AU639699B2 - Dominant negative members of the steroid/thyroid superfamily of receptors - Google Patents

Description

OPI DATE 18/12/90 APPLN. ID 58248 PCT AOJP DATE 97/02/91 PCT NUMBER PCT/US90/03113 INTERNATIONAL APPLICATION PUBLISHED UNDER THE PATENT COOPERATION TREATY (PCT) (51) International Patent Classification 5 (11) International Publication Number: WO 90/14356 C07K 3/00, 13/00, 15/00 Al C07K 17/00, C12P 21/00 (43) International Publication Date: 29 November 1990 (29.11.90) (21) International Application Number: PCT/US90/03113 (74) Agents: WATT, Phillip, H. et al.; Fitch, Even, Tabin Flannery, Room 900, 135 South LaSalle Street, Chicago, (22) International Filing Date: 25 May 1990 (25.05.90) IL 60603 (US).

Priority data: (81) Designated States: AT (European patent), AU, BE (Euro- 358,696 26 May 1989 (26.05.89) US pean patent), CA, CH (European patent), DE (European patent)*, DK (European patent), ES (European patent), FR (European patent), GB (European patent), IT Parent Application or Grant (European patent), JP, LU (European patent), NL (Eu- (63) Related by Continuation ropean patent), SE (European patent), US.

US 358,696 (CIP) Filed on 26 May 1989 (26.05.89) Published With international search report.

(71) Applicant (for all designated States except US): THE SALK Before the expiration of the time limit for amending the INSTITUTE FOR BIOLOGICAL STUDIES [US/US]; claims and to be republished in the event of the receipt of 10010 North Torrey Pines Road, La Jolla, CA 92138 amendments.

(US).

(72) Inventors; and Inventors/Applicants (for US only) EVANS, Ronald, Mark [US/US]; 8615 La Jolla Scenic Drive North, La Jolla, CA 92037 HOLLENBERG, Stanley, Mark [US/ US]; 240 23rd Street, Del Mar, CA 92014 ORO, Anthony. Eugene [US/US]; 4812 Mt. Etna Drive, San Diego, CA 92117 DAMM, Klaus [DE/DE]; Klagenfurterstrasse 25, D-8000 Mfnich HEYMAN, Richard, Alan [US/US]; 147 Honeycomb Court, Encinitas, CA 92024 (US).

(54) Title: DOMINANT NEGATIVE MEMBERS OF THE STEROID/THYROID SUPERFAMILY OF RECEPTORS (57) Abstract Disclosed are novel trans-repressing analog receptors wherein the ligand-binding domain(s) are modified versus wild type receptor, such novel receptors having repressed trans-activation transcription activation properties. Also disclosed are recombinant methods and means for preparing such receptors and assays using such receptors.

See back of page WO 90/14356 PCT/US90/03113 -1- DOMINANT NEGATIVE MEMBERS OF THE STEROID/THYROID SUPERFAMILY OF RECEPTORS Related Applications This is'a continuation-in-part application of U.S. Patent Application Serial No. 07/358,696, filed 26 May 1989, which is in turn, a continuation-in-part application of U.S. Patent Application Serial No.

07/289,561 filed 23 December 1988, the entire contents of which are both hereby incorporated by express reference herein.

Field of the Invention The present invention relates to transrepressing analog receptors of the steroid/thyroid superfamily. In a particular aspect, it relates to the identification and characterization of proteins that function as transcription trans-activation repressors, as well as to their preparation and use, including novel DNA isolates encoding same; expression vectors operatively harboring these DNA sequences; and hosts transfected with said vectors.

In another aspect, the present invention relates to the use of the above-described transcription trans-activation repressors in various assays and screening methods.

Background of the Invention The characterization and preparation of various hormone and hormone-like receptors, including steroid, WO 90/14356 PCT/US90/03113 -2thyroid, and retinoid receptors such as those represented by the glucocorticoid, mineralocorticoid, thyroid, estrogen-related and retinoid classes has been subject of considerable research.

It is known, for example, that the glucocorticoid receptor belongs to a large superfamily of ligand-dependent transcription factors that have themselves diverse roles in homeostasis, growth and development. Comparison of complementary DNAs encoding these receptors, as well as mutational analyses of their coding sequences, have identified certain functional domains within the molecule that are thought responsible respectively for DNA binding, hormone binding and nuclear localization. See Evans, et al., Science 240, 889 (1988) for a review of this subject matter.

In the case of the glucocorticoid receptor, the so-called DNA binding domain spans some sixty-six amino acids and is highly conserved among various species. In addition, this domain has been found to be required in order to activate transcription. See Hollenberg, et al., Cell 49, 39 (1987), Miesfeld, et al., Science 236. 423 (1987), Danielsen, et al., Mol.Endo 1, 816 (1987), Kumar, et al., Cell 51, 941 (1987), Gronemeyer, EMBO J. 6, 3985 (1987), and Waterman, et al., Mol.Endo 2, 14 (1988).

This domain has been found to contain nine invariant cysteine residues. Although the contribution of each cysteine residue to overall function is unknown, as is the actual structure formed by this domain, it has been proposed that these cysteine residues coordinate two zinc ions to form two DNA binding, so-called finger domains, which result in a ternary structure thought responsible for the localization and binding of the glucocorticoid receptor to the requisite DNA site. See Klug, et al., WO 90/14356 WO 9014356PCr/US9O/03 113 -3- Tr.BiocheTn.Sci 12, 464 (1987), Bens, et al., Cell 57, 1 (1988), and Evans, supra.

In a location nearer the carboxyl-terminal end~ distal from the DNA binding region is the so-called ligand binding domain which has the demonstrated ability to block activity of the receptor in the absence of hormone. Thus, presence of the requisite hormone relieves the inhibition of the receptor to activity.

Deletion of this region has been found to produce a hormone-independent transcription activator. See Godowski, et Nature 325 365 (1987), Hollenberg, et al., supra, Kumar, et supra, Danielsen et al., supra, and, Adler et Cll 52, 685 (1988).

In contrast to these two~ domains, the sequences lying towards the amino-terminal region from the DNA binding domain are poorly understood both as to structure, and particularly, function. This region is extremely variable both in size and in composition amongst the various receptors See Evans, supra and may contribute to the heterogeneity of receptor function.

See Kumar eta. supra, and Tora et-al., 333 185 (1988).

Despite extensive analysis, some of which has been reported in the scientific literature, the region(s) that determine(s) trans-activation of transcription initiation remains poorly characterized. Transactivation domains can be defined as poJlypeptide regions that, when combined with the functional DNA binding domain, increase productive transcription initiation by RNA polymerases. See Sigler, Nature 333, 210 (1988), Brent at Cell 4.2, 729 (1985), Hope et Cell 46, 885 (1986), Ma Cell. 48, 847 (1987), Ma et Cell 51, 113 (1987), Lech et Cell 52, 179 (1988), and Hope et al., Nature 333, 635 (1988).

WO 90/14356 PCT/US90/03113 -4- Previous research of the human glucocorticoid receptor by linker scanning mutagenesis identified two regions outside of the DNA binding region having a role in transcription activation. These regions were defined as T7 and T 2 Giguere et al., Cell 46, 645 (1986).

Further research from these laboratories has also resulted in the report of a co-localization of transactivation and DNA binding functions. See Hollenberg et al., supra, Miesfeld, et al., supra, Danielsen et al., supra, and Waterman et al., supra. As a result, this research has given rise to an emerging picture of an increasingly modular molecule with discrete domains, each contributing to the identified properties of ligandbinding, DNA-binding and trans-activation of transcription. Until recently, the region(s) determining the trans-activation activity were not at all well understood. Thus, the picture based upon extant literature lacks an overall appreciation of the dynamic nature of the steroid receptors and how the various domains determine the cascade of events initiated by ligand-binding and consummated by promoter-specific trans-activation.

Further, although previous research has identified functional "domains", there has been little systematic effort to identify amino acids that contribute to the specific activities of the molecule itself. Thus, the previous identification of steroid receptor transactivation regions resulted only from a demonstrated loss of activity via deletion or insertional mutagenesis, but in no case have the properties of the regions themselves been confirmed in assays that reflect a dominant gain of function. See also Ptashne, Nature 335, 683 (1988).

Thus, Godowski et al., Science 241, 812 (1988), report results that show that the glucocorticoid receptor WO 90/14356 PCT/US90/03113 contains at least one "enhancement domain" other than that overlapping the glucocorticoid response element binding region the DNA binding domain) and that the second domain occupies a region near the receptor amino-terminus. Similarly, Webster et al., Cell 54, 199 (1988) report on an inducible transcription activation function of the estrogen and glucocorticoid receptors, and these researchers speculate that the relative positions of the hormone regions ligand and DNAbinding domains) are not important for the transcription induction activity of the receptor. Yet, these researchers admit that they have no definition of the exact location and nature of what they call the hormoneinducible activating domain, to say nothing of its characterization and role in trans-activating potential.

The work by Giguere et al., supra, demonstrated a loss of activity in the glucocorticoid receptor based upon an assay measuring transcription activity, as a result of performing random sitemutagenesis at several locations of the molecule. As a follow-up, Hollenberg et al. deleted regions in the molecule, again demonstrating overall loss of transcription activity induced by such removal of stretches of amino acids.

The human glucocorticoid receptor (hGR) has served as a prototype, model receptor for gene regulation. As noted above, the DNA-binding and ligandbinding functional domains have been defined previously.

Similarly, it has been found that these modular domains of the hGR receptor or other receptors may be moved to other parts of the receptor or attached to heterologous DNA-binding domains and still maintain function.

In contrast, relatively little is known about negative regulation by hGR. This is surprising in light WO 90/14356 PCT/US90/03113 -6of the key role that steroids play in development and negative feedback regulation. Glucocorticoid helps determine neural crest cell fate in the developing sympathoadrenal system, in part by repressing the induction of neural-specific genes [See Stein et al., Dev Bio 127, 316 (1988) and Anderson et al, Cell 47, 1079 (1986)]. Glucocorticoid also modulates the hypothalamicpituitary-adrenal axis by inhibiting second messengerinduced peptide hormone induction. Recently, Akerblom et al. (Science 241., 350 (1988)) showed that the hGR negatively regulates the cyclic AMP-inducible alpha glycoprotein hormone promoter in a steroid and DNAbinding dependent manner. Wild-type expression is exhibited by a promoter of just 168 base pairs (termed alphal68). Basal expression in placental cells is mediated by factors bound to a 36 base pair palindromic cyclic AMP response element (CRE) cooperating with proteins binding to a 25 base pair tissue-specific element (TSE). Expression may be further enhanced through the CRE by the elevation of intracellular cyclic AMP levels. The hGR represses both the basal and cyclic AMP enhanced transcription in a glucocorticoid-dependent fashion. The transacting elements to which the hGR binds have been defined and are related to the consensus GRE sequence for activation. Similar research is reported by Sakai, et al., Genes and Development 2, 1144 (1988).

Summary of the Invention The present invention is the result of a thorough analysis of the structural requirements of hormone receptors for repression. This analysis has revealed an absolute requirement for the DNA binding domain and a role for the carboxyl terminus for repression. Although the DNA-binding domain alone is not WO 90/14356 PCT/US90/03113 -7sufficient for maximal repression, the addition of polypeptides to the carboxyl terminus or other modifications as described herein at the carboxyl terminus leads to the creation of novel fusion proteins having dominant negative repressor activity.

In accordance with the present invention, we have identified, isolated and characterized the domains of intracellular hormone or hormone-like receptors that can be modified so as to repress trans-activation transcription activity. This information has enabled the further characterization of various receptors of the steroid/thyroid superfamily, both in terms of physical attributes and biological function and effect of various domains, particularly that domain capable of being modified to provide repressed transcription activity.

The foregoing has in turn enabled the production of novel analog receptors having repressed transcription activation properties.

It has been determined, based upon the information provided herein, that receptors of the steroid/thyroid superfamily contain domains that function in overall trans-activation transcription activity, even though the three receptor domains, the DNA-binding, the ligand-binding, and the trans-activation transcription domains, are positioned independently of one another and are autonomous in function. In accordance with the present invention, it has been discovered that the carboxyl terminus of a given receptor is that domain responsible for modulating the transactivation transcription activity of said receptor. It has further been found that the DNA-binding domain is a necessary component in any receptor hereof having repressed trans-activation transcription activity.

WO 90/14356 PT/US90/03113 -8- This invention provides for novel hormone or hormone-like analog receptors wherein the transactivation transcription activity is repressed. Such novel analog receptors contain a DNA-binding domain, optionally an N-terminal domain and a C-terminal domain that has been altered so as to provide repressed transactivation transcription activity compared with parental or wild-type receptor. The novel analog receptors hereof may be hybrid receptors wherein the DNA-binding domain, N-terminal domain and C-termirnal domain are provided from different sources. For example, the C-terminal domain of the glucocorticoid receptor can be replaced herein by a portion of the C-terminus of the v-erbA protein.

Alternatively, the C-terminal domain of the glucocorticoid receptor can bet replaced with at least a portion of a polypeptide such as P-galactosidase.

The present invention is further directed to the preparation of such novel analog receptors hereof via recombinant DNA technology in all relevant aspects, including a DNA molecule that is a recombinant DNA molecule or a cDNA molecule consisting of a sequence encoding said analog receptor or a C-terminal modified domain thereof, requisite expression vectors operatively harboring such DNA comprising expression control elements operative in the recombinant hosts selected for the expression, and recombinant host cells transfected with such operative expression vectors.

The present invention is also directed to the use of the novel analog receptors described herein for identifying the response element and/or function of an "orphan" receptor, a receptor for which the associated response element and/or function is not known.

The present invention is further directed to the use of the novel analog receptors described herein in WO 90/14356 PCT/US90/03113 -9an improved assay system for the determination of a specific wild type receptor, wherein more than one response element present in the assay system is capable of interacting with the wild type receptor.

An additional utility for such repressed analog receptors hereof lies in the area of cancer therapy.

Certain cells require augmented levels of hormone in order to become tumorigenic. One example is the elevated estrogen requirement observed in mammary tumors. Indeed, estrigen antagonists, such as tamoxifin, are used in therapy so as to decrease the amount of estrogen available so that the transformation of normal mammary cells into tumorous cells is inhibited. Alternatively, the analog receptors of the present invention can be used for such purpose.

Brief Description of the Figures Figure 1 provides the amino acid sequence of the v-erbA protein.

Figure 2 is a dose response curve for hGRmediated negative regulation.

Figure 3 illustrates the repression by carboxyl terminal fusion proteins.

Figure 4 illustrates several expression and reporter constructs.

Figure 5 shows the structure and activity of rat TRa/v-erbA chimeric proteins.

Figure 6 shows a comparison of T 3 induction of CAT activity by thyroid receptor, v-erbA, and hybrid Tr/v-erbA receptors.

Figure 7 shows the competition of TRa or v-erbA with RA induced transcriptional activation by RARa.

Figure 8 summarizes the structure and activity properties of several RAR hybrids and mutants.

WO 90/14356 PPUS90/03113 Figure 9 shows the competition of RA induction by RARa-v erbA fusion protein.

Figure 10 shows the competition of RA induced transactivation of endogenous RARs in F9 tetracarcinoma stem cells.

Figure 11 summarizes the structure and activity of several GR hybrids and mutants.

Detailed Description of the Invention In accordance with the present invention, there is provided a trans-repressing analog receptor of the steroid/thyroid superfamily of receptors, said analog comprising: a first amino acid sequence which is a DNA-binding domain, through which said analog is capable of binding to a hormone response element of a wild type receptor, and a second amino acid sequence which is positioned at the carboxy-terminal end of the DNA-binding domain, wherein said second sequence is selected from: a polypeptide which has at least about 90 as many amino acids as the ligand binding domain of the carboxy-terminal portion of said wild type receptor; whcrcin aid k a on4 -polypeptide ha~ less than about 60 amino e0 Id identity relative to the carboxy-terminal domain of said wild type receptor over either the entire length of said polypeptide, if shorter than the carboxyterminal domain of said wild type receptor, or (ii) any of said polypeptide segments having the same length as the carboxy-terminal domain of said wild type receptor; or -l1at least the 84 carboxy-terminal amino acids of the carboxyterminal portion of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1; Note that the erb-A am acid numbering will vary by about 255 amino acids depending on whether the gag sequence (the amino-terminal 255 amino acids) is included for purposes of numbering the amino acids of the erb-A protein).

In accordance with another embodiment of the present invention, there is (are) provided expression vector(s) operatively harboring DNA molecule(s) which is (are) recombinant DNA molecule(s) or cDNA molecule(s) encoding the above-described receptor analogs; recombinant host cells transfected with such expression vectors; and cell cultures comprisingsuch cells in an extrinsic support medium.

In accordance with yet another embodiment of the present invention, there is provided a non-human transgenic mammal, having disease symptoms due to the inability to respond normally to a steroid or thyroid hormone, said mammal having at least a subset of its cells capable of expressing an analog receptor for one of said hormones, said analog receptor having 20 trans-repression activity greater than that of its corresponding wild type receptor and trans-activation activity less than that of its corresponding wild type receptor.

In accordance with still another embodiment of the present S. 25 invention, there is provided a method for converting a wild type hormone receptor into a trans-repressing analog receptor of the steroid/thyroid superfamily of receptors, said method comprising: replacing the ligand binding domain of said wild type receptor with 30 at least the 84 carboxy terminal amino acids of the verbA protein as defined by amino acid numbers 313-398 (see Figure 1).

This latter embodiment of the present invention can alternatively be accomplished by replacing the ligand binding domain of said wild type receptor with a polypeptide which has at least 90% as many amino acids a, 1737a/ii -12at least the ligand binding domain of the carboxy-terminus of said wild type receptor; and wherein said polypeptide has less than about 60% amino acid identity relative to the carboxy-terminus of said wild type receptor over either: the entire length of said polypeptide, if shorter than the carboxy-terminus of said wild type receptor, or (ii) any segment of said polypeptide having the same length as the carboxy-terminus of said wild type receptor.

In accordance with a further embodiment of the present invention,there is provided a method for blocking the transcriptional activation by a wild type receptor of a hormone response element in a cell by contacting the cell with an effective amount of an analog receptor as described hereinabove.

In accordance with a still further embodiment of the present invention, there is provided a method for blocking the transcriptional 20 activation by a wild type receptor of the steroid/thyroid superfamily of receptors of a hormone response element present in a cell, said method •comprising substantially deleting the ligand binding domain of said wild 25 type receptor, operatively linking the modified receptor of step at the S.carboxy-terminal end thereof to at least the 84 carboxy terminal amino acids of the verbA protein as defined by amino acid numbers 313-398 (see I: Figure 1) to produce a fusion protein, and thereafter 30 contacting said cell with an effective amount of said fusion :protein.

This latter embodiment of the present invention can alternatively be accomplished by: 1737a/ii -13substantially deleting the ligand binding domain of said wild type receptor; operatively linking the modified receptor of step at the carboxy-terminal end thereof to a polypeptide which has at least about 90% as many amino acids as the ligand binding domain of said wild type receptor, wherein said polypeptide has less than about 60% amino acid identity relative to the ligand binding domain of said wild type receptor over either the entire length of said polypeptide, if shorter than the ligand binding domain of said wild type receptor, or (ii) any segment of said polypeptide having the same length as the ligand binding domain of said wild type receptor, to produce a fusion protein, and thereafter contacting said cell with an effective amount of said fusion protein.

In accordance with another embodiment of the present invention, there is provided a method for identifying the response element and/or function of a receptor for which the associated response element and/or 20 function is not known, said method comprising: comparing the response of a test system having known responsiveness to wild-type receptor to the response of said test system when treated with trans-repressing analog receptor of the steroid/thyroid superfamily of receptors; wherein said trans-repressing analog receptor comprises: a first amino acid sequence which is a DNA-binding domain, through which said analog is capable of binding to a hormone response element of said wild type receptor, and a second amino acid sequence which is positioned at the S" 30 carboxy-terminal end of the DNA-binding domain, wherein said sequence is selected from: a polypeptide which has at least about 90% as many amino acids as the ligand binding domain of the carboxy-terminal portion of said wild type receptor; wherein said polypeptide has less than about 60% amino acid identity relative to the carboxy-terminal domain of said wild type 1737a/ii -14receptor over either: the entire length of said polypeptide if shorter than the carboxy-terminal domain of said wild type receptor, or any segment of said polypeptide having the same length as the carboxy-terminal domain of said wild type receptor; or at least the 84 carboxy-terminal amino acids of the carboxy-terminal portion of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1).

For example, the specific response element(s) with which the orphan receptor interacts can be determined by screening a variety of test systems, each having a single known response element. The response element of the specific test system which is activated by the orphan receptor, but which is repressed by an analog receptor derived from the orphan receptor according to the present invention, is the response element for the orphan receptor.

Similarly, the function of an orphan receptor can be determined by comparing the response of a test system when contacted with the orphan receptor, relative to the response of the same test system when contacted with a trans-repressing analog of the orphan receptor according to the present invention. Differences in the response in the two side-by-side comparisons provide an indication of the functional role of the orphan receptor.

In accordance with yet another embodiment of the present invention, e an improvement is provided for use in assay systems responsive to the presence of a specific wild type receptor, wherein more than one response "element is capable of interacting with said wild type receptor, the improvement comprising inactivating, with respect to said assay, response element(s) which also respond to wild type receptor(s) other than said specific receptor; wherein said response elements are inactivated by adding to said assay system an effective amount of a trans-repressing analog receptor of the steroid/thyroid superfamily of receptors for each of said other receptor(s), wherein each of said trans-repressing analog receptors comprises: 1737a/ii a first amino acid sequence which is a DNA-binding domain, through which said analog is capable of binding to a hormone response element of a receptor other than said specific wild type receptor, and

I

I

a..

a a *a a I I **a

C

a ~av'

/A

I-Q. ~J 4 1737a/ii WO 90/14356 PCT/US90/03113 -16a second amino acid sequence which is positioned at the carboxy-terminal end of the DNA-binding domain, wherein said sequence is selected from: a polypeptide which has at least about 90% as many amino acids as the ligand binding domain of the carboxyterminal portion of said receptor other than said specific wild type receptor; wherein said polypeptide has less than about 60% amino acid identity relative to the carboxy-terminal domain of said receptor other than said specific wild type receptor over either: the entire length of said polypeptide if shorter than the carboxy-terminal domain of said receptor other than said specific wild type receptor, or any segment of said polypeptide having the same length as the carboxy-terminal domain of said receptor other than said specific wild type receptor; or at least the 84 carboxyterminal amino acids of the carboxy-terminal portion of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1).

As employed herein, the term "dominant negative", when used in reference to the analog receptors of the present invention, refers to species which have a negative effect on the transcriptional activation activity of the associated response element, even in the presence of wild-type receptor and its associated ligand.

WO 90/14356 PCT/US90/03113 -17- Amino acid abbreviations employed in the present disclosure use the following standard single- and three-letter designations, i.e.: Asp D Aspartic acid Ile I Isoleucine Thr T Threonine Leu L Leucine Ser S Serine Tyr Y Tyrosine Glu E Glutamic acid Phe F Phenylalanine Pro P Proline His H Histidine Gly G Glycine Lys K Lysine Ala A Alanine Arg R Arginine Cys C Cysteine Trp W Tryptophan Val V Valine Gin Q Glutamine Met M Methionine Asn N Asparagine Receptors employed in the practice of the present invention can be prepared by recombinant techniques, by synthetic chemistry, or the like. The thus produced receptor, in its various forms, is recovered and purified to a level suitable for its intended use.

The existence of a superfamily of ligandinducible trans-acting factors is now recognized, including those for steroid hormones, retinoic acid and vitamin D3, two subtypes (isoforms) of thyroid hormone receptors (termed a and and the like. Mutational analysis and structural comparisons of these hormone receptors has enabled the identification of domains responsible for hormone-binding, DNA-binding and transactivation of gene expression. See Sap et al., Nature 324, 635 (1986), Weinberger et al., Nature 324, 641 (1986) and Evans, Science 240, 889 (1988).

The receptors of the present invention are trans-repressing analogs of hormone or hormone-like receptors which are referred to broadly as members of the WO 90/14356 PCT/US90/03113 -18steroid/thyroid superfamily of receptors, e.g., glucocortoid receptor, mineralocorticoid receptor, progesterone receptor, estrogen receptor, estrogenrelated receptors, vitamin D 3 receptor, thyroid hormone receptor, retinoic acid receptor, aldosterone receptor, androgen receptor, and the like. Receptors of the present invention include functional equivalents of all of the above, including receptors differing in one or more amino acids from the corresponding parent, or in glycosylation and/or phosphorylation patterns, or in bounded conformational structure. The terminology "functional equivalents thereof" refers to transrepressing analog receptors which differ from the previously described analog receptor(s) with respect to one or more amino acids, insofar as such differences do not lead to a destruction in kind of the basic repressed receptor activity or biofunctionality.

It will be understood, therefore, that receptors that are known in the art, whether wild-type, hybrids, or functional equivalents as set forth herein, are suitable as starting materials for the practice of the present invention.

As employed herein, the term "expression vector" includes vectors which are capable of expressing DNA sequences contained therein, where such sequences are operatively linked to other sequences capable of effecting their expression. It is implied, although not always explicitly stated, that these expression vectors may be replicable in host organisms either as episomes or as an integral part of the chromosomal DNA. As employed herein, the term "operative," or grammatical equivalents, means that the respective DNA sequences are operational, that is, work for their intended purposes. In sum, "expression vector" is given a functional definition, and WO 90/14356 PCT/US90/03113 -19any DNA sequence which is capable of effecting expression of a specified DNA sequence disposed therein is included in this term as it is applied to the specified sequence.

In general, expression vectors of utility in recombinant DNA techniques are often in the form of "plasmids" which refer to circular double stranded DNA loops which, in their vector form, are not bound to the chromosome. In the present specification, the terms "plasmid" and "vector" are used interchangeably as the plasmid can be a commonly used form of vector. However, the invention is intended to include such other forms of expression vectors which serve equivalent functions and which become known in the art subsequently hereto.

As employed herein, the term "recombinant host cells" refers to cells which have been transfected with vectors constructed using recombinant DNA techniques.

As employed herein, the term "extrinsic support medium" includes those known or devised media that can support cells in a growth phase or maintain them in a viable state such that they can perform their recombinantly harnessed function. See, for example, ATCC Media Handbook, Ed. Cote et al., American Type Culture Collection, Rockville, MD (1984). A growth supporting medium for mammalian cells, for example, preferably contains a serum supplement such as fetal calf serum or other supplementing component commonly used to facilitate cell growth and division such as hydrolysates of animal meat or milk, tissue or organ extracts, macerated clots or their extracts, and so forth. Other suitable medium components include, for example, transferrin, insulin and various metals.

The vectors and methods disclosed herein are suitable for use in host cells over a wide range of prokaryotic and eukaryotic organisms.

WO 90/14356 PCT/US90/03113 In addition to the above discussion and the various references to existing literature teachings, reference is made to standard textbooks of molecular biology that contain definitions and methods and means for carrying out basic techniques encompassed by the present invention. See, for example, Maniatis, et al, Molecular Clonina: A Laboratory Manual, Cold Spring Harbor Laboratory, New York, 1982 and the various references cited therein, and in particular, Colowick et al., Methods in Enzymology Vol 152, Academic Press, Inc.

(1987). All of the herein cited publications are by this reference hereby expressly incorporated herein.

Non-human transgenic organisms contemplated by the present invention include rodents mice, rats), pigs, sheep, lower eukaryotes Drosophila, Xenopus), and the like.

Transcriptional activation by the steroid receptors and the thyroid hormone receptors has been found to be dependent on the presence and binding of the respective ligand. However, deletion analysis has disclosed that glucocorticoid, estrogen, and progesterone receptors lacking the hormone binding domain still recognize the specific response elements and may function as constitutive activators. Thus, neither the ligand itself nor its binding domain need to participate directly in DNA recognition.

The v-erbA protein has been found to contain an apparently intact DNA-binding domain; however, as a result of amino-acid changes and a deletion in the carboxy-terminal domain (relative to its progenitor, the thyroid hormone receptor), it lacks the ability to bind thyroid hormone. By analogy to mutated steroid hormone receptors, it has been proposed that these mutations, in conjunction with the high level of expression, convert WO 90/14356 PCT/US90/03113 -21the thyroid hormone receptor into v-erbA, which is a hormone independent transcription factor.

Utilizing a gel retardation assay, it has now been demonstrated that both v-erbA and the thyroid hormone receptor recognize and bind to a cognate response element in the absence of ligand. The consequence of this binding is the suppression of a thyroid hormone responsive reporter gene by the non-liganded receptor.

Addition of thyroid hormone results in a 10-100 fold stimulation of transcription from the repressed level.

Surprisingly, v-erbA does not function as an activator as one might expect, but rather as a constitutive repressor of T 3 responsive genes. When co-expressed with the thyroid hormone receptor, and in the further presence of thyroid hormone, the v-erbA repression is dominant, blocking hormone stimulated modulation, thereby demonstrating that v-erbA can function as a potent receptor antagonist.

The present invention thus embraces hormone or hormone-like receptor analogs having the ability to repress trans-activation transcription activity of a promoter with which it is associated, or an extrinsic, operative promoter. Such repression is due, for example, to the intrinsic ability of the analog receptor to competitively bind to a DNA response element of said promoter or by the ability of the analog receptor to competitively-displace other polypeptide(s) that bind to said DNA response element, or a proximate DNA response element, thus creating an overall repression of transactivation transcription activity compared with that of its corresponding parent or wild-type receptor.

WO 90/14356 PCT/US90/03113 -22-

EXAMPLES

The following experimental details set forth the methodology employed in identifying, characterizing and preparing particular novel analog receptors. The art skilled will recognize that by supplying the present information including the location and makeup of transactivation transcription repression domain(s) of a given receptor and how such receptor can be manipulated to produce the novel analog receptors hereof having repressed trans-activation transcription activity, it is not necessary, or perhaps scientifically advisable, to repeat the details described herein to reproduce this work. Instead, they may choose to employ alternative, reliable and kno f"hods. For example, they may synthesize the un _ying DNA sequences encoding a particular novel receptor hereof for deployment within similar or other suitable, operative expression vectors and culture systems. Thais, in addition to supplying details actually employed, the present disclosure serves to enable reproduction of the specific receptors disclosed, as well as others, and fragments thereof, using means within the skill of the art having benefit of the present disclosure. All of such means are included within the enablement and scope of the present invention.

Example 1: Transfections Transfections in JEG-3 human placental cells are performed via the calcium phosphate precipitation method described by Delegeane et al, Mol. Cell. Biol., Vol. 7, pp. 3994-4002 (1987). JEG-3 cells, maintained in Dulbecco's modified Eagle's medium (DMEM), 10 percent defined calf bovine serum (CBS), and 0.4 percent glucose are split 24 hours prior to transfection into 5 percent CBS charcoal-stripped serum plus glucose (Akerblom et WO 90/14356 PCT/US90/03113 -23al., Science 241, pp. 350-353 (1988)). Typically, 2 jg of reporter plasmid and 4 gg of receptor plasmid were used along with 2 gg of a Rous sarcoma virus (RSV)-8galactosidase construct (Hollenberg et al., Cell 49, pp.

39-46 (1987)) as an internal control for transfection efficiency. Dexamethasone and aldosterone M) were added after calcium phosphate treatment. For the Bgalactosidase fusion experiments, the internal control was RSV-luciferase.

Transfections in CV-1 cells are also performed via the calcium phosphate precipitation method. CV-1 cells are maintained in DMEM supplemented with 5% calf bovine serum and transfected at 30%-50% confluency with a total of 20 pg DNA. 5 Ag expression plasmid and 2.5 ig reporter plasmid DNA, together with 2.5 Mg RSV-p-gal as an internal control for transfection efficiency, are cotransfected per 10 cm dish. Transfected cells are grown in 10% resin-treated fetal calf serum [Samuels et al., J. Endocrinolocy, Vol. 105, pp. 80-85 (1979)] in the presence or absence of 10'7M 3,5,3' triiodothyronine (T 3 Cells are harvested 40 hours after the addition of T 3 3galactosidase and CAT-assays are then performed as described in Example 2.

Example 2: Reporter Assays Chloramphenicol acetyltransferase (CAT) assays are performed as described by Hollenberg, et al, Cell, Vol. 49, pp. 39-46 (1987), but with 25 yg of total cell extract protein for 3 hours or less. Thin-layer chromatography (TLC) plates are cut and counted in Econofluor containing 5% dimethyl sulfoxide (DMSO).

P-Galactosidase (P-gal) assays are performed as described by Herbomel, et al. Cell, Vol. 29, pp. 653-662 (1984).

WO 90/14356 PCT/US90/03113 -24- Example 3: hGR Mediated Negative Regulation To demonstrate hGR-mediated repression of the alphal68 promoter (a promoter for the alpha-subunit gene encoding chorionic gonadotropin; which is repressed by GR), a dose response study is conducted of negative regulation of the alphal68-CAT reporter plasmid by the hGR expression plasmid in human placental JEG-3 cells.

Varying amounts of the hGR expression plasmid (driven by the RSV promoter) and alph"168-CAT (reporter) expression plasmid are cotransfected by the calcium phosphate precipitation method. Cotransfections are carried out so as to provide increasing receptor to promoter ratios.

The resultant transient CAT activity ir the presence or absence of the steroid hormone dexameth.isone is then mepsured. Throughout the study, the total amount of transfected RSV promoter DNA is kept constant by substituting an RSV control plasmid, thus controlling for possible titration of transcription factors by RSV DNA.

The CAT activity of the reporter construction is measured as described above.

The alphal68-CAT reporter plasmid is constructed as described by Delegeane et al., supra.

The hGR expression plasmid, driven by the RSV promoter, is constructed as described by Hollenberg et al, Cell, Vol. 49, pp. 39-46 (1987).

Control plasmid RSV contains the rat thyroid hormone receptor coding region in the antisense direction, and is described by Thompson et al, in Science, Vol. 237, pp. 1610-1614 (1987).

Figure 2 shows the effect of the transfection of hGR cDNA on reporter gene expression in the presence and absence of dexamethasone. Open circles in the Figure indicate media without added dexamethasone, while solid WO 90/14356 PCT/US90/03113 circles indicate media with added dexamethasone (10 7

M).

In the particular experiment presented in Figure 2, 2 .g of promoter plasmid is used. The arrow indicates the ratio of receptor to promoter used in subsequent experiments.

Figure 2 shows that increasing amounts of the receptor expression plasmid yield a correspondingly higher steroid-dependent repression. In the absence of receptor cDNA, less than 10 percent of maximal repression is measured. Beginning at a receptor to promoter ratio of 1 and continuing to a ratio of 5, a plateau of repression activity emerges where more receptor plasmid yields no additional steroid-dependent repression. Since the amount of RSV promoter is held constant, this plateau indicates probable saturation of the site of receptor action. For subsequent experiments, a receptor to promoter ratio of 2:1 is used. The steroid-dependent repression varies between 6 and 20 fold with an average of 9 fold as typified in Figure 2. This assay can reliably measure as low as 10 percent of wild type hGR repression.

Figure 3 shows the relative CAT activity of wild-type hGR, truncated hGR's and several hGR fusion proteins. hGR fusion proteins are assayed as described above. For 1582*, wild-type hGR (hgrwt), 1532* and the I532*-B-galactosidase fusion, the hormone used is dexamethasone; for the fusion protein GGM, the hormone used is aldosterone.

In Figure 3B, the general make-up of the various hGR-derived proteins is illustrated. Thus, 1582* is a truncated hGR having only 582 amino acids; 1532* is a truncated hGR having only 532 amino acids; 1532*-Pgalactosidase is a fusion protein comprising amino acids 1-532 from hGR, plus amino acids 8-1025 from 8- WO 90/14356 PCT/US90/03113 -26galactosidase, made by inserting the 3030 bp BamHI LacZ fragment of pBG-1, a derivative of pSK105, in frame into the BamHI site of 1532 (Casadaban et al., Methods in Enzymolory Vol 100, 293, Academic Press (1983)); and GGM is a fusion protein constructed by first introducing an additional XhoI site into both the hGR at nucleotide position 1596 (Hollenberg et al., Nature 318, 635 (1985)) and the human mineralocorticoid receptor (hMR) at nucleotide position 2233 (Arriza et al., Science 237, 268 (1987)) and then inserting the appropriate XhoI fragment of the hMR into the hGR. This gave a receptor with amino acids 1-489 of hGR and 671-984 of hMR.

Figure 3B compares the repression activity of the carboxy terminal fusion proteins on the alphal68 promoter to activation on the mouse mammary tumor virus (MTV) promoter. The percentage of wild type activity is calculated by assigning RSV control plasmid as zero activity and the wild type hGR as 100 percent in each experiment, plus or minus the standard error of the mean.

Control 1 and control 2 refer to transfection with RSV control plasmid. Control 1 plus the next 4 constructions used dexamethasone as the steroid, while control 2 and GGM used aldosterone as the steroid. The numbers on the I532*-B-galactosidase construct refer to 8-galactosidase amino acid numbering from Casadaban et al., Meth.

Enzymol., Vol. 100, pp. 298-308 (1983), and the numbers on GGM refer to hMR amino acid numbering from Arriza et al., supra (1987). An indicates that activity is less than 10 percent wild type repression activity, indicates that activity is less than 1 percent wild type activation activity.

Novel sequence specific repressors can be created by attaching heterologous protein sequences to the carboxyl terminal side of the hGR DNA binding domain.

WO 90/14356 PCT/US90/03113 -27- In one case, E. coli B-galactosidase (B-gal) can be fused in frame to the carboxyl terminal side of the hGR DNAbinding domain and assayed for regulatory properties. On the mouse mammary tumor virus (MTV) promoter, this hybrid functions as a constitutive activator with properties unchanged from that of the parental truncated receptor.

On the alpha168 promoter, the fusion protein is a constitutive repressor whose activity is dramatically increased when compared to the truncated receptor, 1532*.

Thus, the addition of a heterologous E. coli protein sequence to the DNA binding domain of the hGR is sufficient for generation of a functional transcriptional repressor.

The results of this study provide several means to distinguish positive and negative regulatory effects of the hGR. First, the amino terminal domain that contains a potent activator sequence, T1, is not necessary for trans-repression. Indeed, it has been found that deletion of T, engenders a more potent repressor. This argues that even when functioning as a repressor, the amino terminal region of the hGR retains some residual positive activity.

The fact that certain modifications of the hGR produce a receptor that retains normal repressor function but has lost virtually all positive activation capability demonstrates that the process of activation can be mechanistically distinguished from that of repression.

This observation further indicates that the function of the DNA-binding domain is more than simply to locate an appropriate regulatory sequence. Moreover, the result also implies that activation requires an additional event subsequent to DNA-binding that is apparently not critical for negative control.

WO 90/14356 PCT/US90/03113 -28- Yet another distinction between activation and repression is that a B-galactosidase moiety functionally replaces the hGR carboxyl terminus only in repression.

Removal of the carboxyl terminus results in a receptor variant with minimal repression activity.

Example 4: Thyroid Receptor Mediated Negative Regulation Thyroid receptor (TR) differs from the glucocorticoid receptor (GR) in that the TR is able to bind to its cognate response element even in the absence of its associated ligand. In view of these differences, it is of interest to determine if modifications to the TR analogous to the modifications investigated for the GR will provide similar trans-repressing derivatives.

The vectors employed for these studies are prepared as follows: Expression vector RS-rTRa is constructed as described by Thompson et al, Proc. of the Natl. Acad. of Sci., Vol. 86, pp. 3494-3498 (1989).

The expression plasmid RS-v-erbA is generated by excising the coding sequence of the cloned gag-v-erbA gene [Vennstram et al., J. Virol. Vol. 36, pp. 575-585 (1980)], followed by appropriate modification of the and 3' ends and insertion between the Kpnl and BamH1 sites of the vector pRS-hGRN [Giguere at al., Nature 330, 624 (1987)]. Synthetic oligonucleotides encoding a palindromic response element [TREp; TCAGGTCATGACCTGA; see Glass et al., Cell Vol. 54, pp. 313-323 (1988)] flanked by HindIII adaptor sequences are inserted into the unique HindIII cloning site in pBL-CAT2 [Luckow et al., Nuc.

Acids Res. 15, 5490 (1987)]. Plasmids containing one or multiple copies of the TRE are identified by restriction enzyme mapping and sequence analysis. Hybrid genes are constructed using restriction sites common to both rTRa WO 90/14356 PCT/US90/03113 -29- [Thompson et al., Science, Vol. 237, pp. 1610-1614 (1987)] and v-erbA genes [Damm et al., EMBO J. Vol. 6, 375 (1987) and Vennstr6m et al., J. Virol. 36, 575 (1980). A schematic organization of the rTRa and v-erbA protein sequences is given in Figure 3A. cDNAs encoding these proteins are cloned into an RSV expression vector.

The "DNA" and "T 3

/T

4 "designations in the Figure refer to the DNA- and thyroid hormone binding domain, respectively. The 12 amino-terminal amino acids of chicken c-erbA/TRa are replaced by part of the retroviral gag-gene, resulting in the synthesis of a hybrid protein. In addition, v-erbA differs from chicken c-erbA/TRa in 2 amino acids in the DNA binding domain and 9 amino acids plus a 9 amino acid deletion in the hormone binding domain. The comparison to rat TRa shown in Figure 3A reveals an additional 17 amino acid differences that are species specific and also found in comparison between chicken and rat TRa deduced amino acid sequences.

The numbers on top of the constructs shown in Figure 3A indicate amino acid positions.

TR(Al54/317) is created by deleting the Pstl fragment in the ligand binding domain of rTRa.

Replacement of the rTRa carboxy-terminus by a Pstl-Xbal fragment from a v-erbA-neo construct [Sap et al., Nature 324, 635 (1986)] generates TR(A154/317)erbA. The plasmids TR(317)erbA, TR(154erbA and TR(154/317) are generated by reinserting the Pstl fragments from either rTRa or v-erbA into the unique Pstl site of either plasmid. All modifications are performed on a subcloned fragment from the ligand binding domain of rTRa and hybrid expression constructs are generated by replacing the Xba fragment of RS-rTRa with the corresponding chimeric fragments.

WO 90/14356 PCT/US90/03113 COS cells are grown in DMEM with 5% T3 free bovine serum and transfected using DEAE dextran [Giguere et al., Cell 46,'645 (1986)]. After 36 hours, cells are harvested and extracts prepared as described by Kumar et al., in Cell, Vol. 55, 145 (1988), except that the buffer was 20 mM Hepes (pH 0.4 M'KC1, 2 mM dithiothreitol (DTT) and 20% glycerol. DNA binding reactions and gel electrophoresis are performed essentially as described by Glass, et al., in Cell, Vol. 54, 313 (1988). Aliquots containing 6-10 Ag of total protein are diluted so that the final concentration of KC1 is 80 mM, then incubated with 2 Mg of polydeoxycytidylic acid (poly[dC]) for min at room temperature. At this time, 25-50 fmoles of a 32 P-labeled oligonucleotide encoding the palindromic TRE was added. The reaction mixture was incubated at 22'C for 30 min and then loaded on a 5% polyacrylamide gel containing 50 mM Hepes pH 7.8. Competitor DNAs were added prior to the addition of the labeled oligonucleotide.

The DNA binding mutant i95TR was constructed by partial digestion of rTRa with PvuII and adding a BamHI linker (12mer) to restore the open reading frame. The position of the linker wes verified by restriction enzyme mapping and sequence analysis.

The DNA binding mutant i95(154)erbA was constructed by partial digestion of TR(154)erbA with PvuII and adding a BamHI linker (12mer) to restore the open reading frame. The position of the linker was verified by restriction enzyme mapping and sequence analysis.

First, the transcriptional activity of both the rat thyroid hormone a receptor (rTRa) [Thompson, et al., Science 237, 1610 (1987)] and the v-erbA oncogene product is assessed by determining their ability to regulate WOi 90/14356 PCT/US90/03113 -31expression of thyroid hormone responsive reporter genes (see Fig. 4A). The constructs shown in the Figure contain oligonucleotides corresponding to previously identified thyroid hormone response elements [Glass et al., Nature _29, 738 (1987) and Glass et al., Cell A, 313( 1988)] linked to a thymidine kinase-chloramphenicol acetyltransferase (tk-CAT) fusion gene [Luckow et al., Nuc. Acids Res. 15, 5490 (1987)] (see Fig. 4B). The reporter gene constructs used, and shown in Figure 4B, contain oligonucleotides encoding the respective T 3 response elements (TRE) inserted into the HindIII site upstream of the tk promoter-CAT construct.

Expression plasmids encoding rTRa or v-erbA under the transcriptional control of the RSV long terminal repeat are cotransfected with one of the reporter plasmids into CV1 cells, which lack significant levels of endogenous

TR.

CV-1 cells are co-transfected with the reporter construct tk-CAT, tk-TRE-CAT or tk-TRE-,3.CAT; the respective expression plasmid (RS-rTRa or RS-v-erbA) and the internal reference plasmid RSV-PGAL. In the control experiments, a construct carrying the rTRa coding sequences in reverse orientation is used.

Transfection of tk-TRE -CAT or the parental vector tk-CAT results in a high basal level of CAT activity that is only marginally stimulated by the addition of thyroid hormone. In contrast, cotransfection with the TR expression vector RSV-rTRa results in marked effects on tk-TRE -CAT expression. In the absence of thyroid hormone, there is observed an 80% decrease in basal CAT activity (referred to as "low basal level activity"), indicating that TR expression provokes a ligand-independent inhibitory effect on transcription.

Addition of a thyroid hormone, triiodothyronine [T 3 to a WO 90/14356 PCT/US90/03113 -32final concentration of 100 nM results in a 20 fold stimulation of tk-TRE -CAT. This corresponds to a fold stimulation over the basal level of activity obtained in the absence of TR expression.

The regulatory function of the rTRa is not restricted to the TREp; the TRE from the rat growth hormone gene is also able to sustain hormone-independent and dependent transcriptional responses. This functional assay enables a direct determination of the putative transcriptional activity of the v-erbA oncogene product.

Since v-erbA has lost its ability to bind thyroid hormone but retains an intact DNA binding domain, it is reasonable to expect that it would function as a constitutively active TR. Unexpectedly, cotransfection of v-erbA with either reporter plasmid does not stimulate the transcription, but rather resembles the negative regulatory effects of rTRa in the absence of hormone.

Thus, in cells expressing v-erbA, CAT activity is reduced by 80% from the high basal level, and can not be relieved by the addition of T3 To identify the effect of different mutations in v-erbA on the altered properties of this protein (relative to its progenitor, TR), and to further dissect the processes of activation and repression, chimeric receptors of the v-erbA oncogene [see Damm et al., EMBO J. Vol. 6, pp. 375-382 (1987) and Vennstr6m et al. J.

Virol. Vol. 36, pp. 575-585 (1980)], and the rat TRa [Thompson et al., (1987), supra] are constructed. A schematic representation of the structure of the rat TRa/v-erbA chimeric proteins is presented in Figure Numbers on top of each construct indicate amino acid positions. The black bar indicates the deletion of 9 amino acids in v-erbA, resulting in fusion proteins of 401 amino acids compared to the 410 amino acid of the WO 90/14356 PC/US90/03113 -33protein with intact carboxy-terminus. In Figure positive and negative regulation of CAT activity from tk- TRE -CAT is shown, using the indicated expression vectors. The histogram summarizes the average values of 2-6 independent transfection experiments. Stippled bars in the histogram indicate the presence of no hormone; while striped bars in the histogram indicate the presence of 100 nM T 3 From amino acids 154 to 410, v-erbA differs from rTRa in 26 amino acids and a deletion of 9 amino acids close to the carboxy-terminus. Swapping this region of rTRa with v-erbA gives rise to the hybrid TR(154)erbA (Fig. 5A), whose properties are virtually identical to the v-erbA homologue (Fig. 5B). Thus, the mutations in the DNA binding domain and flanking amino acids are not crucial to the v-erbA phenotype.

Substitution of an internal region of rTRa ligand binding domain (at amino acids 154-316), creates a

T

3 responsive hybrid (TR(154/316)erbA) that functions like the natural receptor. In contrast, replacement of the carboxy-terminal 93 amino acids of rTRa with the corresponding sequence of v-erbA, containing the 9-aminoacid deletion and additional 11 amino acid differences, yields the hybrid protein TR(317)erbA with suppressor properties identical to the viral oncogene product.

To examine the effect of v-erbA on the function of endogenous thyroid receptor, cotransfection studies in CV1 cells are performed. The reporter gene tk-TRE -CAT pg) is cotransfected into CV-1 cells with 1 gg of rTRa expression vector and the internal control plasmid RSV-P-GAL. In addition, a 10-fold excess (10 pg) of the expression plasmids indicated in Figure 6 is cotransfected with rTRa. In the run designated by pg of the control plasmid, is used. The WO 90/14356 PCT/US90/03113 -34average values of 2-6 independent transfection experiments are shown in the histogram in Figure 6. All experiments are performed in the presence of 100 nM T 3 As shown in Figure 6A, a 10-fold molar excess of v-erbA reduces by 90% the thyroid hormone dependent induction of the reporter gene transcription by rTRa.

The hybrid constructs TR(154)erbA and TR(317)erbA also exhibit v-erbA-like activities, provoking a virtually complete repression of the T 3 and rTRa induced stimulation of the reporter gene. The activity is, however, dependent on the presence of an intact DNA-binding region since the DNA-binding domain mutant i95(154)erbA, fails to compete.

Similarly, T 3 induction is blocked by v-erbA and the chimeric constructs when the TREE is placed in the context of the MTV promoter. One prediction of these results is that an inverse relationship exists between the activity of the TR and the concentration of the competitor product. To examine this prediction, varying molar ratios of TR(154)erbA and the rTRa expression vector were cotransfected and hormone responsiveness was assessed. Thus, the reporter gene tk-TREp-CAT (0.5 Mg) was cotransfected into CV1 cells with 1 gg of expression vector RS-rTRa and increasing quantities of the nonhormone binding competitor TR(154)erbA. The amounts of RSV-promoter was held constant in all transfections by the addition of the control plasmid, Shown are the average values of three independent transfection experiments. Cells were grown in the presence of 100 nM

T,.

As one might expect, increasing quantities of TR(154)erbA lead to decreasing activity of the rTRa (see Fig. 6B). In this assay, even small amounts of WOo 90/14s356 PCT/US90/03113 TR(154)erbA are potent, with a 3:1 plasmid ratio completely blunting the hormone induced response.

Although it is generally assumed that the ability of a hormone receptor to bind DNA is ligand dependent [Evans, Science 240, 889 (1988)], the results presented herein, as well as previous observations [Lavin et al., J. Biol. Chem. 263, 9418 (1988)] challenge this convention. The downregulation of transcription from the responsive promoter is presumably a consequence of the ability of the receptor to recognize and bond to its cognate response element in the absence of hormone. Two pieces of evidence support this proposal. First, an intact DNA binding domain is necessary for both the gel retardation experiments and the observed transcriptional effects. Second, in the absence of the response element no significant repression is observed, whereas a tandem copy of the TRE potentiates both positive and negative transcriptional effects. Although positive synergism has been observed by glucocorticoid and estrogen receptors and their respective response elements [Schile et al., Science 242, 1418 (1988) and Str&hle et al., EMBO J. 7 3389 (1988)] the negative synergism observed here is without apparent precedent.

Example 5: Retinoic Acid Receptor Mediated Negative Regulation To investigate the possible positive and/or negative interaction of RARa with TRa and v-erbA, CV-1 cells are cotransfected with the RARa and reporter plasmid in the presence of TRa or v-erbA (Fig In the Figure. TRa or v-erbA compete with the RA induced transcriptional activation by RARa. RARa expression vector (1.0 Mg) is cotransfected with reporter plasmid AMTV-TREp-CAT (1.0 Mg) along with TRa (5.0 pg) or with v- WO 90/14356 PCT/US90/03113 -36erbA (5.0 pg). Plasmids are introduced into CV-1 cells (maintained in DME medium supplemented with 5% calf bovine serum), 5 x 105 per 10.0 cm dish, by calcium phosphate precipitation.

The expression plasmids, RARa TRa and v-erbA are under control of the RSV promoter as previously described by Damm et al. in Nature, Vol 339, 593 (1989) and Umesono et al, in Nature, Vol 335, 262 (1988). The cells were cultured for 36 hours in Dulbecco-modified Eagle medium containing 10% resin-charcoal stripped calf bovine serum [Samuels et al, Endocrinology, Vol 105, (1979)] with or without hormone as indicated and subsequently prepared for CAT activity through three cycles of freeze thawing as described by Gorman et al, Mol. Cell. Biol., Vol 2, 1044 (1982). An aliquot of the cell extracts was normalized by p-gal activity prior to carrying out the CAT assay. The hormone is added at a final concentration of 100nM. These data are the average of 4 independent experiments.

A 5 fold molar excess of TRa reduces the RA dependent induction in CAT activity by 90%. No inhibition of transcription is observed when both T3 and RA are added simultaneously. Therefore, in the absence of T3, the TRa prevents the RA induced activation of gene expression by the RARa, whereas in the presence of T3, one observes activation, presumably through the TRa.

Similar observations for the TR inhibition of RAR activation have recently been observed by Graupner et al., Nature, Vol 340, 653 (1989) and Brent et al., The New Biologist, Vol 1, 329 (1989). A similar competition of the RA induced gene activation is observed when v-erbA is cotransfected along with the RARa.

To investigate whether one could confer a dominant negative effect directly onto the RARa, mutant WO 90/14356 PCT/US90/03113 -37- RARas are created consisting of a series of RAR truncations as well as a hybrid RARa-erbA fusion. Since the RAR functions as a ligand dependent transcription factor, a mutant that contains an alteration or deletion in the ligand binding domain but with an intact DNA binding domain may confer a dominant negative phenotype.

In fact, the results set forth in Example 4 above demonstrate that replacing the ligand binding domain, located in the carboxyl terminus of the TRa, with the carboxyl terminus of v-erbA results in a hybrid TRa-erbA molecule that functions as a dominant negative mutant of the TRa. Therefore, a similar fusion between RARa and verbA was constructed by replacing the ligand binding domain of the RARa, located in the carboxyl termiral, with the carboxyl terminus of v-erbA (Fig The hybrid RARa-erbA fusion was constructed by removing the Nterminus and DNA binding domain of TR317-arb-A, see Damm et al, Supra, and replacing it with the cqrresponding Nterminus and DNA-binding domain the RARa (thus creating a RARa-erbA hybrid protein). Truncation mutants give the last amino acid of RARa before the insertion of a translation stop signal at the indicated position. The RAR 185 and 203 truncation mutants were constructed by using unique restriction sites present in the wild type receptor whereas RAR 153 was constructed by inserting a stop codon at position 153 RARa following creation of a unique Xho site at this position. Retinoic acid was added at a final concentration of 100nM.

In tle transactivation experiments, a positive response corresponds to a 25 fold increase in CAT activity upon retinoic acid addition. A negative response corresponds to no transactivation upon retinoic acid addition. In the competition experiments, a positive response corresponds to greater than an 8C.: WO 90/14356 PCT/US90/03113 -38antagonism of the retinoic acid induced transactivation, while a negative response corresponds to no competition upon retinoic acid addition.

The RARa-erbA protein does not act by itself as a transcriptional activator whether hormone is present or absent. However, when the RARa-erbA fusion is cotransfected with the RARa, it functions as a RARa antagonist. See Figure 9.

cells are cotransfected with RAR expression i, tor (1.0 ig), AMTV-TREp-CAT reporter plasmid (see Umesono et al, Supra) (5.0 Mg), reference plasmid mg), along with increasing amounts of RARa-erbA and carrier plasmid up to 20.0 mg. The addition of RSVpromoter is held constant in all transfections with the addition of the carrier plasmid. The maximal response refers to CAT induction observed with the RAR wild type receptor in the presence of 100 nM RA. This activation corresponds to a RA induction of 25 fold. All data are the average of 4 independent experiments.

RARa-erbA is also capable of antagonizing the RA induced activation of the RARP and RARy. Therefore, replacement of the crrboxyl terminus of RARa with the carboxyl terminus of v-erbA confers a dominant negative phenotype onto the RARa. In contrast, RAR mutants consisting of a series of carboxyl terminal truncations do not act as transcriptional activators when transfected by themselves nor do they function as RARa competitors when transfected in conjunction with the RARa. These data on the RAR truncations are in part at odds with the recent observation of Espeseth et al., Genes and Dev., Vol. 3, 1647 (1989), who reported that an RARa mutant, virtually identical to RAR-185, gave rise to a small percentage of stable F9 clones that did not differentiate in response to RA and therefore, hypothesized that this WO 90/14356 PCT/US90/03113 -39- RAR truncation mutant functioned as a dominant negative

RAR.

The properties of the RARa-erbA fusion along with the v-erbA and TRa, are further investigated by examining their ability to repress/antagonize an endogenous RTR(s). F9 cells have been established as a cellular model for RA dependent differentiation. They contain receptors for the a,f, and y subtypes of the RAR family; RARa and RARy are present in undifferentiated stem cell whereas RARP is induced upon RA treatment.

See, for example, Strickland and Mahdavi, in Cell, Vol 393 (1978); Sporn et al, in The Retinoids, Vol 1-2, Academic Press (Orlando, FL, 1984); and Hu and Gudas in Mol. Cell. Biol., Vol 10, 391 (1990). Transfection of either RARa-erbA, TRa or v-erbA into F9 cells results in a strong inhibition of the RA induced transactivation (Fig 10). F9 cells (maintained in DMEM supplemented with CBS) are transfected by the calcium phosphate method with the reporter AMTV-TREp-Luc Mg), 5.0 mg of either the control plasmid (RSV-CAT), or RSV-TRa, RSVverbA or RSV RARa-erbA. 5.0 Mg of reference plasmid and Mg of carrier plasmid. Cells are cultured for 24 hours in the presence or absence of 100 nM RA as indicated. The reporter construct is exactly the same as AMTV-TRE -CAT except that the gene encoding firefly luciferase [see DeWet et al., Mol. Cell. Biol., Vol 7, 725 (1987)] has been substituted for CAT in the reporter plasmid. Cells were cultured for 24 hours following the addition of the hormone and subsequently harvested and assayed for luciferase as described by Hollenberg et al, in Cell, Vol 55, 899 (1988). Data are the average of 4 independent measurements.

WO 90/14356 PCT/US90/03113 Example 6: Glucocorticoid Receptor-v erbA and GR-pGal Fusion Proteins The above observations that the RARa-erbA fusion protein functions as a dominant negative RAR, in conjunction with the ability of TRa-erbA hybrid to function as a TRa inhibitor, suggests that steroid-erbA hybrid receptors may provide a general approach for creating specific hormone receptor antagonists.

Therefore, a fusion between GR and erbA was created (GRerbA, Fig 11) by substituting the carboxyl terminus of verbA for the ligand binding domain of the GR.

In the Figure, wild type GR is shown at top with numbers indicating amino acid positions. The DNA and ligand binding domains are also indicated. The hybrid GR-erbA fusion is constructed by removing the ligand binding domain of the GR and replacing it with the carboxyl terminus of v-erbA, similar to the RAR-erbA fusion protein described above. The three truncation mutants give the last amino acid before the carboxyl terminal nonsense peptides. For GR-532pgal, E.coU igal was fused in frame to position 532 of the GR as described by Oro et al. in Cell, Vol 55, 1109 (1988). It encodes a protein that expresses P-galactosidase as well as the glucocorticoid receptor properties reported by Oro et al, Supra.

For the transactivation experiments, CV-1 cells are cotransfected with expression vector (1.0 the reporter, MTV-Luc (5.0 pg), RSV-CAT (5.0 Ig) as the internal control and carrier plasmid up to a total of 20.0 pg. Data is reported as maximal response and refers to luciferase induction observed with the GR wild type in the presence of the synthetic glucocorticoid, dexamethasone at 1 x 10"M. This activation corresponds to an induction of 3000 fold.

WO 90/14356 PCT/US90/03113 -41- For the competition experiments, CV-1 cells are cotransfected with 1.0 Mg of the GR expression vector, Mg of MTV reporter, 5.0 Ag of competitor along with Ag of RSV-CAT as the internal control and carrier plasmid up to a total of 20.0 Mg. Dexamethasone was added at a final concentration of 1 x 10' 7 M. The response refers to a antagonism of luciferase activity with the wild type receptor in the presence of a nonspecific competitor being equal to All data in Figure 11 are the average of four independent experiments.

The fusion protein does not activate a glucocorticoid responsive reporter gene when glucocorticoids are present or absent. However, when GRerbA is cotransfected with GR, it functions as a GR antagonist; a 5-fold molar excess of GR-erbA reduces by 87% the dexamethasone induction of a reporter gene.

For comparison, the properties are examined of a series of GR truncation mutants to act as dominant negative competitors as well as transcriptional activators. Truncations in the carboxyl terminus result in a mutant receptor that either does not activate transcription (GR487), or result in mutants that are constitutively active receptors (GR515, GR532). When the GR truncations are cotransfected along with the wild type they have either no inhibitory effect or at best a slight suppressive effect upon dexamethasone induced transactivation.

To examine whether the properties exerted by placing the carboxyl terminus of GR could be substituted by another polypeptide, 3-Gal was fused in frame to the carboxyl terminus of the GR at position 532. This GR- 532PGal fusion protein has previously been shown to function as a negative regulator of GR transcription (See WO 90/14356 PCT/US90/03113 -42- Example However, this P-Gal fusion protein is constitutively active similar to the parental truncation and it reduces the GR activation by only 60%. Therefore, the ability and usefulness of the GR truncations and P- Gal fusions to act as dominant negative repressors is diminished primarily by their constitutive activity as well as their weak ability to act as GR competitors. In contrast, GR-erbA is the only mutant receptor that contains no transcriptional activity in the absence of ligand and functions to block the dexamethasone induced transactivation. In addition, GR-erbA is able to act as a very potent competitor against a constitutively active GR receptor, such as GR532, and further supports its potential to function as a dominant negative inhibitor.

The foregoing description details specific methods that can be employed to practice the present invention. Having detailed specific methods initially used to identify, isolate, characterize, prepare and use the receptors hereof, and a further disclosure as to specific entities, and sequences thereof, the art skilled will well enough know how to devise alternative reliable methods for arriving at the same information and for extending this information to other intraspecies and interspecies related receptors. Thus, however detailed the foregoing may appear in text, it should not be construed as limiting the overall scope hereof; rather, the ambit of the present invention is to be governed only by the lawful construction of the appended claims.

Claims (29)

1. A trans-repressing analog receptor of the steroid/thyroid superfamily of receptors, said analog comprising: a first amino acid sequence which iv a DNA-binding domain, through which said analog is capable of binding to a hormone response element of a wild type receptor, and a second amino acid sequence which is positioned at the carboxy-terminal end of the DNA-binding domain, wherein said second sequence is selected from: a polypeptido which has at least about 90 as many amino acids as the ligand binding domain of the carboxy- terminal portion of said wild type receptor; whoroin said polypeptide Q4 la. less than about 60 amino acid identity relative to the carboxy- terminal domain of said wild type receptor over either: the entire length of said polypeptide, if shorter than the carboxy- terminal domain of said wild type receptor, or (ii) any segment of said polypeptide having the same length as the carboxy- terminal domain of said wild type receptor, or at least the 84 carboxy-terminal amino acids of the carboxy-terminal portion of the v-erbA protein as defined by amino acid numbers 555-639 (see Figure 1); provided that when said wild type receptor is a glucocorticoid receptor, said second amino acid \sequence is not •:.!ozIITUTE SHEET WO 90/14356 PCT/US90/03113 -44-

2. An analog receptor according to claim 1 wherein said hormone response element is operatively associated with a promoter which is subject to trans- repression.

3. An analog receptor according to claim 1 having substantially no transcriptional activation activity in the presence or absence of ligand.

4. An analog receptor according to claim 1 wherein said first amino acid sequence is the DNA-binding domain of said wild type receptor.

An analog receptor according to claim 4 wherein said wild type receptor is selected from: a retinoic acid receptor, a thyroid hormone receptor, a vitamin D 3 receptor, a glucocorticoid receptor, a mineralocorticoid receptor, an estrogen receptor, an estrogen-related receptor, an aldosterone receptor, an androgen receptor, or a progesterone receptor.

6. An analog wherein said DNA-binding glucocorticoid receptor, retinoic acid receptor.

7. An analog wherein said DNA-binding glucocorticoid receptor.

8. An analog wherein said DNA-binding thyroid receptor.

9. An analog wherein said DNA-binding retinoic acid receptor.

An analog wherein said first amino said wild type receptor, receptor according to Claim domain is derived from: 1) a 2) a thyroidAreceptor, or 3) a receptor according to Claim 6 domain is derived from the human receptor according to Claim 6 domain is derived from the receptor according to Claim 6 domain is derived from the receptor according to Claim 1 acid sequence is derived from and wherein said analog receptor has substantially no transcriptional activation activity in the presence or absence of ligand.

11. An analog receptor according to Claim 10 wherein said carboxy- terminal domain is obtained from a polypeptide.

12. An analog receptor according to Claim 11 wherein said polypeptide is at least a portion of the beta-galactosidase polypeptide.

13. An analog receptor according to Claim 10 wherein said carboxy-terminal domain comprises at least the 84 C-terminal amino acids of the C-terminal portion of the v-erbA protein.

14. A process for preparing an analog receptor according to Claim 1, said process comprising: expressing, in a recombinant host cell, transfected DNA encoding said receptor.

A process according to Claim 14 further comprising: recovering and purifying said analog receptor to enable its use in assays that measure extrinsically induced biofunctionality of said analog receptor.

16. A DNA molecule that is a recombinant DNA molecule or a cDNA molecule encoding an analog receptor according to Claim 1.

17. An expression vector operatively harboring DNA according to Claim 16.

18. A recombinant host cell transfected with an expression vector 'according to Claim 17.

19. A cell culture comprising cells according to Claim 18 and an S 25 extrinsic support medium assuring the viability of said cells.

A non-human transgenic mammal, having disease symptoms due to an inability to respond normally to a steroid or thyroid hormone, wherein at least a subset of the cells of said mammal are capable of expressing an analog receptor of the steroid/thyroid superfamily of receptors according to any one of Claims 1 to 13 for one of said hormones, and wherein said analog receptor has trans-repression activity greater than that of its corresponding wild type receptor and trans- activation activity less than that of its corresponding wild type receptor.

21. Method for converting a wild type hormone receptor into a trans-repressing analog receptor of the steroid/thyroid superfamily of 4R64/ii -46- receptors according to any one of claims 1 to 13 said method comprising: replacing the ligand binding domain of said wild type receptor with at least the 84 carboxy terminal amino acids of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1).

22. Method for converting a wild type hormone receptor into a trans-repressing analog receptor of the steroid/thyroid superfamily of receptors according to any one of claims 1 to 13 said method comprising: replacing the ligand binding domain of said wild type receptor with a polypeptide which has at least 90% as many amino acids as the ligand binding domain of the carboxy-terminus of said wild type receptor and less than about 60% amino acid identity relative to the carboxy-terminus of said wild type receptor over either: the entire length of said polypeptide, if shorter than the carboxy-terminus of said wild type receptor, or (ii) any segment of said polypeptide having the same length as the carboxy-terminus of said wild type receptor.

23. A method for blocking the transcriptional activation of a hormone response element in a cell, said method comprising contacting said cell with an analog receptor according to Claim 1. 20

24. A method for blocking the transcriptional activation by a wild type receptor of the steroid/thyroid superfamily of receptors of a hormone response element present in a cell, said method comprising substantially deleting the ligand binding domain of said wild type receptor, 25 operatively linking the modified receptor of step at the carboxy-terminal end thereof to at least the 84 carboxy terminal amino Sacids of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1) to produce a fusion protein, and thereafter contacting said cell with an effective amount of said fusion protein.

25. A method for blocking the transcriptional activation by a wild type receptor of the steroid/thyroid superfamily of receptors of a hormone response element present in a cell, said method comprising substantially deleting the ligand binding domain of said wild type receptor, operatively linking the modified receptor of step at the carboxy-terminal end thereof to a polypeptide which has at least about A 90% as many amino acids as the ligand binding domain of said wild type f-o N. receptor, and less than about 60% amino acid identity relative to the 1737a/ii -47- ligand binding domain of said wild type receptor over either the entire length of said polypeptide, if shorter than the ligand binding domain of said wild type receptor, or (ii) any segment of said polypeptide having the same length as the ligand binding domain of said wild type receptor to produce a fusion protein, and thereafter contacting said cell with said fusion protein.

26. A method for identifying the response element and/or function of a receptor for which the associated response element and/or function is not known, said method comprising: comparing the response of a test system which is responsive to wild type receptor when said test system is treated with: wild type receptor, or (II) trans-repressing analog receptor of the steroid/thyroid superfamily of receptors according to any one of claims 1 to 13 wherein said trans-repressing analog receptor comprises: a first amino acid sequence which is a DNA-binding domain, through which said analog is capable of binding to a hormone response element of said wild type receptor, and 20 a second amino acid sequence which is positioned at the carboxy-terminal end of the DNA-binding domain, wherein said second sequence is selected from: a polypeptide which has at least about 90% as many amino acids as the ligand binding domain of the carboxy-terminal portion of said wild 25 type receptor and less than about 60% amino acid identity relative to the carboxy-terminal domain of said wild type receptor over either the entire length of said polypeptide, if shorter than the carboxy- terminal domain of said wild type receptor, or i (ii) any segment of said polypeptide having the same length as the 30 carboxy-terminal domain of said wild type receptor; or at least the 84 carboxy-terminal amino acids of the S 1737a/ii -48- carboxy-terminal portion of the v-erbA protein as defined by amino acid numbers 313-398 (see Figure 1).

27. In an assay system for the determination of the presence of a specific wild type receptor, wherein more than one response element is capable of interacting with said wild type receptor, the improvement comprising inactivating, with respect to said assay, response element(s) which also respond to wild type receptor(s) other than said specific wild type receptor; wherein said response elements are inactivated by adding to said assay system an effective amount of a trans-repressing analog receptor of the steroid/thyroid superfamily of receptors according to any one of claims 1 to 13 for each of said other receptor(s), wherein each of said trans-repressing analog receptors comprises: a first amino acid sequence which is a DNA-binding domain, through which said analog is capable of binding to a hormone response element of a receptor other than said specific wild type receptor, and a second amino acid sequence which is positioned at the carboxy-terminal end of the DNA-binding domain, wherein said second sequence is selected from: a polypeptide which has at least about 90% as many amino acids as the ligand binding domain of the carboxy-terminal portion of said receptor other than said specific wild type receptor and less than about amino acid identity relative to the carboxy-terminal domain of said 5 receptor other than said specific wild type receptor over either the entire length of said polypeptide if shorter than the carboxy-terminal domain of said receptor other than said specific wild type receptor, or (ii) any segment of said polypeptide having the same length as the carboxy-terminal domain of said receptor other than said specific wild type receptor; or S 30 at least the 84 carboxy-terminal amino acids of the carboxy-terminal portion of the v-erbA protein as defined by amino acid inumbers 313-398 (see Figure 1).

28. A trans-repressing analog receptor of the steroid/thyroid superfamily of receptors according to any one of claims 1 to 13 and substantially as herein described with reference to any one of the 4864a/ii -49- Examples or to Figure 1.

29. A process for preparing an analog receptor of the steroid/ thyroid superfamily of receptors according to any one of claims 1 to 13 substantially as hereinbefore described with reference to any one of the Examples or to Figure 1. DATED this TWENTY-SECOND day of FEBRUARY 1993 The Salk Institute for Biological Studies Patent Attorneys for the Applicant SPRUSON FERGUSON e• ee 4i o S t o so a i i 464 */ii /e z i^ e