoter fused to beta-galactosidase as an assayable marker), was also introduced into the cells. As expected, neither the reporter gene alone 87
(YRPE2), nor the wild-type ER (YEPE10) stimulated beta-galactosidase activity in the absence of estradiol. Table IV shows that in contrast to the results obtained with the variant, the transcription from the wild-type ER was maximally active at low concentrations of receptor, in the presence of hormone. In separate experiments, the variant ER exhibited approximately 10-15% of the stimulatory activity of the wild-type, and this activity was hormone independent.
Sequence analysis of the variant ER suggested that the variant might be truncated with the introduction of a stop codon after amino acid 370 of the corresponding wild-type
ER. To determine whether the variant encodes a truncated ER protein, extracts were prepared from yeast cells containing either the variant YEPE9 or the wild-type YEPE10 ER plasmid constructs. These were subjected to Western analysis using the H226 ER antibody which recognizes the amino terminus of the receptor. Fig. 36 shows that the variant migrates at approximately Mr 40,000 compared to the Mr 65,000 for the wild-type ER.
Table IV.
Treatment YRPE2 YRPE2+YEPELO YRPE2+YEPE9
None 0 0 342
+CuSO4 0 0 687
+E2 0 3500 323
+CuSO4, +E2 0 2700 660
Table IV. Beta galactosidaseassay. Transcriptional activities were assayed as described in six separate experiments and are expressed in Miller units (116). One representative assay is shown. 88
Certain mutants lacking virtually all of the hormone-binding domain are unable to
bind ligand and may activate transcription in a constitutive manner (36). In this respect, studies of the ER deletion mutants generated by site-directed mutagenesis are consistent
kD kDa
-- 6 6
-4 3
Figure 36. Western blot analysis. Cytosols prepared from yeast cells containing either the wild-type or variant ER plasmids were analyzed by Western blotting after resolution on a 10% SDS-PAGE. Molecular weight markers (Std.; in thousands) were included. 89 with the yeast expression studies. Mutations within the hormone-binding domains of both
the androgen and the thyroid receptors, as well as altered receptor forms implicated in
hormone resistance, have been shown to be associated with target-tissue hormone resistance (182-183). This is the first report of a variant ER, occurring naturally in a clinical human breast tumor, that can generate transcriptionally active, hormone-independent receptor.
The ER+/PR- tumors, containing estrogen receptor, would be expected to contain progesterone receptor also; but there still are 10% of the tumors that are ER+/PR-. This could be explained by the existence of an ER unable to function as a transcriptional inducer of PR expression.
To address this hypothesis, polymerase chain reaction (PCR) amplification of cDNA was used to detect and clone variant ER transcripts from ER+/PR- tumors. Two different ER variants were cloned from two different tumors. These variants were identical to wild-type sequence for exons 6 and 8 (59), but they missed exon 7 (Fig. 37). In one of the variants there was a point mutation in the hormone binding domain from valine to alanine (at amino acid position 422), in addition to exon 7 deletion. Deletion of exon 7
(nucleotides 1369 to 1548) (59) predicts a reading frameshift with the inclusion of 10 non
ER residues after codon 457. Thus, the ER variants potentially encode truncated ER proteins of approximately 52,000 daltons (Fig. 37).
To test the transcriptional activity of the variant ERs cloned from ER+/PR- tumors yeast expression system (123) was used. A portion of the wild-type ER yeast expression plasmid YEPE10, under the control of metallothionein promoter, was replaced with a
HindIII-BglIJ ER PCR-derived fragments from ER+/PR- tumor expressing the exon 7 ER deletion variants. This generated the variant expression vectors YEPE13 and YEPE15. The 90
variant ER expressed by the plasmid YEPE13 has a point mutation from valine to alanine in
the hormone binding domain. In other respects, this variant ER is similar to the variant
Wild T Variant G Wild T Type T C Type A A A A CTAG A C TAG Ex6 T Ex6 T E7 A T T CT AG Ex7 T cT l...;W G Tr -. g T A G G _ T A A A G A A Ex 7 T Ex8 C Ex8 C G A A A A AlT A A G G G1 G
Exon 6 6/7 Exon7 /8 Exon8 Wild Type TCI r GGA ooo AGU AAC
Variant TCIr GI U AAC
Figure 37. Schematic diagram of the sequencing results for the 6/7 and 7/8 exon boundaries. Dideoxy
sequence analysis of the variant ERs (Upper panel). Wild-type ER fragments and variant ER fragments were prepared from MCF-7 and the ER+/PR- tumors, and cloned into PGEM 7Zf+ (Promega), for dideoxy sequence analysis (Lower panel). 91 expressed by plasmid YEPE15. The variant ER constructs (YEPE13 and YEPE15) and the wild-type ER construct (YEPEIO) were transformed into yeast, along with an estrogen responsive reporter, YRPE2. The ability of the variant receptors to induce transcription from the reporter plasmid was determined. The variants were transcriptionally inactive, and this activity was unaffected by the overexpression of the receptor (at addition of copper), whereas the wild-type ER was transcriptionally active, and its activity was hormone dependent (Table V).
Table V
Treatment YRPE2+YEPE10 YRPE2+YEPE13 YRPE2+YEPE15
None 0 0 0
+CuSO4 0 0 0
+E2 2526 0 0
+CuSO4, +E2 2233 0 0
Table V. Beta galactosidase assay. Yeast cells containing receptor expression plasmids (YEPE10, YEPE13, and YEPE15), and reporter plasmid (YRPE2) were grown in a minimal medium in the presence and absence of estradiol (E2) and copper sulfate (CuSO4). The beta-galactosidase activity measured in yeast extracts is expressed in Miller units per mg protein. 92
To study the interactions between the variants and the wild-type ER, the plasmid
YEPE23 expressing wild-type ER from a constitutive promoter (TDH) was transformed into yeast cells carrying either one the variant ER expression plasmids (YEPE13 and
YEPE15), and a reporter plasmid (YRPE2). In these cells, when both the wild-type and the increasing amount of variant receptors were expressed, the variant YEPE13 and YEPE15 reduced the activity of the wild-type receptor up to 55% (Fig. 38) and 60%, respectively
(Fig. 39). This suggests that the variant receptors cloned from ER+/PR- tumors exert a dominant-negative effect on the wild-type ER activity. The inhibition of the wild-type activity was not observed when the ER cDNA was deleted from the construct (Fig. 40).
To test that these observations were not the result of excessive variant expression,
Western analysis of wild-type ER and variant estrogen receptors was performed (Fig. 41).
This Western immunoblot demonstrated that at maximum copper stimulation, wild-type ER was present at a three fold excess. Thus, variants appear to be potent inhibitors of ER action. They are truncated, and run at 52Kd compared to 65Kd for the wild-type ER receptor.
The dominant-negative effect of the variants could be a result of the constitutive
DNA binding ability of the variant receptors, which interferes with the wild-type ER binding to ERE. Alternatively, it may involve mechanisms unrelated to DNA binding such as formation of an inactive heterodimer. To distinguish between these two mechanisms,
GAL4 transcriptional assay (115) was used to examine the DNA binding ability of the variant receptor proteins. Yeast cells containing the YRPEG3 reporter were transformed with either a receptor expression plasmid expressing wild-type ER or with plasmid expressing variant ER. The cells were grown in galactose either in the absence of hormone or in the presence of estradiol. In the presence of hormone, the wild-type ER reduced
GAL4 activity up to 85% compared to cells grown in the absence of hormone. 93
Activity of Estrogen Receptor In The Presence of YEPE13 (Variant ER)
2000-1
0 Cw 1000
+H
-H 4b p 0 I arz-rzz1 I 0 10 20 30
CuSO4(uM)
Figure 38. Transcriptional activity of wild-type ER in the presence of variant, YEPE13. Yeast cells containing receptor expression plasmids (YEPE13 and YEPE23) were transformed with reporter plasmid
(YRPE2). These cells were grown in minimal medium in the presence and absence of estradiol with increasing concentration of CuSO4. The beta-galactosidase activity measured in yeast extracts is expressed in Miller units per mg protein. 94
Activity of Estrogen Receptor In The Presence of YEPE15 (Variant ER)
2000-1
0 C 1000 0
+H
-H 0 0 10 20 30
CuSO4(uM)
Figure 39. Transcriptional activity of wild-type ER in the presence of variant, YEPE15. Yeast cells containing receptor expression plasmids (YEPE15 and YEPE23) were transformed with reporter plasmid
(YRPE2). The cells were grown in minimal medium in the presence and absence of estradiol with increasing concentration of CuSO4. The beta-galactosidase activity measured in yeast extracts is expressed in Miller units per mg protein. 95
Activity of Estrogen Receptor In The Presence of YEPAR
2000 ,
C S 1000-
-H
0 - I 0 10 20 30
CuSO4(uM)
Figure 40. Transcriptional activity of wild-type ER in the presence of YEPAR. Yeast cells containing plasmid (YRPE2). The receptor expression plasmids (YEPAR and YEPE23) were transfonned with reporter concentration cells were grown in minimal medium in the presence and absence of estradiol with increasing of CuSO4. The beta-galactosidase activity measured in yeast extracts is expressed in Miller units per mg protein. 96
Wild Wild Type Variant Variant Type
kDa 200- 116- 9 7- 6 6-
-S 4 2- 31-
CuSQ4 (25pM) - -+ -+ +
Figure 41. Western blot analysis. Cytosols prepared from yeast cells containing either the wild-type or
variant ER expression plasmids were analyzed by Western blotting after resolution on a 10% PAGE.
Molecular weight markers (Std.; in thousands) were included. 97
In the cells transformed with the plasmid in which the ER coding region had been deleted (YEPAR), no effect was observed on GAL4 activity under any conditions. Whereas
expression of variant YEPE13 reduced GALA activity up to 73% and 70%, and expression
of variant YEPE15 reduced GALA activity up to 83% and 84% in the absence and presence of hormone, respectively (Fig 42).
To confirm that the transcriptional interference observed was due to a direct
interaction of the hormone receptor complex with DNA and was not the result of
squelching of some transcription factor required for GAL4 function (184), an
oligonucleotide that contained an overlap between the GAL4 binding sequence and a
vitamin D response element was generated. It has been shown that the estrogen receptor is unable to interact with this element in vitro (115). Furthermore it will not confer estrogen
responsiveness to the CYC1 promoter. Two copies of this oligonucleotide were inserted
into the vector PC3, and the resulting recombinant (YRPVG2) was transformed into yeast cells containing either the wild type receptor expression plasmid or the variant receptor expression plasmids. As expected, this reporter demonstrates the usual high transcriptional activity in the presence of galactose. The addition of receptor or hormone had no significant effect on GAL4 activity (data not shown).
These data suggest that variant receptors compete for ERE binding with the wild type receptor and exert dominant-negative effect by saturating ERE binding sites. This is the first report of a naturally occurring, alternatively-spliced ER variants which interfere with ER-induced gene transcription. 98
GAL4 TRANSCRIPTIONAL INTERFERENCE ASSAY
12000 10000 -H 0 +H 8000 C
6000
S 4000
2000
0 YEPE10 YEPAR YEPE13(V) YEPE15(V)
Figure 42. GAL4. transcriptional interference assay. Yeast cells containing the YRPEG3 reporter were transformed with either a receptor expression plasmid expressing wild-type ER or with plasmids expressing variant ERs. The cells were grown in galactose either in the absence of hormone or the presence of estradiol with 25uM CuSO4. The beta-galactosidase activity measured in yeast extracts is expressed in Miller units per mg protein. 99
Regulation of osteocalcin gene by estrogen receptor
The human osteocalcin gene is regulated in mammalian osteoblasts by
1,25(OH)2D3 dependent and independent mechanisms (117). The sequences responsible for the basal and hormone-inducible activities of the human osteocalcin gene have been mapped within a region downstream of -1339 (117). It has been shown that this enhancer region functions analogously in Saccharomyces cerevisiae cells engineered to produce active 1,25(OH)2D3 receptor (117).
To address whether the estrogen receptor can stimulate the osteocalcin gene, the
XhoI-SalI fragment of the human osteocalcin gene promoter was fused to the proximal promoter elements of the yeast iso-1 cytochrome c gene, generating Y RPV4 reporter plasmid. This plasmid was transformed into yeast cells containing the estrogen receptor expression plasmid, and cotransformants were selected by tryptophan and uracil prototrophy. The transformants were grown in the presence or absence of estradiol with increasing concentration of copper sulfate. The cytosols were prepared and analyzed for beta-galactosidase activity (Fig. 43). The reporter demonstrated a high basal promoter activity. This activity was elevated further by estrogen. The induction by estrogen was maximally active even at low concentrations of the receptor. These data suggest that the estrogen receptor regulates the osteocalcin gene in a manner similar to 1,25(OH)2D3 receptor.
To reconstitute a regulatable transcriptional unit in yeast, the estrogen receptor expression unit was integrated into the yeast genome. This strain was then transformed with the reporter plasmid YRPV4. The expression of ER in this strain is approximately 10 to 20-fold lower then that in the cells carrying multicopy episomal expression units, and is inducible by copper (Fig. 19). The cells were grown in the presence or absence of hormone with increasing amounts of copper sulfate and the induction of lacZ was measured 100
5000
4000
(0 3000-
Sam * Miller Units * Miller Units 2000
1000 -
H its. " -.- 0 0 10 20 30 40 50 60 CuSO4(uM)
Figure 43. Estradiol dependent and independent activation of transcription of human osteocalcin gene. Cells
containing osteocalcin reporter plasmid (YRPV4) and expression plasmid (YEPE10) were grown in
minimal medium with or without estradiol with increasing concentration of copper sulfate. The cells were
harvested and beta-galactosidase activity were measured. Data are expressed as Miller units per mg protein. 101
M Miller Units 1500 - Miller Units +H
0 1000
500
0- -H
0 0 10 20 30 40 50 60 CuSO4(uM)
Figure 44. Hormone and receptor dependent and independent activation of transcription of human osteocalcin gene. Cells containing reporter plasmid (YRPV4) and integrated receptor expression plasmid (YIPE 10) were grown in minimal medium with or without estradiol with increasing concentration of copper sulfate. Cells were harvested and beta-galactosidase activity were measured. Data are expressed as Miller Units per mg of protein. 102
(Fig. 44). The reporter demonstrated again a basal activity, and this activity was inducible
by hormone. These results also suggest that the induction was directly proportional to the
receptor (copper) concentration.
The above and some previous studies (117) demonstrate that the osteocalcin gene is
regulated by both estrogen and 1 ,25(OH)2D3. The possibility of synergism between
estrogen and vitamin D3 receptors on the osteocalcin gene was further studied. The cells
containing reporter plasmids were cotransformed with vitamin D3 (117) and estrogen
receptor expression plasmids. Cotransformants were selected for tryptophan, uracil, and
leucine prototrophy. The cells were grown at 100 uM CuSO4 with constant estradiol (10-6
M) and increasing concentrations of 1,25(OH)2D3 or constant 1,25(OH)2D3 and
increasing concentrations of estradiol. The cytosols were prepared and the induction of
beta-galactosidase was measured (Fig. 45). Fig. 45 suggests that the transactivation of
osteocalcin gene by estrogen and vitamin D3 receptors is non-synergistic. However, recent
data in our lab has demonstrated that a nuclear accessory factor, NAF, required in
mammalian cells for vitamin D receptor activity is absent in yeast.
In an attempt to define the location of the estrogen receptor binding site in the
osteocalcin promoter region, a lkbp BamHI- BamHI fragment of YRPV4 was cloned into
the cognate site of PGEM-4Z vector in both (+) and (-) orientations. Both 5' and 3'
deletions of the osteocalcin gene promoter were performed using the Promega erase-a-base
system. All deleted fragments were excised from PGEM-4Z vectors by BamHI-HindILI,
repaired by Kienow and inserted into the end-filled XhoI site of the vector PC2 (114, 117).
The reporter plasmids carrying 5' and 3' deleted fragments of osteocalcin promoter were
transformed into yeast cells expressing estrogen receptor under control of the copper metallothionein promoter. The cells containing both receptor expression and reporter plasmids were grown in the presence or absence of hormone and copper sulfate. The 103
Miller Units 1500
1000
500
0 0 10-11 10-10 10-9 10-8 10-7
1, 25(OH)2D3
Miller Units 2000
1000
0. 0 10-11 10-10 10-9 10-8 10-7
Estradiol
Figure 45. Regulation of osteocalcin gene. Yeast cells containing reporter plasmid (YRPV4) were transformed with estrogen (YEPEIO) and vitamin D3 (YEPVI) receptor expression plasmids. Cells were grown in minimal media with lOOuM CuSO4 and constant estradiol (106) and increasing concentrations of 1,25(OH)2D3 or constant 1,25(01-1)2D3 (10") and increasing concentrations of estradiol. The cytosols were prepared and induction of beta-galactosidase was measured. Data are expressed as Miller Units per mg of protein. 104 cytosols were prepared and induction of beta-galactosidase was measured (Fig. 46A and
B). Cells carrying the reporter plasmid containing a 1kb fragment (-1334 to -306) of the osteocalcin promoter demonstrated a high basal promoter activity.
This activity was further elevated by estrogen. On the other hand, cells carrying reporter plasmids containing deleted fragment (both 5' or 3' deletions) exhibited no induction in the transcriptional activity of the lacZ gene in the presence of hormone. These data also suggest the presence of binding sites for two proteins that positively regulate and one that negatively regulates transcription in the osteocalcin promoter. The first positive activator binding site was mapped to the 5' end of the promoter between regions -1183 and -1033. The second site was mapped to the 3' end of the promoter between regions -883 and -583. The negative factor binding site was mapped to the central region of the promoter between regions -1033 and -883. In summary, these studies were unsuccessful in defining the estrogen receptor binding site in osteocalcin promoter.
In conclusion this study was successful in overexpressing the biologically active form of estrogen receptor and in reconstitution of an estrogen responsive transcriptional unit in yeast which has a wide range of applications, as noted several time in this paper. 105
5' Deletion of Osteocalcin Promoter in (+) Orientation 5' Deletion of Osteocalcin Promoter in (+) Orientation 8000, 8000
7000 7000
6000 6000 i 5000 - 3 - 5000
4.. 4000 4000
2 3000 3000
2000- 2000 100:1 1000.
I n AI -1400 -1200 -1000 -000 -600 -41 00 -1400 -1200 -1000 -800 -600 -400 Fragment Size Fragment Size
5' Deletion of Osteocalcin Promoter In (+) Orientation 5' Deletion of Osteocalcin Promoter In (+) Orientation 8000 8000
6000 6000
' 4000-e 4000
2000 2000
-1400 1200 -1000 -800 -00 -400 -1400 -1200 -1000 -800 -600 -400 Fragment Size Fragment Size
Figure 46. Deletion analysis of osteocalcin gene promoter. (A) The reporter plasmids carrying 5' deleted fragments of osteocalcin promoter were transformed into yeast cells expressing estrogen receptor. 106
3' Deletion of Osteocalcin Promoter in (+) Orientation 3' Deletion of Ostecalcin Promoter in (+) Orientation
4000 4000 1
3000 3000
3 2
2000 + 2000
1000- 1000 -
n A- V)--- A -L -1400 -1200 -1000 -800 -400 -400 -200 -1400 -1200 -1000 -800 -600 -400 -200 Fragment Size Fragment Size
3' Deletion of Osteocalcin Promoter in (+) Orientation 3' Deletion of Osteocalcln Promoter in (+) Orientation
4000 4000 -
3000 3000
-J 2000 -i 2000
1000 - 1000
k- 01 - -- -1400 -1200 -1000 -800 -600 -400 -200 -1400 -1200 -1000 -800 -600 -400 -200 Fragment Size Fragment Size
(B) The reporter plasmid carrying 3' deleted fragments of osteocalcin promoter were transformed into yeast cells expressing estrogen receptor. The cells were grown in the presence or absence of CuSO4 (100uM) and estradiol. Cytosols were prepared and induction of beta-galactosidase was measured. Data are expressed as
Miller units per ing of protein. CHAPTER 4
DISCUSSION
Steroid hormones and vitamin D3 exert their biological effect in eukaryotic cells via specific intracellular high affinity receptors (5). These receptors consist of multiple domains, the organization of which are conserved among all members of the steroid receptor family (4). They contain a variable amino-terminal transcriptional activating domain, a central DNA-binding domain, and a carboxyl-terminal ligand-binding domain (4 5).
Conventional receptor isolation techniques can provide only small quantities of purified receptor from a target tissue for biological analysis. Methods such as transient transfection into mammalian cells and in vitro transcription and translation have greatly facilitated the functional dissection of the respective receptor domains, but the amount of receptor material provided by these techniques is not enough for most structural and biochemical analyses. This demonstrated the need for developing a heterologous expression system that could be used to produce large amounts of functionally active receptors.
Bacterial Expression System
Several laboratories, including ours, have attempted to overexpress either truncated or fused steroid receptor genes with variable results (129, 185). In this study, a novel method for overexpression of full length receptor proteins from complementary DNA (cDNA) in E. coli has been developed.
107 108
To express full length human vitamin D3 and estrogen receptors in E. coli, the
carboxyl-terminal glycine of the ubiquitin gene was fused to the amino-terminus of the
receptor genes and the expression of ubiquitin-vitamin D3 receptor, UB-VDR, and
ubiquitin-estrogen receptor, UB-ER, molecules were analyzed. Introduction of the lambda
PL promoter into PUB-VDR and PUB-ER expression vectors, and induction at low
temperature by nalidixic acid (139-140) resulted in the synthesis of biologically active form of UB-VDR and UB-ER proteins.
Previously, the vitamin D3 receptor in E. coli was produced in an unfused form.
This receptor was incapable of binding hormone and was totally insoluble (McDonnell,
Ph.D. Thesis). Fusion of vitamin D3 receptor with ubiquitin and heat induction of UB
VDR protein from the lambda PL promoter resulted in the soluble form of the receptor.
However, as before, this material was unable to bind hormone. Similarly, when the
estrogen receptor gene was fused to ubiquitin, only small quantities of intact receptor were
seen in response to the induction by the 420 C temperature shift and, as with UB-VDR
protein, UB-ER material did not exhibit specific estrogen binding.
These results suggest that the thermal induction is not an appropriate method for
the production of biologically active receptors in E. coli. Apart from the well documented
heat lability of hormone-binding activity, it is also known that a variety of E. coli proteases
are heat inducible; for example, the Ion protease is controlled by the heat shock protein
regulator gene (186). As an alternative approach, thermal induction was replaced by low
temperature (<300C) nalidixic acid induction.
The vitamin D3 and estrogen receptors induced by nalidixic acid bind 1,25(OH)2D3 and estradiol respectively with affinities similar to those reported for the wild-type receptors. Scatchard analysis indicated only one type of binding site for vitamin D3 and estrogen receptors with Kd values of 2.8 X 10-11 M and 3.2 X 1010 M, respectively. 109
Furthermore, the total specific binding activity was ranged from 600 to 800 fmol per mg of
protein for the vitamin D3 receptor and from 400 to 1,000 fmol for the estrogen receptor.
The production of high affinity binding forms of the vitamin D3 and estrogen
receptors in E. coli demonstrates that bacteria possess enzymes for correct post
translational modification of the receptors to a form capable of binding appropriate
hormones with high affinity. Furthermore, it is known that in eukaryotic cells, these
receptors are associated with the heat shock protein 90 (HSP 90) (187). This protein is not
found in prokaryotic cells. It has been postulated that these receptors are maintained in the
untransformed 8S state through selective interaction with HSP 90. Upon binding with the
hormone, the complex dissociates yielding free HSP 90 and transformed receptor (188).
This system is useful to produce receptor proteins which are never exposed to HSP 90.
Ubiquitin fused to the steroid receptor genes is known to increase the yield of
biologically active receptor proteins. It enhances the qay 2 and quantity of the expressed
fusion protein by different mechanisms (143): promotion of proper folding of the fused
protein thus preserving its biological activity, protection of the amino-terminus of the fused
protein from degradation by its compact globular structure, solubilization of high
concentrations (50 mg/ml) of the protein in an aqueous environment by having an outer
hydrophilic surface (189-190). These unique properties of ubiquitin may assure the
appropriate compartmentalization of the fused proteins in the cell and also prevent
denaturation and formation of inclusion bodies (143).
The vitamin D3 and estrogen receptors produced in bacteria as ubiquitin fusion
proteins bind to DNA with wild-type properties. The DNA-binding properties of vitamin
D3 receptor were observed by calf thymus DNA-cellulose chromatography. The DNA binding properties of estrogen receptor were analyzed by gel-retardation assay. The binding of E. coli expressed estrogen receptor to its cognate response elements is independent of 110
either agonist or antagonist. This result is in agreement with other results (121, D. P. McDonnell, pers. comm.).
These studies not only demonstrate that the vitamin D3 and estrogen receptors can be produced in a biologically active form in bacteria, but also indicate that all the processes required for production of the hormone and DNA-binding species are present. They further show that the utilization of the nalidixic acid induction protocol can substantially improve the quality of the recombinant receptors. Thus it may be the protocol of choice for the production of labile proteins in E. coli. The use of this system to over-produce receptors for crystallization studies is not practical due to the low recovery of receptor proteins and due to heterogeneity as a result of proteolysis. However, this system is useful for producing receptor proteins for in vitro phosphorylation studies and some other studies which require receptor that is not modified by eukaryotic cellular processes.
Development of Saccharomyces cerevisiae as an Expression System
The ability to produce large quantities of steroid receptors is important to understand certain fundamental mechanisms of action of the steroid hormone receptors.
The availability of large amounts of purified receptors would allow crystalization of the receptor molecules and definition of the points of contact of ligand with its receptor, and enable physical and biochemical analysis of the receptors. Unfortunately, it has been impractical to examine extensively ligand-receptor interactions or to study the physical changes in the receptor structure upon binding hormone or DNA due to the extremely low abundance of these receptors in the target tissues, their inherent sensitivity to proteolysis and the failure of bacterial expression systems to produce homogenous material.
After the failure of the bacterial expression system, Saccharomyces cerevisiae was the most important candidate for overexpression of steroid receptors. Previously, S. 111 cerevisiae has been used successfully to produce and purify the vitamin D3 receptor in large quantities (191). The quality of this receptor was far superior to that isolated from other heterologous expression systems. In this study, a similar approach was used to express an estrogen receptor in S. cerevisiae. This approach utilized the ubiquitin fusion technology to express the estrogen receptor in yeast. The amino-terminus of the estrogen receptor cDNA was ligated in frame to ubiquitin, creating a synthetic fusion molecule. This fusion cDNA was cloned into a high-copy yeast expression plasmid that contained a CUP1
(metallothionein) promoter. The production of the receptor was initiated by the addition of copper sulfate.
The recombinant protein is produced initially as a fusion molecule, ubiquitin receptor, but soon after translation, a host processing activity (ubiquinase) faithfully removes the ubiquitin, leaving the intact receptor (163). This system enhances the level of production of estrogen receptor, presumably by enhancing stability and solubility and by affording protection from endogenous proteolysis (164). This processing activity is present in the eukaryotic cells only (164). Western blott analysis of yeast-expressed estrogen receptor suggests that this receptor co-migrates with the authentic estrogen receptor produced in mammalian CHO cells. Mammalian ER standards, accurately quantitated by hormone binding, indicates that these yeast cells produce 3-4 picomoles of ER per mg of total protein. The receptor was homogenous and no significant breakdown of the receptor was apparent. The slightly higher molecular weight of the recombinant receptor was due to the 14 amino acids that remained on the receptor following the removal of ubiquitin.
The hormone-binding properties of the recombinant molecules were undistinguishable from those of the wild-type receptor. The DNA-binding characteristics of the receptor were normal, as assayed by calf thymus DNA-cellulose chromatography and band shift analysis (121). 112
The in vitro DNA-binding studies suggest that the recently cloned wild-type hER
expressed in yeast binds to DNA independent of hormone, whereas, under similar
conditions, the mutant ER requires hormone (121,156). These in vitro data appear to
conflict with in vivo data, where hormone is required for DNA binding (114-115, 120,
192). A widely accepted view is that DNA binding itself is not dependent on hormone but
in vivo, hormone is required for the dissociation of the receptor from inhibitory proteins (121).
The system utilized to express the estrogen receptor in yeast indicates that several of
its components can be changed to increase expression levels of the receptor. Such
parameters as: physical changes of the promoter, of the origin of replication, and of the
selectable marker or changes in the protocol for the induction of the receptor are all possible and were evaluated systematically.
Strains of S. cerevisiae defective in leucine biosynthesis have an increased plasmid
copy number as more of the plasmid is required to complement the cellular mutation (165).
This knowledge was used successfully to produce recombinant protein (145). The
inclusion of a defective leu2 allele into the estrogen receptor expression plasmid resulted in
the production of 5-fold more receptor up to a minimal level of 0.2% of the total soluble
yeast protein.
Only a fraction of the receptor produced in yeast is capable of binding hormone as previous reports by our and other laboratories have indicated (113, 116-117). Such results may be artifacts from the use of standard protocols for yeast cell disruption rather than differences in the covalent modifications of receptor. Recent reports on the production of the glucocorticoid receptor in yeast have indicated that the receptor is transcriptionally active
(193). However, there have been difficulties in isolating biologically active receptor from these cells. These problems may be related to the denaturation of the receptor during 113
isolation. The enzymatic extraction procedures significantly improve the hormone-binding
and DNA-binding properties of the estrogen receptor produced in yeast. This suggests that
the yeast system should be suitable for overexpression of other members of the steroid receptor superfamily as well.
Finally, this yeast system has been proven to be cost effective and easy to
manipulate. At present, it is used in our laboratory to generate the milligram quantities of
recombinant receptors required for physical and biochemical analysis.
Development of Estrogen Responsive Transcription Unit in Saccharomyces cerevisiae
A cis-trans transcription assay was developed to test the transcriptional activity of the estrogen receptor produced in yeast. This beta-galactosidase assay consists of determining the ability of the estrogen receptor to activate transcription of reporter in a yeast cell in which the estrogen receptor expression plasmid and an estrogen responsive reporter plasmid were cotransformed.
The reporter plasmid contains the yeast iso-1-cytochrome C proximal promoter elements fused to the E. coli beta-galactosidase gene. The behavior of this particular CYC
Lac Z fusion has been well characterized (166). The CYC I upstream activating sequences have been removed and replaced with synthetic oligonucleotides containing the estrogen response elements.
The transcriptional activity of hER expressed in yeast was absolutely dependent on estrogen. In the presence of hormone, transcription was maximally active even at a low concentration of receptors. These results suggest that the reconstituted estrogen-regulated transcription unit functions in yeast similar to that observed in transfected mammalian cells. 114
Application of Saccharomyces cerevisiae System
The reconstituted hormone responsive transcription unit in yeast will permit an
efficient genetic microdissection of the receptor structures. This host cell system will allow
studies of covalent receptor modifications (e.g. phosphorylation) required for "activation",
as well as studies to elucidate other "helper proteins" involved in the steroid receptor
regulation of the gene expression. The system would also be useful to enhance our
understanding about hormone and antihormone actions. Furthermore, the system can be
used to screen for agonists (hormones) and antagonists (antihormones), for "orphan"
receptors and for other members of the steroid hormone receptor superfamily. The system
has a great variety of other applications in addition to the ones mentioned above.
Effect of antiestrogens on the regulation of human estrogen receptor
In this study, different antihormones were used to examine the regulation of human
estrogen receptor. Data from these studies indicate that in vivo antiestrogens can promote
DNA binding of the wild-type ER without efficiently inducing transactivation. This is an
indication that the activity of ER can be regulated subsequent to DNA binding and reveals a
requirement of hormone to activate the transactivation domains in vivo.
An ER chimera containing a strong unregulated transactivator was then constructed
to examine whether the transactivation functions of the ER are normally under direct
hormonal control. ER-VP16 chimera, in which the acidic region of VP16 was fused to the
N-terminal region of the ER, is still hormone-regulated, although it can be activated by estrogen antagonists. This suggets that both estrogen and antiestrogen can allow the ER to
bind DNA in vivo by inducing conformational changes, but only the agonist promotes a conformation of the wild-type ER that induces gene transactivation in both mammalian and 115
yeast cells. As the hormone binding assays do not reveal significant differences in ligand binding affinities between ER and ER-VP16 in vitro (unpublished data), it is highly
unlikely that the insertion of the VP16 acidic transactivator would alter the structure of the hormone binding domain.
All of the antiestrogens examined can activate ER-VP16 efficiently. ICI 164,384
(194), thought to be a pure antagonist, appears to induce DNA binding as efficiently as the
triphenylethylene antiestrogens tamoxifen and nafoxidine, partial agonists in most cell types (168). As the results suggest, these antagonists differ primarily in the extent to which they allow ER to interact conformationally with the transcriptional machinery.
The result obtained with ICI 164,384 is in agreement with that of Martinez and
Wahli (195) who have shown that this antiestrogen can stimulate the binding of ER to
vitellogenin EREs in vitro. Fawell et al., (196) reported data that suggest inhibition of DNA
binding by ICI 164,384 interfers with receptor dimerization. Although this may appear to contradict the results of this study, the two observations can be reconciled. It is possible
that the VP16 transactivation domain may have a dimerization function that has not been yet
identified, but a more likely explanation is that the dimerization function associated with the
hormone binding domain (94) is not absolutely required for DNA binding, although it may
enhance the binding affinity. It is possible that the concentration of ER in the transiently
transfected cells is sufficiently high to overcome the decrease in affinity resulting from an inhibition of dimerization. It has been shown that an ER derivative, in which the hormone
binding domain and the dimerization motif have been deleted, has some transcriptional activity and even full transcriptional activity in some cell types (167-168). 116
Identification of novel estrogen response elements
Most current methods for the identification of the steroid response elements rely on
a deletion analysis of steroid responsive promoters to define the minimal amount of
sequence required to transfer hormonal regulation from one gene to another (178, 197). An
alternative method employed in this study was the utilization of a reconstituted estrogen
response system in yeast for screening large libraries of possible steroid response elements.
Using this approach, seven estrogen response elements have been identified. A comparison
of these sequences suggests that they are all unrelated to the consensus vitellogenin ERE
(152). The surprising result was the presence of the two half sites arranged as direct repeats
in some of the clones. This indicates that the estrogen receptor can transactivate genes in
yeast via response elements in which the two half sites align as direct repeats. This
hypothesis was confirmed by examining the induction of gene transcription using a
synthetic response element which contained two receptor binding sites as direct repeats.
This study confirmed that the estrogen receptor can activate gene transcription through
response elements where the receptor recognition sites are present as direct or indirect
repeats. This implies that the receptor may have more than one way of interacting with
DNA, and that modulation of these reactions may be another way the cell can generate
specificity in response to steroids.
Although it has been suggested that thyroid hormone receptor, retinoic acid, and
vitamin D3 receptors can bind selectively and transactivate genes from direct repeat
response elements (197), the results presented here contradict those of Naar et al., (198), who indicate that the estrogen receptor cannot induce gene transcription from response elements in which receptor binding sites are aligned as direct repeats. It is quite possible that these differences may be due to cell specific factors which modulate gene induction via 117
protein-protein interaction and thus, the response elements should be tested in a wide variety of eukaryotic cells.
The comparison of the estrogen response elements with the mammalian gene bank
revealed that homologous sequences were present in the promoters of various mammalian
genes. Most of these genes were reported to be regulated by estrogen, but their response
elements have not yet been identified.
The yeast system described in this study is not limited to a single receptor and it can
be used to identify the response elements for any DNA binding protein including those for
the "orphan" members of the steroid receptor superfamily. To confirm this, the system was
also utilized to identify the progesterone receptor response elements (PREs). In view of the
expanding number of sequences in the gene bank, it is possible that this approach may be
useful in identifying new candidate steroid responsive genes.
The newly developed yeast system has major advantages over in vitro methods
(199-201). It does not require purified protein, high affinity antibodies, or a heterologous
expression system. The yeast system also has some advantages over the previously
developed in vivo system that was employed to identify the NGFI-B DNA binding sites
(202). The in vivo system utilized the DNA binding domain of NGFI-B expressed as part
of a LexA-NGFI-B-Gal4 chimeric protein. The LexA DNA binding domain may interfere
with the selection of target sequences whereas the yeast system does not require fused
protein or a two-step selection.
There are some other advantages of the yeast system. First, using a random oligo approach, all possible response elements can be identified, not just those based upon a known sequence. Secondly, the screen can be designed to rank elements relative to activity.
Thirdly, the system is rapid and requires only the expression of the receptor in yeast.
Fourthly, the same oligo library can be used for most steroid superfamily members. 118
Finally, by comparing the sequences of the identified elements, a more informative
consensus binding sequence can be derived.
Study of variant receptors cloned from breast tumors
It is known that the estrogen and progesterone receptors play a major role in the
progression of breast cancer, PR being induced by estrogen. Breast cancer lacking estrogen
receptor would also be expected to lack progesterone receptor as well. Nevertheless, ER
/PR+ tumors do occur. This paradox might sometimes result from a variant ER that is
unable to bind the hormone, but which can still stimulate estrogen-responsive genes, such as PR.
In an experiment using a variant ER cloned from ER-/PR+ human tumors and lacking exon 5 of the hormone binding site, the transcriptional activity of the variant was compared to that of the wild-type ER. The purpose was to determine the consequences of the exon 5 deletion. For this the hormone-responsive transcriptional unit that had been reconstituted in Saccharomyces cerevisiae was employed, and the variant ER exhibited hormone independent transcriptional activity. This suggests that the chronic presence of low levels of constitutively active ER or its overexpression (observed with this variant) as compared to the wild-type ER may be sufficient to activate target gene expressions and account for PR expression in apparently ER-/PR+ breast tumors.
Studies of ER deletion mutants generated by site-directed mutagenesis showed that certain mutants lacking virtually all of the hormone-binding domain are unable to bind ligand and may activate transcription in a constitutive manner (36, 166). This constitutive transcription may be the explanation for the apparent paradox of the ER-/PR+ breast cancer phenotype in some instances, as studies with variant ER (missing exon 5) described here also indicate. 119
The ER+/PR- tumors, containing estrogen receptor, would be expected to contain progesterone receptor as well. The fact that 10% of the tumors are of this type could be explained by the existence of an ER unable to function as a transcriptional inducer of PR expression. Two such variants, lacking exon 7 of the hormone binding domain were cloned from ER+/PR- human tumors. The yeast expression system was employed to test the effect of exon 7 deletion on the transcriptional activities of ERs. The variants were transcriptionally inactive and this activity was unaffected by the overexpression of receptor.
These results suggest that the variant ERs expressed in ER+/PR- tumors can not stimulate the expression of target genes, and that they account for this kind of tumors.
In addition to variant ERs, full-length wild-type ER was also detected in ER+/PR tumors. To study the effect of the variant on the transcriptional activity of the wild-type ER, both the wild-type ER and increasing amounts of variant receptors were coexpressed. The transcriptional activity of wild-type receptor was reduced in this situation, which suggests that the variant receptors exert a dominant-negative effect on the wild-type activity. This effect could be a result of the constitutive DNA-binding ability of the variant receptors, which interferes with the wild-type ER binding to ERE. Alternatively, it may involve mechanisms unrelated to DNA binding, such as the formation of an inactive heterodimer.
The results presented in this study suggest that variant receptors compete with the wild-type receptor for the ERE binding sites and exert their dominant-negative effect by saturating ERE-binding sites. These dominant-negative ER variants in breast tumors could have important effects on the hormonal responsiveness. The ER-/PR+ and ER+/PR tumors would be unresponsive to endocrine therapies directed to the estrogen receptor hormone-binding site, such as estrogen deprivation from ovariectomy or from aromatase inhibitors, and antiestrogen treatments. On the other hand, endocrine therapies not acting 120
directly through the ER, such as pharmacological doses of estrogens or progestins, might still be effective.
This is the first reported study of naturally occurring, alternatively-spliced ER variants that interfere with the ability of wild ER to induce gene transcription.
Regulation of the osteocalcin gene by estrogen receptor
The major noncollagenous protein of bone is osteocalcin, a 49 amino-acid protein
(117). The expression of this protein is restricted to osteoblast or osteoblast-like cells. Its regulation is biphasic and occurs by 1,25(OH)2D3 dependent and independent mechanisms
(117). The sequences responsible for basal and hormone-inducible activities of the human gene have been mapped (117). It has been shown that this enhancer region functions analogously in Saccharomyces cerevisiae cells engineered to produce active 1,25(OH)2D3 receptor (117).
Estrogen plays a crucial role in the regulation of bone metabolism (203). However, the mechanism by which estrogen exerts its effect on bone is unknown. To examine the possible direct effect of estrogen on bone-forming cells, the promoter region of the osteocalcin gene was cloned into the yeast system and the regulation of osteocalcin was examined under estrogen treatment. This study demonstrated that estrogen regulated the expression of osteocalcin gene in a manner similar to that observed for 1,25(OH)2D3
(117).
To test the possibility that both estrogen and 1,25(OH)2D3 receptors can regulate the osteocalcin gene synergistically, the regulation of osteocalcin gene was tested in yeast cells where both receptors were coexpressed. The results suggest that transactivation of the osteocalcin gene by estrogen and by vitamin D3 receptors is not synergistic, and that they are similar to those obtained for rat osteocalcin gene (204). 121
The attempt to define the estrogen receptor-binding site in the osteocalcin promoter was unsuccessful. However, this study defined two positive and one negative activator binding sites in the osteocalcin promoter. The first positive activator binding site was mapped to the 5' end of the promoter. The second site was mapped to the 3' end of the promoter. The negative activator-binding site was mapped to the central region of the promoter. In conclusion, these studies confirm that estrogen regulates bone metabolism by regulating the expression of the osteocalcin gene in bone cells.
In summary, this whole study was successful in overexpressing the biologically active form of the estrogen receptor and in reconstituting an estrogen responsive transcriptional unit in Sacharomyces cerevisiae. This novel system has a wide range of applications, some of them mentioned in this document. REFERENCES
1. Yamamoto, K. R. (1985) Ann. Rev. Genet. 19, 209-252.
2. Charron, J., Richard-Foy, H., Berard, D. S., Hager, G. L., and Drouin, J. (1989) Mol. and Cell. Biol. 9, 3127-3131.
3. Sakai, D. D., Helms, S., Carlstedt-Duke, J., Gustafsson, J-A., Rottman, F. M., and Yamamoto, K. R. (1988) Genes & Dev. 2, 1144-1154. 4. Evans, R. M. (1988) Science 240, 889-895.
5. O'Malley, B. W. (1990) Mol. Endocrin. 4, 363-369. 6. Funder, J. W. (1987) Science 237, 236-237.
7. Rories, C., and Spelsberg, T. C. (1989) Ann. Rev. Physiol. 51, 653-681. 8. Oppenheimer, J. H. (1979) Science 203, 971-979.
9. Ragsdale, Jr. C. W., Petkovich, M., Gates, P. B., Chambon, P., and Brockes, J. P. (1989) Nature 341, 654-657.
10. Durston, A. J., Timmermans, J. P. M., Hage, W. J., Hendriks, H. F. J., de Vries, N. J., Heideveld, M., and Nieuwkoop, P. D. (1989) Nature 340, 140-144.
11. Conneely, 0. M., Sullivan, W. P., Toft, D. 0., Birnbaumer, M., Cook, R. G.,
Maxwell, B. L., Zaruchi-Schuiz, T., Green, G. L., Schrader, W. T., and O'Malley, B. W. (1986) Science 233, 767-770.
12. Gronemeyer, H., Turcotte, B., Quirin-Stricker, C., Bocquel, M. T., Meyer, M. E., Krozowski, Z., Jeltsch, J. M., Lerouge, T., Garnier, J. M., and Chambon, P. (1987) EMBO J. 6, 3985-3994.
122 123
13. Misrahi, M., Atger, M., d'Auriol, L., Loosfelt, H., Meriel, C., Fridlansky, F.,
Guiochon-Mantel, A., Galibert, F., and Milgrom, E. (1987) Biochem. and Biophys.
Res. Comm. 143, 740-748.
14. Koike, S., Masaharu, S., and Maramatsu, M. (1987) Nucleic Acids Res. 15, 2499 2513.
15. Walter, P. W., Green, S., Krust, A., Bornert, J-M., Jeltsch, J-M., Staub, A., Jensen,
E., Scrace, G., Waterfield, M., and Chambon, P. (1985) Proc. Natl. Acad. Sci. USA 82, 7889-7893.
16. Chang, C., Kokontis, J., and Liao, S. (1988) Science 240, 324-327.
17. Lubahn, D. B., Joseph, D. R., Sullivan, P. M., Willard, H. F., French, F. S., and
Wilson, E. M. (1988) Science 240, 327-330.
18. Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G., Oro, A., Lebo, R.,
Thompson, E. B., Rosenfeld, M., G. and Evans, R. M. (1985) Nature 318, 635-641.
19. Miesfeld, R., Rusconi, S., Godowski, P. J., Maler, B. A., Okret, S., Wkstrom, A. C., Gustafsson, J-A., and Yamamoto, K. R. (1986) Cell 46, 389-399.
20. Murry, J. C., Smith, R. F., Ardinger, H. A., and Weinberger, C. (1987) Nucleic
Acids Res. 15, 6765.
21. Arriza, J. L., Weinberger, C., Cerelli, G., Glaser, T. M., Handelin, B. L.,
Houseman, D. E., and Evans, R. M. (1987) Science 237, 268-274.
22. Patel, P. D., sherman, T. G.,Goldman, D. J., and Watson, S. J. (1989) Mol. Endocrin. 3, 1877-1885.
23. Thompson, C. C., Weinberger, C., Lebo, R., and Evans, R. M. (1987) Science 237, 1610-1614.
24. Benbrook, D., and Pfahl, M. (1987) Science 238, 788-791.
25. Giguere, V., Ong, E. S., Segui, P., and Evans, R. M. (1987) Nature 330, 624-629. 124
26. Petkovich, M., brand, N. J., Krust, a., and Chambon, P. (1987) Nature 330, 444 450.
27. Zelent, A., Krust, A., Petkovich, M., Kastner, P., and Chambon, P. (1989) Nature 339, 714-717.
28. Krust, A., Kastner, P., Petkovich, M., Zelent, A., and Chambon, P. (1989) Proc. Natl. Acad. Sci. USA 86, 5310-5314.
29. McDonnell, D. P., Mangelsdorf, D. J., Pike, J. W., Haussler, M. R., and O'Malley, B. W. (1987) Science 235, 1214-1217.
30. Baker, A. R., McDonnell, D. P., Hughes, M., Crisp, T. M., Mangelsdorf, D. J., Haussler, M. R., Pike, J. W., Shine, J., and O'Malley, B. W. (1988) Proc. Natl. Acad. Sci. USA 85, 3294-3298.
31. O'Malley, B. W. (1989) Endocrinol. 125, 1119-1120.
32. Debuire, B., Henry, C., Banaissa, M., Biserte, G., Claverie, J. M., Saule, S., Martin, P., and Stehelin, D. (1984) Science 224, 1456-1459.
33. Giguere, V., Yang, N., Segui, P., and Evans, R. M. (1988) Nature 331, 91-94.
34. Wang, L. H., Tsai, S. Y., Cook, R. G., Beattie, W. G., Tsai, M-J., and O'Malley, B. W. (1989) Nature 340,163-166.
35. Watson, M. A., and Milbrandt, J. (1989) Mol. and Cell. Biol. 9, 4213-4219.
36. Kumar, V., Green, S., Stack, G., Berry, M., Jin, J-R., and Chambon, P. (1987) Cell 51, 941-951.
37. Giguere, V., Hollenberg, S. M., Rosenfeld, M. G., and Evans, R. M. (1986) Cell 46, 645-652.
38. Wrange, 0., and Gustafsson, J-A. (1978) J. Biol. Chem. 253, 856.
39. Boer, W., Bolt, J., Kuiper, G. G. J. M., Brinkman, A. 0., and Mulder, E. (1987) J.
Steroid Biochem., 28, 9-19. 125
40. Rusconi, S., and Yamamoto, K. R. (1987) EMBO J. 6, 1309-1315.
41. Green, S., Kumar, V., Krust, P., Walter, P., and Chambon, P. (1986) Cold Spring
Harbor Symposia on Quantitative Biology, Volume LI, 751-758.
42. Carlstedt-Duke, J., Stromstedt, P-E., Wrange, 0., Bergman, T., Gustafsson, J-A.,
Jornvall, H., (1987) Proc. Natl. Acad. Sci. USA 84, 4437-4440.
43. Maxwell, B. L., McDonnell,D. P., Conneely, 0. M., Schulz, T. Z., Green, G. L.,
and O'Malley, B. W. (1987) Mol. Endocrin. 1, 25-35.
44. White, R., Lees, J. A., Needham, M., Ham, J., and Parker, M. (1987) Mol.
Endocrin. 1, 735-744.
45. Haussler, M. R., Mangelsdorf, D. J., Yanaoka, K., Allegretto, B. S., Terpening, C.
M., McDonnell, D. P., Pike, J. W., and O'Malley, B. W. (1988) Steroid Hormone Action, 247-268.
46. Weinberger, C., Hollenberg, S. M., Rosenfeld, M. G., and Evans, R. M. (1985)
Nature 318, 668-672.
47. Kumar, V., Green, G. L., Staub, A., and Chambon, P. (1986) EMBO J. 5, 223 1 2236.
48. Conneely, 0. M., Dobson, A. D. W., Carson, M. A., Maxwell, B. L., Tsai, M-J.,
Schrader, W. T., and O'Malley, B. W. (1987) Molecular Biology of Hormone Action 624th Meeting, Vol. 16, 683-687.
49. Green, S., and Chambon, P. (1987) Nature 325, 75-78.
50. Miller, J., McLachlan, A. D., and klug, A. (1985) EMBO J. 4, 1609-1614.
51. Freedman, L. P., Luisi, B. F., Korszun, Z. R., Basavappa, R., Sigler, P. B., and Yamamoto, K. R. (1988) Nature 334, 543-546.
52. Umesono, K., and Evans, R. M. (1989) Cell 57, 1139-1146. 53. Berg, J. M., (1989) Cell 57, 1065-1068. 126
54. Hard, T., Kellenbach, E., Boelens, R., Maler, B., A., Dahlman, K., Freedman, L.
P., Carlsted-Duke, J., Yamamoto, K. R., Gustafsson, J-A., and Kapstein, R. (1990) Science 249, 157-160.
55. Green, S., and Chambon, P. (1989) Cancer Res. supply ) 49, 2282s-2285s.
56. Severne, Y., Wieland, Schaffner, and Rusconi, S. (1988) EMBO J. 7, 2503-2508.
57. Huckaby, C., Conneely, 0. M., Beattie, W. G., Dobson, A. D. W., Tsai, M-J., and O'Malley, B. W. (1987) Proc. Natl. Acad. Sci. USA 84, 8380-8384.
58. Misrahi, M., Loosfelt, H., Atger, M., Meriel, C., Zerah, V., Dessen, P., and
Milgrom, E. (1988) Nucleic Acids Res. 16, 5459-5472.
59. Ponglikitmongkol, M., Green, S., and Chambon, P. (1988) EMBO J. 7, 3385-3388.
60. Green, S., and Chambon, P. (1988) TIG 4, 309-314.
61. Green, S., Kumar, V., Theulaz, I., Wahli, W., and Chambon, P. (1988) EMBO J. 7, 3037-3044.
62. Mader, S. Kumar, V., de Verneuil, H., and Chambon, P. (1989) Nature 338, 271 274.
63. Danielsen, M., Hinck, L., and Ringold, G. M. (1989) Cell 57, 1131-1138.
64. von der Ahe, D., Renoir, J-M., T., and Baulieu, E-E. (1986) Proc. Nati. Acad. Sci. USA 83, 2817-2821.
65. Gowland, P.L., and Buetti, E. (1989) Mol. and Cell Biol. 9, 3999-4008
66. Chalepakis, G., Amemann, J., Slater, E., Bruller, H-J., Gross, B., and Geato, M. (1988) Cell 53, 371-382
67. Kumar, V., and Cambon, P. (1988) Cell 55, 145-156.
68. Tremea, F., deMedeiros, S. R. B., Heggeler-Bordier, B. T., Germond, J-E., Seiler Tuyns, A. and Wahli, W. (1989) Mol. Endocrin. 3, 1596-1609.
69. Chandler, V. L., Maler, B. A., and Yamamoto, K. R., (1983) Cell 33, 489-499. 127
70. Ponta, H., Kennedym, N., Skroch, P., Hynes, N. E., and Groner, B. (1985) Proc. Natl. Acad. Sci. USA 84, 1021-1024.
71. Chalepakis, G., Postma, J. P. M., and Beato, M. (1988) Nucleic Acids Res. 16, 10237-10247.
72. Strahle, U., Klock, G., and Schutz, G. (1987) Proc. Natl. Acad. Sci. USA 84, 7871 7875.
73. Payvar, F., Wrange, 0., Carlstedt-Duke, J., Okret, S., Gustafsson, J-A., and Yamamoto, K. R. (1981) Proc. Natl. Acad. Sci. USA 78, 6628-6632. 74. Scheidereit, C., Geisse, S., Westphal, H. M., and Beato, M. (1983) Nature 304, 749-752.
75. Bronnegard, M., Poellinger, L., Okret, S., Wikstrom, A-C, Bakke, 0., and
Gustafsson, J-A. (1987) Biochemistry 26, 1697-1704.
76. Renkawitz, R., Schutz, G., von der Ahe, D., and Beato, M. (1984) Cell 37, 503-510.
77. Danesch, U., Gloss, B., Schmid, W., Schutz, G., Schule, R., and Renkawitz, R. (1987) EMBO J. 6, 625-630.
78. Scheidereit, C., Westphal, H. M., Carlson, C., Bosshard, H., and Beato, M. (1986) DNA 5, 383-391.
79. Peale, F. V., Ludwig, L. B., Zain, S., Hilf, R., and Bambara, R. A. (1988) Proc. Nati. Acad. Sci. USA 85, 1038-1042.
80. Klein-Hitpass, L., Tsai, S. Y., Green, G. L., Clark, J. H., Tsai, M-J., and O'Malley, B. W. (1989) Mol. and Cell. Biol. 9, 43-49.
81. Brent, G. A., Harney, J. W., Chen, Y., Warne, R. L., Moore, D. D., and Larson, P. R. (1989) Mol. Endocrin. 3, 1996-2004.
82. Ham, J., Thomson, A., Needham, M., Webb, P., and Parker, M. (1988) NAR, 16, 5263-5276. 128
83. Otten, A. D., Sanders, M. M., and McKnight, G. S. (1988) Mol. Endocrin. 2, 143 147.
84. Cato, A. C. B., Miksicek, R., Schutz, Arnemann, J., and Beato, M. (1986) EMBO J. 5, 2237-2240.
85. von der Ahe, D., Janich, S., Scheidereit, C., Renkawitz, R., Schutz, G., and Beato, M. (1985) Nature 313, 706-709.
86. Umesono, K., Giguere, V., Glass, C. K., Rosenfeld, M. G., and Evans, R. M. (1988) Nature 336, 262-265.
87. Reichman, M. E., Foster, C. M., Eisne L. P., Eisen, H. J., Torain, B. F., and Simons, S. S. Jr. (1984) Biochemistry 23, 5376-5384.
88. Wrange, 0., Okreat, S., Radojcic, M., Carsted-Duke, J., and Gustaffson, J-A. (1984) J. Biol. Chem. 259, 4535-4541.
89. Dobson, A. D. W., Conneely, 0. M., Bettie, W., Maxwell, B, L., Mak, P., Tsai, M J. Schrader, W. T., and O'Malley, B. W. (1989) J. Biol. Chem. 264, 4207-4211.
90. Carson, M. A., Tsai, M-J., Conneely, 0. M., Maxwell, B. L., Clark, J. H., Dobson,
A. D. W., Elberecht, A., Toft, D. 0., Scharder, W. T., and O'Malley , B, W. (1987) Mol. Endocrin. 1, 791.
91. McDonnell, D. P. Scot, R. A., Kerner, S. A., O'Malley, B. W., and Pike, J. W. (1989) Mol. Endocrin. 3, 635-644.
92. Fuqua, S. A. W., Fitzgerald, S. D., Chamness, G. C., Tandon, A. K., McDonnell, D. P., Nawaz, Z., O'Malley, B. W., and McGuire, W. L. (1991) Cancer Res. 51, 105-109.
93. Godowaski, P. J., Rusconi, S., Miesfeld, R., and Yamamoto, K. (1987) Nature 325, 65-368.
94. Fawell, S. E., Lees, J. A., White, R., Parker, M. G. (1990) Cell 60, 953-962. 129
95. Picard D., and Yamamoto, K. R. (1987) EMBO J. 6, 3333-3340.
9 6. Yamamoto, K. R., Gehring, U., Stamfer, M. R., and Sibley, C. (1976) Rec. Prog.
Hormone Res. 32, 3-32.
9'7. Northrup, J. P., Gametchu, B., Harrison, R. W., and Ringold, G. M. (1985) J. Biol. Chem. 260, 6398-6403.
98. Hope, I. A., and Struhl, K. (1986) Cell 46, 885-894.
99. Keegan, L., Gill, G., and Ptashne, M. (1986) Science 231, 699-704
100. Brent, R., and Ptashne, M. (1985) Cell 43, 729-736.
101. Webster, N. J. G., Green, S., Jin, J. R., and Cambon, P. (1988) Cell 54, 199-207.
102. Tora, L., Grononmeyer, J., Turcotte, B., Gaub, M. P. and Cambon, P. (1988) Nature 333, 185-188.
103. Danielsen, M., Northrop, J. P., Jonklass, J., and Ringold, G. (1987) Mol. Endocrin. 1, 816-822.
104. Hollenberg, S. M., Giguere, V., Sequi, P., and Evans, R. M. (1987) Cell, 49, 39
46.
105. Miesfeld, R., Godowski, P. J., Maler, B. A., and Yamamoto, K. R. (1987) Science 236, 423-427.
106. Godowski, P. J., Picard, D., Yamamoto, K. R. (1988) Science 241, 812-816.
107. Hollenberg, S. M., and Evans, R. M. (1988) Cell 56, 899-906.
108. Tora, L., Mullick, A., Metzger, D., Ponglikitmongkol, M., Park, I., and Chambon,
P. (1989) EMBO J. 8, 1981-1986.
109. Moudgil, V. K., Anter, M. J., and Hurd, C. (1989) J. Biol. Chem. 264, 2203 2211.
110. Yamamoto, K. R., and Alberts, B. M. (1975) Cell 4, 301-3 10.
111. Yamamoto, K. R., and Alberts, B. M. (1976) Ann. Rev. Biochem. 45, 721-756. 130
112. Skafar, D. F., and Notides, A. C. (1985) J. Biol. Chem. 260, 12208-12213.
113. Metzger, D., White, J. H., and Cambon, P. (1988) Nature 334, 31-36.
114. Pham, T. A., Elliston, J. F., Nawaz, Z., McDonnell, D. P., Tsai, M. J. and
O'Malley, B. W. (1991) Pro. Natl. Acad. Sci. USA 88, 3125-3129.
115. McDonnell D. P., Nawaz, Z., and O'Malley, B., W. (1991) Mol. and Cell. Biol. 11, 4350-4355.
116. Mak, P., McDonnell, D. P., Weigel, N. L., Schrader, W. T., and O'Malley, B. W. (1990) J. Biol. Chem. 264, 21613-21618.
117. McDonnell, D. P., Pike, J. W., Drutz, D. J., Butt, T. R., and O'Malley, B. W. (1989) Mol. and Cell. Biol. 9, 3517-3523.
118. McDonnell, D. P., Nawaz, Z., Densmore, C., Weigel, N. L., Pham, T. A., Clark, T. H., and O'Malley, B. W. (1991) J. Steroid Biol. and Mol. Biol. 39, 291-297.
119. Becker, P. B., Gloss, B., Schmid, W., Strahle, U., and Schutz, G. (1986) Nature
324, 686-688.
120. Wijnholds, J., Philipsen, J. N. J., and AB, G. (1988) EMBO J. 7, 2757-2763.
121. Baulieu, E-E. (1987) J. Cell. Biochem. 35, 161-174.
122. Bagchi, M., Tsai, S. Y., Tsai, M. J. and O'Malley, B. W. (1991) Mol. and Cell Biol. 11, 4998-5004.
123. Damm, K., Thompson, C. C., and Evans, R. M. (1989) Nature 339, 593-597.
124. Freedman, L. P., Yoshinaga, S. K., Vanderbilt, J. N., and Yamamoto, K. R. (1989) Science 245, 298-300.
125. Klien-Hitpass, L., Tsai, S. Y., Weigel, N. L., Allen, G., Riley, D., Rodriguez, R., Schrader, W. T., Tsai, M-J. and O'Malley, B. W. (1990) Cell 60, 247-257.
126. Kalff, M., Gross, B., Beato, M. (1990) Nature 344, 360-362. 131
127. Elliston, J. F., Fawell, S. E., Klien-Hitpass, L. K., Tsai, S. Y., Tsai, M. J., Parker,
M. G., and O'Malley, B. W. (1990) Mol. and Cell. Biol. 10, 6607-6612.
128. Dahlman, K., Stromstedt, P-E., Rae, C., Jornvall, H., Hock, J-I., Carlstedt-Duke,
J., and Gustafsson, J-A. (1989) J. Biol. Chem. 264, 804-809.
129. Power, R. F., Conneely, 0. M., McDonnell, D. P., Clark, J. H., Butt, T. R.,
Schrader, W. T., and O'Malley, B. W. (1990) J. Biol. Chem. 265, 1419-1424.
130. Witliff, J. L., Wenz, L. L., Dong, J., Nawaz, Z., and Butt, T. R. (1990) J. Biol. Chem. 265, 22016-22022.
131. Nawaz, Z., Smith, D. W., and McDonnell, D. P. (1991) Proceeding of The 8th
Workshop On Vitamin D, Paris, France, July 5-10.
132. Srinivasan, G., and Thompson, E. B. (1990) Mol. Endocrin. 4, 209-216.
133. Bellingham, D. L., and Cidlowski, J. A. (1989) Mol. Endocrin. 3, 1733-1747.
134. Schena, M., Freedman, L. P., and Yamamoto, K. R. (1989) Genes & Dev. 3, 1590 1601.
135. Wright, A. P. H., Carlsted,-Duke, J. and Gustafsson, J-A. (1990) J. Biol. Chem. 265, 14763-14769.
136. Corthesy, B., Hipskind, R., Theulaz, I. and Wahli, W. (1988) Science 239, 1137 1139.
137. Alberts, B., and Herrick, G. (1971) Methods Enzymol. 21, 198-217. 138. Pike, J. W., and Haussler, M. R. (1979) Proc. Natl. Acad. Sci. USA 76, 5485 5489.
139. Rosenberg, M., Ho, Y., and Shatzman, A. (1983) Methods Enzymol. 101, 123
138.
140. Matt, E., Grant, R., Ho, Y., and Platt, T. (1985) Proc. Natl. Acad. Sci. USA 82,
88-92. 132
141. Ito, H., Fukuda, Y., Murata, K., and Kimura, A. (1983) J. Bacteriol. 153, 163 186.
142. Baker, A. R., McDonnell, D. P., Hughes, M. R., Crisp, T. R., Mangelsdorf, D.
J.,Hussler, M. R., Pike, J. W., Shine, J., and O'Malley, B. W. (1988) Proc. Natl. Acad. Sci. USA 85, 3294-32948.
143. Butt, T. R., Jonnalagadda, S., Monia, B. P., Strenberg, E. J., Marsh, Stadel, J.,
Echer, D. J., and Crooke, S. T. (1989) Proc. Natl. Acad. Sci. USA 86, 2540-2544.
144. Greene, G., Gilna, P., Waterfield, M., Baker, V., Hort, V., and Shine, J. (1986) Science, 231, 1150-1154.
145. Hoylaerts, M., Weyens, A., Bollen, A., Harford, N., and Cabezon, T. (1986) FEBS
Lett. 204, 83-87.
146. Sanger, F., Nicklein, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467.
147. Guarente, L. (1983) Methods Enzymol. 101, 181-191.
148. Elliston, J. F., Tsai, S. Y., O'Malley, B. W., and Tsai, M-J. (1990) J. Biol. Chem. 265, 11517-11521.
149. Wecksler, W. R., and Norman, A. W. (1979) Anal. Biochem. 92, 314-323.
150. Wittliff, J. L. (1987) in Methods in Clinical Chemistry (Pesce, A. J., and Kaplan, L. A., eds), C. V. Mosby Co., St. Louis, pp767-795.
151. Pike, J. W. (1981) Biochem. Biophys. Res. Commun. 100, 1713-1719.
152. Klein-Hitpass, L., Ryffel, G.V., Heitlinger, E., and Cato, A. C. B. (1988) Nucleic Acids Res. 16, 647-663.
153. Weigel, N. L., Schrader, W. T., and O'Malley, B. W. (1989) Endocrinol. 125, 2494-2501.
154. Oliphant, A.R., Hill, D. E., Nussbaum, A. L., Struhl, K. (1986) Gene 44, 177. 133
155. Oliphant, A. R., and Struhl, K. (1987) Methods Enzymol. 155, 568-582.
156. Sentenac, A. (1985) Crite. Rev. Biochem. 18, 30-3 1.
157. Buratowski, S., Hahn, S., Sharp, P., and Guarente, L. (1988) Nature 334, 37-42.
158. Chodosh, L. A., Olesen, J., Hahn, S., Baldwin, A. S., Guarente, L., and Sharp, P.
A. (1988) Cell 53, 25-35.
159. Hitzeman, R. A., Chen, C.-Y., Dowbenko, D. J., Renz, M. E., Liu, C., Pai, R., Simpson, N. J., Kohr, W. J., Singh, A., Chisholm, V., et al., (1990) Methods
Enzymol. 185, 421-440.
160. Clark, S. G., McGrath, J. P., and Levison, A. D. (1985) Mol. and Cell. Biol. 10, 2746-2752.
161. Miyamoto, C., Chizzonite, R., Crow, R., Rupprecht, K., Kramer, R., Schabe, M.,
Kumar, G., Poonian, M., and Ju, G. (1985) Proc. Natl. Acad. Sci. USA 21, 7232 7236.
162. Kakidini, H., and Ptashne, M. (1988) Cell 52, 161-167.
163. Butt, T. R., Khan, M. I., Marsh, J., Ecker, D. J., and Crooke, S. T. (1988) J. Biol. Chem. 263, 16364-16371.
164. Butt, T. R., Khan, M. I., Marsh, J., Ecker, D. J., and Crooke, S. T. (1988) Proc.
Natl. Acad. Sci. USA 86, 2540-2544.
165. Erhart, E., and Hollenberg, C. P. (1983) J. Bacteriol. 156, 625- 635. 166. Shena, M., and Yamamoto, K. R. (1988) Science 241, 965-967.
167. Tora, L., White, J., Bron, C., Tasset, D., Webster, N., Scheer, E., and Chambon , P. (1989) Cell 59, 477-487.
168. Berry, M., Metzger, D., and Chambon, P. (1990) EMBO J. 9, 2811-2818.
169. Simpson, D. M., Elliston, J. F., and Katzenellenbogen, J. A. (1987) J. Steroid Biochem. 288, 233-245. 134
170. Rochefort, H. (1987) Horm. Res. 28, 196-201.
171. Clark, J. H., and Markaverich, B. M. (1988) in The Physiology of Reproduction
(Knobil, E., and Neill, J., eds.), Raven, New York, pp 675-724.
172. Rochefort, H., and Borgna, J. L. (1981) Nature 292, 257-259.
173. Burch, J. B. E., and Weintraub, H. (1983) Cell 33, 66-76.
174. Zaret, K. S., and Yamamoto, K, R. (1984) Cell 38, 29-38. 175. Richard-Foy, H., and Hager, G. L. (1987) EMBO J. 6, 3221-3228.
176. Richard-Foy, H., Sistare, F. D., Riegel, A. T., Simons, S. S., and Hager, G. L.
(1987) Mol. Endocrinol. 1, 659-665. 177. Beato, M. (1989) Cell 56, 335-344.
178. Druege, P. M., Klein-Hitpass, L., Green, S., Stack, G., Chambon, P., and Ryffel,
G. U. (1986) Nucleic Acids Res. 23, 9329-9337.
179. Tsai, S. Y., Carsteds-Duke, J., Weigel, N. L., Dahlman, K., Gustafsson, J. A., Tsai, M.-J., and O'Malley, B. W. (1988) Cell 55, 361-369.
180. Walker, P., Germond, J. E., Brown-Luedi, M., Givel, F., and Wahli, W. (1984)
Nucleic Acids Res. 12, 8611-8626.
181. Klock, G., Strahle, V., and Schutz, G (1987) Nature 329, 734-736.
182. Brown, T. R., Lubahn, D. V., Wilson, E. M., Joseph, D. R., French, F. S., and Migeon, C. J. (1988) Proc. Natl. Acad. Sci. USA 85, 8151-8155. 183. Sakurai, A., Takedo, K., Ain, K., Ceccarelli, P., Nakai, A., Seino, S., Bell, G. I.,
Refetoff, S., and DeGroot, L. J. (1989) Proc. Natl. Acad. Sci. USA 86, 8977-8981.
184. Tasset, D., Tora, L., Fromental, C., Scheer, E., and Chambon, P. (1990) Cell 64,
971-981.
185. Eul, J., Meyer, M. E., Tora, L., Bocquel, M. T., Quirin-Sriker, C., Chambon, P.,
and Gronemeyer, H. (1989) EMBO J. 8, 83-90. 135
186. Goff, S., Canon, L., and Goldberg, A. (1984) Proc. Natl. Acad. Sci. USA 81, 6647-6651.
187. Picard, D., Salser, S. J., and Yamamoto, K. R. (1988) Cell 54, 1073-1080. 188. Pratt, W. B. (1987) J. Cell Biochem. 35, 51-68.
189. Rechsteiner, M. (1987) Annu. Rev. Cell Biol. 3, 1-30.
190. Weber, P. L., Brown, S.C., and Mueller, L. (1987) Biochemistry 26, 7287-7290.
191. Sone, J., McDonnell, D. P., O'Malley, B. W., and Pike, J. W. (1990) J. Biol. Chem. 265, 21997-22003.
192. Bailly, A., LePage, C., Rauch, M., and Milgram, E. (1986) EMBO J. 5, 3235-3241.
193. Wright, A. P. H., Carlstedt-Duke, J., and Gustafsoon, J-A. (1990) J. Biol. Chem. 265, 14763-14769.
194. Wakeling, A. E., and Bowler, J. (1987) Endocrinol. 112, R7-RIO.
195. Martinez, E., and Wahli, W. (1989) EMBO J. 8, 3781-3791.
196. Fawell, S.E., White, R., Hoare, S., Sydenham, M., Page, M., and Parker, M. G.,
(1990) Proc. NatI. Acad. Sci. USA 87, 6883-6887
197. Umesono, K., Murakami, K. K., Thompson, C. C., and Evans, R. M., (1991) Cell 65, 1255-1266.
198. Naar, A. M., Boutin, J. M., Lipkin, S. M., Yu, V. C., Holloway, J. M., Glass, C. K., and Rosenfeld, M. G. (1991) Cell 65, 1267-1279.
199. Oliphant, A. R., Brandl, C. J., and Strhl, K. (1989) Mol. and Cell Biol. 9, 2944 2949.
200. Singer, V. L., Wobbe, R., and Struhl, K. (1990) Genes & Dev. 4, 636-645. 201. Blackwell, T.K., and Weintraub, H. (1990) Science 250, 1104-1110.
202. Wilson, T.E., Fahrner, T. J., Johnston, M., and Milbrandt, J. (1991) Science 252, 1296-1300. 136
203. Colvard, D.S., Eriksen, E. F., Keeting, P. E., Wilson, E. M., Lubahn, D. B., French, F. S., Riggs, B. L., and Spelsberg, T. C. (1989) Proc. Natl. Acad. Sci. USA 86, 854-857.
204. Lian, J. B., Stein, G. S., Stewart, C., Puchaz, E., Mackowiak, S., Aronow, M., Deck, M. V., and Shalhoub, V. (1989) Conn. Tissue Res. 21, 61-69.