WO1992002529A1

WO1992002529A1 - Expression vectors regulated by estrogen and a mutant estrogen receptor

Info

Publication number: WO1992002529A1
Application number: PCT/US1991/005404
Authority: WO
Inventors: David J. Shapiro; David A. Nielsen; Byron W. Kemper; Elzbieta Szczesna-Skorupa; Hong Xing
Original assignee: The Board Of Trustees Of The University Of Illinois; Biotechnology Research And Development Corporation
Priority date: 1990-07-30
Filing date: 1991-07-30
Publication date: 1992-02-20
Also published as: AU8416091A

Abstract

The invention comprises methods and materials for the estrogen-regulated expression of proteins and polypeptides of interest. The invention provides a hybrid mRNA molecule comprising an RNA sequence coding for a protein or polypeptide of interest operatively linked to an RNA sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor. The invention further provides a mutant estrogen receptor with an enhanced ability to activate transcription from an estrogen response element (ERE). The invention also provides vectors which comprise a synthetic estrogen-regulatable promoter operatively linked to a sequence coding for estrogen-stabilizable hybrid mRNA and other vectors which comprise a synthetic promoter comprising two consensus EREs operatively linked to a TATA box. Using the methods and materials of the invention, enhanced production of proteins or polypetides of interest is obtained.

Description

"EXPRESSION VECTORS REGULATED BY ESIHOGEN AND A MUTANT ESTROGEN RECEPTOR" ^•

The research which forms a basis of this in¬ vention was supported in part by grant DCB 86-04109 from the National Science Foundation. The U.S. government may have rights in this invention.

This application is a continuation-in-part of application Serial No. 07/559,853, filed July 30, 1990.

BACKGROUND OF THE INVENTION

This invention relates to methods and materials for producing proteins and polypeptides by recombinant DNA techniques. In particular, the invention relates to a method and materials useful in stabilizing a messenger RNA (mRNA) in the presence of estrogen and estrogen receptor to obtain increased production of the protein or polypeptide coded for by the mRNA. The invention also relates to a mutant estrogen receptor with an enhanced ability to activate transcription from an estrogen response element and to vectors which comprise a powerful, synthetic estrogen-regulated promoter comprising two estrogen response elements linked to a TATA box.

A major goal of the biotechnology industry is the production of large amounts of desired proteins and polypeptides by transformed host cells. Although many proteins and polypeptides have been successfully expressed in prokaryotic host-vector systems, problems of codon usage (Springer and Sugar, Proc. Natl. Acad. Sci. USA. 84. 8961-8965 (1987)), proteolytic cleavage of large polypeptides and the absence of post-translational modification (Bending, In Genetic Engineering. 1_, 90-127 (Rigby ed., Academic Press, London 1988)) mean that there is a large class of important proteins and polypeptides which are not expressed in a biologically active form in prokaryotic host-vector systems. In yeast and baculovirus expression systems, expression of proteins in an insoluble form, incorrect proteolytic processing and incorrect post-translational modification remain common problems. Clark and Kamen, Science. 236. 1229-37 (1987); Rusche, et al. Proc. Natl. Acad. Sci. USA. 84. 6924-28 (1987) . For many proteins and polypeptides (such as cell growth factors, oncogenes and hormone receptors) , expression in vertebrate cells would be the system of choice, but the continuous high level expression of these proteins is lethal to mammalian cells (Bending, supra; Hammerschmidt, et al. J. Virol.. 63. 2469-2475 (1989); Ziegler, et al.. Cell. 27. 477-486 (1981)). Neverthe¬ less, to date, virtually all efforts to produce systems for the expression of proteins and polypeptides of interest have centered around the production of vectors with efficient promoters or with the capacity to exist at high copy number within cells (Bending, supra) . Increasing the stability of mRNA transcripts has been largely neglected.

The intracellular level of an mRNA dictates the amount of protein or polypeptide which can be produced by translation of that mRNA. In turn, the level of the mRNA is regulated by its rate of synthesis and its rate of degradation (its stability) . Regulation of the stability of eukaryotic mRNAs is a common control mechanism, and the processes governing the degradation of the more than 20 eukaryotic mRNAs whose stability is known to be regulated are currently attracting considerable attention. See Cleveland and Yen, New Biol.. 1. 121-126 (1989); Brawerman, Cell. 57. 9-10 (1989); Raghow, Trends Biochem. Sci.. 12. 358-360 (1987) ; Ross, Sci. Am.. 260. 48-55 (1989); Shapiro et al., BioEssays. £, 221-226 (1987) ; Nielsen and Shapiro, Molecular Endo.. 4. 953-57 (1990) .

It has been reported that the stability of a few mRNAs is regulated by their 5' untranslated regions. Morris et al., Proc. Natl. Acad. Sci. USA. SI, 981-985 (1986); Treisman, Cell. 42. 889-902 (1985); Dani et al., Proc. Natl. Acad. Sci. USA. £1, 7046-7050 (1984); Eick et al., EMBO J.. i, 3717-3725 (1985). Indeed, Santos et al. have suggested a role for this region in the estrogen regulation of c-myc mRNA. Santos et al., J. Biol. Chem.. 263. 9565-9568 (1988) .

However, most studies have identified stem and loop structures, A+U-rich sequences and other elements in the 3' untranslated region, and the length of the poly(A) tract available for interaction with the poly(A) -binding protein, as being major determinants of mRNA stability. Mullner and Kuhn, Cell. 53. 815-825 (1988); Pandey and Marzloff, Mol. Cell. Biol.. 7, 4557-4559 (1987); Ross and Kobs, J. Mol. Biol.. 188. 579-593 (1986); Brewer and Ross, Mol. Cell. Biol.. £, 1697-1708 (1988); Shaw and Kamen, Cell. 46. 659-667 (1986); Carrazana et al., Mol. Cell. Biol.. £, 2267-2274 (1988); Muschel et al., Mol. Cell. Biol.. £, 337-341 (1986); Paek and Axel, Mol. Cell. Biol.. 1, 1496-1507 (1987); Shapiro et al., Mol. Cell. Biol.. .8, 1957-1969 (1988) . In particular, Shapiro et al., BioEssays. £, 221 (1987) suggests that the 3' end of vitellogenin mRNA may be capable of conferring the ability to be stabilized by estrogen on a heterologous mRNA.

Translation is required for controlling the stability of j8-tubulin mRNA and histone mRNAs. Yen et al., Nature (London) . 334. 580-585 (1988); Graves et al., Cell. 48. 615-626 (1987)). Premature translation termination results in the degradation of triosephosphate isomerase and yeast URA1 and PGK1 mRNAs. Daar and Maquat, Mol. Cell. Biol.. 8., 802-813 (1988); Pelsy and Lacroute, .Curr. Genet.. §_, 277-282 (1984); Hoekma et al., Mol. Cell. Biol.. 2, 2914-2924 (1987) .

The expression of many genes is regulated by the action of steroid hormones. Evans, Science. 240. 890-895 (1989); Beato, Cell. ≤_£, 336-344 (1989). The intracellular actions of steroid hormones appear to be mediated exclusively by cellular proteins called hormone receptors which bind individual steroid hormones with high affinity and specificity. Evans, Science. 240. 890- 95 (1989) . These steroid hormone receptors have three important properties: (a) specific high affinity binding of their cognate hormone ligand(s); (b) the ability to recognize and bind to specific DNA sequences termed hormone response elements; (c) the ability to undergo a conformational change on hormone binding which enables them to activate transcription from promoters under the regulatory control of their hormone response elements. Id. ; Beato, Cell. £6, 336-44 (1989) .

For instance, the estrogen receptor binds estrogen (e.g.. estradiol-173) , and the ability of the estrogen-estrogen receptor complex to activate transcription depends on its interaction with a DNA sequence termed the estrogen response element (ERE) . The consensus sequence of the ERE is 5' -AGGTCANNNT GACCT-3'

[SEQ ID NO:16] Beato, Cell. 56, 336-44 (1989) .

The 5' flanking regions of estrogen-regulatable genes contain one or more copies of an ERE (a consensus ERE and/or an imperfect ERE) . See, e.g.. Klein-Hitpass et al., Cell. 46. 1053-61 (1986); Martinez et al., The EMBO J.. 6., 3719-27 (1987) . An imperfect ERE is an ERE whose sequence differs by one or more nucleotides from the consensus ERE sequence. It has been reported that two imperfect EREs, which are individually inactive, function synergistically to give levels of transcription activation equal to that of a single consensus ERE. Martinez et al., The EMBO J.. £, 3719-27 (1987); Martinez and ahli, The EMBO J.. £, 3781-91 (1989) . Using synthetic minimal promoters containing only EREs and a TATA box, the transcription activation of two consensus EREs was found to be additive and independent of their relative spacing when they were positioned about 35 bp from the TATA box. Ponglikitmongkol et al., The EMBO J.. £, 2221-31 (1990). However, the two consensus EREs acted synergistically and in a stereoalignment-dependent manner when they were positioned about 175 bp further upstream of the TATA box. Id. Finally, combinations of a consensus ERE and an imperfect ERE acted synergistically when positioned proximate to the TATA box. Id.

Several estrogen receptors have been cloned. For instance, the Xenopus laevis estrogen receptor which binds estradiol-17/5 (Barton and Shapiro, Proc. Natl. Acad. Sci. USA. ££, 7119-23 (1988)) has been cloned (Weiler et al., Molec. Endo.. i, 355-62 (1987).

The human estrogen receptor has also been cloned, and chimeric human estrogen receptors have been prepared which consist of the activation region of the herpes simplex virus transcription factor VP16 inserted into the amino-terminal region of the human estrogen receptor, either containing or lacking its hormone- binding region. Elliston et al., J. Biol. Chem.. 265. 11517-21 (1990) . These chimeric estrogen receptors reportedly are about 10-fold more potent than the wild type estrogen receptor in activating gene expression. They can activate gene transcription in a hormone- dependent manner or in a constitutive, hormone- independent manner when the chimeric receptor lacks its hormone-binding region. The Elliston et al. article also teaches that the seventy-eight amino acid activation region of. the VP16 transcription factor used in the chimeric receptors is highly negatively charged, as are the activation domains of some other transcription activating factors. The Elliston et al. article also indicates that some aspect of structure, such as an alpha-helix, is essential for formation of an activating region. However, Cress and Triezenberg, Science. 251. 87-90 (1991) reports that a putative amphipathic alpha- helix does not appear to be an important structural component of the activation domain of the VP16 transcription factor.

Hormones have also been shown to regulate the stability of several mRNAs, but the absence of functional assays for steroid hormone control of mRNA degradation has greatly hindered studies. As a consequence, the mechanisms controlling mRNA degradation in these systems remain largely unknown.

One mRNA whose synthesis and stability are regulated by a hormone is the mRNA coding for the Xenopus laevis egg yolk precursor protein vitellogenin. The regulation of vitellogenin gene transcription and control of the cytoplasmic stability of vitellogenin mRNA, not changes in the efficiency of nuclear vitellogenin RNA processing, are responsible for the massive estrogen induction of the hepatic mRNA coding for vitellogenin. Blume et al., UCLA Symp. New Ser.. 52. 259-274 (1987); Brock and Shapiro, J. Biol. Chem.. 9_, 5449-5455 (1983); Brock and Shapiro, Cell. 2A, 207-214 (1983) .

The stability of vitellogenin mRNA is greatly increased in the presence of high levels of estradiol- 170. Brock and Shapiro. Cell. 14, 207-214 (1983) . In Xenopus liver fragment cultures, vitellogenin mRNA is essentially stable (with a half-life of more than 500 hours) when estrogen is present. Id. Upon removal of the estrogen, vitellogenin mRNA begins to rapidly degrade with a half-life of 16 hours at 25°C or 30 hours at 20°C. Blume and Shapiro, Nucleic Acids Res.. 17. 9003-9014 (1989) . Vitellogenin mRNA undergoing rapid cytoplasmic degradation can be specifically restabilized by the addition of estrogen to the culture medium [Brock and Shapiro, Cell. 34. 207-214 (1983)], indicating that vitellogenin mRNA stabilization is a reversible cytoplasmic effect of estrogen. In hepatocytes, estrogen elicits a 20-fold increase in the stability of vitellogenin mRNA and decreases the stability of albumin mRNA by 3-fold without affecting the 16-hour half-life (at 25°C) of total poly(A) mRNA. Id. : Reigel et al., Mol. Cell. Endocrinol. M, 201-209 (1986); olffe et al., Eur. J. Biochem.. 146. 489-496 (1985) . The broad spectrum of estrogen effects on mRNA stability in Xenopus liver provides a striking example of the specificity and flexibility of this type of control.

SUMMARY OF THE INVENTION

The invention comprises methods and materials for the estrogen-regulated expression of proteins and polypeptides. First, the invention provides a hybrid mRNA molecule which comprises a sequence coding for a protein or polypeptide of interest operatively linked to a second RNA sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor. The invention further provides a recombinant DNA sequence coding for this hybrid mRNA, a vector comprising the recombinant DNA sequence operatively linked to expression control sequences, and a host cell transformed with the vector. Also provided is a method of stabilizing an mRNA so as to obtain increased produc¬ tion of the protein or polypeptide of interest. The method comprises transforming a host cell with the vector referred to above. If the host cell does not normally produce estrogen receptor, it must also be transformed with a vector coding for estrogen receptor. Finally, the transformed host cell is grown in the presence of estrogen, and the hybrid mRNA is stabilized so that increased production of the protein or polypeptide of interest is obtained. Indeed, the level of the hybrid mRNAs of the invention can be increased by up to 20-fold relative to other mRNAs when the transformed host cells of the invention are cultured in the presence of estrogen.

The invention further comprises an expression vector comprising a synthetic estrogen-regulatable promoter operatively linked to a DNA sequence coding for a hybrid mRNA. The synthetic promoter comprises one or more estrogen response elements, the type and number of estrogen response elements being sufficient so that transcription of the hybrid mRNA is estrogen-regulatable. The promoter preferably comprises two consensus estrogen response elements operatively linked to a TATA box, and most preferably comprises two consensus estrogen response elements spaced 20 nucleotides apart, as measured center to center, and spaced 20 nucleotides upstream from the TATA box. The hybrid mRNA comprises a 5' untranslated region, a coding region coding for a protein or polypeptide of interest, and a 3' untranslated region. The 3' untranslated region preferably contains a sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor (see above) . The invention also provides host cells transformed with the estrogen-regulatable vectors and another method of obtaining enhanced production of a protein or polypeptide of interest. This second method comprises transforming a host cell with the estrogen-regulatable vectors, also transforming the host cell with a vector coding for estrogen receptor if the host cell does not normally produce estrogen receptor, and growing the transformed host cell in the presence of estrogen.

The invention also provides a mutant estrogen receptor having the ability to elicit enhanced production of a protein or polypeptide which is under the control of an estrogen-regulatable promoter. The mutant estrogen receptor is an estrogen receptor which contains one or more copies of a heterologous amino acid sequence coding for a peptide that, should it form an alpha helix, would form an amphipathic helix having one acidic face, one hydrophobic face, and either one uncharged polar face or another acidic face. Also provided is a recombinant DNA sequence coding for the mutant estrogen receptor, a vector comprising the recombinant DNA sequence operatively linked to expression control sequences, and a host cell transformed with the vector. In addition, the invention comprises a method of obtaining enhanced production of a protein or polypeptide of interest comprising transforming a host cell with a vector comprising a DNA sequence coding for the protein or polypeptide of interest operatively linked to an estrogen-regulatable promoter. The method further comprises transforming the host cell with the vector coding for the mutant estrogen receptor, and growing the transformed cells in the presence of estrogen.

Most preferably, the estrogen-regulatable vector, the mRNA stabilization sequence and the mutant estrogen receptor of the invention are used together to obtain maximum production of a desired protein or polypeptide. Thus, an estrogen-regulatable expression vector of the invention containing sequences that provide for mRNA stabilization of the hybrid mRNA coded for by the vector and a vector coding the mutant estrogen receptor are used to transform the same host cell which is then cultured in the presence of estrogen. In this manner, efficient, high-level transcription of the hybrid mRNA coding for the protein or polypeptide of interest is obtained, and the mRNA is stabilized. As a result, large quantities of the desired protein or polypeptide are produced.

The invention is particularly suitable for the production of large quantities of proteins and polypeptides that are toxic to cells since the transformed host cells can be cultured in the absence of estrogen until a large number of cells have been grown, at which time estrogen is added to the culture medium to elicit a large burst of protein synthesis.

The use of the present invention will greatly decrease the cost of producing many proteins and polypeptides by genetic engineering techniques since the number of cells and the amount of cell culture medium required to produce a given amount of protein or polypeptide will be substantially reduced. Further, the protein or polypeptide will be present at a higher concentration relative to other cell proteins, and the number and complexity of the purification procedures required to obtain the protein or polypeptide of interest in usable form will be reduced. This simplification of the purification procedures will also lower the cost of producing the proteins and polypeptides.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure IA: Diagram of pMV5, MV5 mRNA and the MV5 probe.

Figures 1B-D: Flow diagrams illustrating the preparation of pMV5.

Figures 2A-B: Shows the results of assays of the stabilization of MV5 mRNA by estradiol and the estrogen receptor.

Figure 3: Schematic of the MV5 and TK probes and the sizes of the parts of the probes protected by MV5, vitellogenin Bl and TK mRNAs. Figure 4: Flow diagram illustrating the prepa¬ ration of pXER.

Figure 5: Flow diagram illustrating the preparation of pMV8.

Figure 6A-6F: Sequences of the relevant portions of pMV5, pMV8, pMVIO, pMVll, pMV12 and pMV13.

Figure 7: Schematic representations of pMV8, pMVIO, pMVll, pMV12 and pMV13, and a summary of the results of mRNA stabilization assays using these vectors.

Figure 8: Flow diagram illustrating the preparation of pXER-lAH and pXER-2AH.

Figure 9: Proposed structure of the peptide coded for by OligoAH as an amphipathic helix.

Figure 10: Schematic representations of the domain structure of native estrogen receptor and of two mutant estrogen receptors having one or two copies of an amphipathic helix peptide inserted in the hinge region (domain D) .

Figure 11: Schematic diagram of pVITCAT.

Figure 12: Flow diagram illustrating the preparation of p2EREXER.

DETAILED DESCRIPTION OF THE PRESENTLY PREFERRED EMBODIMENTS

The invention can be used to produce large quantities of any desired protein or polypeptide. The protein or polypeptide may be one that is normally made by the host cell (a "homologous" protein or polypeptide) or may be one that is not normally made by the host cell (a "heterologous" protein or polypeptide) . In this man¬ ner, even an mRNA coding for a homologous protein or polypeptide may be stabilized and production of the homologous protein or polypeptide enhanced.

Suitable proteins and polypeptides include: human and other animal hormones such as human insulin, human growth hormone, bovine growth hormone, swine growth hormone, thyroid stimulating hormone, follicle stimulating hormone, vasopressin and prolactin; cell growth factors; the various interferons; hormone recep¬ tors; oncogenes; blood factors such as Factor VII, Factor VIII, erythropoietin and tissue plasminogen activator; lymphokines; globulins such as immunoglobulins; albumins; endorphins such as beta-endorphin and enkephalin; enzymes; viral antigens such as influenza antigenic protein and hepatitis core and surface antigens; and other useful proteins and polypeptides.

The invention is particularly suitable for the production of proteins and polypeptides whose biological activity makes them toxic to cells when they are produced continuously. Examples of such toxic proteins are cell growth factors, oncogenes and hormone receptors.

A. Stabilization of Hybrid mRNAs

The RNA sequence that provides for stabilization of the hybrid mRNAs of the invention may be derived from any mRNA which is normally stabilized by estrogen and estrogen receptor. Such mRNAs include those coding for vitellogenin, conalbumin, ovalbumin, and apo VLDL II. The stabilization sequence may comprise one or more portions of these mRNAs which are sufficient to stabilize the hybrid mRNA.

The stabilization sequence preferably comprises one or more portions of a vitellogenin mRNA sufficient to stabilize the hybrid mRNA. In Example 1, it was found that an mRNA having the following X. laevis vitellogenin Bl mRNA sequences was stabilized: a first portion comprising the entire 5' untranslated region (13 nucleotides) linked to 81 nucleotides coding for the 27 amino-terminal amino acids of vitellogenin Bl; and a second portion comprising 270 nucleotides coding for the carboxy-terminal 90 amino acids of vitellogenin Bl linked to the entire 3' untranslated region (165 nucleotides). A hybrid mRNA according to the invention can be prepared using these sequences as follows. The first portion of the vitellogenin Bl sequences would be located 5' of the sequence coding for the desired protein or polypeptide, and the second portion would be located 3' of this sequence. Of course, all sequences would be operatively linked so that the hybrid mRNA would be stabilized in the presence of estrogen and estrogen receptor and would be translated to produce a protein or polypeptide of interest having the correct amino acid sequence.

Most preferably the stabilization sequence is the 3' untranslated region of a vitellogenin mRNA, such as the 3' untranslated region of the Xenopus laevis vitellogenin Bl mRNA, or one or more portions thereof sufficient to stabilize the hybrid mRNA. In Example 3, data are presented which establish that the 5' untranslated and coding regions of the X. laevis vitellogenin Bl mRNA are not essential for mRNA stabilization. Thus, the 3' untranslated region of this mRNA is sufficient to stabilize mRNAs in the presence of estrogen and estrogen receptor, and a hybrid mRNA according to the invention can be prepared by operatively linking the 3' untranslated region of the X. laevis vitellogenin Bl mRNA to an RNA sequence coding for a protein or polypeptide of interest.

When, the stabilization sequence is located downstream of the coding sequence in the hybrid mRNAs of the invention, it may comprise two sub-sequences. The first sub-sequence is the sequence responsible for the estrogen stabilization of the hybrid mRNA. The second sub-sequence is located downstream from the first sub¬ sequence and stabilizes the hybrid mRNA by providing for correct processing of the 3' end of the hybrid mRNA.

A preferred first sub-sequence is the 96 nucleotides at the start (i.e.. the upstream end) of the 3' untranslated region of X. laevis Bl mRNA (see Example 3) , or a portion thereof sufficient to provide for estrogen stabilization of the hybrid mRNA. Deletion of these nucleotides from the 3' untranslated region of X. laevis Bl mRNA produces mRNAs that are not stabilized in the presence of estrogen and estrogen receptor (see Example 3) . Also, it has been found that the basal level of an mRNA having this deletion was about the same as the basal level of an mRNA identical in other respects but containing the 96 nucleotides. These results indicate that one or more sequences in this region provide for the specific estrogen stabilization of the hybrid mRNA (see Example 3) .

Preferably the second sub-sequence is: 5'-UNUAAAUGUR UAUU-3' [SEQ ID NO:13] wherein N is any nucleotide and R is any purine. This sequence is a conserved sequence found in the 3' untranslated regions of many mRNAs, including all vitellogenin mRNAs and other estrogen-stabilized mRNAs (see Example 3) . The data presented in Example 3 indicate that this sequence plays a role in the proper processing of the 3' ends of mRNAs. In particular, the deletion or replacement of this sequence resulted in the production of mRNAs shorter than expected and with new 3' ends. The data also show that deleting or replacing this sequence produces mRNAs that are not stabilized. The failure of the mRNAs having this sequence deleted or replaced to be stabilized is likely due to the formation of shorter mRNAs with different 3' ends and 3' untranslated regions.

It is contemplated that the first sub-sequence and the second sub-sequence may be from the same source or from different sources. For instance, the entire 3' untranslated region of the X. laevis vitellogenin Bl mRNA, which contains sequences responsible for estrogen stabilization and sequences which stabilize mRNA by providing for correct processing of the mRNA, may be used as the stabilization sequence. Alternatively, the sequences of the X. laevis vitellogenin Bl mRNA responsible for estrogen stabilization (such as the 96 nucleotides at the start (upstream end) of the 3' untranslated region, or portions thereof sufficient to stabilize an mRNA) could be combined with sequences from another mRNA (whether estrogen stabilized or not) which provide for correct processing of the 3' end of the mRNA or could be combined with a synthetic sequence comprising the sequence:

5'-UNUAAAUGUR UAUU-3' [SEQ ID NO:13].

The hybrid mRNAs may further comprise a third sequence which codes for a chemical or enzymatic cleavage site so that the protein or polypeptide of interest may be cleaved from any unnecessary or interfering amino acid sequences. For instance, the stabilization sequence may contain short coding sequences. It may be desirable or necessary to cleave the protein or polypeptide of interest from any such coding sequences.

In Examples 1 and 3, vectors are described which contain two small portions of the coding region of the vitellogenin Bl gene. With respect to these vectors, two sequences coding for cleavage sites may be necessary: a third sequence coding for a cleavage site may be included between the first portion of the vitellogenin Bl mRNA sequence and the sequence coding for the protein or polypeptide of interest; and a fourth sequence coding for a cleavage site may be included between the sequence coding for the protein or polypeptide of interest and the second portion of the vitellogenin Bl mRNA sequence.

The cleavage site may be a site for proteolytic cleavage. Many such sites and RNA and DNA sequences coding for them are well-known. Alternatively, where the selected protein or polypeptide of interest does not contain any methionine residues, the cleavage site may be methionine (encoded for by an ATG codon in DNA) . The protein or polypeptide of interest may then be cleaved from the amino acids coded for by the second RNA sequence at the methionine residue by treatment with cyanogen bromide. Gross, Methods in Enzvmology. 11. 238-55 (1967) .

The hybrid mRNAs containing the mRNA stabilization sequences are produced by the transcription of a recombinant DNA sequence that codes for the hybrid mRNA, and the invention also comprises this recombinant DNA sequence and a vector in which the recombinant DNA sequence is operatively linked to expression control sequences. The recombinant DNA sequence can be prepared and incorporated into the vector using conventional genetic engineering techniques well known to those skilled in the art.

First, DNA sequences coding for the mRNA stabilization sequence and for the protein or polypeptide of interest are prepared. This may be accomplished in a variety of ways. For instance, a cDNA or a genomic DNA library can be constructed and screened for the DNA sequence of interest using appropriate hybridization probes. Of course, many genes and DNA sequences useful in the practice of the invention have already been isolated and cloned and are readily available. For example, the gene coding for vitellogenin has been cloned (see below) , as have those coding for many proteins and polypeptides of interest such as the interferons and human insulin. If only portions of genes are desired, fragments may be prepared from genes or larger fragments using appropriate restriction enzymes. Further, many desired DNA sequences may be prepared by chemical synthesis if the DNA or amino acid sequence is known. DNA sequences may be altered by techniques such as site directed mutagenesis. Other conventional manipulations may be employed in the preparation of desired DNA sequences. The recombinant DNA sequence of the invention coding for an mRNA containing a stabilization sequence is prepared by linking the DNA sequence coding for the mRNA stabilization sequence to the DNA sequence coding for the protein or polypeptide of interest. The DNA sequence coding for the cleavage site, if used, is built into the protein or polypeptide produced by the hybrid mRNA by constructing the recombinant DNA sequence so that it has one or more codons that code for the desired cleavage site located between the DNA sequence encoding the mRNA stabilization sequence and the DNA sequence encoding the protein or polypeptide of interest. All of the DNA sequences are operatively linked so that the hybrid mRNA produced by transcribing the recombinant DNA sequence is stabilized by estrogen and estrogen receptor and is translated so as to produce a protein or polypeptide of the proper amino acid sequence.

The invention also includes a vector capable of expressing the protein or polypeptide of interest in an appropriate host. The vector comprises the recombinant DNA molecule linked to appropriate expression control sequences. Methods of effecting this operative linking, either before or after the recombinant DNA sequence is inserted into the vector, are well known.

Expression control sequences include promoters, activators, enhancers, ribosomal binding sites, start signals, stop signals, cap signals, polyadenylation signals, and other signals involved with the control of transcription or translation. Expression control sequences suitable for use in the invention are well known. They include sequences known to control the expression of genes of eukaryotic cells, their viruses, or combinations thereof.

The vector must contain a promoter and a transcription termination signal, both operatively linked to the recombinant DNA sequence. The promoter may be any DNA sequence that shows transcriptional activity in the host cell and may be derived from genes encoding homologous or heterologous proteins (preferably homo¬ logous) and either extracellular or intracellular pro¬ teins, such as thymidine kinases, vitellogenins, amylases, glycoamylases, proteases, lipases, cellulases and glycolytic enzymes.

The promoter may include upstream activator or enhancer sequences. In particular, estrogen response elements may be used. In this manner, the estrogen and estrogen receptor that are used to stabilize the hybrid mRNA would also induce the transcription of the hybrid mRNA. Such estrogen-regulatable promoters are preferred promoters, and particularly preferred are estrogen- regulatable promoters comprising two consensus estrogen response elements operatively linked to a TATA box which are very powerful promoters. Suitable and preferred estrogen-regulatable promoters are further discussed in section C below.

The vector should also have a translation start signal immediately preceding the DNA sequence coding for the protein or polypeptide of interest. There should be no stop signal between the start signal and the end of the coding sequence.

The vector may also contain one or more origins of replication which allow it to replicate in the host cells. The vector should further include one or more restriction enzyme sites for inserting the recombinant DNA and other DNA sequences into the vector, and a DNA sequence coding for a selectable or identifiable phenotypic trait which is manifested when the vector is present in the host cell ("a selection marker") .

Suitable vectors for use in the invention are well known. They include vectors suitable for use in eukaryotic cells such as viral vectors (retroviral vec- tors, vaccinia vectors), other vectors described in the Examples below, and derivatives of these vectors.

In a preferred embodiment, a DNA sequence encoding a signal or signal-leader sequence, or a func¬ tional fragment thereof, is included in the recombinant DNA vector between the translation start signal and the portion of the recombinant DNA sequence coding for the protein or polypeptide of interest. A signal or signal- leader sequence is a sequence of amino acids at the amino terminus of a polypeptide or protein which provides for secretion of the protein or polypeptide from the cell in which it is produced. Many such signal and signal-leader sequences are known.

By including a DNA sequence encoding a signal or signal-leader amino acid sequence in the vectors of the invention, the protein or polypeptide encoded by the recombinant DNA sequence may be secreted from the cell in which it is produced. Preferably, the signal or signal- leader amino acid sequence is cleaved from the fusion protein during its secretion from the cell. If not, the protein or polypeptide should preferably be cleaved from the signal or signal-leader amino acid sequence after isolation of the fusion protein.

Signal or signal-leader sequences suitable for use in the invention include: signal sequences which are normally part of precursors of proteins or polypeptides such as the precursors of vitellogenin, parathyroid hormone, and interferon (see U.S. patent No. 4,775,622); synthetic signal-leader sequences; Saccharomyces cerevisiae alpha factor (see U.S. patents Nos. 4,546,082 and 4,870,008); fragments of S. cerevisiae alpha factor; and the yeast BAR1 secretion system (see U.S. patent No. 4,613,572) .

Particularly suitable signal sequences are mutant parathyroid hormone signal sequences such as those described in Cioffi et al., J. Biol. Chem.. 264. 15052- 15058 (1989) . The following is a particularly preferred mutant parathyroid hormone signal sequence:

Met Met Ser Ala Lys Asp Met Val Lys Val Met lie Val

5 10

Met Leu Ala lie Leu Ala Cys Ala Arg Ser Asp Gly 15 20 25

[SEQ ID NO:14]

This sequence differs from the native sequence in that the cysteine residue at position 18 and the leucine residue at position 20 in the native sequence have been switched. Another preferred mutant parathyroid hormone signal sequence has the cysteine residue at position 18 replaced by a leucine residue.

The resulting vector having the recombinant DNA sequence of the invention thereon is used to transform an appropriate host. This transformation may be performed using methods well known in the art.

The selection of a particular host is dependent upon a number of factors recognized by the art. These include, for example, compatibility with the chosen expression vector, toxicity to it of the protein or polypeptide of interest encoded for by the recombinant DNA sequence, rate of transformation, ease of recovery of the protein or polypeptide of interest, expression characteristics, ability to grow in serum-free medium, biosafety and costs. A balance of these factors must be struck with the understanding that not all hosts may be equally effective for the expression of a particular recombinant DNA sequence.

Useful hosts for practicing the present in¬ vention are eukaryotic cells. It is not believed that prokaryotes can serve as hosts in the present invention. Preferred host cells are amphibian and mammalian (includ- ing human) liver cells, particularly X. laevis liver cells. Most preferred are the XL110 cell line described below and other similar cell lines which can be prepared from X. laevis liver by the method described below for preparation of the XL110 cell line.

If the host cell does not normally produce estrogen receptor, it must be transformed with a vector which codes for estrogen receptor. Genes coding for estrogen receptors have been cloned. See Weiler et al., Molec. Endo. 1, 355-362 (1987) . Vectors suitable for transforming host cells so that they produce estrogen receptor may be prepared by methods well known in the art (see discussion above) . A particularly preferred estrogen receptor is the mutant estrogen receptor described in the next section.

Next, the transformed host cell is cultured in the presence of estrogen so that the hybrid mRNA is transcribed and stabilized. In this manner, production of the desired protein or polypeptide is enhanced.

In culturing the host cells, normal culture medium and conditions may be used. However, it is desirable to use serum-free culture medium or serum which has been treated to remove endogenous estrogen, since long-term exposure to the estrogen-estrogen receptor complex is toxic to cultured cells. Alternatively, a mutant estrogen receptor having a single amino acid change may be used. See Tora et al., "The Cloned Human Estrogen Receptor Contains A Mutation Which Alters Its Hormone Binding Properties," The EMBO Journal. ___, 1981 (1989) . This mutation reduces the affinity of the estrogen receptor for estrogen by approximately 10 fold and makes the receptor much less sensitive to activation by trace levels of estrogen in serum.

The type of estrogen used in the culture is not critical. Estradiol-17/? is preferred. The amount of estrogen to be added to the cultures is known for many types of cells. If not, it can be determined by a standard dose response experiment.

Finally, after fermentation, the protein or polypeptide is purified. Methods of purifying proteins and polypeptides from cultures are well-known in the art.

B. Mutant Estrogen Receptor

As explained in the Background section, estrogen regulates the transcription of target genes through the interaction of the estrogen-estrogen receptor complex with estrogen response elements on the target gene. A mutant estrogen receptor having the ability to elicit enhanced transcription of mRNAs under the transcriptional control of an estrogen-regulatable promoter has been developed. Increased production of the protein and polypeptides coded for by the mRNAs is obtained as a result of the enhanced transcription caused by the mutant estrogen receptor.

The mutant estrogen receptor of the invention contains one or more copies of an amino acid sequence coding for a peptide that, should it form an alpha helix, would form an amphipathic helix having one acidic face, one hydrophobic face, and a third face which is either an uncharged polar face or another acidic face. This amino acid sequence will be referred to herein as the "amphipathic helix peptide". The amphipathic helix peptide may, or may not, actually form an alpha helix by itself or when part of an estrogen receptor but, when the peptide is modeled as an alpha helix, it will have the characteristics described above.

Suitable amphipathic helix peptides are those found in the activating regions of some DNA-binding regulatory proteins as described in Giniger and Ptashne, Nature, 330. 670 (1987) . Also, and preferably, synthetic sequences are designed which code for amphipathic helix peptides having the above characteristics. See id. Amphipathic helices and methods of designing them are known in the art. Particularly preferred is the following amino acid sequence:

Glu Leu Gin Glu Leu Gin Glu Leu Gin Ala Leu Leu Gin

5 10

Gin Gin 15

[SEQ ID NO:15]

which, should it form an alpha-helix, would form an amphipathic helix having the characteristics described above (see Figure 9) .

The mutant estrogen receptors of the invention are prepared using genetic engineering techniques. These techniques are conventional and well known.

First, a DNA sequence coding for an estrogen receptor is isolated or synthesized as described in the previous section. Any estrogen receptor may be used. Many estrogen receptors have been cloned, and the sequences of many estrogen receptors are known and can be found in, e.g.. the GenBank database. See also Weiler et al., Molec. Endo.. 1, 355 (1987).

The sequences of estrogen receptors exhibit a substantial degree of homology. See i≤. Thus, mutant estrogen receptors having the ability to enhance transcription from an estrogen-regulatable promoter can be prepared starting with the DNA sequence of any estrogen receptor.

Next, the sequence coding for the amphipathic helix peptide is isolated or synthesized using the techniques described in the previous section. Then, the DNA sequence coding for the estrogen receptor is cut with a restriction enzyme, and one or more copies, preferably one or two copies, of the sequence coding for the amphipathic helix peptide are inserted into the restriction site. As needed, linkers may be added to one or both of the sequences. In this manner, a recombinant DNA sequence coding for the mutant estrogen receptor of the invention is prepared, and this recombinant DNA sequence is also part of the invention.

Preferably only one or two copies of the sequence coding for the amphipathic helix are used. Evidence indicates that the estrogen receptor can bind to DNA even when it is not complexed to estrogen, and it is believed that mutant estrogen receptors containing more than two copies of the amphipathic helix peptide may non- specifically activate transcription. If the protein or polypeptide of interest is toxic to cells, then this nonspecific activation could cause cell death.

Estrogen receptors share a common domain structure (see Figure 10; see also Weiler et al., Molec. Endo., 1 , 355 (1987)). The point in the sequence coding for the estrogen receptor at which the sequence coding for the amphiphatic helix peptide is inserted is not critical, except that it should not be inserted into the sequences coding for the domains responsible for DNA binding and ligand binding (domains C and E—see Figure

10) . The hinge region (domain D see Figure 10) is a particularly preferred region for the insertion of the amphipathic helix peptides since it is located near other activator sequences.

A vector capable of expressing the mutant estrogen receptor is then prepared by operatively linking the recombinant DNA sequence coding for the mutant estrogen receptor to expression control sequences. Suitable methods, vectors and expression control sequences are described in the previous section. However, estrogen-regulatable promoters and signal sequences should not be used.

This vector is then used to transform a host cell which is cultured under conditions allowing for expression of the mutant estrogen receptor. Suitable transformation methods, host cells and culture conditions were described in the previous section. When estrogen is added to the culture medium, the mutant estrogen receptor binds to the estrogen and this complex then interacts with estrogen response elements on estrogen-regulatable promoters within the host cell. As a result, enhanced transcription of mRNAs under the control of the estrogen- regulatable promoters and, therefore, increased production of proteins and polypeptides coded for by the mRNAs, are obtained.

C. Estrogen-Regulatable Vector Comprising A Promoter Having Two Estrogen Response Elements Linked To A TATA Box

The invention also comprises an estrogen- regulatable expression vector comprising a synthetic (i.e.. not a wild type) promoter operatively linked to a DNA sequence coding for a hybrid mRNA. The promoter comprises one or more estrogen response elements, the type and number of estrogen response elements being sufficient so that transcription of the hybrid mRNA is estrogen-regulatable. Suitable synthetic promoters can be prepared using conventional genetic engineering techniques or can be chemically synthesized. Suitable types of estrogen response elements, numbers of estrogen response elements, spacings between estrogen response elements, relationships between estrogen response elements and other promoter elements and activators are described in, e.g.. Klein-Hitpass et al.. Cell. 46. 1053- 61 (1986); Martinez et al.. The EMBO J.. £, 3719-27 (1987) ; Martinez and Wahli, The EMBO J.. £, 3781-91 (1989); Ponglikitmongkol et al., The EMBO J.. £, 2221-31 (1990); and Chang et al., Molec. Endo. (in press).

The synthetic promoter used in the vectors of the present invention preferably comprises two consensus estrogen response elements operatively linked to a TATA box. Eukaryotic promoters generally contain a short consensus sequence, TATA(A/T)A(A/T) , about 25-30 bp upstream from the transcription start point. This sequence is termed the TATA box. The TATA box is required for accurate initiation of transcription.

The consensus sequence of the estrogen response element has been identified as

5'-AGGTCANNNT GACCT-3'

[SEQ ID NO:16] . Beato, Cell. 56. 336-44 (1989) . The sequence

5'-GGTCANNNTG ACC-3'

[SEQ ID NO:17] functions as well as the full-length consensus sequence, and this shorter consensus sequence is the preferred consensus sequence.

A particularly preferred promoter has the two copies of the consensus estrogen response elements spaced so that their centers are about 20 nucleotides apart. The identity of the spacer nucleotides used to achieve this separation is not critical.

Also, the two copies of the consensus estrogen response elements are most preferably spaced about 20 nucleotides upstream from the TATA box . This spacing is chosen since it represents about two turns of an alpha helix. Promoters in which the estrogen response elements are more than 20 nucleotides from the TATA box will also be functional, although the activity of the promoter will decrease as the spacing increases. The activity of a promoter where the spacing is about 500 nucleotides is expected to be about 5 fold less than that of a promoter in which the spacing is 20 nucleotides. It is believed that promoters in which the spacing is less than 20 nucleotides may not be functional due to steric hindrance which may prevent the binding of proteins to the estrogen response elements and TATA box. When short spacings (about 20 to 150 nucleotides) are used, the identity of the spacer nucleotides is not critical. When larger spacings (over about 150 nucleotides) are used, the intervening sequences will influence the activity of the promoter.

Thus, in the most preferred promoter, the centers of the estrogen response elements are about 20 nucleotides apart, and the estrogen response elements are spaced about 20 nucleotides upstream of the TATA box. This promoter exhibits negligible activity in the absence of estrogen and is a very powerful promoter in the presence of the estrogen-estrogen receptor complex.

Other suitable synthetic promoters are desribed in Chang et al., Molec. Endo. (in press) which was referred to above. This article reports the results of investigations into the role of the two imperfect estrogen response elements (ERE1 and ERE2) located at -302 to -334 in the 5' -flanking region of the estrogen- regulated Xenopus laevis vitellogenin Bl gene. Deletion of either ERE1 or ERE2 effectively abolished estrogen- dependent transcription. Surprisingly, neither replacement of the two imperfect estrogen response elements with a single consensus estrogen response element at -334, nor insertion of 1 or 2 consensus estrogen response elements at -359, restored full estrogen responsiveness to the mutant promoter. Only when 4 consensus estrogen response elements were present at the -359 site did estrogen responsiveness approach that of the wild type promoter. Estrogen response elements at the -359 site can exert biological activity, since an additional consensus estrogen response element at -359 greatly enhanced estrogen responsiveness of the promoter when the imperfect estrogen response elements at -302/-334 were present. The addition of a second consensus estrogen response element at the -359 site did not further increase promoter activity. Further, a consensus estrogen response element at -359 acted synergistically with ERE1 to restore activity when ERE2 was deleted. The authors concluded that their data demonstrated that the DNA context in which the estrogen response elements are embedded exerts important effects on the efficiency with which the estrogen response elements function as hormone-dependent transcription activators, but the precise nature of the DNA context which potentiates the activity of the imperfect Bl estrogen response elements was not eluciated.

The Chang et al. article further reports that deletion of all of the sequence between the imperfect EREl and ERE2 and the TATA box significantly increased the estrogen inducibility of the Bl promoter. The mutant promoter in which an additional estrogen response element was inserted at -359 exhibited increased estrogen inducibility relative to a construct which contained this same deletion but lacked the additional estrogen response element. Moving the set of three estrogen response elements closer to the TATA box did not decrease estrogen inducibility and greatly increased promoter activity.

Finally, the Chang et al. article reports the results obtained using a simplified promoter containing only estrogen response elements and a TATA box. It was found that a single consensus estrogen response element inserted at -74 relative to the TATA box was sufficient to confer estrogen-dependent expression on this promoter. When this estrogen response element was moved to a site at -182, estrogen inducibility was lost, and the estrogen response element became a constitutive activator. When two consensus estrogen response elements were present at the -74 site, they acted synergistically to produce a powerful promoter which exhibited substantially higher expression than a construct which contained the two imperfect Bl estrogen response elements (EREl and ERE2) close to the TATA box. The highest overall activity was exhibited by a construct in which a single consensus estrogen response element was able to interact with the two imperfect estrogen response elements close to the TATA box.

The hybrid mRNA coded for by the vector comprises a 5' untranslated region, a coding region coding for a protein or polypeptide of interest, and a 3' untranslated region. The 3' untranslated region preferably contains a sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor. Suitable stabilization sequences for use in the 3' untranslated region are those discussed in section A above.

The 5' untranslated region may be any 5' untranslated region, but is preferably one that provides for efficient translation. The 5' untranslated regions that provide for efficient translation are the 5' untranslated regions of mRNAs that code for proteins that are normally produced at high levels. For instance, suitable 5' untranslated regions are those of the vitellogenin mRNA, the Herpes thymidine kinase mRNA, and the Tobacco Mosaic Virus mRNA. Particularly preferred is a 5' untranslated region containing a combination of the 5' untranslated regions of X. laevis estrogen receptor mRNA and Herpes thymidine kinase mRNA (see Example 6) .

The proteins and polypeptides coded for by the coding region may be any protein or polypeptide such as those described above. The coding region may also comprise a DNA sequence coding for a signal peptide to provide for secretion of the protein or polypeptide. Suitable signal sequences are discussed above in section A.

The vector may be prepared using genetic engineering techniques. These techniques are conventional and well known and are described above. Then, the estrogen-regulatable vector of the invention.is used to transform a host cell. Suitable hosts and transformation methods are described above.

When the transformed host cell is cultured in the presence of estrogen (as described above) , estrogen- regulatable transcription from the promoter of the vector is obtained. When the 3' untranslated region contains a stabilization sequence, it provides for stabilization of the hybrid mRNA coded for by the vector in the presence of estrogen and estrogen receptor. In this manner, inducible, enhanced production of a protein or polypeptide is obtained using the vectors of the invention. A further improvement in protein/polypeptide production can be obtained by using the very powerful preferred promoters of the invention comprising two estrogen response elements operatively linked to a TATA box. Yet a further improvement in production can be obtained by co-transforming the host cell with a vector that codes for the mutant estrogen receptor of the invention.

Thus, the invention provides an integrated estrogen-regulatable expression system that allows for production of large amounts of proteins and polypeptides of interest. The invention is particularly well adapted for the production of proteins and polypeptides that are normally toxic to cells.

EXAMPLES In the Examples which follow, standard tech¬ niques for working with nucleic acids, such as restric¬ tion mapping, agarose and acrylamide gel electrophoresis, isolation of DNA fragments from gels, filling in and blunt-ending protruding ends of restriction fragments, ligation of DNA fragments, preparing plasmids for trans¬ formation by calf intestinal alkaline phosphatase diges¬ tion, growth and transformation of _____ coli. preparing plasmid minipreps and large-scale plasmid preparations (by the alkaline lysis method) , etc. were carried out as described in Maniatis et al., Molecular Cloning (Cold Spring Harbor Laboratory, New York 1982) . DNA sequencing of double-stranded plasmid DNA was done as described in Chen and Seeberg, PJJA, 1165-170 (1985).

Unless otherwise noted, the enzymes used in the following Examples were obtained from Bethesda Research Laboratories. They were used according to the manufac¬ turer's instructions.

EXAMPLE 1: Construction of plasmid pMV5 A hybrid mini-vitellogenin gene (MV5) derived from the X. laevis vitellogenin Bl gene was constructed. Normally, vitellogenin genes encode mRNAs of over 5,600 nucleotides [Gerber-Huber et al., Nucleic Acids Res.. 15. 4737-4760 (1987)], but the MV5 gene codes for an mRNA lacking 5,075 nucleotides of the internal vitellogenin- coding region. The MV5 mRNA transcript contains the following portions of the vitellogenin Bl mRNA: 1) the entire 5' untranslated region (13 nucleotides); 2) 81 nucleotides co.ding for the 27 amino-terminal amino acids, including the signal peptide; 3) 270 nucleotides coding for the carboxy-terminal 90 amino acids; and 4) the entire 3' untranslated region (165 nucleotides) (see Figure IA) . The processing and polyadenylation signals from the vitellogenin Bl gene are positioned in MV5 at the 3' end so that the MV5 transcripts will be properly polyadenylated in Xenopus cells.

The MV5 gene was constructed so that it coded for regions of the vitellogenin mRNA that have been reported to be important in controlling the stability of other mRNAs (see Background) . In particular, the 5' and 3' untranslated regions, the signal sequence, short 3' and 5' coding regions, and genomic sequences specifying processing and polyadenylation of vitellogenin mRNA were included. Of course, it could not be predicted in advance whether the resultant mRNA would be stabilized since the mechanism of mRNA stabilization by hormones is largely unknown.

Plasmid pMV5 was prepared for expression of the MV5 gene (see Figures 1B-D) . In this plasmid, expression of the MV5 gene is driven by the herpes simplex virus (HSV) thymidine kinase (TK) promoter, a promoter that is efficiently expressed in Xenopus cells and is not regulated by estrogen. McKnight et al., Cell. 25. 385- 398 (1981); see below. The 5' untranslated and coding regions of MV5 were constructed by cloning four overlap¬ ping synthetic oligonucleotides after the TK promoter so as to direct the transcription of an mRNA with an authentic vitellogenin Bl 5' end. The 3' coding and untranslated regions are from a vitellogenin Bl cDNA clone and a genomic vitellogenin Bl clone. To maintain translation of the MV5 transcript, the 3' coding region was fused in frame with the 5' coding region. This construct was inserted into the pTZ19R vector. Finally, the simian virus 40 poly(A) region was placed upstream of the MV5 gene to prevent readthrough transcription into the MV5 gene. The details of the construction of the pMV5 mini-vitellogenin plasmid are as follows.

First, the 1235 base pair (bp) Hind III to Hind III fragment containing the-vitellogenin Bl 3' end was excised from the genomic clone Xlvl4.3 (see Figure IB) . Clone XIvl4.3 was kindly provided by M. Hayward and D. Lew of the University of Illinois, Champaign-Urbana, Illinois, and was prepared as described in D. Lew, 1988 Ph.D. Thesis, University of Illinois, Champaign-Urbana, Illinois. This fragment, was made blunt ended by filling in the sticky ends with Klenow fragment, and was then inserted into the Hpa I site of pRSV-CAT to yield pRCVll (see Figure IB) . Plasmid pRSV-CAT is available from Bethesda Research Laboratories and Pharmacia. Next, a 2728 bp Nco I to Pst I fragment from pRCVll was ligated to a 758 bp Bam HI to Nco I fragment from pTKICO and a 2800 bp Pst I to Bam HI fragment from pTZ19R to form pTCVl (see Figure IB) . Plasmid pTKICO, was a gift of Keith Yamamoto, University of California, San Francisco, Ca. Its preparation is described in DeFranco and Yamamoto, Molec. Cell. Biol.. £, 993-1001 (1986) (referred to as OTCO in the article) . Plasmid pTZ19R was prepared as described in Mead et al., Protein Eng.. 1, 67-74 (1986) .

Plasmid pTCVl was cleaved with Pst I in the SV40 poly(A) region, the sticky ends removed with mung bean nuclease (obtained from Stratagene) and religated to yield pTCV2 (see Figure IC) . Then, a 203 bp Pst I to Bgl II fragment containing 3' vitellogenin Bl cDNA from pXlvc56 was inserted into the Pst I to Bgl II sites of pTCV2 to yield pMVl (see Figure IC) . Clone pXIvc56 was kindly provided by M. Hayward and D. Lew, University of Illinois, Champaign-Urbana, Illinois and was prepared as described in D. Lew, 1988 Ph.D. Thesis, University of Illinois, Champaign-Urbana, Illinois.

Next, a 137 bp fragment containing the SV40 early termination signal was excised from pRCVlO with Bglll and BamHI, and the fragment was then inserted into the Bam HI site of pMVl to form pMV2 (see Figure IC) . Plasmid pRCVlO was derived by the insertion of Bgl II linkers (obtained from New England Biolabs Inc.) into the Hpa I site of pRSV-CAT (see Figure IB) .

The following four overlapping oligonucleotides were synthesized on an Applied Biosystems Model 380B DNA synthesizer using J-cyanoethylphosphoramide technology:

[SEQ ID NO:18] 5'AGGCCTCGAA TTCGCCATCA CCATGAGGGG AATCATACTA GCTCAGCTTC TCGCTCTAG 3'

5'AAGCTGAGCT AGTATGATTC CCCTCATGGT GATGGCGAAT TCGAGGCCTT GCA 3'

[SEQ ID NO:19] [SEQ ID NO:20] 'CGGGAAGTGA AAAGTCACAA TATGAACCTT TTTTCAGTGA GAGCCTGCA 3'

5'GGCTCTCACT GAAAAAAGGT TCATATTGTG ACTTTTCACT TCCCGCTAGA GCGAG 3'

[SEQ ID NO:21]

The oligonucleotides were then hybridized to yield the following two overlapping fragments:

[SEQ ID NO:18]

5 ' AGGCCTCGAATTCGCCΛT<__ACCΑTGAGGGGAATCATACTAGCT 3 ' ' ACGTTCCGGAGCTTAAGCGGTAGTGGTACTCCCCΓΓAGTATGATCGAGTCGAA 5 '

[SEQ ID NO:19]

[SEQ ID NO:20]

5' CGGGAAGTGAATΛGTCIAC_AATATGAACCTTTTΓTCAGTGAGAGCC_>GCA 3' GAGCGAGATCGCCCTTC_ACTTTTC_AGTGTTATACTTGGAAAAAAGTCACTCTCGG 5'

[SEQ ID NO:21]

These were inserted into the Pst I site of pMV2 to create pMV3 (see Figure ID) . Then, the 27 bp fragment of pMV3 from between the Hae III site 6 bp upstream of the tk mRNA start site and the Stu I site located in the synthetic oligonucleotide was deleted to yield pMV4 (see Figure ID) . The sequence of pMV4 was verified by dideoxy sequencing of all the junctions and the synthetic sequence. Finally, 583 bp of pMV4 from between the Stu I site 3' of the vitellogenin gene in the genomic region and the Hind III site were deleted to yield pMV5 (see Figure ID) . The nucleotide sequence of the relevant portion of pMV5 is shown in Figure 6A [SEQ ID NO:l] .

EXAMPLE 2 : Requirement for estrogen and estrogen receptor for MV5 mRNA St frilizfr ipn j ra^feς^ed ce g

A. C$1 J- 3-ine

XLllO cells were transfected with pMV5 and other plasmids as described below. XLllO cells are a clonal line of partially de-differentiated X. laevis liver cells.

This cell line was developed as follows. The liver of an adult male Xenopus laevis. which had never been exposed to estradiol, was perfused with an ice-cold solution of Barth X (Barth and Barth, J. Embyol. and Experimental Morphology. 2_. 210-222 (1959)) without calcium, and then with a collagenase solution (Wangh et al., Develop. Biol.. 70. 479-489 (1979)). The perfused liver was excised, minced and then digested again with collagenase (Folger, et al., J. Biol. Chem.. 258. 8908- 8914 (1983)). The dissociated cells from one liver were plated into one flask and maintained in medium containing fructose instead of glucose and 20% dialyzed fetal bovine serum (fibroblasts lack the enzymes to convert fructose to glucose) . One of the two clones which grew out was recloned and designated XLllO (also sometimes referred to as A110I) cells.

Both the selection and standard growth of XLllO cells is on 0.6 X Higuchi's medium (Higuchi, J. Cell Physiol.. 23., 65-72 (1969)), containing 20 mM HEPES (N-2- hydroxyethylpiperazine-N'-2-ethanesulphonic acid), pH 7.4. During cell isolation the cells were maintained in 0.6 X Higuchi's medium supplemented with 20% dialyzed fetal bovine serum. The cells were isolated by growth at 20°C in sealed flasks in air. In normal growth of the cells, they are maintained in 0.6 X Higuchi's medium supplemented with either 10%. fetal bovine serum or 10% dextran-charcoal treated fetal bovine serum (Eckert and Katzenellenbogen, J. Biol. Chem.. 257. 8840-8846 (1982)). The cells are harvested by standard trypsin digestion using trypsin in 0.6 X PBS (lOmM sodium phosphate, 150 mM NaCl, pH 7.4) according to the supplier's (GIBCO) directions. Standard plating of the cells is approxima¬ tely 1:10.

Charcoal-dextran treatment of serum to remove endogenous estrogen was performed as follows. A 500 ml suspension was prepared containing 50 grams Norit A charcoal, 5 grams dextran (Sigma, No. D-4751) , 4.4 grams NaCl and remainder water. When ready to treat serum, the charcoal was resuspended thoroughly, and a 25 ml aliquot removed. The charcoal in the aliquot, was pelleted at 3000 rpm for 5 minutes at room temperature. The super¬ natant was poured off, and the volume brought to 25 ml by the addition of 0.15 M NaCl. The resuspended charcoal solution was added to 500 ml of fetal bovine serum, and the serum was inactivated at 55°C for one hour, with swirling of the bottle every 15 minutes. Next, the serum was aliquoted into sterile 50 ml tubes, and the charcoal was pelleted by centrifugation at 3000 rpm for 20 minutes at room temperature. The supernatants were harvested, and the centrifugation step repeated using fresh tubes. Finally, the serum supernatants were filtered through a 0.2 micron filter and were stored at -20°C until needed. The entire procedure should be carried out using sterile solutions and plasticware.

B. Transfection and induction with estrogen

The XLllO cells were grown at 20°C in 0.6x phenol-red-free Higuchi medium containing 10% charcoal- dextran-treated fetal bovine serum. The cells were grown in air in sealed flasks or in Petri dishes sealed with parafilm.

One day prior to transfection, one million cells were seeded into a 100 mm culture disk containing 9 ml of medium. At 2-4 hours before transfection, the medium was changed.

DNA was transfected into the XLllO cells by a modification of the calcium phosphate coprecipitation method described in Parker and Stark, J. Virol.. 31. 360- 369 (1979) . To prepare calcium phosphate coprecipitates, approximately 30 ug of the DNA of interest (in TE at 1-3 ug of DNA/ul) was added to 0.5 ml of 2X HBS in a 15 ml conical tube. TE is a solution containing 10 mM Tris and 1 M EDTA, pH 7.6, and 2X HBS is a solution containing 280 mM NaCl, 50 mM HEPES, pH 7.09, and 1.5 mM sodium phosphate. The DNA solution was mixed on a vortex mixer at moderate speed, and an equal volume of 0.2 M CaCl₂ was added dropwise to the tube. After 5-10 minutes to allow the crystals to form, the DNA-CaP0₄ co-crystals were added to the dish of cells with gentle mixing. The cells were then incubated for 24 hours.

A sterile 15% solution of glycerol in cell culture medium was prepared. The cells were glycerol shocked by adding 2 ml of this glycerol solution to each plate, and incubating the cells for 3 minutes. The medium was aspirated, and the plates rinsed 2 times with 0.6 X PBS to remove all glycerol.

After glycerol shock, the cells were removed from the plates with trypsin, pooled, replated, and maintained for 72 hours in medium containing either 100 nM estradiol-17J and 0.1% ethanol or in medium containing 0.1% ethanol alone. The medium was changed at 24-hour intervals. The purpose of pooling and replating the cells was to ensure accurate comparisons between trans¬ fected cells maintained in the presence and absence of estrogen.

C. Plasmids In addition to pMV5, XLllO cells were sometimes transfected (as described above) with other plasmids as specified below. These other plasmids are described below.

1. Plasmid pXER

Plasmid pXER is a Xenopus estrogen receptor [XER] expression plasmid in which synthesis of XER mRNA is driven from the herpes virus thymidine kinase (TK) promoter. Plasmid pXER (also sometimes referred to as pTKXER or pTKXERO) was prepared as shown in Figure 4.

The XER gene was cloned as described in Weiler et al., Molec. Endo.. i, 355-62 (1987). Briefly, a XER cDNA clone was isolated by screening a lambda gtlO cDNA library prepared from poly(A) mRNA from the livers of male Xenopus laevis which had received multiple injections of estradiol-17S (which induces the XER) . The isolated done was designated XXER4. The insert from λXER4 was removed by digestion with EcoRl, and the isolated insert was subcloned into the EcoRl site of the pGEM3 vector (Promega Biotech) . The resulting clone was designated pXER4.4.

Next plasmid pTKXER2 was prepared as described in D. Lew, 1988 Ph.D. Thesis, University of Illinois, Champaign-Urbana, Illinois. First, the CAT gene was removed from pTKICO. As noted above, the vector pTKICO was a gift from Keith Yamamoto. Vector pTKICO contains the herpes virus TK promoter and the chloramphenicol acetyl transferase (CAT) structural gene. A similar vector (pTK-luciferase) is available from the American Type Culture Collection with the luciferase structural gene in place of the CAT gene (Nordeen, Biotechniques. 6., 454-458 (1988)). Since the CAT gene was removed in the construction of pXER, the identity of the segment to be deleted is not relevant.

To remove the CAT structural gene from pTKICO, the vector was digested with Bglll and Hpal. The pro¬ truding ends on the restriction sites were removed by treatment with mung bean nuclease (Stratagene) as described by the manufacturer.

Plasmid pXER4.4 was digested with Eco Rl, and the 2,418 nucleotide Eco Rl fragment, which contains the entire protein coding region of the XER gene, was isolated by gel electrophoresis. The protruding ends on the isolated restriction fragment were filled in by treatment with the Klenow fragment of DNA polymerase.

The XER fragment from pXER4.4 was inserted at the Bgl II/Hpa I junction in the blunt ended TK1CO plasmid by blunt end ligation. The resulting vector, designated pTKXER2, was characterized by restriction mapping and DNA sequencing.

The remaining steps of the preparation of pXER are described in T.C. Chang, 1990 Ph.D. Thesis, University of Illinois, Champaign-Urbana, Illinois. First, plasmid pTKXER2 was digested with Cla I, and the 3,419 base-pair Cla I fragment was isolated by gel electrophoresis. This fragment contains the entire TK promoter, the TK 5' -untranslated region, the protein coding region of the XER gene and the SV40 polyadenyla¬ tion signal. The vector pTZ18U (Pharmacia) was digested with Ace I. The Cla I fragment from pTKXER2 was inserted into the Ace I site of pTZ18U by ligation to produce pXER which was characterized by restriction digestion and DNA sequencing.

2. Plagmjg T£7

Plasmid pTK7 synthesizes thymidine kinase (TK) mRNA, and this mRNA is used as an internal standard in some experiments. When pTK7 was used, a TK cRNA probe was also used simultaneously with the MV5 cRNA probe in the Si nuclease protection assay (see below and Figure 3) .

Plasmid pTK7 was prepared by cloning the 2061 bp Bam Hi/Hind III fragment of pJC331 containing the HSV TK gene into the Bam Hi/Hind III sites of pTZ19U (Mead et al., Protein Eng.. 1, 67-74 (1986)). Plasmid pJC331 was obtained from Bernard Roizman, The University of Chicago, Chicago, Illinois. It was prepared by cloning 2040 bp Pvu II/Pvu II fragment containing the HSV TK gene into the Hinc II site of pUC9 (Vieira and Messing, Gene. 19. 259 (1982) ) . The original TK gene description may be found in Post et al., Cell. 24. 555-565 (1981).

3. Plasmid pTKCAT

Plasmid pTKCAT is a TK promoter-chlor- amphenicol acetyltransferase gene fusion plasmid that was added in some experiments as carrier DNA to maintain a constant TK promoter concentration. This plasmid is the same as pTKICO described above.

4. TK-luciferase vector

This vector was described above. As noted above, it is available from the American Type Culture Collection, and its preparation is described in Nordeen, Biotechniques. 6., 454-458 (1988) .

5. Plasmid pTK-β-gal. Plasmid pTK-S-gal is a TK promoter- β- galactosidase gene fusion plasmid. It was constructed by inserting the TK promoter in place of the SV40 promoter in pCHHO. This was accomplished by digesting pCHHO with Kpnl and PvuII and blunt ending the sites by mung bean digestion. The TK promoter was excised from pXER by digestion with EcoRl and Pstl. The ends were made blunt by mung bean digestion, and the fragments were joined by blunt end ligation. Plasmid pCHHO is a SV40-/S- galactosidase fusion plasmid available from Pharmacia. D . RNA isolation

RNA was isolated from transfected XLllO cells at the indicated times after glycerol shock by the guanidine thiocyanate-phenol-chloroform method which is described in Chromczynski and Sacchi, Anal. Biochem.. 162. 156-159 (1987) . Xenopus liver RNA, isolated from uninduced livers and from livers induced for 12 days with estradiol, was isolated as previously described in Shapiro and Baker, J. Biol. Chem.. 252. 5244-5250 (1977) .

E. Solution hybridization and SI nuclease protection assay

Solution hybridization and the SI nuclease protection assay were performed as described previously in Salton et al., Mol. Endocrinol.. g, 1033-1042 (1988), with the following modifications. A 2 ng sample of ³²P- labeled cRNA probe (2.7 x 10⁶ cpm) was hybridized to the indicated amount of RNA at 50°C overnight. SI nuclease (1,000 U; Pharmacia, Inc.) was used for digestion at 43°C for 90 min.

After SI nuclease digestion, samples were electrophoresed in a 4% polyacrylamide gel, using the 1- kb DNA ladder ^'(Bethesda Research Laboratories) as a molecular size standard. Autoradiograms were scanned by using an LKB UltroScan XL densitometer.

Figure 3 presents a schematic of the probes used in the SI nuclease protection assays and the sizes of the parts of the probes protected by MV5, vitellogenin Bl, and TK mRNAs. The 1036-nucleotide MV5 RNA probe used to detect MV5 mRNA in the SI nuclease protection assay is also shown in Figure IA above the pMV5 construction. The probes were synthesized as described in Salton et al., Mol. Endocrinol.. 2 , 1033-42 (1988) . F. Results

The results of the transfections and the SI nuclease assays are shown in Figures 2A-B and in Tables 1 and 2. In Figure 2A, the results are shown of an assay in which four plates of XLllO cells were each transfected with 20 μg of pMV5 and 20 μg of pXER. RNA isolated from these cells was analyzed by the Si nuclease protection assay using only the MV5 cRNA probe (see Figures IA and 3).

In Figure 2A, the lanes contained the following materials: Lanes 1 and 2 30 μg of RNA isolated from transfected XLllO cells grown in the absence (Lane 1) or presence (Lane 2) of 100 nM estradiol-173; Lane 3—10 μg of RNA from uninduced male Xenopus liver; Lane 4 1 ng of RNA from Xenopus liver induced for 12 days with estrogen plus 10 μg of uninduced Xenopus liver RNA.

In Figure 2B, the RNA analyzed was from cells transfected as follows: Lanes 1 and 2—10 μg each of pMV5, pXER, and pTK7; Lanes 3 and 4—10 μg each of pMV5, pTK7, and pTKCAT; Lane 5—50 ng of HSV-infected Vero cell RNA (obtained from Bernard Roizman, The University of Chicago Medical School, Chicago, Illinois) plus 10 μg of uninduced Xenopus liver RNA; Lane 6 1 ng of RNA from a Xenopus liver induced for 12 days with estradiol plus 10 μg of uninduced Xenopus RNA; Lane 7, 10 μg of RNA from uninduced male Xenopus liver. The RNA electrophoresed in Lanes 1 and 3 was from cultures grown for 72 hours in the absence of hormone, and that in Lanes 2 and 4 from cultures grown for 72 hours in 100 nM estradiol - 11 β . The autoradiograms for Lanes 1, 2, 5, 6, and 7 were exposed for 3 days; those for Lanes 3 and 4 were exposed for 22 hours.

Table 1 presents the MV5 mRNA content found in transfected XLllO cells cultured in the presence of 100 nM estradiol-170 relative to the MV5 mRNA content in cells cultured in the absence of hormone. The MV5 mRNA content is expressed as mean ± standard error of the mean, and the figures given are the means of nine inde¬ pendent transfection experiments. In each experiment, XLllO cells were transfected with 10 or 20 μg of pMV5 and an equal amount of pXER. RNA isolated from the trans¬ fected cells was analyzed by the Si nuclease protection assay using only the MV5 cRNA probe (see Figures IA and 3).

TABLE 1 Control of MV5 mRNA content by estradiol

Relative content of MV5 mRNA Culture Condition (mean ± SEM)

No hormone 100

100 nM estradiol 164 ± 13

Shown in Table 2 are the results of co- transfecting XLllO cells with 10 μg each pMV5, pTK7 and pXER or with 10 μg each pMV5, pTK7 and pTKCAT. After 72 hours, RNA was isolated from the transfected cells, and the levels of MV5 and TK mRNAs were simultaneously determined by the SI nuclease protection assay. The ratio of MV5 mRNA content to TK mRNA content was used to calculate the content of MV5 mRNA relative to that in cells transfected with pMV5, pXER and pTK7 and grown in no hormone. Plasmid pTK7 was used to provide an internal standard that would correct for any potential estrogen induction of transcription from the TK promoter. When pXER was not used, plasmid pTKCAT was used to keep the level of TK promoter the same in all cells. TABLE 2

Control of MV5 mRNA content by estradiol and estrogen receptor

Relative content of MV5 mRNA DNA" Culture Condition (mean ± SEM) pMV5, pXER No hormone TO!) pMV5, pXER 100 nM estradiol 163 ± 19 pMV5, pTKCAT No hormone 62 ± 12 pMV5, pTKCAT 100 nM estradiol 72 ± 12

'All cells were also transfected with pTK7.

G. Discussion of results 1. Production of MV5 mRNA Cotransfection of pMV5 and pXER into Xenopus XLllO liver cells resulted in the production of MV5 mRNA. A 534-nucleotide region of the MV5 cRNA probe, corre¬ sponding to the 529-nucleotide MV5 mRNA and the first 5 nucleotides of the poly(A) tail, was protected in the SI nuclease assay (see Fig. 2A and 2B) . Uninduced Xenopus liver RNA (see Fig. 2A, lane 3; Fig. 2B, lane 7) and mock-transfected XLllO cell RNA (data not shown) produced no protected bands in this size range. Vitellogenin mRNA from estradiol-induced Xenopus liver protected a 440- nucleotide band, corresponding to the 3' 435 nucleotides of vitellogenin mRNA and the first 5 nucleotides of the pol (A) tail (see Fig. 2A, lane 4; Fig. 2B, lane 6) . The small 90-nucleotide fragment corresponding to the region of the probe protected by the 5' end of full-length vitellogenin mRNA was not seen because it was run off the gel. 2. Role of estrogen in stabilization of MV5 mRNA

To test whether the MV5 mRNA could be stabil¬ ized by estrogen, pMV5 and pXER were transfected into XLllO cells, and the cells were maintained for 72 hours in the presence or absence of estrogen. MV5 mRNA was stabilized by estrogen (see Fig. 2A, Lanes 1 and 2) . When the data from nine separate experiments were aver¬ aged, the MV5 mRNA content of cells maintained in estradiol was increased by 64 ± 13% (see Table 1) com¬ pared to cells grown in the absence of hormone.

The relative level of MV5 mRNA expected after estrogen stabilization can be calculated by using data obtained from studies of the stability of full-length vitellogenin mRNA in Xenopus liver fragment cultures (Brock and Shapiro, Cell. 34. 207-214 (1983); Blume and Shapiro, Nucleic Acids Res.. 17. 9003-9014 (1989)) and from transfection data on the accumulation of MV5 mRNA and chloramphenicol acetyltransferase enzymatic activity at various times after transfection. Those studies indicate that at 20°C there is no significant transcrip¬ tion for 12 hours after transfection into XLllO cells and that transcription for the next 12 hours is at a reduced rate (data not shown) . Assuming that transcription then occurs at a constant rate between 24 and 72 hours post- transfection and that the MV5 mRNA exhibits a half-life of 30 hours at 20°C in the absence of estrogen and 500 hours (infinite in this calculation) in the presence of estrogen, then we would expect a 1.68-fold increase in MV5 mRNA content due to stabilization by estrogen 72 hours after transfection. This value is in excellent agreement with the average 1.64-fold increase in MV5 mRNA levels that was observed (see Table 1) . Because of the relatively long half-life of vitellogenin mRNA in the absence of estrogen, even if transcription of the MV5 gene occurred at a constant rate throughout the 72 hours, estrogen stabilization would result in only a 1.9-fold increase in MV5 mRNA content.

Although these data provide strong support for the view that the estradiol-estrogen receptor complex regulates the stability of the MV5 mRNA, they do not explicitly exclude the possibility that this complex induces increased transcription of the MV5 gene. Direct evidence that transcription of the TK promoter used in the pMV5 vector is not regulated by estrogen was obtained from experiments in which the same TK promoter in a TK- luciferase vector and plasmid pXER were cotransfected into XLllO cells. In 16 transient cotransfection experiments carried out under conditions similar to those described above, the average level of expression of the TK-luciferase plasmid was 665 luminescence units per μg in the absence of estradiol and 670 luminescence units per μg in the presence of estradiol. These values differ by <1%. This result provides compelling evidence that expression of the TK promoter is not induced by the estradiol-XER complex.

The related possibility that the vitellogenin DNA sequence cloned into the pMV5 vector contains an estrogen response element (ERE) that functions as a 3' enhancer is excluded by examination of the sequence of pMV5. The entire nucleotide sequence of the MV5 DNA and the 3'-flanking genomic fragment inserted into the pMV5 clone has been determined. There is no sequence either identical to the consensus ERE or diverging from the ERE by one, two, or even three nucleotides. All known func¬ tional EREs contain zero to two nucleotide changes relative to the consensus ERE, indicating that there is no functional estrogen receptor-binding site in the MV5 DNA and its 3'-flanking sequence.

Therefore, three lines of evidence support the conclusion that the estradiol-estrogen receptor complex stabilized MV5 mRNA: (1) the failure of the estradiol- estrogen receptor complex to induce transcription from the TK promoter used in the pMV5 plasmid; (2) the absence of any sequence related to the ERE with the potential to serve as a 3' transcription enhancer and estrogen receptor-binding site in the MV5 DNA and its flanking DNA; and (3) the excellent agreement between the observed level of MV5 mRNA and the predicted level for an mRNA undergoing estrogen-mediated stabilization.

3. Requirement for estrogen receptor for MV$ mRNA stabilization

In the experiments described above, high levels of estrogen receptor were synthesized from the trans¬ fected pXER plasmid. To determine whether the estrogen receptor is required for stabilization of MV5 mRNA, cells were transfected with pMV5 and pXER or with pMV5 and pTKCAT. When pMV5 and pXER were cotransfected, MV5 mRNA was stabilized by estrogen (see Fig. 2B, Lanes 1 and 2) . After transfection of pMV5 without the XER expression plasmid, MV5 mRNA was observed at levels slightly lower than those seen in the presence of unliganded estrogen receptor (see Fig. 2B; compare Lanes 3 and 4 with Lane 1) . These data indicate that estradiol and estrogen receptor are both required to obtain the increased level of MV5 mRNA seen on estrogen stabilization.

To provide an internal standard that would also correct for any potential estrogen induction of tran¬ scription from the TK promoter, pMV5 and pTK7 were cotransfected into XLllO cells. As noted earlier, plasmid pTK7 contains the HSV TK structural gene driven by the HSV TK promoter. The levels of MV5 and TK mRNAs were simultaneously determined by the SI nuclease pro¬ tection assay. The results from three experiments in which the TK mRNA level was used as an internal standard to correct the level of MV5 mRNA confirm a requirement for estrogen receptor in the stabilization of MV5 mRNA (see Table 2) . Without estrogen receptor, estradiol was unable to.stabilize MV5 mRNA. The data obtained in experiments using TK mRNA as an internal transfection and assay standard and in the nine experiments summarized in Table 1, in which an internal standard was not used, are in excellent agreement, providing additional evidence that the increased level of MV5 mRNA in estrogen-treated cells is due to post-transcriptional mRNA stabilization rather than an increase in transcription from the TK promoter.

As shown in Table 2, MV5 mRNA content in the absence of estrogen receptor was slightly lower than when estrogen receptor was present without added estradiol. This result may have been due to partial stabilization resulting from traces of estrogen sulfate remaining in the charcoal-dextran-treated serum (Brown et al., Proc. Natl. Acad. Sci. USA. £1, 6344-6348 (1984)). The actual level of estrogen-dependent stabilization would then be somewhat greater than is presented in Tables 1 and 2. Alternatively, the high level of unliganded estrogen receptor in pXER-transfected cells may elicit partial stabilization of the MV5 mRNA. Since a TK internal standard was used in these experiments, even if this high level of unliganded receptor induced increased synthesis from the TK promoter, this effect would be corrected for in the data. Direct evidence that high levels of unliganded estrogen receptor do not induce increased expression of the TK promoter was obtained in additional experiments in which the TK-luciferase vector (see above) was transiently cotransfected with either pXER or pTK-_/8- gal. Plasmid pTK-0-gal was added to maintain a constant TK promoter concentration when pXER was not used. The average levels of expression of the TK-luciferase vector in four separate experiments, performed similarly to those reported in Table 2, differed by <1% in the presence or absence of pXER. This evidence and the use of the pTK7 internal standard in these experiments rule out the possibility that this result was due to increased transcription of the TK promoter induced by high levels of unliganded XER.

Since estrogen receptor does not destabilize MV5 mRNA in the absence of estradiol, the data exclude the possibility that MV5 mRNA is a stable mRNA that is destabilized by unliganded estrogen receptor. Maximum stabilization of MV5 mRNA occurred only when both the estrogen receptor and estradiol were present (Table 2) , indicating that the estradiol-estrogen receptor complex plays a positive role in stabilizing MV5 mRNA. This result is consistent with our observation that in the absence of estrogen, vitellogenin mRNA and total poly(A) mRNA exhibit the same half-life values. Brock and Shapiro, Cell. 24.:207-214 (1983) . These data support the hypothesis that in the absence of estrogen, vitellogenin mRNA is degraded by normal cellular mechanism. Blume et al., UCLA Symp. New Ser.. ≤2, 259-274 (1987).

In pMV5, whose preparation is described in Example 1 above, there is a Pstl site located at the junction of the two small vitellogenin coding sequences (see Figure ID) . A DNA sequence coding for a protein or polypeptide of interest can be inserted into this Pstl site. DNA sequences coding for cleavage sites can also be inserted between the DNA sequence coding for the protein or polypeptide of interest and each of the vitellogenin coding regions in pMV5. Alternatively, the remaining small vitellogenin coding regions can be eliminated from pMV5 so that the DNA sequence coding for the protein or polypeptide of interest is linked directly to the DNA coding for the 3' and 5' untranslated regions. Methods of preparing the DNA sequences and of inserting them into, or deleting them from, vectors such as pMV5 are well-known.

The plasmid resulting from the above manipula¬ tions of pMV5 can be used to transform an appropriate host (such as the XLllO cell line; see Example 2) . If necessary, the host cells may also be transformed with a vector coding for estrogen receptor (see Example 2) . When the transformed host cells are cultured in the presence of estrogen, the content of the hybrid mRNA coded for by the plasmid would increase substantially (up to 20-fold) as compared to the mRNA content of cells maintained in the absence of estrogen. As a consequence, enhanced production of the protein or polypeptide of interest coded for by the hybrid mRNA would be obtained.

EXAMPLE 3: Sequences Required For mRNA Stabilization

A. Plasmids

A series of plasmids was prepared to investigate which of the sequences of the X. laevis vitellogenin Bl mRNA was required for mRNA stabilization.

This series of plasmids was derived from pMV5, whose preparation is described in Example 1. First, the TK promoter of pMV5 was replaced by the more powerful

Adenovirus major late (ML) promoter to produce pMV8.

Then, sequences in pMV8 coding for portions of the 5' and

3' untranslated regions of the mini-vitellogenin mRNA were deleted or replaced by site directed mutagenesis as described below.

1. Plasmid pMV8 The preparation of pMV8 is illustrated in

Figure 5. In Figure 5, "MV" designates the mini- vitellogenin sequence.

First, pMV5 was partially digested with EcoRl.

The fragment containing the mini-vitellogenin sequence was isolated by size fractionation, and the EcoRl site was filled in with the Klenow fragment. The ends of the fragment were then ligated with bacteriophage T4 DNA ligase to produce pMV6 (see Figure 5) .

To remove the TK promoter from pMV6, the plasmid was digested with EcoRl and gaπiHI. The restriction cleavage sites were made blunt ended by filling-in with Klenow fragment. The resulting 3848 bp fragment was isolated by gel electrophoresis.

The ML promoter fragment was excised from plasmid pMLCAT by digesting the plasmid with Sstl and BamHI. The cleavage sites were made blunt ended by treatment with mung bean nuclease (Stratagene) . A 303 bp fragment containing the ML promoter was isolated by size fractionation on a polyacrylamide gel.

Plasmid pMLCAT was obtained from Dr. Philip Sharp, Department of Biology and Center for Cancer Research, Massachusettes Institute of Technology, Cambridge, Ma. Expression vectors driven from the identical ML promoter are available from the American Type Culture Collection, and the ML promoter can also be isolated from these vectors using standard techniques.

The 3848 bp fragment containing the mini- vitellogenin sequence and the 303 bp fragment containing the ML promoter were joined by blunt-end ligation with bacteriophage T4 DNA ligase. The plasmid with the ML promoter in the correct orientation with respect to the mini-vitellogenin sequence was identified by restriction mapping and was designated pMV7. In pMV7, the ML promoter is flanked by BasHI and £SQRI sites (see Figure 5).

Finally, pMV7 was digested with I≤sRI and BamHI (see Figure 5) . The resulting 3848 bp fragment containing the mini-vitellogenin sequence was dephosphorylated by treatment with calf intestinal alkaline phosphatase, and the dephosphorylated fragment was isolated.

The fragment of the pMV7 plasmid containing the ML promoter (see Figure 5) was amplified by the polymerase chain reaction (PCR) (Roth et al., BioTechniques. 2_. 746 (1989)). The 5' primer used was the Universal Primer (available from Pharmacia and Promega) . Its sequence is:

5'-GTAAAACGAC GGCCAGT-3'

[SEQ ID NO:11]

The 3' primer was:

5' -CCTTGAGAAT TCGGACGAAC G-3'

[SEQ ID NO:12]

The 362 nucleotide amplified fragment was cut with BamHI and EcoRI. and the 182 nucleotide fragment containing the ML promoter was isolated by polyacrylamide gel electrophoresis.

The 182 bp fragment was ligated to the 3848 bp fragment, and the resulting vector was designated pMV8 (see Figure 5) . The nucleotide sequence of the relevant portion of pMV8 is shown in Figure 6B [SEQ ID NO:2] .

B. Plasmids pMVlO-13 Site directed mutagenesis was used to delete or replace sequences in pMV8 coding for portions of the 3' and 5' untranslated regions of the mini-vitellogenin mRNA. It should be noted that the mini-vitellogenin sequence in pMV8 codes for the complete 3' and 5' untranslated regions of the X. laevis vitellogenin Bl mRNA (see Example 1) . Site directed mutagenesis was carried out using single-stranded synthetic oligonucleotides as described by Chang and ^"Shapiro, J. Biol. Chem.. 265. 8176 (1990) and Kunkel, Proc. Nat'l Acad. Sci. USA. §_ , 488 (1985) with the modification that single-stranded DNA was isolated by NACS chromatography as follows. A NACS column (BRL) was attached to a 1 ml. rainin micropipet and equilibrated by pipeting up and down three times with a buffer containing 10 mM Tris, pH 7.8, 1 mM EDTA, 500 mM NaCl. The column was detached from the micropipet, and the DNA (dissolved in the above buffer) was loaded onto the column by gravity feed. The column was washed with 1 ml. of the above buffer, and the single-stranded DNA was eluted by washing the column three times with a buffer containing 10 mM Tris, pH 7.8, 1 mM EDTA and 2 M NaCl. The eluted DNA was precipitated with ethanol and used for site directed mutagenesis.

The sequences of the oligonucleotides used for site directed mutagenesis to produce plasmids pMV10-13 are as follows. The sequence used to produce pMVIO was:

5' -GGATGCAAAT ATTTATTTAA GCTTGAGATC AGTTTATCAT C-3'

[SEQ ID NO:7]

The sequence used to produce pMVll was:

5' -ATACATTTGG TTTGAAAGCT TCATGCTGCA CATTTCAG-3'

[SEQ ID NO:8]

The sequence used to produce pMV12 was:

5'-CGCCTGTGCC CTCGAACCTA AGCTTACCAT GAGGGGAAT-3'

[SEQ ID NO:9] The sequence used to produce pMV13 was:

5'-GAATTGTTTA CAATAGTATT CAAGCTTGGT CGTTTGAAAT TGAGATCAGT-3'

[SEQ ID NO:10]

The complete nucleotide sequences of the relevant portions of pMVIO, pMVll, pMV12 and pMV13 are shown in Figures 6C-6F [SEQ ID NOS:3-6] .

Plasmid pMV12 codes for a mini-vitellogenin mRNA in which eight of the bases of the 5' untranslated region have been replaced (see Figure 7 and compare Figures 6B and 6E) . Plasmid pMVll codes for a mini- vitellogenin mRNA in which 96 bases have been deleted from the start (upstream end) of the 3' untranslated region (see Figure 7 and compare Figures 6B and 6D) .

Plasmids pMVIO and pMV13 were designed to examine the role in mRNA stabilization of a short sequence in the 3' untranslated region of the mini- vitellogenin mRNA. This sequence is 5' -CAAATGTATA TT-3'

[SEQ ID NO:22] (nucleotides 835-846 in Figure 6B) . Plasmid pMVIO codes for a mini-vitellogenin mRNA in which 43 bases have been deleted from the 3' end of the 3' untranslated region (see Figures 6C and 7) . The deleted region contains this sequence. Plasmid pMV13 codes for a mini-vitellogenin mRNA in which this sequence has been replaced (see Figure 7 and compare Figures 6B and 6F) .

The short sequence referred to above is a conserved sequence found in the 3' untranslated regions of all four of the estrogen-stabilized vitellogenin mRNAs. It was found by comparing the sequences of the 3' untranslated regions of the four vitellogenin mRNAs. The sequences of the 3' untranslated regions of the four Xenopus vitellogenins were determined by standard techniques in our laboratory (see supra) . The other sequences have been published in the scientific literature and can also be found in the GenBank database. The DNA sequences coding for the conserved region of the estrogen-stabilized mRNAs are shown below in Table 3.

TABLE 3

Distance Conserved mRNA Sequence to poly (A) nucleotides

Vitellogenin Al CtAATGTATG TT [SEQ ID NO:23]

Vitellogenin A2 TtAATGTGTA ac [SEQ ID NO:24]

Vitellogenin Bl CAAATGTATA TT [SEQ ID NO:22]

Vitellogenin B2 CAAATGTATA TT [SEQ ID NO:25]

Conalbumin TGAATGTGgc TT [SEQ ID NO:26]

Ovalbumin TGAATGTGCt cT [SEQ ID NO:27]

Apo VLDL II TtAATGTAac Ta [SEQ ID NO:28]

Chicken Vitello¬ gAAATGTtaA aT

genin [SEQ ID NO:29]

C. Assay For mRNA Stabilization Plasmids pMV8, pMVIO, pMVll, pMV12 and pMV13 were transfected into XLllO cells as described in Example 2. Each plasmid was transfected into the cells with or without plasmid pXER (preparation described in Example 2) . The transfected XLllO cells were cultured with estradiol as described in Example 2. RNA was isolated and analyzed for the level of the mini-vitellogenin mRNA in cells containing and lacking estrogen receptor. RNA analysis was by the Si nuclease protection assay as described in Example 2.

The results are summarized in Figure 7. For each assay, the data represent the average of at least two determinations. The mRNA was considered to be stabilized if the fold change in its level was comparable to that exhibited by MV5 mRNA (i.e.. at least a 1.5 fold induction see Example 2) .

The data demonstrate that the replacement of the thymidine kinase promoter in pMV5 with the more powerful ML promoter (pMV8) does not alter the ability of the estradiol-estrogen receptor complex to stabilize the mini-vitellogenin mRNA. The data further demonstrate that the sequences responsible for cytoplasmic stabilization of mRNA are located outside the protein coding region of the mRNA. In particular, the data show that the 3' untranslated region contains determinants essential to the stabilization of vitellogenin mRNA, and that the 5' untranslated region is not essential for mRNA stabilization.

Specifically, the data show that MV11 mRNA was not stabilized by the estradiol-estrogen receptor complex. This indicates that all or some of the 96 bases deleted from the start (upstream end) of the 3' untranslated region of the mini-vitellogenin mRNA are essential for mRNA stabilization. The level of MV11 mRNA in the absence of estrogen receptor is comparable to the level of MV8 mRNA under the same conditions. This indicates that the deletion of this large segment of the 3' untranslated region did not alter the basal stability of the mRNA. This result also indicates that sequences in this region stabilize mRNA through a specific effect of the estrogen-estrogen receptor complex rather than as a result of some nonspecific mechanism.

The MV10 mRNA also was not stabilized by the estradiol-estrogen receptor complex. This indicates that all or some of the 43 bases deleted from the 3' end of the 3' untranslated region of the mini-vitellogenin mRNA are essential for mRNA stabilization. Further, MV13 mRNA, in which the short conserved sequence had been replaced, was not stabilized by the estradiol-estrogen receptor complex. The results obtained with pMVIO and pMV13 show that it is likely that the short conserved sequence identified above plays a role in mRNA stabilization. Also, it is likely that this^' sequence is responsible for the loss of mRNA stabilization when all 43 bases are deleted from the 3' end of the 3' untranslated region.

It was also found that the mRNAs produced by pMVIO and pMV13 were considerably shorter than expected and contained new 3' ends. The new 3' ends were located in the region of 96 nucleotides deleted in pMVll approximately 20-40 nucleotides downstream from a second copy of the conserved sequence. This result indicates that the conserved sequence plays a role in the selection of the correct 3' ends of mRNAs. Thus, the failure to stabilize the MV10 and MV13 mRNAs is likely due to the formation of shorter mRNAs with very different 3' ends and 3' untranslated regions.

The conclusion that the conserved sequence plays a role in the selection of the correct 3' ends is supported by two other pieces of information. First, a small cytoplasmic 4.5S RNA contains a sequence complementary to the conserved sequence. This small cytoplasmic RNA is found bound to polyadenylated mRNAs, and other small cytoplasmic RNAs play a role in poly(A) addition.

Second, a search was made of Xenopus. human and mouse cDNA sequences in the GenBank database. The conserved sequence was found upstream of the poly(A) tract in about 30 out of the 50 sequences analyzed. In all of the cDNA sequences coding for alternately polyadenylated mRNAs (about 6 out of 50) , the conserved sequence was found upstream of each polyadenylation site.

Taking into consideration the results of the search of the GenBank database and the DNA sequences coding for the estrogen-stabilized mRNAs given above, a consensus RNA sequence was determined. It is:

5'-UNUAAAUGUR UAUU-3' ,

[SEQ ID NO:13] wherein N is any nucleotide and R is any purine.

Finally, the MV12 mRNA was stabilized (see Figure 7) by the estradiol-estrogen receptor complex. This shows that the 5' untranslated region of the mini- vitellogenin mRNA is not essential for mRNA stabilization.

EXAMPLE 4: Preparation Of Plasmids Coding For Mutant Estrogen Receptors

The starting material for the preparation of the mutant estrogen receptors was the wild type X. laevis estrogen receptor (XER) . The preparation of plasmid pXER which codes for XER was described in Example 2.

The plasmids coding for the mutant XERs were designated pXER-lAH and pXER-2AH. They are the pXER plasmid with either one or two copies of a synthetic DNA sequence cloned into the XER cDNA at a BamHI site. The inserted synthetic DNA sequence codes for a sequence of 15 amino acids which form an amphipathic α-helix. This amino acid sequence has been previously described in Giniger and Ptashne, Nature. 330. 670 (1987) . Plasmids pXER-lAH and pXER-2AH were prepared as follows.

Plasmid pXER contains four BamHI sites (see Figure 10) , so partial digests were necessary to obtain a fraction of the vector with a single BamHI cut at the correct site in the XER sequence. To do so, approximately 100 μg of pXER DNA was digested with 100 units of BamHI in a final volume of 480 μl. Aliquots containing 150 μl of the reaction mix were removed after incubating for 20 min. at 37θC and after 30 min. at 37θC. The remainder of the reaction mix was allowed to digest for a total of 60 minutes. The reaction in each aliquot was stopped by adding EDTA to a final concentration of 20mM. The partially-digested samples obtained after 20 min. and 30 min. of digestion were pooled and analyzed by gel electrophoresis. The linearized plasmid was separated from multiply-cut and circular plasmid forms by agarose gel electrophoresis using a 0.7% gel prepared from Genetic Technology Grade agarose (Marine Colloids) , and the linearized plasmid DNA was isolated from the gel. To decrease the ability of the linearized BamHI-digested pXER to religate to itself and recircularize, the 6.3Kb linearized pXER fragment was dephosphorylated by incubation with calf intestinal alkaline phosphatase. A synthetic oligonucleotide was prepared by synthesizing two single-stranded oligonucleotides on an applied Biosystems Model 380B DNA synthesizer using β- cyanoethylphosphoramide technology and then annealing them. The sequences of the two single-stranded synthetic oligonucleotides are shown below:

AHl:

-GATCCAGAAT TGCAAGAGCT GCAGGAACTG CAAGCTCTGC TGCAACAGCA ACAA.-3'

[SEQ ID NO:30]

AH2:

5' -GATCTTGTTG CTGTTGCAGC AGAGCTTGCA GTTCCTGCAG CTCTTGCAAT TCTG-3'

[SEQ ID NO:31]

Each of the two single-stranded synthetic oligonucleotides AHl and AH2 was purified by passage through oligonucleotide purification cartridges (Applied Biosystems) in the manner recommended by the manufac¬ turer. These cartridges contain hydrophobic resins that bind to the trityl blocking groups on the oligonucleo¬ tides. The bound oligonucleotides are detritylated and eluted. Next, the oligonucleotides AHl and AH2 were incubated with polynucleotide kinase and annealed by standard methods. The resulting double-stranded oligo¬ nucleotide was designated OligoAH. Its sequence is shown below:

AHl

5 ' -GATCC_ftGAA_CTGC-AAGAGCTGC_AGGAACTGC_AAG

3 ' -GTCTTAACGTTCTCGACGTCC TGACGTTCGAGACGACGTTGTCGTTGTTCTAG- 5 '

AH2

OligoAH contains a BamHI site at its 5' -end and a BamHI-compatible BglII site at its 3' -end. OligoAH was designed to contain an internal Pstl site (which is underlined in AHl above) , so that recombinant plasmids containing OligoAH could be identified by digestion with Pstl.

The amino acid sequence encoded by OligoAH is:

Glu Leu Gin Glu Leu Gin Glu Leu Gin Ala Leu Leu Gin Gin

5 10

Gin 15

[SEQ ID NO:15]

A potential structure of this peptide as an amphipathic helix is shown in Figure 9.

An aliquot of the double-stranded OligoAH was used in the production of pXER-lAH.

To prepare the oligonucleotide used in the construction of pXER-2AH, OligoAH was ligated to itself with bacteriophage T4 DNA ligase. The ligated oligonucleotide was digested with BamHI and Bglll. These enzymes will cleave oligonucleotides which have been joined BamHI-BamHI or Bglll-Bglll. Only oligonucleotides joined head-to-tail (i.e.. BamHI to Bglll) will fail to re-form a site recognized by either restriction enzyme and will not be cut. Thus, a population of OligoAH multimers containing variable numbers of copies of OligoAH joined so that they code for two or more copies of the amphipathic helix was prepared. The correct reading frame and amino acid sequence are preserved in the OligoAH multimers.

The linearized, phosphatase-treated pXER (see above, this example) was mixed with OligoAH or the mixture of OligoAH multimers, and the DNAs were joined by ligation (at 16θC for 16 hours) . £. coli NM522 was transformed by the ligation mixtures, and ampicillin- resistant colonies were picked from a plate and grown up in 5 ml cultures. DNA was prepared from each potential clone, and the plasmid mini-preps were analyzed by digestion with Pstl. Since there is a Pstl site in OligoAH, pXER-lAH and pXER-2AH show 3 bands on agarose gels after digestion with Pgtl. while the wild type pXER only shows two. (Note: The small fragment between one OligoAH Pstl site and the second OligoAH Pstl in pXER-2AH is too small to be seen on an agarose gel of the Pstl- digested DNA. Plasmids containing one and two copies of OligoAH, therefore, show the same number of bands after Pstl digestion.)

The potential recombinant clones were sequenced by the dideoxy method using sequenase (United States Biomedical Corp.). Plasmids with the correct map and sequence which were found to contain one or two copies of OligoAH were designated pXER-lAH and pXER-2AH, respectively. Plasmids pXER-lAH and pXER-2AH contain the TK promoter controlling the transcription of mRNAs coding for mutant XERs. These mutant XERs contain one or two copies of the synthetic amphipathic helix inserted into the hinge region of the estrogen receptor.

EXAMPLE 5: Enhanced Transcription Using pXER-lAH and pXER-2AH

Plasmids pXER-lAH and pXER-2AH prepared in

Example 4 were tested for the effect of the mutant estrogen receptors produced by them on transcription of the chloramphenicol acetyl transferase (CAT) gene under the control of an estrogen regulated promoter. To do so, XLllO cells were transfected with either pERE-VIT-CAT or p2ERECAT and one of pXER, which produces the wild type estrogen receptor, pXER-lAH or pXER-2AH. The preparation of pXER and of the XLllO cell line are described in Example 2, as is the method of transfection of these cells.

A. Preparation Of pERE-VIT-CAT

Plasmid pERE-VIT-CAT contains a 618 bp fragment (from -598 to +21) from the X. laevis vitellogenin Bl 5' flanking region linked to the CAT gene. To increase its activity, a copy of the consensus estrogen response element (ERE) was inserted at -359.

Plasmid pERE-VIT-CAT was prepared as follows. First, a genomic clone containing the X. laevis vitellogenin Bl gene was isolated as described in Denise Lew, Ph.D. Thesis, 1988, University of Illinois, Urbana, Illinois. See also Wahli et al., Proc. Nat'l Acad. Sci. USA. 79. 6832 (1982) (describes the structure of the same vitellogenin Bl gene) and Walker et al., Nucleic Acids Research. 12. 8611 (1984) (gives an identical sequence for the 5' flanking region of the vitellogenin Bl gene as determined by Lew) .

Briefly, the genomic clone was isolated as follows. Xenopus red blood cell DNA was partially digested with EcoRI, and 20-24 kb fragments were packaged in bacteriophage lambda vector L47.1 which was used to transform E. coli strain M4126. The library was screened with labelled vitellogenin cDNA probes. Four positive clones were identified, and DNA from three of these isolates was purified and analyzed by agarose gel and Southern blot. These revealed that all three clones were identical to each other. Further characterization showed that the three clones contained fragments which included the vitellogenin Bl gene. One of the three clones was chosen as the representative clone for further processing. Library preparation and screening were by standard techniques as described in Sambrook et al., Molecular Cloninσ (2d ed. 1989) .

Also as described in the Lew Thesis,

_* restriction analysis of this clone revealed a Hinfl fragment in the vitellogenin insert that would contain 597 nucleotides of 5' flanking region plus 21 nucleotides of transcribed downstream region. This Hinfl fragment was isolated, filled in with Klenow fragment, and Hindlll linkers were ligated to the ends. The resulting fragment was inserted into the Hindlll site of the SVOCAT vector (available from the American Type Culture Collection) to yield pVITCATl (called 597-3 vitCAT in the Lew Thesis) .

Next, the 2.2 kb Hindlll/BamHI fragment from pVITCATl, which contains the 618 bp vitellogenin fragment and the CAT structural gene, was subcloned into the Hindlll/BamHI cut phagemid vector pTZ18U (prepared as described in Mead et al., Protein Engineering. 1, 67-74 (1986) ) . Oligonucleotide-directed site-specific mutagenesis was used to delete the nucleotides between the start codon of the vitellogenin Bl gene at +14 and the start codon of the CAT gene. The resulting plasmid was designated pVITCAT. A schematic drawing of pVITCAT is shown in Figure 11. The preparation of plasmid pVITCAT is also described in Chang and Shapiro, J. Biol. Chem.. 265. 8176 (1990) (designated therein pTZVITCAT3) and Chang, 1990 Thesis, University of Illinois, Urbana, II (designated therein pTZ18U-BlVITCAT) .

To increase the activity of pVITCAT, a copy of a consensus ERE sequence 5' -GGTCACAGTG ACC-3'

[SEQ ID NO:32] was inserted into the Clal site at -359 in the vitellogenin 5' flanking sequence. The consensus ERE sequence was synthesized on an Applied Biosystems Model 380B DNA synthesizer using beta-cyanoethyl-phosphoramide chemistry. The resulting plasmid was designated pERE- VIT-CAT. Plasmid pERE-VIT-CAT contains a total of 4 EREs (the inserted consensus ERE and the three imperfect EREs normally found in the vitellogenin 5'flanking sequence; see Figure 11) .

B. Preparation of Plasmid p2ERECAT

Plasmid p2ERECAT contains a synthetic promoter comprising two copies of the consensus ERE linked to a TATA box. The promoter is regulated by the estradiol- estrogen receptor complex and is a very powerful promoter. The CAT gene is cloned downstream of this promoter.

To prepare plasmid p2ERECAT, plasmid pVITCAT (prepared as described above) was cut with Bglll and Ncol which removed the sequence from -42 to +528 which contains the vitellogenin TATA box at -30 and vitellogenin sequences from +14 to -42. This segment of the vitellogenin promoter contains no known transcription activation sequences and simply provides a TATA box and flanking DNA known to function in estrogen-regulated transcription. The remaining sequences from +15 to +528 are CAT sequences (the vitellogenin sequence is directly fused to the CAT gene translation start site see above) .

The promoterless vector SVOCAT (available from the American Type Culture Collection) was digested with Bglll and Ncol. The -42 to +528 fragment was also isolated by size fractionation on an agarose gel and then inserted into the cut SVOCAT by T4 DNA ligation. The resulting plasmid was designated pTATACAT. Next, an oligonucleotide containing a consensus ERE was synthesized on an Applied Biosystems Model 380B DNA synthesizer using 5-cyanoethyl^phosphoramide chemistry. The oligonucleotide had the following sequence:

5' -CGATAGGTCA CAGTGACCAT-3'

[SEQ ID NO:33]

The ERE sequence is underlined.

The oligonucleotide was annealed to itself by heating the deprotected oligonucleotide to 90°C in lOmM Tris, pH 7.8, ImM EDTA, and then allowing the tubes to cool overnight. The resulting double-stranded DNA was ligated to itself with T4 DNA ligase to form multimers of the ERE-containing sequence.

Vector pTATACAT was digested with Sail, and the multimerized oligonucleotides were inserted into the Sail site of the digested vector at position -53. Clones were identified by screening with labelled nucleotides having the above sequence, followed by dideoxy DNA sequencing. A clone with 2 copies of the consensus ERE was identified and designated p2ERECAT. The net result of these manipulations was to place two copies of the consensus ERE 20 nucleotides upstream of the TATA box.

C. Expression of the CAT Gene of p2ERECAT

Plasmids pXER (prepared as described in Ex¬ ample 2) and p2ERECAT were used to co-transfect XLllO cells as described in Example 2. Cells were cultured as described in Example 2. CAT activity was assayed as described in Nielsen, et al., Anal. Biochem.. 179. 19-23 (1989) , except that the amount of ³H-acetyl CoA substrate used was reduced by 50%.

The results are shown in Table 4 below. Levels of CAT activity in cells maintained in the presence (+E) and absence (-E) of estradiol-173 are shown. The fold induction (+E/-E) is the ratio of the levels of CAT activity in the presence and absence of estradiol. As shown in Table 4, estrogen-regulated production of CAT was obtained. Indeed, the production of CAT was substantially enhanced in the presence of estradiol.

TABLE 4

Transfection Estradiol CAT Fold

Activity Induction (cpm/ug) (+E/-E)

1 + 1,444 22

70

2 + 1,411 23

61

3 + 908 15

60

D. Regulation of the Transcription of p2ERECAT and pERE-VIT-CAT by Mutant Estrogen Receptor

XLllO cells were co-transfected with either p2ERECAT or pERE-VIT-CAT and one of pXER, pXER-lAH or pXER-2AH. The transfection and culturing of cells was carried out as described in Example 2. CAT activity was assayed as described above.

The results are shown in Table 5 below. To calculate the numbers shown in Table 5, the CAT activity produced by the wild type estrogen receptor in the pres¬ ence of estrogen was set equal to one for each experi¬ ment. The activity of the mutant estrogen receptors in the presence and absence of estrogen were calculated relative to the activity of the wild type estrogen receptor in the presence of estrogen. The relative values for the different experiments were then averaged, and these averages are reported in Table 5. The data represent the mean ± standard error of the mean for at least 5 separate experiments for pERE-VIT-CAT and for at least 3 separate experiments for p2ERECAT.

TABLE 5

Estrogen Relative CAT Activity

Receptor pERE-VIT-CAT p2ERECAT

XER (wild type) 1 1

XER + 7.1 + 4.0 1.3 ± 0.3

1 amphipathic helix

XER + 4.8 + 1.6 2.0 + 0.5

2 amphipathic helices

Table 5 shows that more CAT activity was detected when the two mutant estrogen receptors were used, as compared to when the wild type estrogen receptor was used, with either pERE-VIT-CAT or p2ERECAT. This demonstrates that the mutant estrogen receptors produced by pXER-lAH and pXER-2AH cause substantially enhanced transcription from estrogen-regulated promoters as compared to the wild type estrogen receptor.

EXAMPLE 6: Construction Of Vector Having Two EREs Linked To TATA Box

Plasmid p2EREXER was prepared. It contains a synthetic promoter comprising two copies of the consensus estrogen response element (ERE) linked to a TATA box.

The promoter is regulated by the estradiol-estrogen receptor complex and is a very powerful promoter. The gene coding for X. laevis estrogen receptor (XER) is cloned downstream of this promoter in plasmid p2EREXER. Thus, plasmid p2EREXER places expression of the toxic estrogen receptor protein under the control of an estrogen-regulated promoter. The preparation of p2EREXER is illustrated in Figure 12.

To prepare plasmid p2EREXER, four oligonucleo¬ tides were synthesized on an Applied Biosystems Model 380B DNA synthesizer using 3-cyanoethylphosphoramide chemistry. The oligonucleotides were designated A2ERES1 and A2ERES2 for the upper and lower strands, respec¬ tively, of Oligo 1, and BlTKl and B1TK2 for the upper and lower strands, respectively, of Oligo 2. The sequences of the four oligonucleotides used are as follows:

A2ERES1:

5'-TCGACCTAGA TCACATTAGG TCACAGTGAC CTTCAACAAG GTCACAGTGA CCTTGAACAA T-3'

[SEQ ID NO:34]

A2ERES2:

5'CGATTGTTCA AGGTCACTGT GACCTTGTTG AAGGTCACTG TGACCTAATG TGATCTAGG 3'

[SEQ ID NO:35]

BlTKl:

5' CGATATAAAT ACACTACAGT ATCTCATCGA ACACCGAGCG ACCCTGCA 3'

[SEQ ID NO:36]

B1TK2:

5' GGGTCGCTCG GTGTTCGATG AGATACTGTA GTGTATTTAT AT 3'

[SEQ ID NO:37]

The sequences of Oligos 1 and 2 are shown below: Qliαo 1 :

5 ' A2ERES1

I

TCGACCTAGATCACATTAGGTCACAGTGACCTTa^C_ftAGGTC_AC_^ '

3 ' GGATCTAGTGTAATCC_AGTGTC_ACTGGAAGTTGTTCCA^ '

A2ERES2

Olicro 2 :

BlTKl

5 ' CGATATAARTAC_RCTACLRGTAT(_rrCATCGAA__ACCGAGCGACCC^ 3 ' 3 ' TATATTTAITJI-GATGTCΛTAGAGTAGCTTG^ 5 '

B1TK2

The left end of Oligo 1 (as pictured above) is a cohesive end for Sail, and the right end of Oligo 2 is a cohesive end for Pstl. The right end of Oligo 1 and the left end of Oligo 2 anneal to form a Clal site. In Oligo 1, the palindromic segments of the two consensus estrogen response elements are underlined. Oligo 2 contains a consensus TATA sequence (underlined) (Lewin, Genes IV. pp. 545-61 (1990)) and a portion of the 5' untranslated region of the Herpes thymidine kinase gene.

Each of the four synthetic oligonucleotides was purified by passage through oligonucleotide purification cartridges (Applied Biosystems) as recommended by the supplier (see Example 4) . Oligonucleotides A2ERES1 and A2ERES2 were further purified by electrophoretic isolation on 15% denaturing acrylamide gels using standard sequencing buffers (Maxam and Gilbert, Methods Enzvmol.. 65. 499 (1980)). BlTKl, the upper strand oligonucleotide for Oligo 2, was phosphorylated on the 5' end by polynucleotide kinase as described in Sambrook, ≤t. al., Molecular Cloning. A Laboratory Manual page 5.68 (Cold Spring Harbor Laboratory Press, New York, 1989) .

Oligos 1 and 2 were formed by annealing equal molar amounts of the appropriate oligonucleotides by heating to boiling and cooling to room temperature over a period of 60 min. Five μg of Oligo 1 was ligated to 3 μg of Oligo 2 (id. at 1.68), and the fragments were sepa¬ rated by electrophoresis on a nondenaturing 8% acrylamide gel with Tris-borate-EDTA buffer (id. at 6.39). The region of the gel containing the desired 101 bp ligated Oligo 1-2 fragment was excised, and the DNA extracted by soaking in 100 μl of 0.5 M ammonium acetate, 1 mM EDTA, as recommended for isolation with Gel/X by its supplier, Genex Corporation. The gel was removed by centrifugation in a Brinkman microfuge for 5 min and washed in 100 μl of the same buffer. The DNA was separated from the elution buffers by precipitation with ethanol.

Plasmid pTZ18R was constructed as described in Mead, et al., Protein Engineering. 1 , 67 (1986). It is also available from several commercial sources, including United States Biochemical, Inc. and Pharmacia PL Biochemicals. Plasmid pTZ18R was digested sequentially with Pstl and Sail. About 20 ng of Oligo 1-2 were ligated to about 0.25 μg of the digested pTZ18R DNA (Sambrook et al., Molecular Cloning. A Laboratory Manual, p. 1.68) .

E. coli NM522 was transformed by the ligation mixture (Cohen et al., Proc. Natl. Acad Sci USA.. 74. 5463 (1977) ) , and recombinant white colonies were picked and grown up in 1 ml cultures. E. coli NM522 is avail¬ able commercially from a number of companies, including Bethesda Research Laboratories. It is a recA⁺ strain and is capable of α-complementation of the jS-galactosidase gene.

DNA was isolated and analyzed by digestion with Clal and with PvuII. Colonies 1, 2 and 4, which were linearized with Clal and generated fragments of about 450 bp after PvuII digestion, were selected for further analysis. DNA from colony 2 was isolated and sequenced by the dideoxynucleotide method using the M13 reverse sequencing primer (Sanger, et al., Proc. Natl. Acad. Sci. USA. 74. 5463 (1977)). The oligonucleotide used for the reverse sequencing primer was obtained from Promega. The sequence of the inserted Oligo 1-2 was as expected. This plasmid was designated pTZERE.

For isolation of the Oligo 1-2 fragment, pTZERE DNA was isolated from E. coli NM522 grown overnight at 370C in 150 ml of 2 x TY medium (16 g. tryptone powder, 10 g. yeast extract and 5 g. NaCl per liter) . About

100 μg of plasmid DNA was sequentially digested to completion with Pstl and Sail. The fragments were separated by non-denaturing electrophoresis on 4% acrylamide gels with Tris-borate-EDTA buffers, and the

101 bp Oligo 1-2 fragment was isolated as described above.

Next, plasmid pXER (prepared as described in Example 2), which contains two Pstl sites, was partially digested with Pstl. This partial digestion was carried out as follows. About 100 μg of pXER DNA was digested for 15 min. with 10 units of Pstl. and linearized plasmid was isolated by electrophoresis in 1.2% low melting agarose (Weislander, L. Anal. Biochem.. 98. 305 (1979)). The linearized vector was then digested with Sail, and the desired larger DNA fragment was isolated from agarose gels after electrophoresis. Agarose was obtained from International Biotechnologies, Inc. The effect of these steps was to delete the TK promoter in pXER.

The isolated Oligo 1-2 fragment and the Pstl- Sall-digested pXER were ligated (Sambrook et al., Molecular Cloning. A Laboratory Manual at p. 6.39), and the ligation mixture was used to transform E. coli DH5c_ as recommended by Bethesda Research Laboratories, Inc. IS. coli DH5o_ was obtained as transformation competent cells from Bethesda Research Laboratories. DH5α is a recA^" strain and is capable of α-complementation of the jS-galactosidase gene.

DNA samples from six colonies were digested with Clal. and colonies 3 and 4 were linearized as expected for the desired construction. Digestion with the combination of Pstl and Sail and with PvuII confirmed the identity of these clones as probable p2EREXER plasmids.

DNA from colony 3 was prepared in large scale, including isolation through two CsCl gradients. The DNA was sequenced by the dideoxynucleotide method using the T7 RNA polymerase promoter oligonucleotide (obtained from Pharmacia PL Biochemicals) as a primer. The sequence confirmed the identity of p2EREXER. The sequence of p2EREXER from the Sail site to the start codon of the XER sequence is as follows:

GTCGACCTAG ATCACATTAG GTCACAGTGA CCTTCAACAA GGTCACAGTG

ERE ERE

ACCTTGAACA ATCGATATAA ATACACTACA GTATCTCATC GAACACCGAG

TATA CGACCCTGCA GCGACCCGCT TAACACCGTC AACAGCGTGC CGCAAATTCC

|→ TK «-TK|→ER

GGCACAAACT AGCTGGAACA GTGGACAGCC ATG

[SEQ ID NO:38]

The EREs and TATA box are underlined. The sequences coding for the combined thymidine kinase and estrogen receptor 5' untranslated region are also indicated.

In conclusion, two plasmids were generated. The first is pTZERE, which contains cassettes of: 1) the 2 consensus EREs; 2) the TATA-TK sequence; and the fused 2 ERE-TATA-TK sequence. This vector is derived from pUC and grows to high copy number in E. coli. providing large yields of plasmid DNA. It contains the bacterial /3-lactamase gene so that bacteria containing the plasmid are resistant to ampicillin.

The second plasmid is the estrogen-regulated expression vector, p2EREXER, which encodes the Xenopus estrogen receptor. Like pTZERE, this vector is a high copy number plasmid and contains the S-lactamase gene for ampicillin selection. This vector provides for estrogen- regulated expression of the estrogen receptor in cells that contain low concentrations of estrogen receptor, either endogenously or by transfection with a plasmid that expresses estrogen receptor constitutively.

EXAMPLE 7: Expression of Xenopus Estrogen Receptor Using P2EREXER

Plasmids pXER-lAH (prepared as described in Example 4) and p2EREXER (prepared as described in Ex¬ ample 6) were used to co-transfect XLllO cells. Low levels of the mutant estrogen receptor were produced by pXER-lAH, which in turn elicited estrogen-dependent tran¬ scription of p2EREXER and production of estrogen receptor mRNA and protein.

A transient co-transfection of XLllO cells with these two plasmids was carried out as described in Ex¬ ample 2. Seventy-two hours after glycerol shock, the cells were harvested in 1.0 ml TEN (40 mM Tris, pH 7.4, 1 mM EDTA, 150 mM NaCl, no trypsin) by rocking the flasks occasionally during approximately 5 min. at room temperature. The flasks were also struck on the sides to dislodge the cells. The cells in the flask were brought up to a volume of 10 ml by addition of 9 ml of 0.6X PBS (10 mM sodium phosphate, 150 mM NaCl, pH 7.4) .

The cell suspension was transferred to a polypropylene tube (never polystyrene since the hormone may stick to polystyrene) . The cells were counted and sedimented by centrifugation at 1500 RPM, for 5 min., at 40C, and resuspended in 2.0 ml of phenol-red-free, serum- free medium.

Prior to use, ³H-estradiol-173 (specific radioactivity 92.5 Ci/mmol) obtained from New England Nuclear was dried under nitrogen and resuspended in ethanol to a final concentration of 2xl0'^s M (1.85 μCi of estradiol-17/3/μl) . This solution is designated "hot." The "cold" solution contains estradiol-17/3 at a concentration of 2xl0^"3 M.

To measure estrogen receptor levels, one ml. of each 2 ml sample was placed into each of two tubes. One tube designated "hot" received 2μl (3.7 μCi) of ³H-estradiol-17/3 (final concentration of 40 nM estradiol- 17/3) . The other tube (designated "hot + cold") received the same amount of ³H-estradiol - 11 β plus sufficient unlabeled estradiol to give a final concentration of 2x10^" ⁶ M estradiol (a 50-fold excess of unlabelled estradiol- 173) in the same volume (2μl) . The purpose of the hot + cold tubes was to measure non-specific binding which was subtracted from total binding to determine specific binding.

The cells were incubated with the estradiol-17/3 at 20OC for one hour, with gentle shaking every 15 minutes. The cells were pelleted by centrifugation at 1500 rpm as described above, and the cell pellet was resuspended in 5 ml of 0.6 X PBS plus 1% Tween 80. The washing procedure in PBS/Tween was carried out a total of 3 times.

The final cell pellet was resuspended in 100 μl of cold 0.6X PBS and transferred to a scintillation vial. The tube was rinsed with an additional 50 μl of 0.6X PBS which was added to the same scintillation vial. Then, 10 ml of scintillation fluor (250 ml. Triton X100, 750 ml. xylene, 5 g. paraphenyloxazole) were added to each tube, and the samples were counted. The amount of estrogen receptor per cell was calculated from the specific radioactivity of the ³H-estradiol - 11 β, the number of counts of labelled estradiol specifically bound to the cells, and the number of cells used in the assay. Since some estrogen receptor was produced from the pXER-lAH vector, and the purpose of this experiment was to determine the level of exprecsion of estrogen receptor as a result of the expression of plasmid p2EREXER, the level of estrogen receptor determined in a parallel transfec¬ tion using empty vector in place of p2EREXER was sub¬ tracted from the results obtained using p2EREXER.

In three separate transient transfections, estrogen receptor levels in the total cell population averaged 468,000 molecules per cell. In a transient transfection most of the cells do not take up the DNA. The upper limit of our estimate of transfection efficiency in these experiments was 10%. Thus, the level of expression in the cell actually expressing the DNA is at least four million molecules per cell. This represents on the order of 0.5 of total cell protein. Since estradiol is not added to the cells until after the transfection, the efficiency of transfection is identical in the cells which will eventually receive estradiol and in the cells which will not receive the hormone.

Claims

WE CLAIM:

1. A hybrid mRNA molecule comprising a sequence coding for a protein or polypeptide operatively linked to a sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor.

2. The hybrid mRNA molecule of Claim 1 wherein the stabilization sequence comprises the 3' untranslated region of a vitellogenin mRNA, or one or more portions of the 3' untranslated region sufficient to stabilize the hybrid mRNA.

3. The hybrid mRNA molecule of Claim 2 wherein the stabilization sequence comprises the 3' untranslated region of Xenopus laevis vitellogenin Bl mRNA, or one or more portions of the 3' untranslated region sufficient to stabilize the hybrid mRNA.

4. The hybrid mRNA molecule of Claim 3 wherein the stabilization sequence comprises the first 96 nucleotides at the upstream end of 3' untranslated region of Xenopus laevis vitellogenin Bl mRNA, or one or more portions thereof sufficient to stabilize the hybrid mRNA.

5. The hybrid mRNA molecule of Claim 1 wherein the stabilization sequence is located downstream of the coding sequence and comprises: a first sub-sequence which is the sequence responsible for the estrogen stabilization of the hybrid mRNA, and a second sub-sequence located downstream from the first sub-sequence which stabilizes the hybrid mRNA by providing for correct processing of the hybrid mRNA's 3' end.

6. The hybrid mRNA of Claim 5 wherein the first sub-sequence comprises the first 96 nucleotides at the upstream end of the 3' untranslated region of Xenopus laevis vitellogenin Bl mRNA, or a portion thereof sufficient to provide for estrogen stabilization of the hybrid mRNA.

7. The hybrid mRNA of Claim 6 wherein the second sub-sequence comprises the sequence

5'-UNUAAAUGUR UAUU-3' [SEQ ID NO:13]

wherein R is any purine and N is any nucleotide.

8. The hybrid mRNA molecule of Claim 1 further comprising a third sequence located between, and oper¬ atively linked to, the coding and stabilization sequences, the third sequence coding for a cleavage site.

9. A recombinant DNA molecule coding for the hybrid mRNA of Claim 1, 2, 3, 4, 5, 6, 7 or 8.

10. A vector comprising the recombinant DNA mole¬ cule of Claim 9 operatively linked to expression control sequences.

11. A host cell transformed with the vector of Claim 10.

12. A method of stabilizing an mRNA so as to obtain increased production of a protein or polypeptide coded for by the mRNA comprising: transforming a host cell with the vector of Claim 10; also transforming the host cell with a vector coding for estrogen receptor if the host cell does not normally produce estrogen receptor; and growing the transformed host cell in the presence of estrogen.

13. An expression vector comprising a synthetic promoter operatively linked to a DNA sequence coding for a hybrid mRNA, the promoter comprising one or more estrogen response elements, the type and number of estrogen response elements being sufficient so that transcription of the hybrid mRNA is estrogen-regulatable, the hybrid mRNA comprising a 5' untranslated region, a coding region coding for a protein or polypeptide of interest, and a 3' untranslated region which contains a sequence which is capable of stabilizing the hybrid mRNA in the presence of estrogen and estrogen receptor.

14. The vector of Claim 13 wherein the promoter comprises two consensus estrogen response elements operatively linked to a TATA box.

15. The vector of Claim 14 wherein the centers of the estrogen response elements are spaced about 20 nucleotides apart.

16. The vector of Claim 14 wherein the two consensus estrogen response elements are spaced about 20 nucleotides from the TATA box.

17. The vector of Claim 13 wherein the 5' untranslated region is one which provides for efficient translation of the hybrid mRNA.

18. The vector of Claim 17 wherein the 5' untranslated region is that of Herpes thymidine kinase mRNA, a vitellogenin mRNA, or a combination thereof.

19. The vector of Claim 13 wherein the 3' untranslated region is the 3' untranslated region of a vitellogenin mRNA, or a portion thereof sufficient to provide for estrogen stabilization of the hybrid mRNA.

20. The vector of Claim 19 wherein the 3' untranslated region is the 3' untranslated region of Xenopus laevis vitellogenin Bl mRNA, or a portion thereof sufficient to provide for estrogen stabilization of the hybrid mRNA.

21. The vector of Claim 13 wherein the stabilization sequence comprises: a first sub-sequence which is the sequence responsible for the estrogen stabilization of the hybrid mRNA, and a second sub-sequence located downstream from the first sub-sequence which stabilizes the hybrid mRNA by providing for correct processing of the hybrid mRNA's 3' end.

22. The vector of Claim 21 wherein the first sub¬ sequence comprises the first 96 nucleotides at the upstream end of the 3' untranslated region of Xenopus laevis vitellogenin Bl mRNA, or a portion thereof sufficient to provide for estrogen stabilization of the synthetic mRNA.

23. The vector of Claim 21 wherein the second sub¬ sequence comprises the sequence

5'-UNUAAAUGUR UAUU-3'

[SEQ ID NO:13] wherein R is any purine and N is any nucleotide.

24. The vector of Claim 13 wherein the coding region of the hybrid mRNA also codes for a signal peptide operatively linked to the protein or polypeptide of interest, the signal peptide being capable of providing for secretion of the protein or polypeptide from a host cell transformed with the vector.

25. The vector of Claim 24 wherein the signal peptide is a mutant parathyroid signal peptide.

26. A host cell transformed with the vector of Claim 13.

27. A method of obtaining enhanced production of a protein or polypeptide of interest comprising: transforming a host cell with the vector of Claim 13; also transforming the host cell with a vector coding for estrogen receptor if the host cell does not normally produce estrogen receptor; and growing the transformed host cell in the presence of estrogen.

28. A mutant estrogen receptor which is an estrogen receptor containing one or more copies of an amphipathic helix peptide.

29. The mutant estrogen receptor of Claim 28 wherein the amphipathic helix peptide has the sequence:

Glu Leu Gin Glu Leu Gin Glu Leu Gin Ala Leu Leu Gin Gin

5 10

Gin 15

[SEQ ID NO:15] .

30. The mutant estrogen receptor of Claim 29 wherein the specified sequence is located in the hinge region of the estrogen receptor.

31. The mutant estrogen receptor of Claim 28 which is the Xenopus laevis estrogen receptor containing one or more copies of the amphipathic helix peptide.

32. The mutant estrogen receptor of Claim 31 wherein the amphipathic helix peptide has the sequence:

Glu Leu Gin Glu Leu Gin Glu Leu Gin Ala Leu Leu Gin Gin

5 10

Gin 15

[SEQ ID NO:15]

33. The mutant estrogen receptor of Claim 32 wherein the specified sequence is located in the hinge region of the estrogen receptor.

34. A recombinant DNA sequence coding for the mutant estrogen receptor of Claim 28, 29, 30, 31, 32 or

33.

35. A vector comprising the recombinant DNA sequence of Claim 34 operatively linked to expression control sequences.

36. A host cell transformed with the vector of Claim 35.

37. A method of obtaining enhanced production of a protein or polypeptide of interest comprising: transforming a host cell with a vector containing a DNA sequence coding for the protein or polypeptide operatively linked to an estrogen-regulatable promoter; transforming the host cell with the vector of Claim 35; and growing the transformed host cell in the presence of estrogen.

38. A method of obtaining enhanced production of a protein or polypeptide of interest comprising: transforming a host cell with the vector of Claim 13; also transforming the host cell with the vector of Claim 35; and growing the transformed host cell in the presence of estrogen.

39. An estrogen-regulatable expression vector comprising a synthetic promoter operatively linked to a DNA sequence coding for a hybrid mRNA, the promoter comprising two consensus estrogen response elements operatively linked to a TATA box, the hybrid mRNA comprising a 5' untranslated region, a coding region coding for a protein or polypeptide of interest, and a 3' untranslated region.

40. The vector of Claim 39 wherein the centers of the estrogen response elements are spaced about 20 nucleotides apart.

41. The vector of Claim 39 or 40 wherein the estrogen response elements are spaced about 20 nucleotides upstream from the TATA box.

42. The vector of Claim 39 wherein the 5' untranslated region is a 5' untranslated region that provides for efficient translation.

43. The vector of Claim 39 wherein the coding region of the hybrid mRNA also codes for a signal peptide operatively linked to the protein or polypeptide of interest, the signal peptide being capable of providing for secretion of the protein or polypeptide from a host cell transformed with the vector.

44. A host cell transformed with the vector of Claim 39.

45. A method of obtaining enhanced production of a protein or polypeptide of interest comprising; transforming a host cell with the vector of Claim 39; also transforming the host cell with a vector coding for estrogen receptor if the host cell does not normally produce estrogen receptor; and growing the transformed host cell in the presence of estrogen.

46. The method of Claim 45 wherein the host cell is transformed with the vector of Claim 35 coding for a mutant estrogen receptor.