Extension of a Protein-Protein Interaction Surface to Inactivate the Function of a Cellular Protein

The present invention relates to the production of sequence-specific DNA binding proteins which function as eukaryotic transcription factors, i.e., transcription regulatory proteins. The invention more particularly relates to the generation of multimeric proteins having nucleic acid (i.e., DNA or RNA) binding domains in which the binding domain or protein interaction surface is engineered or modified to be acidic in nature. Such a nucleic acid binding protein having an acidic multimerization domain is capable of regulating the function of a target nucleic acid sequence or gene to which it is bound, thereby acting as potent dominant-negative regulators of gene transcription, and cell growth and proliferation.

Dominant-negative proteins are capable of inhibiting the binding of nucleic acid binding proteins, i.e., DNA binding proteins, such as transcription regulatory proteins, to target DNA sequences to inactivate gene function. (I. Herskowitz, 1987, Nature, 329:219-222).

The basic-region leucine zipper ("bZIP") DNA binding proteins are a family/class of nucleic acid binding proteins, which are eukaryotic transcription regulatory proteins that regulate transcription of genes by binding as dimers to specific DNA sequences. bZIP proteins characteristically possess two domains--a leucine zipper structural domain and a basic domain that is rich in basic amino acids (C. Vinson et al., 1989, Science, 246:911-916). The two domains are separated by a short segment known as the fork. Two bZIP proteins dimerize by forming a coiled coil region in which the leucine zipper domains dimerize. The basic regions then interact with the major groove of the DNA molecule at a specific DNA sequence site. The binding to DNA stabilizes the dimer. The dimerization and DNA-interaction event regulates eukaryotic gene transcription.

The leucine zipper motif is common to the primary structure of a number of DNA binding proteins, including the yeast transcription factor GCN4, the mammalian transcription factor CCAAT/enhancer-binding protein C/EBP, and the nuclear transforming oncogene products, Fos and Jun, and is characterized by a repeat of leucine amino acids every seven residues (i.e., a heptad repeat); the residues in this region can form amphipathic .alpha.-helices. The leucine-rich amphipathic helices interact and form a dimer complex, called a leucine zipper, at the carboxyl terminus (W. H. Landschultz et al., 1988, Science, 240:1759-1764; A. D. Baxevanis and C. R. Vinson, 1993, Curr. Op. Gen. Devel., 3:278-285), such that the dimerization region forms a coiled coil (E. K. O'Shea et al., 1989, Science, 243:538-542).

Another class of DNA binding proteins, which have similarities to the bZIP motif, are the basic-region helix-loop-helix ("bHLH") proteins (C. Murre et al., 1989, Cell, 56:777-783). bHLH proteins are also composed of discrete domains, the structure of which allows them to recognize and interact with specific sequences of DNA. The helix-loop-helix region promotes dimerization through its amphipathic helices in a fashion analogous to that of the leucine zipper region of the bZIP proteins (R. I. Davis et al., 1990, Cell, 60:733-746; A. Voronova and D. Baltimore, 1990, Proc. Natl. Acad. Sci. USA, 87:4722-4726). Nonlimiting examples of hHLH proteins are myc, max, and mad; myc and mad are known to heterodimerize.

The existence of the leucine zipper in the dimerization region of bZIP proteins allows for a high degree of biological control through the formation of both homodimers and heterodimers. For example, heterodimers are known to form between Fos and Jun (D. Bohmann et al., 1987, Science, 238:1386-1392), among members of the ATF/CREB family (T. Hai et al., 1989, Genes Dev., 3:2083-2090), among members of the C/EBP family (Z. Cao et al., 1991, Genes Dev., 5:1538-1552; S. C. Williams et al., 1991, Genes Dev., 5:1553-1567; and C. Roman et al., 1990, Genes Dev., 4:1404-1415), and between members of the ATF/CREB and Fos/Jun families (T. Hai and T. Curran, 1991, Proc. Natl. Acad. Sci. USA, 88:3720-3724). In general, dimerization of bZIP proteins depends upon the ability of both of the individual carboxyl terminal .alpha.-helices to line up in correct register with one another and to generate a symmetric coiled coil. This, in turn, places the amino terminal basic regions in a symmetric orientation, thus allowing them to interact with DNA (A. D. Baxevanis and C. R. Vinson, 1993, Curr. Op. Gen. Devel., 3:278-285). It has been shown that the ability of the helices within the coiled coil to find the proper register with respect to one another is controlled inherently by the individual helices themselves, and not by the placement of the basic region with respect to the DNA (W. Pu and K. Struhl, 1993, Nucleic Acids Research, 21:4348-4355). However, it will be appreciated that the generation of a symmetric coiled coil structure is not a mandatory requirement for the interaction of the multimerization or dimerization domains of various types of nucleic acid binding proteins.

The bZIP proteins are highly conserved throughout the eukaryotic kingdom and have been isolated and identified in yeast, plants, and mammals. These proteins mediate a variety of biological processes, including oncogenesis, memory, segmentation, and energy regulation (R. Boussoudan, 1994, Cell, 79:59-68; S. Cordes, and G. Barsh, 1994, Cell, 79:1025-1034; S. McKnight et al., 1989, Genes Dev., 3:2021-2024; and I. Verma, 1986, Trends in Genetics, 2:93-96.). Therefore, the ability to inhibit the activity of those proteins associated with oncogenesis or abnormal cell growth and proliferation, for example, is a desirable goal in the field.

In addition, inhibition of the production or function of other cellular proteins that are detrimental, or that influence unwanted or inappropriate phenotypes, in cells, tissues, and, ultimately, the whole organism, is an aim for practitioners in the art.

Of the nearly 70 bZIP proteins that have been identified to date, (H. Hurst, 1994, Protein Profiles, 1:123-168), most can be categorized into one of five major subfamilies on the basis of their DNA recognition properties and amino acid sequence similarities (P. F. Johnson, 1993, Mol. Cell. Biol., 13:6919-6930). These bZIP subgroups include the AP-1, CREB/ATF, C/EBP, PAR, and plant G-box proteins. The proteins in each subfamily recognize highly similar or identical DNA sites whose consensus sequences are 9- or 10-base pair palindromes composed of two 5-base pair half-sites. Binding sites for the various classes of bZIP proteins may differ either by their half-site sequences or their half-site spacing properties. AP-1 proteins, such as Fos, Jun, and GCN4 bind to a 9-base pair pseudopalindromic sequence that can be viewed as two half-sites that overlap by a single base pair, while the consensus binding sites for the other four families have directly abutted pairs of half-sites (N. B. Haas et al., 1995, Mol. Cell. Biol., 15:1923-1932). In addition, thyrotroph embryonic factor (TEF), a transcription factor expressed in the developing anterior pituitary gland, and the liver-enriched albumin D box-binding protein (DBP), (C. R. Mueller et al., 1990, Cell, 61:279-291), have been reported to constitute another class of bZIP proteins (D. W. Drolet et al., 1991, Genes Dev., 5:1739-1753).

bZIP proteins lacking the transactivation domain are naturally occurring dominant negatives that are generally produced by a genetic deletion of the transactivation domain (A. Clark and K. Dougherty, 1993, Biochem J., 296:521-541; P. Descombes and U. Schibler, 1991, Cell, 67:569-579; N. Foulkes et al., 1991, Cell, 64:739-749; and J. Yin et al., 1994, Cell, 79:49-58). These truncated bZIP proteins are able to dimerize and bind to DNA, and if overexpressed, can act as dominant negatives, presumably by competing with the endogenous bZIP protein for its promoter DNA binding site. Accordingly, the truncated bZIP proteins act by mass action to occlude the normal transactivator from the DNA. In addition, it is possible that the deletion of the transactivation domain could also produce a protein having increased, rather than decreased, DNA binding properties. If this were the case, then this type of truncated and naturally occurring dominant negative would not have to be overexpressed to generate particular phenotypes (A. Braiser and A. Kumar, 1994, J. Biol. Chem., 269:10341-10351).

Needed in the art are proteins, expressed and operative in cells, having dominant-negative function to control the transcription of genes or which regulate RNA production and function in a cell. Such expressed proteins can be used for regulating abnormal cell growth in a variety of eukaryotic organisms, including plants, animals, mammals, including humans, insects, microorganisms, and viruses. The present invention provides to the art proteins which can be modified in a particular way to control gene regulation. The particular type of modification may control gene function, for example, to inhibit abnormal or cancer cell growth and proliferation, to inhibit pathogenic diseases caused by microorganisms, particularly eukaryotic microorganisms, such as yeast, and the like, or viruses and may be used as therapeutics for treating pathological diseases and cancer.

The present invention provides multimeric nucleic acid (i.e., DNA or RNA) binding proteins, such as transcription regulating proteins, which have been engineered to contain in their multimerization or protein interaction domains at least one amino terminal acidic amino acid residue. The acidic nature of such nucleic acid binding proteins affects the binding of the proteins to other proteins, e.g., forming heterodimers or heteromultimers, and, ultimately the binding of the proteins to a target DNA or RNA sequence or gene. Nucleic acid binding proteins containing an extension of acidic amino acid residues have an extended protein interaction surface or multimerization or dimerization interface. DNA binding proteins are a particular example of nucleic acid binding proteins suitable for the invention. RNA binding proteins are also suitable for use in the invention. In accordance with the invention, the acidic nature of the protein increases the stability of heteromultimeric or heterodimeric complexes that are formed.

It is an object of the present invention to provide an acidic extension to a protein-protein interaction surface or dimerization interface to inactivate the function of a cellular protein. In accordance with the invention, proteins which are useful as drugs, inhibitory molecules, or growth-controlling agents or compounds are provided which, when used both in vitro and in vivo, can inhibit the expression and activity of cellular proteins, the effects of which can be harmful, deleterious, and even lethal, to cell growth and survival.

It is another object of the invention to add acidic amino acid residues to create a multimerization or dimerization surface or extension onto a multimeric complex, particularly a protein, polypeptide or peptide having basic regions that bind to nucleic acids, such as DNA or RNA. More particularly, the acidic extension appended onto the multimeric protein can replace the basic region of such proteins to create molecules that regulate and control cell growth.

It is another object of the invention to regulate gene transcription and expression by providing suitable dominant negative mutant molecules having an acidic phenotype that specifically heterodimerize with native proteins to disrupt the normal action of the native proteins in vivo, thereby causing the subsequent inactivation of cellular gene products.

It is another object of the invention to create dominant negative transcription regulatory proteins by extending the protein-protein interaction surface, making it acidic in nature, to inactivate the function of a cellular protein by specifically and stoichiometrically displacing the native protein from its normal binding with DNA and by inhibiting transactivation and, ultimately, gene transcription and protein production.

It is another object of the invention to provide improved, genetically engineered dominant negative transcription factor proteins that stoichiometrically inhibit the DNA binding of DNA binding proteins, such as bZIP and bHLH proteins, and structurally related types of proteins, to limit the pleiotropic effects normally associated with naturally occurring dominant negative overexpression. The dominant negative mutant proteins of the invention are created by the addition of acidic residues the N-terminus of a multimeric protein, most preferably a DNA binding protein.

It is another object of the invention to provide methods and rationally-designed constructs suitable for producing and expressing the dominant negative proteins described herein for the specific inactivation of cellular gene products and for use in gene therapy techniques.

It is another object of the invention to utilize the acidic extension of the engineered DNA binding proteins to stabilize a variety of different basic regions of proteins and to create robust dominant negative protein members of various families of DNA binding proteins.

It is yet another object of the invention to provide transgenic animals harboring at least one genetically engineered plasmid construct or vector containing a DNA sequence encoding a DNA binding protein that is acidic in nature to control gene expression in a tissue specific manner.

It is a further object of the invention to provide transgenic animals harboring at least one genetically engineered plasmid construct or vector containing a DNA sequence encoding a DNA binding protein that is acidic in nature which behaves as a dominant negative to the wild type protein and provides viable phenotypes to evaluate and assess the in vivo effects of the protein. Such animals may also be used for rational drug design (i.e., with the dominant negatives considered as drugs) and for testing and evaluating additional or supplemental treatments, drugs, therapies, and the like, and for ameliorating or alleviating the produced transgenic phenotypes.

It is another object to provide dimeric and multimeric nucleic acid binding proteins of plant and animal origin and having acidic multimerization domains allowing the nucleic acid binding protein to bind to a target DNA or RNA sequence, i.e., a specific gene, thereby regulating the function of the gene to which it is bound. Such proteins are expressed, as described herein, from expression eukaryotic or prokaryotic vector constructs molecularly engineered to contain isolated DNA sequences encoding a nucleic acid binding protein having an acidically extended dimerization or multimerization domain.

It is a further object of the invention to control the regulation of a gene through the type of acidic modification that is made to a nucleic acid binding protein, thereby providing therapeutic applications where the target gene or DNA or RNA sequence is present in abnormal or diseased cells and tissues and their normal counterparts. The engineered, acidic nucleic acid binding proteins provide tools for use in cancer therapeutics, diseases caused by eukaryotic microorganisms, for example, yeast, protozoans, algae, parasites, or by viruses, as well as tools for drug development, rational drug design, and drug and gene therapies.

Further objects and advantages afforded by the invention will be apparent from the detailed description hereinbelow.

FIG. 1 shows thermal melts of chimeric proteins containing the CEBP leucine zipper and three bZIP basic regions derived from CEBP, VBP and GBF (CEBP, VBP-CEBP, and GBF-CEBP) mixed with the dominant negative 4heptad-F. The results show that the 4heptad acidic extension interacts similarly with three, different basic proteins, i.e., the ellipticity for all three is 65.degree. C.

FIG. 2 shows the protein sequence of the 4heptad extension (top line), SEQ ID NO:53, and a modified version of the 4heptad extension (new extension:4heptad) which has one amino acid change (N to L), SEQ ID NO:54, which produces a better dominant negative to the proteins Jun and CREB. Specifically, an asparagine in the a position of the first acidic heptad has been changed to a leucine, resulting in a more potent dominant negative to bZIP proteins containing a hydrophobic amino acid in the same a position, e.g., Jun, CREB, opaque, and ATF2. The basic region of the bZip protein JUN is shown (SEQ ID NO:55) and the basic region of the bZip protein C/EBP is shown (SEQ ID NO:56).

FIG. 3 demonstrates thermal melting curves of CREB. As shown, the melting temperature of unmodified CREB homodimerization is 47.degree. C. (Tm=47.degree. C.) and the melting temperature of CREB containing 4 heptad amino acid residue repeats appended at the amino terminus of the CREB protein (4heptadCREB homodimerization) is 22.degree. C. (Tm=22.degree. C.). The melting temperature of dimerization of the CREB protein with 4heptadCREB is 53.degree. C. (Tm=53.degree. C.). An acidically extended CREB was designed to contain a new acidic extension (New4hepCREB) in which the asparagine was changed to a leucine as described in FIG. 2. As shown, CREB+New4heptadCREB have a melting temperature of 69.5.degree. C. (open circles). This dramatic increase in thermal stability relative to the asparagine-containing acidic extension suggests that different classes of acidic extensions can be used depending on the exact bZIP protein to be inactivated.

FIG. 4 demonstrates thermal melting curves showing the extent of heterodimerization between the oncogenic DNA binding proteins of the bZIP class, c-Fos and c-Jun, as a consequence of engineering or modifying the expressed c-Fos protein to contain an amino terminal extension of acidic amino acid residues. As shown in FIG. 4, the mixture of unmodified c-Fos and c-Jun melted as 50.degree. C. (closed circles). An engineered c-Fos protein having the basic region deleted (0heptad-Fos) increased the melting temperature of the Fos-Jun complex to 53.degree. C. (open squares). An engineered c-Fos protein having the basic region replaced with an acidic region containing four heptad repeats (4heptad-Fos) increased the melting temperature of the Fos/Jun complex to 61.degree. C. (open circles). The plasmid carrying DNA encoding 4heptad-Fos has been deposited with the American Type Culture Collection, 10801 University Blvd., Manassas, Va. 20110, under ATCC Designation No. 97583. An engineered c-Fos protein in which one amino acid in the acidic region was changed from an asparagine to a leucine, as described for FIG. 2, called new4heptadFos (or N4heptad-Fos), resulted in a higher melting temperature of 72.degree. C., and more stable heterodimerization with c-Jun (closed squares). The plasmid carrying DNA encoding N4heptad-Fos has been deposited with the American Type Culture Collection, 10801 University Blvd., Manassas, Va. 20110 under ATCC Designation No. 97584. Also as shown in FIG. 4, homodimerization of unmodified c-Jun had a melting temperature of 29.4.degree. C., and homodimerization of acidically modified c-Fos (having four acidic repeats appended at the amino terminus of the c-Fos protein) had a melting temperature of 25.degree. C.

FIGS. 5A and 5B: FIG. 5A shows that the specificity of the basic zipper region (bZIP) of the protein CEBP and the acidic extension on the N-terminus of a leucine zipper of the protein 4heptadFos is determined by the leucine zipper for DNA binding proteins having basic and acidic zipper regions (ZIP classes of proteins). A ZIP DNA binding protein having an acidic extension appended N-terminally is an acidically modified protein and is termed aZIP herein. As shown in FIG. 5A, mixing a bZIP CEBP protein and an aZIP 4heptadFos protein with incompatible zippers produces no interaction, i.e., 4heptadFos does not interact with CEBP.

FIG. 5B shows that mixing a bZIP VBP protein and an aZIP 3heptadF (CEBP specific) protein also produces no heterodimeric interaction. In FIGS. 5A and 5B, the solid lines represent the simple sum of the two homodimer curves. The fact that the actual mixture gives identical results to the sum curve demonstrates that no heterodimers are formed.

FIG. 6 shows the results of transient transfections of human cells with expression constructs designed to contain cytomegalovirus (CMV) promoter and DNA sequences encoding various DNA binding proteins engineered to have acidic amino acid extensions resulting in expressed acidically modified DNA binding proteins. The constructs harbored DNA sequences encoding the cloned and isolated Fos gene having no acidic extensions (0hep-fos), or the Fos gene sequence modified to encode from 1-4 acidic extensions appended thereto (i.e., 1hep-fos, 2hep-fos, 3hep-fos, and 4hep-fos, respectively). Other constructs contained DNA encoding unmodified Fos (bZIP-Fos); CREB having 4 acidic extensions (4hep-CREB); VBP having 4 acidic extensions (4hep-VBP); and Jun having 4 acidic extensions (4hep-Jun). In these experiments, human hepatoma cells (HepG2) were transiently transfected with up to three plasmids, e.g., a chloramphenicol acetyl transferase (CAT) expression plasmid driven by a single AP1 cis transactivation element, and different DNA binding protein-encoding DNAs. CMV566 AND CMV500 are plasmids containing the cytomegalovirus promoter driving the expression of DNA binding proteins that contain at their N-termini, either the Hemagglutinin or the FLAG epitopes, respectively (Example 4). The histogram presents the extent of transactivation observed in the presence of different combinations of expressed transactivator and various acidically extended, expressed dominant-negative (DN) DNA binding proteins. In the absence of any DN, AP1 transactivation was from approximately 10 to 30 fold. Transactivation was inhibited by 1-3heptad-Fos. The results in FIG. 6 show that an AP1 cis element is activated by Jun transactivation. This activation is inhibited by the acidic extension on the Fos zipper protein (aFos). The acidic extension appended to other leucine zipper proteins did not inhibit AP1 transactivation, thus indicating the specificity of the inhibition.

FIG. 7 shows relative CAT activity in an additional cell system (Jurkat) demonstrating that an expressed Fos protein containing an acidic extension (aFos) is a better dominant negative than an expressed Fos protein containing no acidic extension (bFos). Jurkat cells (a B cell model) have a high level of endogenous Fos/Jun (AP1) activity. Cells were transiently transfected with two plasmids (1 .mu.g), namely, a reporter construct containing AP1 cis elements and a second construct containing the CMV500 promoter driving the expression of either a truncated form of Fos (bZipFos) or the Fos leucine zipper containing the 4heptad acidic extension (aFos). AP1 activity was induced by the addition of 12-O-tetradecanoylphorbol (TPA) which results in a 40-fold induction of the reporter gene. The transactivation of the reporter gene occurs because of endogenous AP1 activity. As demonstrated in FIG. 7, expression of the bZipFos protein by the bZipFos construct results in dominant negative activity; however, expression of the aZipFos protein by the aZipFos construct is more efficacious and has a more complete inhibition of AP1 activity.

FIG. 8 represents the Jurkat cell system as described in FIG. 7, with the addition that the leucine zipper ("zip") construct alone is also transfected into cells. Jurkat cells were transfected with two plasmids and stimulated with TPA. The efficacy of three potential dominant negatives containing the Fos leucine zipper (i.e., Bzip, zip, and Azip) were compared. These three dominant negatives are the Fos Bzip domain without a transactivation domain (Bzip), the Fos leucine zipper with no basic region (zip), and the Fos leucine zipper containing the 4heptad acidic extension (Azip). The data show that expressed aFos inhibits AP1 transactivation significantly better than does an expressed Fos zipper protein or an expressed bFos protein in the Jurkat cell system.

FIG. 9 shows the results of appending an acidic extension onto the dimerization domain of Max, a DNA binding protein of the bHLH class. The 2heptadMax proteins used in the heterodimerization studies shown in FIG. 9 are expressed Max proteins having acidic extensions as follows:

DPDLEKEAEELEQENAELELEDSF, called 2heptadMax(783); (SEQ ID NO:1). The plasmid carrying DNA encoding this acidically extended Max protein has been deposited with the American Type Culture Collection, under ATCC Designation No. 97585.

DPDLEKEAEELEQENAELEELEDSF, called 2heptadMax(784); (SEQ ID NO:2). The plasmid carrying DNA encoding this acidically extended Max protein has been deposited with the American Type Culture Collection, under ATCC Designation No. 97582.

DPDLEKEAEELEQENAELEELEDSF, called 2heptadMax(785); (SEQ ID NO:3). The plasmid carrying DNA encoding this acidically extended Max protein has been deposited with the American Type Culture Collection, under ATCC Designation No. 97586. AcidMax is an expressed Max protein having appended N-terminally an acidic extension as follows: DPDEEEDDEEELEELEDSF; (SEQ ID NO:4). The plasmid carrying DNA encoding this acidically extended Max protein has been deposited with the American Type Culture Collection, under ATCC Designation No. 97581.

0heptadMax is an expressed Max protein having an acidic extension as follows:

DPDLEELEDSF (SEQ ID NO:5).

As can be seen from FIG. 9, an expressed Max protein having an appended polyglutamic acid sequence (i.e., AcidMax) heterodimerized better with c-Myc than did the expressed 2heptadMax proteins having the above-described acidic extensions.

FIGS. 10A-10D are schematic depictions of the bZIP protein C/EBP dimerization both with and without DNA. The relative positions of the activation region, basic region, and leucine zipper within the protein are shown. The NH.sub.2 (amino) terminus of the protein is the activation domain. The dashed line between the amino acids in the d positions of the leucine zipper represent a physical interaction. FIG. 10A: C/EBP dimerization. The basic region, presented as a bold line, is not helical. FIG. 10B: C/EBP binding to DNA. The presence of DNA induces the basic region to form an .alpha.- which increases the stability of the complex approximately 10-fold. FIG. 10C: Heterodimerization between wild type C/EBP and a C/EBP mutant having an F zipper which preferentially heterodimerizes with the C/EBP leucine zipper. FIG. 10D: Heterodimerization between C/EBP and 0heptad-F (i.e., the F zipper as shown in FIG. 10C, but having the basic region removed).

FIGS. 11A-11C demonstrate that acidic helical extension of the leucine zipper stabilizes dimerization with wild type bZIP protein. FIG. 11A shows the amino acid sequences of the 4heptad acidic helical extension, SEQ ID NO:57, and the three different basic regions of bZIP proteins that were examined, i.e., C/EBP, SEQ ID NO:58; VBP, SEQ ID NO:59; and GBF1, SEQ ID NO:60. The first leucine position of the zipper, the invariant asparagine and arginine of the basic region, and the tryptophan in the VBP bZIP protein are presented in bold-faced type. The coiled coil nomenclature of the basic region extending from the leucine zipper is indicated below with the hydrophobic a and d positions in bold type. FIG. 11B depicts a proposed heterodimeric coiled coil structure between the 3heptad acidic helical extension and the C/EBP basic region; both nonequivalent sides are shown. The f position amino acids are shown in both views. The amino acids in the 3heptad extension that have been changed from the parental C/EBP basic region are shown in bold type. The first leucine position of the zipper and the invariant asparagine and arginine of the basic region are stippled. The designed g.revreaction.e' and g.revreaction.a' electrostatic interactions are indicated by a black bar between the amino acids. The right-most d amino acid shown, indicated by an arrow, is the first d position of the leucine zipper (T for C/EBP and L for 3heptad) The coiled coil heptad letter designations (a,b,c,d,e,f,g) are shown to the right. The supercoiling of the two helices is not depicted. FIG. 11C is a coiled coil helical wheel diagram of the interaction shown in FIG. 11B, with a view from the C-terminus toward the N-terminus. The coiled coil sequence reads outward from the C- to N-termini around the wheel, starting at the e position following the first d position of the leucine zipper. Presumed electrostatic interactions between g.revreaction.e' and g.revreaction.a' are indicated.

FIG. 12 shows the oligomerization of C/EBP determined in the presence or absence of 3heptad-F using sedimentation equilibrium analysis at 37.degree. C. C/EBP alone (open squares) behaves as a single species with an apparent MW=33,000 Daltons, suggesting a homodimer structure. An equimolar mixture of C/EBP and 3heptad-F (closed circles) has an apparent MW=28,000 Daltons, indicating heterodimer formation.

FIGS. 13A and 13B: FIG. 13A shows the circular dichroism (CD) spectra of the C/EBP bZIP domain (4 .mu.M), 0heptad-F (4 .mu.M), and the equimolar mixture of C/EBP and the F zipper containing different length acidic extensions (0,1,2,3,4,) (4 .mu.M+4 .mu.M). The minima at 208 and 222 nm indicate .alpha.-helices. The sum of the spectra of C/EBP and 0heptad-F, which would be observed if the two samples did not interact, is shown as a dotted line. The increase in the 222 nm signal for the different mixtures suggests an increase in .alpha.-helical content. The noise below 210 nm is due to the absorbance of dithiothreitol (DTT). FIG. 13B shows CD thermal melting curves at 222 nm of the samples shown in FIG. 13A. The fitted curve through each of the data sets was used to calculate T.sub.m. The mixtures were more stable than C/EBP alone suggesting a stabilizing interaction between the two proteins. The bar shows the amount of ellipticity at 222 nm generated by extending the leucine zipper one heptad on both proteins of the heterodimer.

FIGS. 14A and 14B show that acidic extension inhibits the binding of C/EBP to DNA. FIG. 14A: 3heptad-F inhibits C/EBP DNA binding. The left panel shows a gel retardation assay of C/EBP binding to a specific DNA probe 28 bp in length. Protein was diluted four-fold in the successive lanes. The right panel shows the same assay as the left, except that 3heptad-F was added at the same concentration as C/EBP, resulting in a total absence of C/EBP DNA binding. FIG. 14B: Displacement of VBP-C/EBP from DNA by five dominant negatives (DNs) assayed by fluorescence. Fluorescence of a single tryptophan located in the basic region of VBP-C/EBP increased upon DNA binding (minutes 5-10). The addition of 1.1 molar equivalents of the F zipper or of the four different acidic extension caused a decrease in fluorescence. After 20 minutes, 10 more molar equivalents were added, causing an additional displacement of VBP-C/EBP from the DNA. 4heptad-F, shown in bold, causes the most dramatic displacement of VBP:C/EBP. The bottom line shows that mixing of VBP-C/EBP with 4heptad-F results in a new base line. This is not observed for the other four mixtures which return to the VBP-C/EBP-only baseline.

FIG. 15 demonstrates that acidic extension inhibits C/EBP transactivation. Human hepatoma cells (HepG2) were transiently transfected with three plasmids, a chloramphenicol acetyl transferase (CAT) expression plasmid driven by a single C/EBP cis element, the C/EBP transactivator, and different DNs. The histogram presents the extent of transactivation observed in the presence of different combinations of transactivator and dominant-negatives. In the absence of any DN, expressed C/EBP transactivated a single C/EBP site approximately 10 fold. This transactivation was partially inhibited by a C/EBP lacking the transactivation domain (.DELTA.C/EBP), the heterodimerizing zipper (0heptad-F), and the acidic extension appended to the regular C/EBP zipper (3heptad-C/EBP). Transactivation was completely inhibited by expressed 3heptad-F. The final two bars of the histogram show transactivation of an expressed C/EBP protein containing an alternate leucine zipper, the GCN4 zipper. This protein was able to transactivate a C/EBP containing promoter and was not inhibited by 3heptad-F.

FIGS. 16A and 16B are schematic diagrams of plasmid vectors used in the transfection studies as described. FIG. 16A shows the vector pET3b.seq into which the DNA encoding DNA binding proteins (with and without acidic extensions) are inserted between the BamHI and the HindIII sites. FIG. 16B shows the vector CMV500-junbZip comprising the CMV promoter, and the FLAG and JunbZip sequences inserted between the NcoI and HindIII sites. The sequence range for this vector is from 1 to 5809. Restriction enzyme cleavage sites, the number of cutting sites per enzyme, and their sequence positions are as follows: BamHI: 3 cuts at positions 965, 1571, and 3615; EcoRI: 1 cut at position 2106; HindIII: 1 cut at position 1245; NcoI: 4 cuts at positions 610, 898, 2342, and 3077; NdeI: 2 cuts at positions 484 and 928; and XhoI: 2 cuts at positions 1562 and 3628. FIG. 16C shows the vector CMV500 -CREBbZip comprising the CMV promoter, and the FLAG and CREBbZip sequences inserted between the NcoI and HindIII sites. The sequence range for this vector is from 1 to 5746. Restriction enzyme cleavage sites, the number of cutting sites per enzyme, and their sequence positions are as follows: BamHI: 3 cuts at positions 965, 1508, and 3552; EcoRI: 1 cut at position 2043; HindIII: 1 cut at position 1182; NcoI: 4 cuts at positions 610, 898, 2279, and 3014; NdeI: 2 cuts at positions 484 and 928; and XhoI: 2 cuts at positions 1499 and 3565.

FIG. 17 depicts the isolated nucleotide sequence (SEQ ID NO:6) and the translated amino acid sequence (SEQ ID NO:7) of a CREB protein having an appended acidic extension. A 5' BamHI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, e.g., the pET3b.seq vector of FIG. 16A.

FIG. 18 depicts the isolated nucleotide sequence (SEQ ID NO:8) and the translated amino acid sequence (SEQ ID NO:9) of CMV500-CREBbZIP protein which has no appended acidic extension. A 5' NcoI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, e.g., the CMV500-CREBbZip vector of FIG. 16C.

FIG. 19 depicts the isolated nucleotide sequence (SEQ ID NO: 10) and the translated amino acid sequence (SEQ ID NO: 11) of CMV500-4heptadCREB (called New4hepCREB; FIG. 3) protein which has a 4heptad appended acidic extension as described in FIG. 2. A 5' NcoI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as a vector similar to that shown in FIG. 16C.

FIG. 20 depicts the isolated nucleotide sequence (SEQ ID NO:12) and the translated amino acid sequence (SEQ ID NO:13) encoding human c-Fos protein, as used in the experiments described and shown in FIGS. 4-6. A 5' BamHI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as a vector similar to that shown in FIG. 16A.

FIG. 21 depicts the isolated nucleotide sequence (SEQ ID NO:14) and the translated amino acid sequence (SEQ ID NO:15) of CMV500-FosbZIP(MO) protein, as used in the experiments described and shown in FIG. 6. A 5' NcoI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as a vector similar to that shown in FIG. 16C. The nucleotide sequence as shown contains nucleotide sequence encoding the FLAG epitope, nucleotide sequence encoding .phi.10, and the nucleotide sequence encoding FosbZIP.

FIG. 22 depicts the isolated nucleotide sequence (SEQ ID NO:16) and the translated amino acid sequence (SEQ ID NO:17) encoding the 4heptadFos protein, containing the 4heptad acidic extension as described in FIG. 2 and as used in the experiments described and shown in FIGS. 4, 5A, 6, and 7.

FIG. 23 depicts the isolated nucleotide sequence (SEQ ID NO:18) and the translated amino acid sequence (SEQ ID NO:19) of CMV500-4heptadFos leucine zipper protein, containing the 4heptad acidic extension as described in FIG. 2 and as used in the experiments described and shown in FIG. 6. A 5' NcoI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as a vector similar to that shown in FIG. 16C. The sequence as shown contains sequence encoding the FLAG epitope, sequence encoding .phi.10, sequence encoding the 4heptad acidic extension, and sequence encoding the Fos leucine zipper.

FIG. 24 depicts the isolated nucleotide sequence (SEQ ID NO:20) and the translated amino acid sequence (SEQ ID NO:21) of CMV500-JunbZIP protein. A 5' NcoI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as the vector shown in FIG. 16B. The sequence as shown contains sequence encoding the FLAG epitope, sequence encoding .phi.10, sequence encoding the JunbZip protein.

FIG. 25 depicts the isolated nucleotide sequence (SEQ ID NO:22) and the translated amino acid sequence (SEQ ID NO:23) encoding AcidMax, the bHLH Max protein containing an acidic extension comprising a tract of polyglutamic acids appended to the dimerization domain of Max as shown.

FIG. 26 depicts the isolated nucleotide sequence (SEQ ID NO:24) and the translated amino acid sequence (SEQ ID NO:25) encoding 2heptadMax(783), the bHLH Max protein containing two amphipathic acidic extensions, as described in FIG. 9 and in Example 6, appended to the dimerization domain of Max as shown.

FIG. 27 depicts the isolated nucleotide sequence (SEQ ID NO:26) and the translated amino acid sequence (SEQ ID NO:27) encoding 2heptadMax(784), the bHLH Max protein containing two amphipathic acidic extensions, as described in FIG. 9 and in Example 6, appended to the dimerization domain of Max as shown.

FIG. 28 depicts the isolated nucleotide sequence (SEQ ID NO:28) and the translated amino acid sequence (SEQ ID NO:29) encoding 2heptadMax(785), the bHLH Max protein containing two amphipathic acidic extensions, as described in FIG. 9 and in Example 6, appended to the dimerization domain of Max as shown.

FIG. 29 depicts the isolated nucleotide sequence (SEQ ID NO:30) and the translated amino acid sequence (SEQ ID NO:31) encoding murine c-Myc bHLH protein, as used in the experiments described and shown in FIGS. 4-6. Murine, rat and human c-myc sequences are highly conserved and virtually identical. A 5' BamHI restriction site and a 3' HindIII restriction site serve as sites for insertion of the nucleotide sequence into an expression vector, such as a vector similar to that shown in FIG. 16A.

FIG. 30 is a schematic diagram of a plasmid vector (approximately 11.9 kb) which was linearized by cutting with the restriction enzymes HindIII and NotI, as indicated by the asterisks (*). The resulting linearized plasmid DNA contained the 422 or aP2 promoter (approximately 7.6 kb) contained within the HindHIII-PstI site as shown. The 422 or aP2 promoter is known to be associated with the regulation of the expression of adipocyte-specific genes during adipocyte (fat cell) differentiation. The 422/aP2 gene encodes the adipose fatty acid-binding protein. The 5' flanking sequence of the 422/aP2 gene contains regulatory regions for adipose-specific expression (Bernlohr et al., 1985, J. Biol. Chem., 260:5563-5567; Spiegelman et al., 1983, J. Biol. Chem., 258:10083-10089; Cook et al., 1988, Proc. Natl. Acad. Sci. USA, 85:2949-2953). The plasmid also contained, operably linked therewith, nucleotide sequence (approximately 0.3 kb) encoding the 3heptadC/EBP dominant negative (Flag.phi.10 3heptadF), in accordance with the invention as described hereinabove, contained within the KpnI-SmaI site as shown, and a fragment of approximately 1 kb which contained the polyadenylation site from the early region of SV40 and a splice-donor acceptor sequence from the small t antigen intron from SV40 (Gorman et al., 1982, Mol. Cell. Biol., 2:143-190). The insert containing the 422/aP2 promoter, the Flag.phi.10 3heptadF C/EBP, the polyadenylation and splice sites was cloned between the HindIII and BamHI sites of the polylinker of the vector Bluescript KS+ (Stratagen). The linearized plasmid was used to transfect mice for transgenic mouse production and breeding, employing techniques routinely known in the art.

FIG. 31 shows the results of the transgenic mouse experiments using the plasmid DNA construct described and shown in FIG. 30. Two litter mates resulting from successful breeding of transgenic animals are shown: the smaller mouse near the left side of the cage exhibits the transgenic phenotype and has the scruffy, scrawny appearance of a skinny or thin mouse, relative to the normal-sized, non-skinny littermate on the right, which does not carry the transgene. The thin, scruffy transgenic mouse as shown in the figure was found to possess several copies of the transgene as assayed by Southern Blot hybridization.

The present invention provides the production of novel multimeric nucleic acid (i.e., DNA or RNA) binding proteins and transcription regulatory proteins which function as potent dominant negative proteins to regulate gene expression and to inhibit cellular protein production and function. Those skilled in the art will appreciate that DNA and RNA are also referred to as polynucleic acids to which the proteins of the invention can bind. These novel proteins are produced by adding acidic amino acid residues (e.g., glutamic acid, aspartic acid) to a dimeric or multimeric protein, e.g., bZIP or bHLH transcription regulatory proteins, such that an engineered acidic extension, comprised of acidic amino acids, is appended onto the protein. For example, a basic region of a multimeric DNA or RNA binding protein can be replaced with an acidic region to yield a dominant negative function. Alternatively, for example, a plurality of acidic amino acids can be appended to a nucleic acid binding protein, preferably onto the N-terminus of the protein, to afford an acidic nature to the resulting protein. As will be appreciated by those having skill in the art, and as exemplified and described herein, the acidically extended nucleic acid binding proteins of the invention are produced using conventional molecular techniques by manipulating isolated DNA sequence encoding a nucleic acid binding protein such that the DNA sequence encoding the dimerization or multimerization domain of the protein contains nucleotides which encode one or more acidic amino acid residues. Also, DNA encoding nucleic acid binding proteins having acidic extensions, or parts of such DNA, e.g., DNA encoding an acidic extension, can be synthetically produced using DNA synthesis techniques conventionally known in the art.

The proteins to which an acidic region has been extended or added can be derived from a variety of protein types, nonlimiting examples of which include DNA or RNA binding proteins, transcription regulatory proteins, such as the bZIP and the bHLH classes of proteins, e.g., fos, jun, opaque, DBP, ATF2, CREB, Max and Myc (e.g., N-myc, C-myc, and v-Myc). Several members of the bZIP class of transcription factors, or structurally similar types of proteins, have functional roles in activating the transcription of genes, the protein products of which are subsequently expressed and function during the growth response of cells. Some of these transcription regulatory proteins are closely linked to tumorigenesis, oncogenesis, and to mediating the production of proteins involved in a variety of aspects of cell growth and differentiation. Such transcription factors are prime candidates for conversion into dominant negatives to control the activity of their wild type counterparts and, ultimately, to regulate or inactivate the function of particular cellular protein products. It will be apparent to those in the art that the terms wild type protein, native protein, naturally occurring protein and non-mutant or non-mutated protein are synonymous as used herein. It will also be understood by those in the art that the nucleic acid binding proteins of the invention may be derived from plants, animals, including mammals and humans, microorganisms such as yeast and fungi, protozoans, algae, and parasites, as well as from RNA and DNA viruses.

More specifically, those skilled in the art are aware that non-mammalian eukaryotic organisms, such as plants, have nucleic acid binding proteins, which are involved with plant physiology. GBF-1 is an example of a plant nucleic acid binding protein of the bZIP class of DNA binding proteins, which binds to the most divergent cis element of the bZIP proteins. Acidic extensions to plant nucleic acid binding proteins are encompassed by the invention to modulate or regulate gene transcription. Indeed, an acidic extension appended onto GBF-1 resulted in stable heterodimerization (see Example 11 and Table 2). In other organisms, preferably eukaryotic and mammalian organisms, tissue specific regulation can be effected by appending an acidic extension onto a molecule such as a protein or polypeptide in accordance with the invention to cause controlled regulation, in particular cell and tissue types, of a cellular protein that is associated with the physiology and endogenous growth of the organism.

Applications for use of the acidically extended proteins in accordance with the invention include the growth of specific organs and tissues by selective regulation of proteins expressed during plant development. For example, storage proteins may modified to achieve healthier and more vigorously-growing plants, vegetables and crops. As another nonlimiting example, plant or crop growth can be augmented or otherwise modified by appending acidic extensions in accordancewith the invention onto suitable endogenous regulatory molecules to increase carbohydrate production, for example, in potato tubers, and the like. The use of acidic extensions appended onto specific proteins, polypeptides, and cellular regulatory molecules is intended to have broad application throughout the plant kingdom. For example, it is envisioned that both monocotyledonous and dicotyledonous plants can be used, as well as stem and leaf vegetables (e.g., broccoli, lettuce, spinach, cabbage), fruit and seed vegetables (e.g., tomato), fiber crops and cereals (e.g., corn, oats, wheat), and forest and ornamental crops (e.g., cotton). In addition, other uses include controlled regulation of proteins to either enhance or decrease the growth of specific plant organs and tissues by reducing the expression or effectiveness of endogenous growth-associated proteins, for example, in fruits (e.g., grapes, citrus fruits, apples, pears, apricots, and the like), or to regulate leaf, root, stem or petiole growth (e.g., cabbage, spinach, celery, beets, soybeans, sugarcane, flower stalks). Plant tissues whose growth and development may be regulated in accordance with the invention can result in the modification of various traits, such as durability, size, succulence, texture, and longevity. The control of the expression of certain genes in accordance with the invention may reduce the expression of growth regulatory genes or genetic control elements that regulate gene expression to result in such nonlimiting and exemplary uses as the dwarfing of stems for durability or enhanced mechanical stability, genetic pruning, stunting, or the elimination of undesirable plant organs.

Examples of candidate nucleic acid binding and transcription regulating proteins suitable for creating dominant negatives to use as inhibitory and regulatory compounds or drugs in accordance with the invention include any dimeric or multimeric DNA binding protein to which an acidic extension may be appended or which can be manipulated to contain acidic amino acid residues. Nonlimiting examples of such proteins are the bZIP family of transcription regulatory proteins having basic and leucine-zipper domains, especially, proteins in the Jun and Fos families, LRF-1 (liver regenerating factor 1), (J. C. Hsu et al., 1991, Proc. Natl. Acad. Sci. USA, 88:3511-3515); C/EBP, GCN4, DBP, CHOP-10, GBF-1; and the bHLH family having basic domains and helix-loop-helix motifs, such as ID, MyoD1, E12, c-myc, n-myc, i-myc, max, AP-4, TFE3, USF, and FIP (A. D. Baxevanis and C. R. Vinson, 1993, Curr. Op. Gen. Devel., 3:278-285); as well as other structurally similar proteins.

In a general sense, the invention includes the production of acidically extended nucleic acid binding multimeric complexes, e.g., proteins, polypeptides, or peptides, which bind to a nucleic acid, i.e., DNA or RNA. The products are nucleic acid binding proteins having an acidic extension of amino acids in the dimerization or multimerization domain of the multimeric complex. The multimeric complex will usually have a basic region, which may be replaced with the acidic extension to create molecules which are capable of controlling cell growth and proliferation when such molecules are introduced into and expressed in cells. This is achieved by engineering the nucleotide sequences encoding nucleic acid binding proteins such that nucleotides encoding acidic amino acid residues are present in N-terminal extensions to the basic region or multimerization or dimerization domains of the proteins ultimately expressed.

Dominant negatives (DN) to both bZIP and bHLH proteins containing basic regions, with and without a leucine zipper motif, have been developed and genetically engineered in accordance with the invention by altering and replacing the normal basic domain of the wild type protein with one or more acidic regions. Expression constructs containing DNA sequences which produce the acidically modified DN proteins that function to regulate or control cell growth (e.g. by inhibiting gene transcription or gene activation) when introduced into cells are described.

In one embodiment of the invention, an acidic extension is appended onto an expressed nucleic acid binding protein. Particularly useful nucleic acid binding proteins for use in the invention are DNA binding proteins. In general, one or more acidic amino acid residues comprise the appended acidic extension. The extension can comprise repeating domains or regions, wherein each region or domain comprises a sequence of amino acids, some or all of which are acidic amino acid residues. Each domain or region comprising the appended extension can comprise from two to about one hundred total amino acids, with each domain or region engineered to contain at least one, more preferably two or more, acidic amino acids. An appended acidic extension generally comprises two or more acidic domains or regions comprising a plurality of amino acids appended to the DNA binding protein. The number of acidic domains or regions comprising the acidic extension appended to the protein, as well as the total number of amino acid residues comprising each domain or region, can be modified as desired or required so as to produce an optimum dominant negative or inhibitory function. In general, an appended extension having two or more acidic regions (e.g., heptads) with at least one (e.g., two to five) acidic amino acids in each region, results in more potent dominant negative or inhibitory function. As a particular but nonlimiting example, an acidic extension can contain four heptad amino acid regions or domains comprising twenty-eight total amino acid residues, with at least two amino to four amino acids in each region or domain being acidic (FIG. 2).

In accordance with the invention, a sequence of amino acid residues, at least one of which, preferably two or more, is an acidic amino acid residue, can be appended onto a DNA binding protein to yield an extended protein interaction surface or an extended dimerization interface that is acidic in nature. The invention encompasses an acidic extension which comprises a sequence of acidic amino acid residues, all of which are acidic, e.g., glutamic acid and/or aspartic acid. The invention also encompasses an acidic extension which comprises a sequence of acidic amino acid residues, some of which are acidic. As a general guide, when all of the amino acids in the extension are not acidic, then about 1% to about 98% of the amino acids in the appended sequence can be acidic. Thus, the appended extension can comprise a sequence of amino acids which are not necessarily grouped in regions or domains, some or all of which are acidic residues. As a simple, but nonlimiting example, the appended acidic extension can comprise fifteen to twenty glutamic acid residues appended onto the multimerization or dimerization domain of the nucleic acid binding protein. Accordingly, the total number of amino acids in the sequence of amino acid residues extended onto the protein can be of variable length, for example, from about two to one hundred amino acid residues, preferably from about three to fifty amino acid residues, more preferably from about three to thirty amino acid residues, most preferably from about four to twenty-eight amino acid residues, some or all of which are acidic amino acids. It is to be understood that the length of the acidic extension may impact upon the specificity of dimerization and DNA binding function of the resulting DNA binding protein. In addition, it will be appreciated by those in the art that acidic extensions of varying length (i.e., long and short extensions) can be appended onto a protein to increase or decrease the potency of the resulting dominant negative protein; i.e., to make a stronger or more toxic "drug" comprising the acidically-extended DNA binding proteins of the invention, the number of acidic amino acids in the extension can be increased as required. Those skilled in the art will appreciate that the number of acidic amino acids can be changed to change the strength of an acidically extended protein to suit the needs of the practitioner. For example, a dominant negative DNA binding protein having a short acidic extension (e.g., three to thirteen amino acids) can be created, as well as a dominant negative DNA binding protein having a long acidic extension (e.g., twenty to one hundred amino acids). In general, a longer acidic extension may be expected to provide increase potency, toxicity, or stability to the dominant negative DNA binding protein to which it is appended. The ability to change the strength of the drug afforded by the acidically extended DNA binding proteins of the invention by increasing the number of acidic amino acids comprising the extension allows the resulting proteins to have differential strength as regulators and controllers of cell growth and proliferation.

In another embodiment, the acidically extended DNA binding protein is an expressed bZIP transcription regulatory protein having a basic region and a leucine zipper. In accordance with the invention, the protein is engineered to contain a designed acidic amino acid sequence which extends N-terminally from the leucine zipper region of the expressed bZIP protein. The extension of amino acids can comprise from two to one hundred amino acid residues, preferably from about three to fifty amino acid residues, more preferably from about three to thirty amino acid residues, some or all of which are acidic residues, thereby making the appended amino acid extension acidic in nature. Alternatively, the acidic extension can comprise a sequence of amino acids, all of which are acidic in nature, e.g., glutamic acid, in which the acidic amino acid sequence comprises from about two to eighty amino acids, preferably from about three to fifty amino acids, and more preferably about five to thirty amino acids. It is to be understood that all of the amino acid residues comprising the acidic extension of the bZIP protein can be acidic amino acids or that at least one amino acid residue in each of the amino acid regions comprising an appended extension is an acidic amino acid (e.g., at least one amino acid residue is an acidic residue in each of four heptad repeats comprising the extension, FIG. 2). More frequently, two or more amino acid residues within a given region of an appended extension of the bZIP protein are acidic (FIG. 2; FIG. 11A)

In another embodiment, the extension of amino acids can be amphipathic. As described above, the extension can also be designed to occur in one or more repeating domains or regions. Preferred are acidic extensions comprising one or more amino acid domains or regions, wherein each domain or region of amino acid residues comprises at least one acidic amino acid residues, preferably two or more acidic residues, among the amino acids comprising each region. For example, an acidic extension containing four repeating regions with each region comprising a sequence of seven amino acids comprised of acidic amino acid residues is a 4heptad repeat appended onto the protein in accordance with the invention.

As described, the acidic amino acid sequence can extend for one or more heptad repeats from the N-terminus of the protein and can have an N-terminal cap, which may be comprised of three glycine residues (e.g., DP-GGG). As described herein (Example 1F), the N-terminal cap may also be comprised of other sequences, namely, DP- (aspartic acid and proline alone); DP-D (aspartic acid and proline and aspartic acid); DP-EE (aspartic acid and proline and two glutamic acid residues), and DP-DEEE (aspartic acid and proline and three glutamic acid residues) to replace the DP-GGG cap.

In another embodiment of the invention, the acidically extended DNA binding protein is an expressed transcription regulatory bHLH protein having a basic region and a hexlix-loop-helix structure. In accordance with the invention, the bHLH protein is engineered to contain a designed acidic amino acid sequence which extends N-terminally from the multimerization or dimerization domain of the protein. In these proteins the N-terminal DNA binding amino acids are deleted and replaced with the acidic extension. The acidic extension comprises repeating domains or regions comprised of a sequence of amino acids having at least one acidic amino acid in each repeating region. Alternatively, the acidic extension comprises a sequence of amino acids, some or all of which are acidic, e.g., glutamic and/or aspartic acid. Such an acidic extensions can comprise from two to eighty amino acids, preferably from three to fifty amino acids, and more preferably four to thirty amino acids. It is also to be understood that an acidic extension can simply comprise a sequence of amino acids appended onto the N-terminal multimerization or dimerization domain of a nucleic acid binding protein, wherein some or all of the amino acid residues in the appended extension are acidic. As but one example, an acidic extension appended to a bHLH protein is comprised of fifteen repeating glutamic acid residues. In another example, as described for the bHLH protein, Max (FIG. 9 and Example 6), it was found that an expression construct resulting in the expression of a Max protein having a stretch of polyglutamic amino acid residues appended N-terminally thereto resulted in superior heterodimerization with c-myc protein. Plasmid vectors harboring the DNA sequence encoding acidically modified Max proteins were created using the N-terminal DPD amino acid sequence, which is a BamHI cloning site (Example 6).

The designed acidic region appended onto the protein, e.g., a leucine zipper protein, functioned as a dominant negative to transcription factors comprised of basic regions and leucine zipper regions; basic regions and helix-loop-helix structures, and other DNA binding proteins that are capable of dimerization and/or of binding to a target DNA sequence or gene and regulating or controlling the function of the DNA target sequence or gene to which the DNA binding proteins are bound. In general, specificity is derived from the dimerization domain in the dominant negatives. It will be appreciated that the production of nucleic acid binding proteins which are acidic in nature is applicable to protein-protein interacting systems other than those particularly described and exemplified herein. As but one nonlimiting example, the bHLH family of transcription factors are considered to be good candidates for extending the dimerization motif, because the basic region is an extension of the dimerization region, in a manner that is similar to that of the bZIP protein family.

In another embodiment, the present inventors have designed constructs which contained and expressed isolated DNA encoding a bZIP leucine zipper-containing protein (called the F zipper or heterodimerizing zipper) that preferentially heterodimerized with the wild type bZIP transcription factor C/EBP, a heat-stable DNA-binding protein present in rat liver (W. H. Landschultz et al., 1988, Genes Dev., 2:786-800; C. R. Vinson et al., 1993, Genes Dev., 7:1047-1058). The F-zipper was created by placing charged amino acids in the e and g positions of the coiled coil structure, thereby resulting in the preferential formation of heterodimers over homodimers, as described in C. R. Vinson et al., 1993, Ibid. Interestingly, the heterodimer formed by the complexation of native C/EBP and the F zipper constructed without a basic region (i.e., C/EBP:0heptad-F) was more stable than the wild type or native C/EBP homodimer in the absence of DNA. However, the heterodimer was not as stable as the native C/EBP homodimer bound to DNA, and thus, an equimolar amount of the heterodimerizing F zipper without an acidic amphipathic extension (0heptad-F) could not displace native C/EBP from its binding to DNA. Thus, the earlier-produced F zipper was not considered to be a good or useful dominant negative at stoichiometric concentrations.

In accordance with the present invention, the creation of novel multimeric nucleic acid binding proteins containing an acidic extension allows the extension to stretch the protein-protein multimerization or dimerization interface or surface. A heterodimer complex ultimately formed can be stabilized by over one hundred fold (i.e., 2.5 kcal/mol). For proteins having basic regions, such as bZIP and bHLH proteins, acidic extension allows the extension to increase the protein-protein interaction interface into the basic region of the protein. The acidic extension also can increase the stability of interaction of a heterodimeric complex between the engineered protein containing the acidic extension and the native protein. The increase in heterodimer stability makes the dominant negatives of the invention effective stoichiometric competitors of the wild type or native protein products.

In accordance with the invention and as a particular nonlimiting example, the stability of the heterodimer formed between the wild type C/EBP protein and the F zipper protein was increased by appending a designed three-heptad-long amphipathic .alpha.- to the N-terminus of the F zipper (abbreviated 3heptad-F herein), (FIGS. 11A-C). In this way, the acidic extension could electrostatically mimic DNA, thus providing the C/EBP basic region with an alternative interaction surface. In the particular case of C/EBP, the extension of the C/EBP leucine zipper into the basic region would create new a and d positions containing hydrophobic amino acids. For other proteins, such as the bHLH proteins and the like, the acidic extension provides an alternative or extended interaction surface which ultimately controls the extent of binding of a protein to a target DNA sequence (i.e., a gene).

For bZIP proteins, the amino acid sequence of the acidic extension forming a protein interaction surface for binding with the basic region of C/EBP is graphically presented from both a side and an end view (FIGS. 11B-11C). The N-terminal extension was designed to prevent unfavorable stearic or electrostatic interactions with the basic region, and to form attractive electrostatic and hydrophobic interactions that would result in the formation of an appropriate interaction surface. Commercially-available CPK molecular models were used to aid in the molecular design. Leucine was placed in all the d positions of the extension to drive potential hydrophobic interactions. Because the native C/EBP basic region contains two asparagines in the a position, asparagine was placed in the acidic extension to create favorable a.revreaction.a' interactions between the asparagines of the basic region and the acidic extension, as reported for GCN4 leucine zipper structures (T. Ellenberger et al., 1992, Cell, 71:1223-1237). In the second and fourth a positions of the native C/EBP basic region is an arginine; the arginine in the second position is conserved in all bZIP proteins. In accordance with the invention, an alanine was placed in the opposing a position of the acidic extension to avoid possible stearic clashes. A glutamic acid was put in the g position to create an attractive g.revreaction.a' electrostatic interaction, to parallel the type of interaction that occurs in the zipper of the c-Fos:c-Jun heterodimer (M. Glover and S. Harrison, 1995, Nature, 373:257-261). In addition, glutamic acid was placed in the e and g positions to create favorable g.revreaction.e' interactions. Additional glutamic acids were included in the b and c positions to help electrostatically neutralize the heterodimeric structure. The resulting protein comprising the three heptad N-terminal extension and the F zipper is referred to as 3heptad-F. The hyphen designates the junction between the basic region and the leucine zipper. For heterodimers, the names of the individual proteins comprising the heterodimer are separated by a colon, e.g., C/EBP:3heptad-F.

Also in accordance with the invention, acidic extensions were appended onto the dimerization domain of expressed bHLH proteins, e.g., Max, and heterodimerization with expressed c-myc protein was observed (FIG. 9, Example 6). Increased heterodimerization stability was found when the acidic extension was appended onto the bHLH protein in any of three different orientations, thereby showing the universal nature of the acidic extension to extend protein-protein interaction interfaces of the nucleic acid binding domains of nucleic acid binding proteins, particularly DNA binding proteins.

In accordance with the invention, it was found that the addition of one or more heptads of acidic amino acid sequences onto the N-terminus of a leucine zipper protein stabilized heterodimer formation, as determined by circular dichroism (CD) spectra and thermal melting assays as described herein. This acidic amino acid extension provided an extended protein interaction surface with the basic region of the wild type or native protein containing the leucine zipper and allowed the creation of a dominant negative to the native protein that stoichiometrically displaced the native protein from DNA and inhibited transactivation by the native protein.

It will be appreciated by those in the art that at least one (1heptad) and preferably two to four or more heptads, i.e., 2heptad, 3heptad, 4heptad, and the like (a maximum number can be routinely and empirically determined by those having skill in the art) can be appended to the N-terminal region of appropriate DNA binding proteins to achieve optimal heterodimer binding stability of the proteins produced in accordance with the present invention. However, it is also noted that longer acidic extensions may lead to nonspecificity as a result of an increased interaction with more of the same or similar nucleic acid binding proteins. Such nonspecificity may be alleviated by expressing less protein and/or by employing fewer acidic sequence extensions. A combination of these strategies will allow for the generation of a dominant negative with minimal nonspecific or pleiotropic effects.

In accordance with the invention, plasmid constructs or vectors are provided for the production and expression of the dominant negative nucleic acid binding proteins described. These nucleic acid binding proteins can be expressed, if desired, in a variety of expression systems, particularly eukaryotic and mammalian systems, e.g., yeast, insect, human, hamster, mouse, rat, and the like, using reagents and methods known to those in the art. The constructs for use in both in vitro and in vivo systems are designed to contain at least one promoter, an enhancer sequence (optional, for mammalian expression systems), and other sequences as necessary or required for proper transcription and regulation of gene expression (e.g., transcriptional initiation and termination sequences, origin of replication sites, polyadenylation sequences). As will be appreciated by those skilled in the art, the selection of the appropriate vector and plasmid components for proper transcription, expression, and isolation of proteins produced in eukaryotic and prokaryotic expression systems is known and routinely determined and practiced by those having skill in the art.

Ultimately, the constructs containing the nucleic acid sequences coding for the dominant negative nucleic acid binding proteins of the invention are introduced into cell types of interest, having the appropriate milieu for transcription of the gene(s) whose transcription regulatory proteins and products are to be inhibited or negatively regulated by the dominant negatives of the invention. The constructs can be designed to contain the appropriate and necessary DNA elements for expression of the dominant negative protein in a given cell type, if desired.

For example, the expression of a dominant negative nucleic acid binding protein (e.g., DNA binding protein) having an acidic amino acid extension can be placed under the control of promoters such as viral promoters, e.g., cytomegalovirus (CMV), Rous sarcoma virus (RSV), phosphoglycerol kinase (PGK), thymidine kinase (TK), or the .alpha.-actin promoter. Further, a regulated promoter, such as the glucocorticoid response element (GRE) of mouse mammary tumor virus (MMTV), would confer inducibility by glucocorticoids (V. Chandler et al., 1983, Cell, 33:489-499). Alternatively, tissue-specific promoters or regulatory elements can be used (G. Swift et al., 1984, Cell, 38:639-646), non-limiting examples of which include the following: the N-CAM promoter (specific for brain and central nervous system); the PIT-1 promoter (pituitary specific transcription factor); the crystalline promoter (specific for regulating protein expression in the lens of the eye); the keratin promoter (specific for regulating protein expression in the skin); the albumin promoter (liver-specific); the alpha- or beta-globin promoters (specific for red blood cells); the Ig enhancer (specific for B lymphocytes); the T cell receptor .alpha.- or .beta.-promoters (specific for T cells); the insulin promoter (specific for pancreatic cells); the gastrin promoter (specific for cells of the stomach); the cardiac actin promoter (specific for heart); the tropomyosin promoter (specific for skeletal muscle); and the lactalbumin promoter and the whey acidic protein (WAP) promoter (A. C. Andres et al., 1987, Proc. Natl. Acad. Sci. USA, 84:1299-1303; C. W. Pittius et al., 1988, Proc. Natl. Acad. Sci. USA, 85:5874-5878), (specific for expression in breast tissue).

Regulated and tissue specific promoters allow the expression of the mutant dominant negative phenotype to be conditional, in general, thereby being exhibited if and when expression of the gene is induced. This would allow the propagation of dominant negative nucleic acid binding proteins of the invention, in the absence of expression of the dominant negative phenotype, so that the inducible dominant negatives can be used in a manner similar to that of temperature sensitive mutations. In addition, since the modification made to DNA binding proteins are dominant and the phenotype is exhibited in the presence of functional wild type genes, the consequences of inactivating the function of genes present in multiple copies can be assessed without having to inactivate each copy of the gene (I. Herskowitz, 1987, Nature, 329:219-222).

As mentioned above, those skilled in the art will appreciate that a variety of promoters, enhancers, and genes are suitable for use in the constructs of the invention, and that the constructs will contain the necessary initiation, termination, and control sequences for proper transcription and processing of the gene of interest when the construct is introduced into a cell. The constructs may be introduced into cells by a variety of gene transfer methods known to those skilled in the art, for example, conventional gene transfection methods, such as calcium phosphate co-precipitation, liposomal transfection, microinjection, electroporation, and infection or viral transduction. In addition, it is envisioned that the invention can encompass all or a portion of a viral sequence-containing vector, such as those described in U.S. Pat. No. 5,112,767 to P. Roy-Burman and D. A. Spodick, for targeted delivery of genes to specific tissues. The choice of the method is within the competence of the skilled practitioner in the art. It will be apparent to those skilled in the art that one or more constructs carrying DNA sequences for expression in cells can be transfected into the cells such that expression products are subsequently produced in and/or obtained from the cells.

In the constructs of the invention, the nucleic acid sequences encoding nucleic acid binding proteins with acidic extensions can be obtained and isolated from natural sources or by synthetic means. For example, using conventionally known techniques in molecular biology, the appropriate DNA sequences can be excised and isolated and purified from genomic clones or from cDNA clones generated from the diverse species known to express nucleic acid binding proteins and the like; the nucleotide sequences can be genetically engineered to encode amino acid residues comprising the acidic extensions as described, and the resulting nucleic acid can then be ligated with at least one of the appropriate or desired segments of DNA sequence to construct a plasmid or vector, comprising, for example, a promoter, an enhancer, terminator site, and polyadenylation site. Alternatively, the nucleic acid sequences can be synthesized for use in the constructs according to all or part of the sequences provided in the sequence ID numbers herein, and in the accompanying figures, by conventional techniques of DNA synthesis, such as the phosphite triester chemistry method (for example, see U.S. Pat. No. 4,415,732 to Caruthers et al.; and Sinha, N. D. et al., 1984, Nucl. Acids Res., 12:4539-4557).

In another aspect of the invention, the acidically extended dominant negatives constructed in accordance with the invention are tested for their in vitro and in vivo functions in inactivating the relevant family of transcription factors in an in vitro or in an in vivo milieu. With regard to in vitro studies, the ability of expressed nucleic acid binding proteins produced in accordance with the invention to inhibit cellular proteins that are associated with cell growth can be assayed using the transient transfection and transformation techniques as described. Such techniques and assays can also be used to screen the engineered, acidically extended nucleic acid binding proteins of the invention for use as candidate anti-cancer or disease drugs and in therapeutic applications.

With regard to in vivo studies, constructs as described for the transient transfection experiments (Examples 1F and 8C) are used to construct transgenic animals, e.g., mice, as described in Example 13. The plasmid constructs carrying the CMV promoter to drive expression of the DNA sequence coding for the DNA binding proteins, e.g., bZIP and bHLH proteins (e.g., 4heptadFos, 4heptadCREB, 2heptadMax, and 3heptad-F proteins), are introduced into animals. Thereafter, the effects of the expression of the acidically extended dominant negatives are monitored in animals harboring the DN transgenes. The animals are assayed for the presence of the transgene and for the expression of the DN protein in cells. Alternatively, DNA sequences encoding the acidically extended dominant negative proteins of the invention are constructed in a vector containing a tissue specific promoter to target the expression of this heterologous gene to specific tissues of the transgenic animals, along with the other components required for a complete vector construct.

Another aspect of the invention provides tissue specific constructs which will target those tissues in which the acidic extension-containing nucleic acid binding proteins are to be expressed. Such constructs are designed to contain tissue-specific promoters such as those enumerated hereinabove, and a DNA sequence encoding a DNA binding protein having acidic extensions, e.g., the 4heptadFos, 4heptadCREB, 3heptadMax, or other transcription factors, the expression of which will be driven by the tissue-specific promoter, and the appropriate initiation and termination sites for transcription, and polyadenylation sites, if necessary, for expression in a given cell and tissue type. Thus, the invention allows the selective expression and the determination of the function of DNA binding protein family members, such as those of the bZIP and bHLH families, in individual tissues by means of the production of transgenic organisms and via genetic therapy approaches known and used in the art.

For gene therapy applications of the present invention, constructs designed to carry the acidically extended nucleic acid binding proteins, e.g. DNA binding transcription regulatory proteins, RNA binding proteins, and the like, as described herein can supply the described DN protein inactivation function to a variety of cell types, including stem cells, differentiated cells, germ cells, and oncogenic or transformed cells. Supplying such an inactivating or inhibitory function by the dominant negatives of the invention can lead to and/or cause the suppression of neoplastic growth in a variety of cell types, especially those cells which have been growth-altered, for example, as a consequence of the interaction and activity of Fos-Jun dimerization (Examples 2 and 4 and FIGS. 1 and 4) or Myc-Max dimerization (Examples 6 and 7 and FIG. 9). In addition, because of the divergent biological roots of the DNA binding proteins, the DN constructs afforded by the invention are applicable to the areas of human biology, non-human mammalian biology, insect biology, and plant biology. In the latter case, the dominant negatives can modulate the expression of valuable gene products related to plant stress, anti-fungal agents, insecticides, and pesticides.

The dominant negatives, or a dominant negative-functioning part or fragment thereof, in accordance with the invention, can be introduced into a cell in a vector such that the DNA segment remains extrachromosomal and will be expressed by the cell from the extrachromosomal location in the cell. Vectors for the introduction of genes and DNA segments suitable for both recombination and extrachromosomal maintenance are known in the art. Cells transformed or transfected with the DN-containing constructs of the invention can be used as model systems to study cancer or tumor remission.

Gene transfer systems known in the art may be useful in the practice of the invention. Both viral and non-viral methods are suitable. Numerous viruses have been used as gene transfer vectors, including papovaviruses, e.g., Simian Virus 40, (SV40), adenovirus, vaccinia virus, adeno-associated virus, herpes viruses, including Herpes Simplex Virus, (HSV) and Epstein Barr Virus, (EBV), and retroviruses of avian, murine, and human origin. As is appreciated by those in the art, most human gene therapy protocols have been based on disabled murine retroviruses. Recombinant retroviral DNA can also be employed with amphotropic packaging cell lines capable of producing high titer stocks of helper-free recombinant retroviruses (e.g., R. Cone and R. Mulligan, 1984, Proc. Natl. Acad. Sci. USA, 81:6349).

Receptor-mediated gene transfer methods allow targeting of the DNA in the construct directly to particular tissues. This is accomplished by the conjugation of DNA (frequently in the form of covalently-closed supercoiled plasmid) to a protein ligand via poly-lysine. The appropriate or suitable ligands are selected on the basis of the presence of the corresponding ligand receptors on the cell surface of the target cell or tissue type. These ligand-DNA conjugates can be injected directly into the blood, if desired, and are directed to the target tissue where receptor binding and DNA-protein complex internalization occur. Co-infection with adenovirus to disrupt endosome function can be used to overcome the problem of intracellular destruction of DNA.

Nonviral gene transfer methods known in the art include chemical techniques, such as calcium phosphate co-precipitation, direct DNA uptake and receptor-mediated DNA transfer, and mechanical means, such as microinjection and membrane fusion-mediated liposomal transfer. In addition, viral-mediated gene transfer can be combined with direct in vivo gene transfer using liposomes, thereby allowing the delivery or the viral vectors to tumor cells, for example, and not to surrounding non-proliferating cells. The retroviral vector producer cell line can be injected directly into specific cell types, e.g., tumors, to provide a continuous source of viral particles, such as has been approved for use in patients afflicted with inoperable brain tumors.

An approach that combines biological and physical gene transfer methods utilizes plasmid DNA of any size combined with a polylysine-conjugated antibody specifically reactive with the adenovirus hexon protein. The resulting complex is bound to an adenovirus vector. The trimolecular complex is then used to infect cells. The adenovirus vector allows efficient binding to the cell, internalization, and degradation of the endosome before the coupled DNA can be damaged.

The described constructs may be administered in the form of a pharmaceutical preparation or composition containing a pharmaceutically acceptable carrier and a physiological excipient, in which preparation the vector may be a viral vector construct, or the like, to target the cells, tissues, or organs of the recipient organism of interest, including human and non-human mammals. The composition may be formed by dispersing the components in a suitable pharmaceutically-acceptable liquid or solution such as sterile physiological saline or other injectable aqueous liquids. The amounts of the components to be used in such compositions may be routinely determined by those having skill in the art. The compositions may be administered by parenteral routes of injection, including subcutaneous, intravenous, intramuscular, and intrasternal. Other modes of administration include intranasal, intrathecal, intracutaneous, percutaneous, enteral, and sublingual. For injectable administration, the composition is in sterile solution or suspension or may be emulsified in pharmaceutically- and physiologically-acceptable aqueous or oleaginous vehicles, which may contain preservatives, stabilizers, and material for rendering the solution or suspension isotonic with body fluids (i.e. blood) of the recipient. Excipients suitable for use are water, phosphate buffered saline, pH 7.4, 0.15 M aqueous sodium chloride solution, dextrose, glycerol, dilute ethanol, and the like, and mixtures thereof. Illustrative stabilizers are polyethylene glycol, proteins, saccharides, amino acids, inorganic acids, and organic acids, which may be used either on their own or as admixtures.

It is also envisioned that the nucleic acid binding proteins of the invention can be used to screen compounds using the expressed acidically extended nucleic acid binding proteins or a functional binding fragment thereof in a variety of drug screening methods. For example, an engineered and expressed DNA binding protein or fragment employed in such a test may either be flee in solution, affixed to a solid support, or displayed on a cell surface. One drug screening technique utilizes eukaryotic or prokaryotic host cells which are stably transformed with recombinant polynucleotides expressing the protein or fragment, preferably in competitive binding assays. Such cells, either in viable or fixed form, can be used in standard binding assays. For example, the formation of complexes between a nucleic acid binding protein or fragment and the agent being tested may be quantified, or the degree to which the formation of a complex between a nucleic acid binding protein or fragment and a known ligand is interfered with by the agent being tested can be examined.

Thus, the present invention provides methods of screening for drugs comprising contacting such an agent with a nucleic acid binding protein of the invention or fragment thereof and assaying (i) for the presence of a complex between the agent and the nucleic acid binding protein or fragment, or (ii) for the presence of a complex between the nucleic acid binding protein or fragment and a ligand, by methods well known in the art. In such competitive binding assays, the nucleic acid binding protein or fragment is typically labeled. Free nucleic acid binding protein or a fragment is separated from that present in a protein:protein complex, and the amount of free (i.e., uncomplexed) label serves as a measure of the binding of the agent being tested to the nucleic acid binding protein or of its interference with the nucleic acid binding protein:ligand binding, respectively.

Another technique for drug screening provides high through screening for compounds having suitable binding affinity to the DNA binding proteins and is described in detail in Geysen, European Patent Application 84/03564, published on Sep. 13, 1984. In brief, large numbers of different small peptide test compounds are synthesized on a solid substrate, such as plastic pins or some other surface. The peptide test compounds are reacted with the DNA binding protein and washed. After washing, bound DNA binding protein is then detected by methods well known in the art. Alternatively, purified DNA binding protein can be coated directly onto plates for use in the aforementioned drug screening techniques. However, non-neutralizing antibodies to the polypeptide can be used to capture antibodies to immobilize the DNA binding protein on the solid phase.

The invention also contemplates the use of competitive drug screening assays in which antibodies capable of specifically binding the DNA binding protein compete with a test compound for binding to the DNA binding protein or fragments thereof. In this manner, the antibodies can be used to detect the presence of any peptide which shares one or more antigenic determinants with the DNA binding protein.

Rational drug design is also envisioned as a use for the present invention. In general, the goal of rational drug design is to produce structural analogs of biologically active polypeptides and proteins of interest or of small molecules with which they interact (e.g., agonists, antagonists, inhibitors) in order to fashion drugs which are, for example, more active or stable forms of the protein or polypeptide, or which, for example, enhance or interfere with the function of a protein or polypeptide in vivo. See, e.g., Hodgson, 1991. In one approach, the three-dimensional structure of a protein of interest (e.g., an acidically extended DNA binding protein of the invention) is first determined by x-ray crystallography, by computer modelling or most typically, by a combination of approaches. Less often, useful information regarding the structure of a polypeptide may be gained by modeling based on the structure of homologous proteins. The production and design of nucleic acid binding proteins having acidic amino acid extensions can lead to the expression, production and detection of products having an increased affinity for binding to cognate cellular proteins than do the normal, unmodified cellular counterparts. Upon complexation with its naturally occurring cognate cellular nucleic acid (e.g., DNA) binding protein, an acidically extended nucleic acid binding protein of the invention can inhibit or prevent the function of its cognate protein and ultimately affect cellular gene transcription and expression. Thus, the products designed and created in accordance with the invention can yield more potent dominant negatives of wild type or naturally occurring proteins.

Accordingly, drugs may be designed which have, e.g., improved DNA binding function, activity or stability or which act as more potent dimerization molecules, or as inhibitors, agonists, antagonists, etc. of DNA protein binding activity. By virtue of molecularly synthesizing the DNA binding protein sequences, sufficient amounts of the proteins may be made available to perform such analytical studies as X-ray crystallography. In addition, DNA sequence of the DNA binding protein will allow computer modeling techniques to be employed in place of or in addition to x-ray crystallography.

Active nucleic acid binding protein molecules can be introduced into cells by microinjection or by use of liposomes, for example. Alternatively, some functioning molecules may be taken up by cells, actively or by diffusion. Extracellular application of the gene product of the acidically extended DNA binding proteins of the invention may be sufficient to affect tumor growth, provided that the gene products become localized in the nucleus. The supply of molecules with the activity of an acidically extended DNA binding protein should lead to partial reversal of the neoplastic state. Other molecules with acidically extended nucleic acid binding protein activity (for example, peptides or functional portions thereof) may also be used to effect such a reversal. Modified polypeptides having substantially similar function are also used for peptide therapy.

The examples herein are meant to exemplify the various aspects of carrying out the invention and are not intended to limit the invention in any way.

A. Nucleic Acids, Plasmids, and Proteins

Plasmid constructs harboring DNA encoding the nucleic acid binding proteins, both with and without acidic extensions, for cell expression were produced using routine methods, protocols and reagents in the art. Constructs for the expression of the 4heptad-F zipper protein were built to contain the cytomegalovirus (CMV) promoter, the FLAG epitope (Invitrogen), and the nucleic acid sequence shown in SEQ ID NO:32. The DNA sequence depicted by SEQ ID NO:32 contains an N-terminal leader sequence (described below in SEQ ID NO:33), three glycine (GGG) residues (nucleotides 43-51), which, when deleted, produce a more effective dominant negative, as described herein (see Example 1F), the 4heptad sequence depicted in SEQ ID NO:38, and the F-zipper DNA sequence embraced by SEQ ID NO:34.

SEQ ID NO:32 is as follows:

The protein sequences expressed from the constructs transformed into E. coli, produced in vivo, and isolated and purified as described herein are listed below. In each protein, the initiator methionine is processed off in E. coli. The conventional one-letter symbols for abbreviating amino acid residues are used herein and are as follows: A (alanine); C (cysteine); D (aspartic acid); E (glutamic acid); F (phenylalanine); G (glycine); H (histidine); I (isoleucine); K (lysine); L (leucine); M (methionine); N (asparagine); P (proline); Q (glutamine); R (arginine); S (serine); T (threonine); V (valine); W (tryptophan); and Y (tyrosine).

All proteins, except for C/EBP and GBF-C/EBP, have a 13 amino acid N-terminal leader sequence as follows: ASMTGGQQMGRDP-(SEQ ID NO:33).

The F-zipper amino acid sequence is:

GGGTQQEVLELESRNDRLRKEVEQLERELDTLRGIFRQLPESSLVKAMGNCA (SEQ ID NO:34).

The amino acid sequences of the multiheptad-containing zippers are provided below. The L in the penultimate position in the heptad sequences is the first "L" position of the F leucine zipper. Those skilled in the art will be aware that the nucleic acid sequences, i.e., the triplet codons, corresponding to the various amino acid sequences presented herein and for use in the presently claimed invention can be determined in a conventional manner in accordance with the genetic code.

1heptad-F GGGLEQENAELE (SEQ ID NO:35);

2heptad-F GGGLEKEAEELEQENAELE (SEQ ID NO:36);

3heptad-F GGGLARENEELEKEAEELEQENAELE (SEQ ID NO:37);

4heptad-F GGGLEQRAEELARENEELEKEAEELEQENAELE (SEQ ID NO:38);

3heptad-F (cap)GGGLARNNIAVRKSRDKAKQRNVELE (SEQ ID NO:39);

3heptad-F (cap, a d) GGGLARNNIALRKSADKLKQRNVELE (SEQ ID NO:40);

3heptad-F (cap, e g) GGGLARENIAVEKERDKAEQENVELE (SEQ ID NO:41);

3heptad-F (cap, b c) GGGLARNNEEVRKSREEAKQRNAELE (SEQ ID NO:42).

Three bZIP domains described and utilized herein are shown in 1)-3) below. Unless otherwise noted, the nucleic acid sequences corresponding to the protein sequences of the chimeric bZIP proteins are those reported in the referenced publications.

1) The chimeric sequence of GCN4-C/EBP protein is depicted with the GCN4 sequence in italics:

SEYQPSLFALNPMGFSPLDGSKSTNENVSASTSTAKPMVGQLIFDKFIKTEEDP GKAKKSVDKNSNEYRVRRERNNIAVRKSRDKAKQRNVE TQQKVLELTSDNDRLRKRVEQLSRELDTLRGIFRQLPESSLVKAMGNCA, (SEQ ID NO:43). Cloning of GCN4 and its corresponding nucleic acid sequence have been described by A. G. Hinnebusch, 1984, Proc. Natl. Acad. Sci. USA, 81:6442. The isolation and nucleic acid sequence of the gene encoding C/EBP are described in W. H. Landschulz et al., 1988, Genes Dev., 2:786-800.

2) The polypeptide sequence of the GBF-C/EBP chimeric protein:

PVKDERELKRQKRKQSNRESARRSRLRNEAECEQTQQKVLELTSDNDRLRKR VEQLSRELDTLRGIFRQLPESSLVKAMGNCA (SEQ ID NO:44). Cloning of the Arabidopsis thaliana-derived bZIP protein, GBF, and its corresponding nucleic acid sequence have been described in U. Schindler et al., 1992, The EMBO J., 11:1261-1273.

3) The polypeptide sequence of the VBP-C/EBP chimeric protein:

ASMTGGQQMGRDPLEEKVFVPDEQKDEKYWTRRKKNNVAAKRSRDARRLK ENQTQQKVLELTSDNDRLRKRVEQLSRELDTLRGIFRQLPESSLVKAMGNCA (SEQ ID NO:45). Cloning of the chicken vitellogenin gene-binding protein VBP and its corresponding nucleic acid sequence have been described by S. Iyer et al., 1991, Mol. Cell. Biol., 11:4863-4875.

B. Protein Expression and Purification

Proteins with and without acidic extensions were expressed in E. coli and isolated and purified from culture supernatants. More particularly, proteins were synthesized in E. coli using the phage T7 expression system (F. Studier and B. Moffatt, 1986, J. Mol. Biol., 189:113-130. Bacterial cultures (about 400 mL) were induced with 1 mM isopropyl-b-D-thiogalactopyranoside) at an optical density of 0.6 at 600 nm for about 2 hours. Cells were then recovered by centrifugation, resuspended in 6 mL of lysis buffer (50 mM Tris-HCl, pH 8.0, 1 mM EDTA, 1 mM benzamidine, 1 mM dithiothreitol (DTT), and 0.2 mM phenylmethylsulfonyl fluoride (PMSF)), frozen, thawed, and gently brought to 1 M KCl by the addition of 2 mL of 4 M KCl. The sample was centrifuged at 25,000 rpm in a Beckman T42 rotor, and the supernatant was removed and isolated. The isolated supernatant was then heated to 65.degree. C. for 10 minutes and centrifuged, and the supernatant was isolated. The proteins that were capable of binding DNA were diluted to 100 mM KCl and purified over a heparin-agarose column as described by D. Krylov et al., 1994, EMBO J., 13:1849-1861, and were subsequently purified over a Rainin HPLC system. The proteins lacking DNA binding domains were purified over a Mono Q Sepharose column and eluted with 200 mM, 400 mM, and 600 mM KCl. The fractions enriched in the proteins were then purified on a Rainin HPLC system using a C18 column chromatographed from 0% to 100% acetonitrile in 0.1% trifluoroacetic acid. The molar concentrations were calculated using the molar extinction at 230 nm of amino acids (300/residue). The contribution of tyrosine (4980) and tryptophan (6818) to the absorbance at 230 nm were included if necessary (C. Cantor and P. Schimmel, 1980, Biophysical Chemistry. W. H. Freeman and Co., New York; G. Fasman, 1976, CRC Handbook of Biochemistry and Molecular Biology, 3rd Edition. CRC Press).

C. Analytical Ultracentrifugation

Analytical ultracentrifugation was performed as described in D. Krylov et al., 1994, EMBO J., 13:2849-2861, except that the protein absorbance was monitored at 260 nm. Protein samples at three concentrations were centrifuged at 24,000 rpm until equilibrium and data were collected after 48 hours. The C/EBP sample (MW=16,104), which contains one tyrosine, absorbed at a wavelength of 260 nm, while 3heptad-F (MW=9,847), which does not contain any aromatic amino acids, does not absorb at 260 nm. This absorption difference allowed the monitoring of the oligomerization state of C/EBP in the presence or absence of 3heptad-F.

D. Circular Dichroism (CD)

CD analyses were performed as described in Materials and Methods of D. Krylov et al., 1994, EMBO J., 13:2849-2861. T.sub.m values were calculated as described in D, Krylov et al., 1994, Ibid, converted to K.sub.d(25) and .DELTA.G(25) using a .DELTA.C.sub.p of -0.96 kcal/mol .degree. C..sup.-1 calculated from a T.sub.m versus .DELTA.H plot for all of the proteins as described herein. All thermal melts were reversible. Table 2 shows the stability of the mixtures of heterodimerizing zipper (F) with C/EBP under different salt conditions.

Table 1 presents the melting temperature (T.sub.m, .degree. C.), .DELTA.G and k.sub.d at 25.degree. C. for CD thermal melts of a variety of proteins either as homodimers or an equimolar mixture of two proteins. If the mixture had a higher melting temperature than either homodimer alone, it was inferred that the mixture sample was composed of heterodimers. The (L) denotes `low` salt conditions (50 mM KCl, 25 mM Tris pH 8.0, 0.5 mM EDTA, 1 mM DTT) and the (H) denotes regular salt conditions (12.5 mM potassium phosphate, pH 7.4, 150 mM KCl, 0.25 mM EDTA, and 0.5 mM DTT). The latter conditions were used in Table 2. "0heptad-F" represents the F zipper with the basic region deleted. The RMS error in .DELTA.G is 0.2 kcal/mol and 0.5 kcal/mol for homo- and heterodimers respectively.

E. Tryptophan Fluorescence

Fluorescence of a tryptophan in the VBP-C/EBP basic region was monitored using an Aminco SLM8000 photon counting spectrofluorimeter in slow kinetics mode 25.degree. C., with excitation and emission at 296 and 342 nm, respectively. Samples contained the same salt conditions used as described for the CD measurements. Samples containing 10.sup.-6 M VBP:C/EBP monomer were mixed with a DNA probe, 28 base pairs in length and containing the C/EBP binding site (0.5.times.10.sup.-6 M of duplex), to which a sequence the VBP basic region binds specifically (N. B. Haas et al., 1995, Mol. Cell. Biol., 15:1923-1932). Haas, et al. 1995) The sequence of the probe is 5' GTCAGTCAGATTGCGCAATATCGGTCAG 3' (SEQ ID NO:46). The consensus C/EBP DNA binding site is shown underlined and in bold.

The K.sub.d 's of homo- and heterodimers calculated from the CD thermal melts provided a determination of the fraction of VBP-C/EBP displaced from DNA when the DN was added in the fluorescence equilibrium experiments. For example, the VBP-C/EBP DNA binding constant was calculated by determining how effective the DN was at displacing VBP-C/EBP from DNA using the following equation:

wherein K.sub.hetero is the heterodimer dissociation constant, K.sub.DN is the DN dissociation constant, both measured from CD thermal melts, and K.sub.DNA is the DNA binding constant of a transcription factor.

The VBP-C/EBP DNA binding constant was calculated as 10.sup.-10 when data for 0heptad-F in 11 molar excess is used, or 1heptad-F, 2-heptad-F or 3heptad-F in 1.1 molar excess is used. The consistency of the calculated DNA binding constants suggests that the .DELTA.G's for the heterodimers as determined by CD thermal melts are valuable indicators of efficacy.

F. Transient transfections

The C/EBP family of transactivator plasmids (pMEX-C/EBP.alpha., pMEX-C/EBP.beta., pMEX-C/EBP.delta.) containing the MSV promoter used herein were as described by S. Williams et al., 1991, Genes and Devel., 5:1553-1567. These plasmids were modified to include a FLAG epitope (Invitrogen) at the N-terminus. The C/EBP.alpha.-GCN4 chimeras contained the leucine zipper from the bZIP protein GCN4. The CAT reporter plasmid contained a consensus C/EBP binding site (i.e., ATTGCGCAAT, SEQ ID NO:47) in front of the minimal promoter. All the DNs were cloned into pRc/CMV (Invitrogen) modified to contain the N-terminal FLAG epitope with a nuclear localization sequence (MDYKDDDDK-KKRK, SEQ ID NO:48). C/EBP.alpha. was also cloned into pRc/CMV and gave results similar to those obtained using pMEX-C/EBP.alpha. for both the transactivation and inhibition assays. Similar results were obtained for plasmids modified with the FLAG epitope and without the FLAG epitope.

In these experiments, 10 .mu.g of reporter plasmid, 0.3 .mu.g of transactivator, and 5 .mu.g of DN were added to each transient transfection. 3heptad-F with the DPD cap totally inhibited C/EBP transactivation in a 1:1 molar ratio. Sheared salmon sperm DNA was added to bring the DNA concentrations to 20 .mu.g. Calcium phosphate transfections were performed following the manufacturer's instructions (Gibco).

Western blotting of cellular extracts of generated from the transient transfections indicated that 3heptad-F containing a FLAG epitope placed 5' or N-terminal of the acidic extension was not seen on the blot, even though the construct completely inhibited C/EBP transactivation. A possible explanation was that the three glycine N-terminal cap used to stop the acidic extension was a proteolysis-sensitive protein sequence. Consequently, four additional sequences were employed as N-terminal caps, namely, DP-, DP-D, DP-EE, and DP-DEEE, replacing DP-GGG), (J. Richardson and D. Richardson, 1988, Science, 240:1648-1652). Their thermal stability as heterodimers with C/EBP was monitored, and their ability to be detected using a Western blot of cell extracts obtained from the transient transfections was determined. All of the sequences were detected in Western blots. Interestingly, the DP- and DP-D caps were three-fold more stable than the DP-GGG cap.

It was found by the present inventors that transactivation of the bZIP protein C/EBP was not totally inhibited either by the heterodimerizing F zipper (i.e., 0heptad-F) or by a three heptad acidic extension appended onto the C/EBP leucine zipper (i.e., 3heptad-C/EBP). Notwithstanding, the 3heptad-C/EBP did inhibit to a measurable degree. Changes in both the leucine zipper and the acidic extension generated a protein (3heptad-F) that formed a sufficiently stable heterodimer with C/EBP to inhibit competitively both C/EBP DNA binding and transactivation. However, it will be appreciated that for other proteins in the bZIP family and in other structurally similar families of proteins, the acidic extension alone may contribute enough stabilizing energy to create a dominant negative without the additional stability conferred by the heterodimerizing zipper. This will depend on the relative stabilizing effect of the DNA binding for particular bZIP proteins. Based on the present invention, acidic extensions can be designed that dimerize even more strongly with the basic region, thus allowing the creation of a dominant negative to any bZIP protein by simply appending an acidic extension onto the N-terminus of any bZIP leucine zipper.

Acidic Extensions Increase Heterodimerization of the bZIP Class of Proteins

Experiments were conducted to determine the effect of acidic extensions on DNA binding proteins of the bZIP class, c-Fos, CREB, and CEBP. c-Fos and CREB proteins were designed to contain an acidic 4 heptad repeat (FIGS. 3 and 4). An acidicaUy extended CREB was designed to contain a new acidic extension (New4hepCREB) in which the asparagine was changed to a leucine as described in FIG. 3. As shown, CREB+New4heptadCREB had a melting temperature of 69.5.degree. C. (open circles). This dramatic increase in thermal stability relative to the asparagine-containing acidic extension suggests that different classes of acidic extensions can be used to append to the N-terminal dimerization domain of bZIP proteins, depending on the exact bZIP protein to be inactivated.

In FIG. 4, an engineered c-Fos protein having the basic region deleted, called 0heptad-Fos, increased the melting temperature of the Fos/Jun complex to 53.degree. C. (open squares). An engineered c-Fos protein having the basic region replaced with an acidic region (containing four acidic repeats, called "4heptadFos") increased the melting temperature of the Fos/Jun complex to 61.degree. C. (open circles). An engineered c-Fos protein in which one amino acid in the acidic region was changed from an asparagine to a leucine, as described for FIG. 2, called new4heptadFos, resulted in a higher melting temperature of 72.degree. C., and more stable heterodimerization with c-Jun (closed squares). Also as shown in FIG. 4, homodimerization of unmodified c-Jun had a melting temperature of 29.4.degree. C., and homodimerization of acidically modified c-Fos (having four acidic repeats appended at the amino terminus of the c-Fos protein) had a melting temperature of 25.degree. C. Also in FIG. 4, the Kd(37.degree. C.) for the c-Jun homodimer was 20 .mu.M; 0.2 .mu.M for the c-Jun/c-Fos heterodimer; and 6 nM for the c-Jun/4heptadFos heterodimer.

Specificity of a bZIP and the Acidic Extension is Determined by the Zipper

The specificity of the basic region of a bZIP protein and the acidic extension on the N-terminus of a leucine zipper (i.e., called aZIP) was found to be determined by the leucine zipper. As shown in FIG. 5A, mixing a bZIP CEBP protein and an aZIP 4heptadFos protein with incompatible zippers produces no interaction, i.e., 4heptadFos does not interact with CEBP. FIG. 5B shows that mixing a bZIP VBP protein and an aZIP 3heptadF-(CEBP specific) protein also produces no heterodimeric interaction. In FIGS. 5A and 5B, the solid lines is the simple sum of the two homodimer curves. The fact that the actual mixture gives identical results demonstrates that no heterodimers are formed.

Transactivation by an AP1 Cis Element is Inhibited by Acidic Extension Appended onto a DNA Binding Protein

AP1 activity was demonstrated to be inhibited by several acidic extensions appended onto the bZIP DNA binding protein c-Fos in transient transfection experiments using human hepatoma cells (HepG2) and AP1-dependent CAT reporter expression constructs and measuring CAT activity (FIG. 6). HepG2 cells were transfected with up to three plasmid constructs for the expression of c-Fos proteins. The three plasmids were: a construct for expression of the Jun transactivator, a reporter construct containing an AP1 cis element driving CAT expression, and a construct containing DNA encoding the dominant negative and appropriate expression and regulatory sequences. CMV566 AND CMV500 are plasmids containing the cytomegalovirus promoter driving the expression of DNA binding proteins that contain at their N-termini, either the Hemagglutinin (YPYDVPDYA) (SEQ ID NO:49) or the FLAG (DYKDDDDK) (SEQ ID NO:50) epitopes, respectively. c-Fos proteins were designed to contain no acidic heptad repeats appended N-terminally from the basic region (0hep-Fos), one acidic heptad repeat appended N-terminally from the basic region (1hep-Fos), two acidic heptad repeats appended N-terminally from the basic region (2hep-Fos), three acidic heptad repeats appended N-terminally from the basic region (3hep-Fos), or four acidic heptad repeats appended N-terminally from the basic region (4hep-Fos). Transfections were carried out using either 3 or 15 .mu.g of plasmid containing CMV566 or CMV500 and DNA sequence encoding bZIP having N-terminal extensions as shown in FIG. 6. The results of these experiments show that AP1 transactivation by expressed c-Jun protein is inhibited by expressed c-Fos protein having appended acidic extensions. The results also demonstrate that the inhibition of AP1 transactivation is specific, since acidic extensions appended onto other bZIP proteins (i.e., CREB, VBP, or Jun) did not inhibit AP1 transactivation.

A bZIP Protein Containing an Acidic Extension (aZIP) is a Better Inhibitor of AP1 Activation Than a bZIP Protein Containing no Acidic Extension

Experiments similar to those described in Example 4 were performed in an additional transient cell system using Jurkat cells extension (bFos). Jurkat cells (a B cell model) have a high level of endogenous Fos/Jun (AP1) activity. Cells were transiently transfected with two plasmids (1 .mu.g each), namely, a reporter construct containing AP1 cis elements and a construct containing the CMV500 promoter driving the expression of either a truncated form of Fos (bZipFos) or the Fos leucine zipper containing the 4heptad acidic extension (aFos). AP1 activity was induced by the addition of 12-O-tetradecanoylphorbol(TPA) which resulted in a 40-fold induction of the gene. The transactivation of the reporter gene occurred because of endogenous activity. As demonstrated in FIGS. 7 and 8, expression of the bZipFos protein encoded by a bZipFos DNA sequence construct resulted in dominant negative activity; however, expression of aZipFos protein encoded by an aZipFos construct was more efficacious and showed a significantly high level of inhibition of AP1 activity. In these experiments, the aZIP protein c-Fos was designed to contain 4 acidic heptad repeats appended N-terminally (4heptad-Fos). The inhibitory activity of the expressed aZIP protein was compared with that of a c-Fos protein that was not engineered to contain an acidic extension (bZIP).

Acidic Extensions Appended to DNA Binding Proteins Provide Compounds for the Control and Regulation of Gene Function

An acidic extension was appended onto the dimerization domain of a DNA binding protein of the bHLH class (FIG. 9). Replacing the basic region of an expressed bHLH protein Max having an amphipathic acidic extension (i.e., 2heptadMax) increased heterodimerization of Max with c-Myc, irrespective of the orientation of the acidic extension as appended to the HLH protein. The 2heptadMax proteins used in the heterodimerization studies shown in FIG. 9 are expressed Max proteins having amphipathic acidic extensions as follows:

DPDLEKEAEELEQENAELELEDSF, called 2heptadMax(783), (SEQ ID NO:1);

DPDLEKEAEELEQENAELEELEDSF, called 2heptadMax(784), (SEQ ID NO:2); and

DPDLEKEAEELEQENAELEELEDSF, called 2heptadMax(785), (SEQ ID NO:3).

AcidMax is an expressed Max protein having appended N-terminally an acidic extension as follows: DPDEEEDDEEELEELEDSF (SEQ ID NO:4). 0heptadMax is an expressed Max protein having the sequence DPDLEELEDSF (SEQ ID NO:5) as its acidic extension. OheptadMax is the dimerization domain of Max without an acidic extension. Because of gene cloning, this protein as expressed has the short acidic extension DPDLEELEDSF (SEQ ID NO:5). The increase in the dimerization with c-Myc by the addition of the acidic extension on Max shows the utility of the approach of the invention.

In addition, the expressed AcidMax protein having a polyglutamic acid sequence heterodimerized with c-Myc better than did Max protein containing any of the three amphipathic extensions. In FIG. 9, the Kd(M) for the acidically modified and unmodified HLH proteins is as follows:

In the plasmid vectors designed to carry the Max DNA sequence, the "DPD" amino acid sequence is a Bam cloning site. Immediately N-terminal of the BamHI cloning site DPD, the protein constructs contain a 13 amino acid sequence that is from the .PHI.10 phage (ASMTGGQQMGR) (SEQ ID NO:51). The eukaryotic expression vectors can contain an additional N-terminal protein sequence 5' of the .PHI.10 protein sequence. This additional N-terminal protein sequence is an epitope used to monitor expression in mammalian cells. Its sequence is YPYDVPDYA (SEQ ID NO:49).

The last amino acid of the Max extensions is a phenyalanine (F) in 1 of the dimerization domain of Max. This amino acid is conserved in most bHLH proteins, thus providing a convenient landmark in the Max protein structure from which to know where the designed acidic sequences were appended onto the dimerization domain of Max. In the experiments described in this Example, an amphipathic acidic protein sequence was added onto the N-terminus of the Max dimerization domain. The linker between the acidic extension and the dimerization domain contained different numbers of glutamic acids. This difference in the length of the linker had the effect of rotating the acidic amphipathic sequence relative to the dimerization domain.

These results demonstrate the universal nature of the acidic extension for allowing a DNA binding protein from different classes ultimately to control and regulate gene function.

Effects on Cell Growth and Transformation Following Stable Transfection of Cells with Constructs Harboring DNA Binding Proteins Containing Appended Acidic Extensions

Cell transfection assays were performed in which cells (e.g., C3H10T1/2) were plated and transfected with 50 ng DNA per plate. The foci assay used in these experiments was as described by S. Min et al., 1993, Oncogene, 8:2691-2701. Two experiments with the potential DNs were conducted. First, cells were transformed with the constructs as described and were plated in conventional tissue culture plastic dishes (e.g., Petri dishes). These assays examined the effects that the constructs containing DNA encoding acidically extended DNs had on cell viability. It was found that cell viability was not affected by the constructs. Next, the same cells were transfected with constructs harboring Max DN DNA (described in Example 6) in conjunction with plasmids harboring Ras and/or Myc-encoding DNA, and the ability of the cells to grow in a contact-inhibited independent fashion was examined. The cells were plated in plastic dishes at high concentrations and their ability to overcome contact inhibited growth was evaluated. When the Max DNs of the invention (e.g., acidicauy extended Max) were co-transfected with Ras (10 ng) and/or Myc (300 ng), the action of the DN was clearly observed. OheptadMax (1 .mu.g) inhibited foci formation from 73 colonies to 34 colonies, while 2heptadMax decreased the number of colonies down to zero. A polyacidic Max construct is likely to inhibit foci formation in these assays as well as, or better than, the inhibition activity of 2heptadMax. The ras construct used in the experiments contains a ras gene which contains a mutation which keeps it constitutively active as described in E. Taparowsky et al, 1982, Nature, 300:762-765. The ras gene is driven by its normal promoter. The myc construct is a regular myc gene driven by the RSV LTR. After 3 days, plates were fixed and were stained after 14 days. Also, after 14 days, G418 resistant colonies were quantified and the results are shown below.

As shown below, the same types of experiments performed using the Fos constructs of the invention showed an inhibition of foci formation. In contrast to Max, the Fos expression constructs were able to inhibit both colony formation and foci formation in these assays. In the colony assay, results are also shown for a control plasmid construct (designated pKOneo) carrying only the neomycin gene for the selection of cells carrying this gene and expressing its product; the control plasmid resulted in 127 colonies. As known by those in the art, the neo gene is a conventional marker gene, built into constructs containing other expressible genes and used as a selection marker for expression of the neo gene as determined by resistance of the cells to neomycin following transformation or transfection of cells with a plasmid construct as described.

Competition Assays

Three types of competition experiments demonstrated that the acidic extension created a robust DN to the bZIP protein C/EBP's activity. Two biochemical assays addressed DNA binding (i.e., gel shift and fluorescence assays) and one biological assay addressed transactivation function (i.e., transient transfection assay).

A. Gel Shifts

Gel shift experiments were undertaken to examine if 3heptad-F inhibited the binding of C/EBP to its cognate DNA sequence. Gel shift data indicated that C/EBP bound to DNA with a K.sub.d =10.sup.-9 M (Z. Cao et al., 1991, Genes Dev., 5:1538-1552). The CD experiments presented herein indicated that 3heptad-F heterodimerized with C/EBP with K.sub.d =4.times.10.sup.-11 M. Therefore, an equimolar mixture of DNA, C/EBP dimers and 3heptad-F dimers was expected to prevent C/EBP from binding DNA, because of the formation of the C/EBP:3heptad-F heterodimer. The gel shift data confirmed that this equimolar mixture completely inhibited C/EBP binding to a consensus radiolabeled oligonucleotide (FIG. 13A).

B. Fluorescence

The second competition assay monitored the intrinsic fluorescence of the only tryptophan found in the basic region of VBP. DNA binding of the VBP-C/EBP chimera increased the fluorescence of this tryptophan by approximately 40%. The increase in fluorescence is to be expected if the tryptophan is in a more constrained environment, a situation which is expected upon DNA binding (Cantor and Schimmel, 1980, Biophysical Chemistry. W. H. Freeman and Co., New York), (FIG. 13B). This change in the intrinsic fluorescence of VBP-C/EBP upon DNA binding allowed the examination of the kinetic and equilibrium consequences of adding the F zipper-acidic extension series to this bZIP DNA complex. It was found that VBP-C/EBP bound to DNA with a fast on-rate (.about.5 seconds), and that the addition of the F zipper-acidic extension proteins resulted in a decrease in tryptophan fluorescence. This is expected if the F zipper-acidic extension proteins heterodimerized with VBP-C/EBP, thus preventing DNA binding.

Proteins containing different length acidic extensions appended to the F zipper were mixed with the VBP-C/EBP DNA complex in a 1 to 1.1 molar ratio. The longer acidic extensions caused a greater decrease in the fluorescence, thus indicating more effective heterodimerization with VBP-C/EBP. Ten additional molar equivalents of DN were added after the reaction came to equilibrium in about 20 minutes. 0heptad-F and 1heptad-F were not able to totally displace VBP-C/EBP from DNA, even in 11-fold excess. However, at an 11-fold molar excess, 2heptad-F and 3heptad-F were able to totally displace VBP-C/EBP from DNA. 4heptad-F was the most potent, as 1.1 molar equivalents displaced 85% of VBP-C/EBP from its binding to DNA.

The location of the tryptophan residue in the VBP basic region is four heptads from the leucine zipper. The tryptophan would be expected to be in a random coil conformation as a heterodimer with the shorter acidic extensions, such as those having one, two, or three heptad extensions (i.e., 0-3heptad-F), but would be expected to be in a coiled coil conformation as a heterodimer with more acidic extensions (i.e., 4heptad-F). The tryptophan would be expected to be part of the hydrophobic interface, i.e., it would occupy the a position, and, consequently, might have a greater fluorescence. This was observed. The tryptophan fluorescence of VBP-C/EBP in mixtures with 0, 1, 2, and 3heptad-F is similar to that of VBP-C/EBP alone which we interpret as the tryptophan fluorescence in the random coil state. In the mixture with 4-heptad-F, the fluorescence of VBP-C/EBP increases by about 10% suggesting that the tryptophan is more constrained as would happen if it is in the hydrophobic interface of a coiled coil

C. Transient Transfections

The third competition assay involved the transient transfection of constructs harboring the acidic extension and bZIP DNA sequences into a hepatoma cell line (HepG2) to monitor the transactivation properties of C/EBP on a promoter containing a single C/EBP binding site (see also Example 1F). C/EBP was shown to be able to transactivate this promoter 10-fold (FIG. 14). Constructs containing four, different, potential DNs were also able to inhibit C/EBP transactivation in a 15 to 1 ratio. These four DNs are a truncation of the transactivation domain (.DELTA.C/EBP), 0heptad-F, 3heptad-F, and 3heptad-C/EBP. Constructs carrying .DELTA.C/EBP, 0heptad-F, and 3heptad-C/EBP inhibited C/EBP transactivation only slightly. In marked contrast, 3heptad-F was able to inhibit completely C/EBP-mediated transactivation. This total inhibition was also observed at a ratio of C/EBP plasmid to 3heptad-F plasmid of 1 to 1. Thus, both the heterodimerizing F zipper and the acidic extension were discovered to be essential for total inactivation of C/EBP. In addition, the 3heptad-F-containing construct was found to inhibit the transactivation activity of all the C/EBP family members examined (C/EBP.alpha., C/EBP.beta., C/EBP.delta.).

To investigate if 3heptad-F inhibition of C/EBP transactivation was dependent on the C/EBP leucine zipper, C/EBP transactivating proteins were generated in which the C/EBP leucine zipper was replaced with the GCN4 leucine zipper. This chimeric protein was able to transactivate a minimal promoter in a C/EBP cis element-dependent manner. This transactivation was not inhibited by 3heptad-F, thus indicating that the DN properties of 3heptad-F are dependent on specific leucine zipper interactions.

The Acidic Extension Forms a Coiled Coil with the Basic Region of Leucine Zipper Containing Proteins, e.g., bZIP

It was found that the addition of one or more heptads of acidic amino acid sequences onto the N-terminus of the leucine zipper stabilized heterodimer formation, as determined by circular dichroism (CD) spectra and thermal melting assays as described herein. This acidic protein (or polypeptide or peptide) extension formed a heterodimeric coiled coil with the basic region of the wild type or native protein containing the leucine zipper and allowed the creation of a dominant negative to the native protein that stoichiometrically displaced the native protein from DNA and inhibited transactivation by the native protein.

Sedimentation equilibrium analysis was carried out in order to determine the oligomerization state of an equimolar mixture of C/EBP and 3heptad-F (FIG. 11). The data were fitted to a single molecular species. The calculated molecular weight for C/EBP is 33,000 Daltons (dimer MW is about 32,208) and for the C/EBP--3heptad-F mixture is 28,000 (heterodimer MW is about 26,200). These data are consistent with C/EBP forming a heterodimer when mixed with 3heptad-F.

Circular dichroism (CD) was used to determine the amount of .alpha.-present in mixtures of C/EBP and different length acidic extensions appended to the F zipper (FIG. 12A) The mixtures of C/EBP with progressively longer acidic extensions showed a progressive increase in .alpha.-content. The increase in .alpha.-helical content of the heterodimerizing mixtures as described herein reveals that the leucine zipper dimerization interface is extending into the basic region: the longer the acidic extension, the farther the coiled coil structure extends from the leucine zipper into the basic region. The CD spectra of the acidic extensions series (0heptad-F to 3-heptad-F) as homodimers at 6.degree. C. displayed a similar amount of .alpha.-helical content, thus indicating that the progressively longer acidic extensions were not helical. The CD melt of 4heptad-F was biphasic; this is believed to indicate that the acidic extension forms a coiled coil which melts at 13.degree. C., independently of the leucine zipper.

The stability of the mixtures of C/EBP and the different length acidic extensions was determined by CD thermal melts (Table 2 and FIG. 12B). All mixtures had a higher T.sub.m than the individual homodimers, thereby suggesting that heterodimers were formed. The 0heptad-F:C/EBP heterodimer was more stable than the C/EBP homodimer, but the former did not contain any additional helical content. This suggests more stability for the leucine zipper structure. The progressively longer acidic extensions created progressively more stable complexes. These extensions raised the T.sub.m about 10.degree. C., which corresponded to 2.5 kcal/mol or over 100 fold stabilization.

The energetic and structural consequences of adding up to four heptads of acidic protein sequence onto the N-terminus of the F leucine zipper was assayed by circular dichroism (CD) spectra and thermal melts. The first acidic heptad appended to the F zipper stabilized the heterodimer with C/EBP by over 1 kcal/mol. This increase in stability was accompanied by an increase in ellipticity, indicative of .alpha.-formation. Each additional acidic heptad extension produced an increase in ellipticity consistent with extending the leucine zipper coiled coil by one heptad. However, the contribution to stability became successively less with increased heptad extensions. The combination of the increase in ellipticity and the progressive decrease in the contributed stability suggests that the coiled coil structure between the C/EBP basic region and the acidic extensions is weaker as the structure extends further from the leucine zipper. Similar results were obtained if the C/EBP basic region was replaced with other basic regions, e.g., the VBP or GBF basic regions. On the basis of these findings, extensions can be designed and created to dimerize specifically with a subset of basic regions in bZIP proteins and structurally similar proteins.

Stability of the Heterodimerizing Zipper (F) with C/EBP

The thermal stability of the heterodimers formed between C/EBP and the F leucine zipper produced as described by C. Vinson et al., 1993, Ibid., either as a chimera with the C/EBP basic region (C/EBP-F) or alone (0heptad-F), was determined (see Table 1 and FIGS. 9C-9D). Using gel shift conditions (50 mM KCl), a chimera containing the F zipper and the C/EBP basic region (C/EBP-F) formed a heterodimer with C/EBP that was five times more stable than the C/EBP homodimer. However, using more physiological salt conditions (150 mM KCl), C/EBP-F formed a heterodimer with C/EBP that was only twice as stable as the C/EBP homodimer. Deletion of the basic region (0heptad-F) increased the stability of the heterodimer with C/EBP by 0.5 kcal/mol, but the heterodimer was still not as stable as the wild type C/EBP homodimer bound to DNA. These results showed that initial attempts to create a DN that could displace C/EBP from DNA by making modifications only in the leucine zipper were unsuccessful.

The wild type C/EBP homodimer was less stable by 4.degree. C. in 50 mM KCl compared with 150 mM KCl, presumably because of repulsion between the basic regions. This decrease was not observed for a C/EBP:C/EBP-F heterodimer, because the increase in the electrostatic attraction between the F zipper and the wild type C/EBP zipper apparently compensated for the increased repulsion between the two C/EBP basic regions. Similarly, the increased stability of the C/EBP:0heptad-F heterodimer relative to that of the C/EBP:C/EBP-F heterodimer may have been due to the same interactions, because the 0heptad-F mutant lacked the repulsive basic region.

The Acidic Extension Forms a Coiled Coil with Two Additional bZIP Basic Regions

To investigate if the acidic N-terminal extension formed a coiled coil with other bZIP basic regions, chimeras were generated containing two additional basic regions attached to the native C/EBP leucine zipper. The first basic region was derived from the chicken vitellogenin gene-binding protein (VBP) (S. Iyer et al., 1991, Mol. Cell. Biol., 11:4863-4875), the chicken equivalent of TEF (D. Drolet et al., 1991, Genes Dev., 5:1739-1753), and the second was derived from GBF-1, a plant bZIP protein known to bind to the most divergent cis element of the bZIP proteins (U. Schindler et al., 1992, EMBO J., 11:1261-1273). The results of thermal melts of these chimeras (i.e., VBP-C/EBP and GBF-C/EBP) as homodimers showed a T.sub.m within 1.degree. C. of C/EBP, indicating that dimerization strength was determined by the leucine zipper (Table 2). Table 2 presents the stability of heterodimers between the F zipper with different length acidic extensions and chimeras containing three different bZIP basic regions appended to the C/EBP zipper.

As shown, four acidic heptad extensions were able to interact favorably with the three different bZIP basic regions that were examined. Initially, this was found to be surprising since it was assumed that the C/EBP basic region would be a better candidate for forming a coiled coil with the acidic extension than the GBF and VBP basic regions. Most bZIP proteins, including GBF and VBP, contain a basic amino acid in the first d position N-terminal of the leucine zipper, while C/EBP contains an alanine. It was expected that the basic amino acids would disfavor coiled coil formation while the alanine would be more favorable. CD thermal melts indicated that this was true; C/EBP heterodimerized with 1heptad-F 1.7.+-.0.5 kcal/mol better than with 0heptad-F, while the VBP and GBF basic regions were only stabilized 0.9.+-.0.5 and 1.3.+-.0.5 kcal/mol, respectively. Interestingly, this difference in stability among the three basic region became less pronounced as the basic region was "zippered" up by longer acidic extensions.

The acidic extension to the leucine zipper domain was successfully stabilized by a variety of different basic regions, thus suggesting that this protein sequence can be used to create dominant negatives for a variety of different types of proteins comprised of basic domains and leucine zipper domains. The biological assays exemplified herein demonstrate that C/EBP transcription factor proteins designed and produced to contain the acidic extension are robust dominant negatives of the C/EBP family, which is considered to be representative of other protein members in the bZIP family or in structurally similar protein families.

The two chimeras were mixed with the four different length acidic extensions appended to the F zipper and their stability was determined from CD thermal melts. Interestingly, the two additional basic regions were also stabilized by the acidic extensions. The degree of stabilization was the same for the three basic regions examined, thus suggesting that the interaction occurred with the more conserved part of the basic region.

Thus, two lines of evidence suggested that the acidic extension formed a coiled coil with the basic region in the proteins expressed and assayed in accordance with the invention. The first was that in .alpha.-helical content increased when the acid extensions were mixed with the basic regions to form chimeric structures. The second was determined from the fluorescence analysis of exemplary heterodimers formed between chimeras of the VBP and C/EBP bZIP proteins and the F zipper-acidic extension series (See Examples 1E and 8B). The tryptophan in the VBP-C/EBP chimera is located in the basic region, four heptads from the leucine zipper. If a coiled coil were formed between the basic region of VBP and the acidic extension, then only the heterodimer between the 4heptad-F and VBP-C/EBP would be expected to place the tryptophan in the coiled coil structure, thereby causing a new fluorescence signal. Indeed, this result was observed. However, such a change in the fluorescence signal was not observed in heterodimers which contained the shorter acidic extensions.

The Acidic Extension is a DNA Mimetic

The importance of different coiled coil positions (a,b,c,d,e,f,g) in the acidic extension to heterodimer stability with native C/EBP was examined by producing four mutant proteins. Each protein comprised a subset of the mutations to the C/EBP basic region that was used to create 3heptad-F. The largest contribution to heterodimer stability was provided by the glutamic acids placed in the e and g positions. These negatively charged amino acids presumably interact with the positively charged amino acids found in abundance in the basic region, thus suggesting that the acidic extension acts as a DNA mimetic by matching the electrostatic properties of DNA. The three additional mutant proteins demonstrated the contributions of the N-terminal helical cap consisting of three glycines, the cap plus the hydrophobic core (a and d), and the cap plus general electrostatic and forming concerns (b, and c). All of the mutant proteins contributed some stability to the heterodimer with the wild type C/EBP basic region (see Example 1F).

Transgenic Mouse Studies

Transgenic animals, and in particular mice, are created using the constructs described for the transient transfection assays in Examples 1F and 8C. Briefly, the constructs are introduced into an animal or an ancestor of the animal at an embryonic stage, i.e., the one-cell stage, or generally not later than about the eight-cell stage. Transgenic animals carrying the constructs of the invention can be made by several methods known to those having skill in the art. One method involves transfecting a retrovirus constructed to contain the DNA sequence encoding a dominant negative leucine-zipper containing protein to provide a complete shuttle vector harboring the dominant negative nucleic acid sequence as a transgene. Another method involves directly injecting a transgene into the embryo. A third method involves the use of embryonic stem cells. Examples of animals into which the transgenes or constructs may be introduced include, but are not limited to, mice, rats, other rodents, sheep, pigs, and primates (see "The Introduction of Foreign Genes into Mice" and the cited references therein, In: Recombinant DNA, Eds. J. D. Watson et al., W. H. Freeman and Company, New York, N.Y., pp. 254-272). Transgenic animals carrying and stably expressing transgenes encoding the DNs of the invention can be used as biological models for the study of cancer and tumor regression or atrophy, for example. The appropriate tissues of the transgenic animal can be monitored for the integration of the construct, or components thereof, by assaying for the presence of RNA or DNA in the cells using established methods in the art.

In addition, transgenic mice can be produced in which tissue specific promoters are used in constructs containing DNA sequences encoding acidic extensions to the zipper of bZIP, or structurally similar proteins, as well as other pertinent sequences, to act as dominant negatives to one or more of the three C/EBP isoforms expressed during adipose conversion of 3T3-L1 cells (Z. Cao et al., Genes Dev., 5:1538-1552). The production of functional dominant negatives in this system will allow the inactivation of gene(s), or mutants thereof, involved in the conversion of cells to adipose cells. The constructs as described may provide the potential to reduce or alleviate the conditions of morbid obesity and type II diabetes associated with the syndrome of morbid obesity in model animals and in humans.

Tissue specific promoters such as the whey activating protein (WAP) which targets mammary tissue, can be used to create transgenic animals in which the DN is localized to and functions specifically in breast tissues. For example, a construct can be designed to include the mouse WAP promoter and isolated DNA encoding the DN proteins of the invention, in addition to the appropriate elements for proper transcription of the gene and protein expression in cells. The activity of the DN to inactivate the abnormal activity of cellular proteins can be monitored in mammary tissue of the female transgenic mice.

In addition, the constructs can be used in human gene therapy treatments as known in the art to inactivate cellular protein products the regulation of which is controlled by or linked to transcription regulatory proteins possessing leucine zipper structures. Gene therapy techniques which can be modified as required for use of the constructs of the invention are described, for example, in U.S. Pat. No. 5,399,346 to W. French Anderson et al.

Transgenic Mouse Studies using a Dominant Negative C/EBP to Create Phenotypically "Skinny" Mice

As set forth in Example 13, transgenic mouse studies were carried out and mouse pups were born which exhibited a transgenic phenotype showing that the dominant negatives produced and expressed in accordance with the invention functioned in an in vivo environment in an animal. The transgenic mice were produced by routine methods practiced in the art of transgenic animal production.

In these studies, transgenic mice were produced by injecting early-stage mouse embryos with the construct described in FIG. 30. As described, this plasmid construct was designed to contain the 422/aP2 promoter (adipose fatty acid-binding protein), DNA sequence encoding the dominant negative 3heptadF C/EBP of the invention, which produces, upon translation and expression, an acidically extended bZIP C/EBP protein, that functions as a dominant negative to the non-mutant C/EBP protein, and other regulatory sequences as described for FIG. 30. C/EBP has been implicated in the development and differentiation of adipose (fat) cells. The dominant negative 3heptadF C/EBP sequence was placed under the control of the 422/aP2 adipocyte-specific promoter, and also contained the FLAG.phi.10 epitope as described hereinabove.

Four initial founder mice were produced by injecting mouse embryos with the construct capable of expressing the dominant negative C/EBP protein. The founder mice were used to produce offspring carrying one or more copies of the transgene. One of the offspring of a founder mouse showed a thin, skinny phenotype, which would be expected if expression of the dominant negative affected the expression of normal C/EBP protein and resulted in negative regulation and influence of adipose or fat cell and tissue development and differentiation.

More specifically, the male founder transgenic mouse impregnated two female mice. The two litters produced by these two females contained two phenotypically distinct types of pups: the first types were scrawny, small-sized mice with scruffy fur and the second types were normal-sized with normal-looking fur. The small-sized, scrawny and scruffy pups all carried copies of the transgene, while their normal littermates did not. These results indicate that constructs designed to contain DNA encoding acidically extended dominant negatives and expressed as transgenes in animals in vivo can influence, e.g., via inactivation, the expression and/or function of a normal counterpart protein encoded by the normal gene. In addition, such dominant negatives can demonstrably affect the phenotypes of the animals carrying the transgene, which were bred from the transgenic founder animals.

The contents of the patents, articles, texts, and references contained herein are hereby incorporated by reference in their entirety.

Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be obvious to those skilled in the art that certain changes and modifications may be practiced without departing from the spirit and scope thereof, as described in the specification and as defined in the appended claims.