WO1993010250A1 - Transcription adaptors in eukaryotes - Google Patents

Abstract

The present invention relates to a protein or protein complex which stimulates activated DNA transcription in eukaryotic cells, genes encoding the protein or proteins, and uses therefor. The protein or protein complex, referred to herein as a transcriptional adaptor, is associated with two other factors, each of which, in turn, binds to a region of a eukaryotic promoter. The first of these factors, referred to herein as an activator, binds to a region of the promoter DNA sequence termed the UAS binding site or enhancer. The second factor or factors, referred to herein as a general transcription factor (GTF), binds to second region of the promoter DNA sequence, termed the TATA box, which is located downstream from the UAS binding site. The transcriptional adaptor specifically binds to an acidic activation domain (i.e., one which is highly enriched in acidic amino acids, such as aspartic acid and glutamic acid) on the activator protein which, in addition, includes a DNA binding domain.

Description

TRANSCRIPTION ADAPTORS IN EUKARYOTES

Description

BACKGROUND OF THE INVENTION

The activation of DNA transcription is of central importance in the regulation of genetic programs that range from control of cell growth to embryonic development in eu aryotes. Transcription begins with the synthesis of an RNA chain encoding one strand of a DNA duplex. The reaction takes place by complementary base pairing and covalent joining of ribonucleotides and is catalyzed by the enzyme RNA polymerase. Activation is not initiated by the chance encounter of the DNA and the polymerase, however, the enzyme is directed to a specific site on the DNA. In eukaryotes, two cis-acting elements on a target DNA, which together make up a promoter, have been implicated as protein binding sites essential for the initiation of transcription. It has been demonstrated that activation requires binding of specific proteins at both of these sites and only when these two sites are bound, does transcription occur. The first binding site is generally located 25-30 nucleotides upstream of the transcriptional start site and is termed the TATA box. The second binding is at a distal location from the TATA box and is termed the upstream activation site (UAS) , or enhancer.

It has been demonstrated that activation of transcription is orchestrated by two classes of protein ^•■factors" that are essential for the initiation of activation and which bind to the promoter at its two binding sites. It is believed that RNA polymerase II (RNAP II) , which is involved in the transcription of messenger RNA, is directed to the TATA box binding site by the first class of factors, referred to as general transcriptional factors (GTF) .

Further investigations have found that the second class of factors, termed activators, bind to the UAS binding site. Activators have been found to contain two domains: a DNA-binding domain and an activation domain. It has been suggested that there is a third factor involved in the DNA transcriptional machinery which mediates the interaction between GTFs and activators, termed mediators or adaptors, but the responsible protein has not been identified. There is considerable conservation in eukaryotes of factors involved in transcription, as evidenced by the fact that yeast activators work in mammalian cells and, conversely, mammalian activators function in yeast. However, the central question remains, how activators can transmit their signal to the GTFs that comprise the transcription initiation complex, even when the activator binds to DNA at significant distances from the initiation site.

SUMMARY OF THE INVENTION

The present invention relates to a protein or protein complex which functions in stimulation of activated DNA transcription in eukaryotic cells, genes encoding the protein or proteins, and uses therefor. The protein or protein complex, referred to herein as a transcriptional adaptor, is associated with two other factors, each of which, in turn, binds to a region of a eukaryotic promoter. The first of these factors, referred to herein as a transcriptional activator or activator protein, binds to a region of the promoter DNA sequence termed the UAS binding site or enhancer. The second factor or factors, referred to herein as a general transcription factor (GTF) , binds to a second region of the promoter DNA sequence, termed the TATA box, which is located downstream from the UAS binding site. The transcriptional adaptor specifically binds to an acidic activation domain (i.e., one which is highly enriched in acidic a ino acids, such as aspartic acid and glutamic acid) on the transcriptional activator protein, which, in addition, includes a DNA binding domain. The subject invention, thus, relates to novel transcriptional factors, such as factors from yeast, human and other eukaryotic cells, which allow the transcription machinery or transcription complex of eukaryotic cells to respond to acidic activation domains of a transcriptional activator.

The present invention further relates to purified transcriptional adaptors obtained from sources in which it naturally occurs; DNA or RNA encoding the adaptor; recombinantly produced transcriptional adaptor; antibodies specific for the adaptor; methods of producing the adaptor; and adaptor inhibitors and their use. The present invention further relates to isolated yeast genes, referred to as ADA2 and ADA3 (ADA stands for "alteration/deficiency in activation") , which encode products necessary for function of acidic activation domains of transcriptional activators in vivo (i.e., genes which encode proteins with the biological properties of transcriptional adaptors) , and to DNA from other eukaryotes, which encode transcriptional factors, referred to herein as transcriptional adaptors, required for function of acidic activation domains of transcriptional activators. As described herein, the ADA2 gene encodes a protein required for the function of some, but not all, acidic activation domains of transcriptional activators. The sequence of the ADA2 gene has been shown not to correspond to any known gene, either from yeast or other species. The ADA2-encoded protein acts as a transcriptional adaptor or as a component of an adaptor complex. The nucleotide sequence of the ADA2 gene and the deduced amino acid sequence of the encoded product are presented herein. The ADA2 gene has been shown not to be essential for cell growth. As described herein the ADA3 gene encodes a protein which allows the general transcription machinery to respond to acidic activation domains of transcriptional activators. The ADA3 transcriptional factor or transcriptional adaptor is a actor in the general transcription machinery. Experimental evidence presented herein, reasonably suggests that ADA3 is associated with a general transscriptional factor which is bound at the TATA box, such as TATA box-binding protein (TBP) . The ADA3 gene is not essential for cell growth, but ADA3 gene disruption mutants grow more slowly and are temperature sensitive. The nucleotide sequence of the ADA3 gene and the deduced amino acid sequence of the encoded product are presented herein; the nucleotide sequence has been shown not to correspond to any known gene, either from yeast or other species. The present invention further relates to a method of identifying inhibitors or enhancers of the transcriptional adaptor, inhibitors or enhancers of the transcriptional adaptors described herein and a method of altering, particularly inhibiting, transcription in cells. For example, the adaptor has been shown to be inhibited by GAL4-VP16. Therefore, in one embodiment of the method, cells expressing cloned adaptor can be used to assay agents, such as GAL4-VP16 analogues, for their ability to inhibit the transcriptional adaptor; other agents can also be assessed for their ability to inhibit transcriptional adaptors. A transcriptional adaptor inhibitor interferes with the activation of transcription and, thus, inhibits transcription of RNA. The resulting reduction can be assayed in vivo in cells in which a reporter, such as a CYCl-lacZ reporter, is expressed, by the diminished expression of the reporter gene.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is the nucleic acid sequence of the ADA2 gene (SEQ ID NO. 1) , including an open reading frame which encodes a 434 amino acid residue protein, and the deduced amino acid sequence (SEQ ID NO. 2) of the ADA2 protein represented in single letter code.

Figure 2 is the nucleic acid sequence of the ADA3 gene (SEQ ID NO. 3) , including the 702 codon open reading frame (ORF) and the deduced amino acid sequence (SEQ ID NO. 4) of the ADA3 protein represented in single letter code. Figure 3 is a schematic representation of a model illustrating the toxicity of GAL4-VP16.

Figure 4 shows the results of an analysis of the ADA2 amino acid sequence (SEQ ID NO. 2), which revealed several motifs of possible significance.

Figure 5 is a comparison of a 12-amino acid stretch (295-306) from ADA3 (SEQ ID NO. 3) with similar sequences in several other genes.

Figure 6 is a schematic representation of a model for ADA2 and ADA3 in transcriptional activation, in which the transcriptional activator is represented by GAL4 (DNA binding domain) and VP16 (acidic activation domain) .

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to the discovery of a protein or protein complex, referred to as a transcriptional adaptor, which are transcriptional factors which function in activation of DNA transcription in eukaryotic cells. The present invention relates to transcriptional adaptors which have a role in the stimulation of activated DNA transcription in eukaryotic cells. The transcriptional adaptor is associated with two other factors, each of which in turn binds to a DNA sequence, the UAS binding site and the TATA box binding site. The TATA box binding site is located several kilobases downstream from the UAS binding site. The adaptor specifically binds to an acidic activation domain (i.e. highly enriched in acidic amino acids, such as aspartic acid and glutamic acid) on the transcriptional activator protein which, in addition, includes a DNA binding domain. The subject invention also relates to two genes, referred to as the ADA2 gene and the ADA3 gene, which encode proteins (transcriptional adaptors) which allow for the normal function of acidic activation domains.

As used herein, the term transcriptional adaptors refers to a protein of eukaryotic origin which has a function similar to that of the ADA2 protein or the ADA3 protein and hybridizes to all or a portion of the ADA2 gene or the ADA3 gene or is recognized by an antibody specific for the ADA2 protein or the ADA3 protein.

A transcriptional adaptor of the present invention, as well as equivalents and analogues thereof, are useful in stimulating transcription and, thus, associated protein production. ADA2 and/or ADA3 or their analogues or equivalents can be used to enhance transcription and associated protein production by increasing the activation signal between general transcriptional factors (GTFs) and transcriptional activators. In addition, inhibitors of these proteins can be valuable in reducing protein production by disrupting or inhibiting the transcription machinery of a target cell. Such inhibitors include ADA2 and ADA3 mimics, which are molecules (e.g., proteins or small organic molecules) which bind or interact with a GTF or a transcriptional activator, but not both, thus preventing linking of the transcriptional activator and the GTF, which is necessary for transcriptional activation to occur. Alternatively, the inhibitor can be a molecule, such as a protein or small organic molecule, which binds the transcriptional activator or GTF, again preventing their functional linking through the transcriptional adaptor. An antibody which binds the transcriptional adaptor, the transcriptional activator or the GTF can also serve as an inhibitor. For example, viral transactivators like VP16 activate transcription by using adaptors. Therefore, a defective ADA3 protein or analogue would reduce the activity of viral transactivators and, thus, reduce viral growth by binding the acidic activation domain, interfering with the endogenous adaptor and, as a result, with the transcription machinery.

The identification of the ADA2 and ADA3 genes makes it possible to isolate DNA encoding transcriptional adaptors from other organisms (e.g., other eukaryotes, including mammalian, particularly human cells) by using nucleic acid probes which hybridize to all or a portion of the ADA2 gene or the ADA3 gene and known hybridization methods. See Maniatis et al. , Molecular Cloning, A Laboratory Manual. 2d, Cold Spring Harbor Laboratory Press, page 16.54 (1989) . Alternatively, primers can be designed based on knowledge of the ADA2 and the ADA3 gene sequences and used in known amplification methods, such as polymerase chain reaction (PCR) to identify similar or equivalent genes in other cell types, such as mammalian cells, including human cells.

In addition, antibodies can be raised against either the ADA2 protein, the ADA3 protein or both and used to isolate and identify equivalent proteins in eukaryotes other than yeast, such as mammalian, particularly human, cells. Antibodies to the ADA2 protein or the ADA3 protein can be used in immunoprecipatation procedures known by those skilled in the art to isolate equivalents in other eukaryotes. Both polyclonal antibodies and monoclonal antibodies can be produced using methods known by those Skilled in the art. Once a candidate or potential transcriptional adaptor-encoding gene has been identified by methods described above, the mutational assay, described herein, can be used to verify that it encodes a transcriptional adaptor. In brief, the mutational assay is based on the experiments, described herein, that show that the transformation of a host cell with an expressible plasmid containing a gene for a transcriptional activator (i.e., a transcriptional adaptor containing an acidic activation domain) can significantly inhibit growth of the transformed host cell by sequestering the host cell's endogenous transcriptional adaptor, preventing normal transcription from occurring. If the addition of the new transcriptional adaptor to the transformed host cell reverses the growth inhibition of the transformed host cell, this is further proof that newly isolated protein functions as a transcriptional adaptor. For example, a plasmid expressing the fusion protein GAL4-VP16 (GAL 4 is an acidic yeast activator and VP16 is an acidic activation domain of herpes simplex virus protein) can be used as a transcriptional inhibitor. A transcriptional adaptor candidate that reverses a growth inhibition caused by GAL4-VP16 in a yeast cell would be further proof that the newly isolated protein is a transcriptional adaptor.

Transformed microorganisms or other host cells which express ADA2 or ADA3 as recombinantly produced ADA2 or reco binantly produced ADA3 can be created by various methods known to those skilled in the art. For example, the ADA2 gene or the ADA3 gene can be introduced into a host cell by transfection, infection or electroporation. Suitable host cells include yeast cells, mammalian cells, including human cells, and bacterial cells.

Enhancement of transcription in a microorganism can be accomplished by introducing the ADA2 gene or ADA3 gene or analogues thereof into the microorganism. Inhibition of transcription in a microorganism can be carried out by introducing a defective ADA2 gene or a defective ADA3 gene or defective analogues into the host cells. Introduction of a construct containing the ADA2 gene or the ADA3 gene into an individual can be accomplished a variety of routes of administration. For example, administration can be done by various routes of injection, such as intravenous, intradermal, intraperitoneal or subcutaneous.

The following is a description of the isolation and characterization of the yeast transcriptional adaptors ADA2 and ADA3. It is shown that the ADA2 protein is associated with a specific subset of acidic activation domains from different activators. ADA 3 protein is shown to be a general adaptor, associated with acidic activation domains from all transcriptional activators tested. The ADA2 gene and the ADA3 gene are not essential for growth; however, organisms with ADA3 gene mutations grow slowly and are heat sensitive. A data base search indicated that the 702 codon ORF of the ADA3 gene did not correspond to any known yeast gene, and ADA2, as well, did not correspond to any known gene. Isolation and Characterization of the Adaptor Protein

As previously discussed, transcription involves two components on a target DNA, collectively termed the promoter: the TATA box and the UAS site or enhancer. The activation site (UAS) is located at a range of several kilobases from the TATA box. Benoist et al.. Nature. 290:304-310 (1981); and Guarente, L. , Annu.Rev.Genet.. 21:425-452 (1987).

Activation of transcription is directed by two classes of "factors" that are required for activated transcription and which bind to the two sites of the promoter. RNA polymerase II (RNAP II) which transcribes messenger RNA is directed to the basal promoter by GTFs which directly bind to the TATA box. GTFs have been identified by the fractionation of the mammalian in vitro transcription system, and include factors termed TFIID, TFIIA, TFIIB, and TFIIE. Matsui et al.. J. Biol. Che .. 255:11992-11996 (1980).

The second class of factors, termed activators, bind to the UAS site on the promoter. Activators have been found to contain two domains; a DNA binding domain and an activation domain. Brent et al..Cell. 4.:729-736 (1985); and Hope and Struhl, Cell. 16:885-894 (1986).

Suggestions that a mediator is required for the interaction between activation domains and the GTFs has been presented previously. Flanagan et al.. Nature. 350_f 436-4438 (1991) ; Kelleher et al^.. Cell. 61, 1209-1215 (1990); Pugh et al^., Cell. 61:1187-1197 (1990); and Berger et al-_s., Cell. 61-H99-1208 (1990). It has been demonstrated that a cloned GTF, cloned TFIID, supported basal transcription but did not respond to an activator. Pugh et al.. (1990) . It was therefore inferred that a "coactivator" was required for full transcription.

As described below, a transcriptional adaptor has been isolated and shown to be part of the activated transcription. The newly isolated protein or protein complex appears to be associated with the activator by binding to the acidic activation domain of the activator. The adaptor allows for the linkage of an activator and a GTF, stimulating activated transcription. A chimeric protein which mimics a transcriptional activator including a DNA-binding domain of the yeast GAL4 protein fused to the acidic activation domain (i.e., a domain highly enriched in acidic amino acids, such as aspartic acid and glutamic acid) from Herpes virus VP16 protein was used to provide further evidence of the existence of a transcriptional adaptor. The chimeric protein consists of the first 147 amino acid residues of the yeast GAL4 activator, which contains the DNA-binding domain, fused to 78 amino acid residues of the Herpes simplex virus VP16 activator, which contains a strong acidic transcriptional activation domain. Sadowski et al.. Nature. 135:563-564 (1988).

It was found that inhibition by the chimeric protein in a yeast in an in vitro transcription system was reversed by a partially purified yeast nuclear extract divided on a DEAE-rSephacel column. Kelleher et al.. (1990) . A more recent study confirmed this finding. Flanagan et al.. (1991) .

Previous work had addressed the question of whether an adaptor was involved- in the transcription machinery in vitro. Berger et al.. 1990. In that work, the chimeric protein GAL4-VP16 was altered to prevent the binding of the DNA by blocking the GAL4 DNA-binding site with a specific oligonucleotide encoding the GAL4 binding site. The altered chimeric protein did not inhibit basal transcription (i.e., transcription from the TATA box), but inhibited the ability of the dA/dT activator bound to the UAS site to stimulate activated transcription. The results suggested that the altered chimeric protein sequestered the transcriptional adaptor preventing activated transcription but not basal transcription.

As described herein, the transcriptional adaptor(s) in yeast has been identified, through the use of a genetic selection procedure established to do so (Example 1) . In the work described herein the chimeric protein GAL4-VP16 was used to inhibit transcription in yeast. It was found that GAL4-VP16 was toxic at high dosages, but not at low. Mutations of the acidic activation domain of VP16 of the chimeric protein reduced its toxicity. In addition, the mutated protein had a reduced activation potency when tested for its ability to activate transcription in vivo. Thus it was believed that the toxicity in vivo was due to the sequestration of the VP-16 target (i.e., the adaptor) from the transcriptional system.

In general, mutations that weaken the interaction between the chimeric protein GAL4-VP16 and the GTFs will prevent the sequestration of the transcriptional adaptor and spare toxicity. These mutations could reside in the basal factor that is the ultimate target of the activation signal, or in any adaptors that connect the GAL4-VP16 to the general factors. GAL4-VP16 resistant yeast mutants were isolated. Rare mutants were found which did not bear lesions in the plasmid, but rather in the host. These host mutants were tested to determine whether resistance to toxicity was dominant or recessive, by crossing them to a wild type strain. In all cases, resistance was recessive to the wild type. Complementation studies indicated that the mutations lay in three particular genes termed ADAl, ADA2, and ADA3. It appears that the ADAl mutant may resist toxicity by reducing the level of expression of GAL4-VP16, by reducing the expression of the ADAl promoter, which drives the synthesis of GAL4-VP16. However, mutations in ADA2 and ADA3 did not reduce the levels of GAL4-VP16, and were thus characterized further. The ADA2 and ADA3 mutants grow much more slowly than the wild type strain. Further, the ability of GAL4-VP16 to activate transcription was reduced in the ADA2 strain, even though the levels of this protein were identical to the wild type strain. The potency of acidic activation domains per se was shown to be reduced in the ADA2 and ADA3 strains, using a chimeric protein consisting of the bacterial repressor, LexA, fused to an acidic activation domains belonging either to GCN4 or HAP4 (Example 2) . The ability of both of the LexA fusions to activate transcription (from a promoter containing a LexA binding site and TATA box) was greatly reduced in the mutants. Thus, the ADA2 and ADA3 gene products are necessary for the normal function of acidic domains of transcriptional activators in vivo. Since mutations in ADA2 and ADA3 prevent toxicity by GAL4-VP16, prevent normal activation by GAL4-VP16 and prevent normal activation by other acidic activation domains, it appears these genes encode proteins that are necessary in the normal function of acidic activation domains. Because all three ADA- mutants are recessive to wild type, it was possible to clone the genes from a yeast library by rescue of the slow growth phenotype. The gene were sequenced and the sequences of ADA2 and ADA3 do not correspond to any known yeast gene, nor are they similar over an extended region to any genes in data bases. The sequences of the ADA2 and the ADA3 genes, as well as the amino sequence encoded by each, are shown in Figure 1 (SEQ ID NO. 1) and Figure 2 (SEQ ID NO. 3) , respectively. The lack of similarity between these two sequences appears to be due to the fact that this is the first isolation of genes encoding proteins with the biological properties of adaptors.

The present invention will now be illustrated by the following examples, which are not intended to be limiting in any way.

Example 1 ISOLATION OF TRANSCRIPTIONAL ADAPTOR ADA2 EXPERIMENTAL PROCEDURES A GAL4 deletion strain: The yeast strain BP1 was derived from BWG1-7A (MATA ura3-52 leu2-3, 2-112 his4-519 adel-100, Guarente and Mason, Cell. 11:1173-1181, 1983) by deleting the GAL4 gene as follows. A BamHI/EcoRI fragment from pMA441 (Ma and Ptashne, Cell. .50:137-142, 1987) was introduced into the corresponding backbone of pMA242 (Ma and Ptashne, Cell. 48.:847-853, 1987) cut with BamHI and EcoRI. The resulting construct has an unique EcoRI site from the GAL4 gene. It was thereafter cut with Xhol (+220) and EcoRI, filled-in, and a Not! linker was introduced. The HIS4 gene, flanked by NotI linkers, was finally cloned into this unique NotI site to generate PBPDGAL4. B G 1-7A was transformed with the BamHl/MIuI fragment from pBPΔGAL4 and His+/gal- transformants isolated.

GAL4-VP16 expression plasmids: Plasmids for expressing GAL4-VP16 fusion proteins in yeast were constructed as follows. To express GAL4-VP16 from a high copy plasmid containing a strong constitutive promoter, we first generated pDB20L by cleaving plasmid pDB20 (2lm origin, ADH1 promoter; Becker, et al.. PNAS. 88.:1968-1972, 1991) with BamHI and moving the ADH promoter/terminator cassette into Yep 351. pDB20L was then modified by inserting a Bglll linker in place of a small Hindlll fragment. Restriction fragments from pJL2 (an E. coli expression vector for GAL4-VP16; Chasman, et al.. Mol. Cell Biol.. £:4746-4749, 1989), extending from an NIalll site (at codon 1 of GAL4) to a Hindlll site (520 bp 3' of the VP16 stop codon) were cloned into pEMBL19+ (Triezenberg, et al. , Genes Dev. 2:730-742. 1988), generating plasmid pJR3. A BamHI fragment from pJR3, bearing 24 bp of polylinker, the GAL4-VP16 fusion gene, and 119 bp of 3' flanking sequence, was then ligated into the Bglll site of the modified pDB20, generating plasmid pSB201. Selected missense mutations within the full-length VP16 activation domain (J.L.R. and S.J.T., unpublished data) were first subcloned into an E. coli expression vector analogous to pJL2 (A. Cress and S.J.T., unpublished data), and were then inserted into pJR3 by replacing the wild type Xhol to Hindlll fragment. Mutant derivatives of pSB201 were then made using the corresponding BamHI fragments from the pJR3 derivatives.

To express GAL4-VP16 derivatives from a low copy (ARS-CEN) plasmid, BamHI fragments from wild type (WT) or mutant pSB201 plasmids (containing the ADH promoter, the fusion gene, and the ADH transcriptional terminator) were ligated into the BamHI site of pRS315 (Sikorski and Hieter, Genetics 122; 19-27, 1989), generating pSB202 and mutant derivatives. Orientations of the inserts were determined by digestion with Xhol and agarose gel electrophoresis.

The plasmids expressing GAL4 DNA binding domain mutants fused to wild type VP16 activation domains were constructed either by site-directed mutagenesis, or by ligations using fragments containing missense mutations. In the first case, a twenty-seven nucleotide (nt) oligomer (nt 527 to 553 in the numbering system of Laughon and Gesteland, 1984) containing the Leu32Pro mutation (Johnston and Dover, Proc. Natl. Acad. Sci. USA 84_.:2401- 2405, 1987) and an engineered Xcml restriction site was hybridized to single-stranded pJR3 containing wildtype GAL4-VP16. Double-stranded DNA was generated using the Klenow fragment of E. coli DNA polymerase I. Leu32Pro mutation-containing plasmids were identified using the novel Xcml site. Subsequent cloning steps into the high copy yeast expression plasmids, to generate pSB203, were identical to those outlined above. The second method of generating mutations in the GAL4 DNA binding domain was by three-way restriction fragment ligation. The mutations Pro42Leu (pBM850) , Cysl4Tyr (pBM848) , and Lysl7Glu (pBM792) were contained in the parent plasmid pBM292 (a gift of M. Johnston; see Johnston and Dover, 1987) . These plasmids were digested with SphI and Xhol and the 194 nt fragment containing the GAL4 DNA binding domain was gel-purified. Two fragments from pJR3, a 1 kb Xhol to Hindlll and a 4 kb SphI to Hindlll were also gel-isolated and the three fragments were ligated.

Correct plasmids were identified using additional internal restriction sites to discern orientations. Subsequent cloning into expression plasmids was the same as described above.

Construction of GAL4-HAP4: The 2-micron GAL4-VP16 plasmid was cleaved with X al removing the VP16 domain and the ADH terminator. The HAP4 activation domain was released from pSF 06 (Forsburg and Guarente, Annu. Rev. Cell Biol. 5_:153-180, 1989) as a Hpal fragment (containing codons 330-554 of HAP4) and a Xmal linker attached. Following digestion with Xmal this fragment was ligated into the 2-micron backbone. The orientation of the insert was confirmed by further digests. The GAL4-HAP4 junction was determined to be in frame by DNA sequencing. The fusion gene was also subcloned into the E. coli expression vector used for producing GAL4-VP16 fusion proteins (Chasman et al.. Mol. Cell Biol. 9:4746-4749. 1989). Construction of GAL4 fusion genes under control of UAS2UP1: GAL4-VP16 was excised from pJR3 on a BamHI fragment and inserted into pLG265UPl-ATG (made by cleaving pLG265UPl (Guarente et a , Cell 16:503-511, 1984) with Xho and Sad (removing the CYCl ATG) and inserting the

Xho-SacI fragment from pLGSD5-ATG) . GAL4-HAP4 was excised from the ADH 1 promoter plasmid on a Sall-SSstI fragment and swapped into the aforementioned plasmid (containing GAL4-VP16 under control of UAS2UP1 ) . The final construct contains the GAL4 fusion gene driven by UAS2UP1 of CYCl on a 2-micron plasmid with the URA3 marker.

Selection of GAL4-VP16 resistant mutants: The Δgal4 strain, BP1, was transformed with the GAL4-VP16 high expression construct and Leu-f transfor ants were selected. The vast majority (99.9%) were tiny colonies, due to the toxicity of GAL4-VP16. Thirty seven larger colonies were picked and purified. Genetic tests were carried out on these candidates, including curing and retransforming to test whether resistance were linked to GAL4-VP16, and mating to an alpha tester (PSY316, MATA ade2-101 his3-del.200 leu2-3, 2-112 lys2 ura3-53) to determine whether resistance were recessive. We found that two of the 37 strains bore host mutations, both of which were recessive. By crossing the mutants and conducting subsequent complementation tests, both mutations were placed in the same complementation group, ADA 1. The selection was repeated, this time mutagenizing BP1 with EMS prior to transformation with the GAL4-VP16 vector (Sherman et al. , Cold Spring Harbor Laboratory. 1986) . Forty one resistant colonies were picked and purified. 38 out of the 41 resistant colonies bore recessive host mutations. Complementation in crosses between mutants indicated that the EMS-derived mutants included eight more alleles of ADA 1. Remaining mutants were grouped by complementation tests using cloned genes that were subsequently isolated. These tests identified two alleles of ADA2, and one allele of ADA3. The remainder of the mutants were either unstable, or had phenotypes too weak to score in complementation tests.

Cloning and seguencing ADA2: The ADA2-1 mutant was cured of the GAL4-VP16 plasmid and transformed with a yeast genomic library on an ARS-CEN vector ( Rose et al. , Gene €>p_:237-243, 1987) and faster growing colonies isolated. Two clones with overlapping inserts containing the desired gene were identified by their ability to restore growth and sensitivity to GAL4-VP16 when retransformed into ADA2-1. The shorter of the two inserts (Al, 7.1 Kb) was used thereafter. Tests described in the text showed that this insert mapped at the ADA2 locus in the yeast genome. A 3.4 Kb EcoRV/Hindlll fragment from clone Al when cloned into the yeast ARS/CEN vector pRS31 6 (Sikorski and Hieter, Genetics 122:19-27. 1989) complemented both ADA2 mutants. This fragment was used for nested (Erase-a-Base, Promega) and internal deletions, both in pRS316 and BlueScript (Stratagene) , in order to locate the complementation activity and to sequence one strand of the gene (Sequenase version 2, U.S. Biochemical). The other strand was sequenced by using selected oligonucleotides complementary to the first strand. The DNA sequence identified an ORF of 434 codons. This ORF was isolated on a DNA fragment generated by the polymerase chain reaction and subcloned into the yeast expression plasmid pDB20 (Becker et al. , 1991 ) . The subclone complemented the ADA2 mutation, verifying that it encodes ADA2 function.

Deletion of the ADA2 gene: A derivative of BlueScript containing a 3.4 kb insert of the ADA2 gene was cleaved with Sail, filled in using Klenow fragment, and ligated to remove the Sail (and AccI) site in the polylinker. The resulting plasmid was cleaved with AccI, which cleaves at codon 41 of ADA2 and MscI, which cleaves just past the stop codon of ADA2, and filled in using Klenow fragment. Into this backbone was inserted a filled in Bglll-BAMI fragment of pNK51 containing the hisG cassette (and URA3 gene) . The resulting plasmid was cleaved with BAMI and Xhol (leaving 430 base pairs of ADA2 upstream DNA and about 1600 base pairs of downstream DNA) and transformed into strains BWG1-7A and PSY316. Slow growing transformants predominated and were shown to be resistant to GAL4-VP16. Transformants were isolated and plated on medium with 5-fluoroorotic acid to select for Ura- excision of the URA3 gene. The resulting strains were still slow growing and resistant to GAL4-VP16. These strains were transformed with the ADA2 clone on a URA3 plasmid and were restored to wild type growth. Further, sensitivity to GAL4-VP16 was also restored.

Media and beta-galactosidase assay: Appropriate yeast strains were transformed with the indicated lacZ reporters (2-micron, URA3) and, in some cases, GAL4- or LexA-fusion effector plasmids (either 2-micron or ARS/CEN, LEU2) using the lithium acetate protocol (Sherman et al.. 1986) . Cells were grown for A-galactosidase assays in SD minimal medium supplemented with amino acids and adenine. The assay was carried out either by permeabilizing whole cells with chloroform and SDS (Guarente and Mason, 1983) or by breaking cells with glass beads (Rose and Botstein, 1983) . In the former case, units of activity are normalized to cell OD₆₀₀, while in the latter case to protein concentrations (Bradford assay, BioRad) .

Gel shift assav: To prepare whole cell yeast extracts, 100 mis of culture were grown under conditions to select for the presence relevant of plasmids. At OD 600 of 1.0, the cultures were harvested, washed in extraction buffer (200 mM Tris-Cl, [pH 8], 400 mM (NH₄)₂S0₄, 10 mM MgCl₂, 1 mM EDTA, 10% glycerol, 1 mM PMSF, 1 mg/ml leupeptin and pepstatin, 7 mM 2-mercaptoethanol) , pelleted, and resuspended in 400Ϊ1 of storage buffer (20 mM HEPES, pH 8, 5 mM EDTA, 20% glycerol, 7 mM 2-mercaptoethanol, 1 mM PMSF and 1 mg/ml of both leupeptin and pepstatin) . Glass beads were added and the mixture was vortexed 7 times for one minute each, followed by a 5 minute Eppendorf centrifugation to clear the lysate of cell debris. The supernatant was then centrifuged for one hour and the resulting supernatant was stored at -70"C. The protein concentration of these extracts was about 7 mg/ml.

DNA binding and gel retardation were done using a double-stranded oligomer containing a consensus GAL4 binding site: 5•-CGGAGGGCTGTCAACCCG-3*. .Extract (50 ig) and purified GAL4-VP16 (5 ng about 0.1 pmol) was incubated at 25|C with 0.05 ng of radiolabeled DNA probe for 20 minutes in binding buffer (200 mM NaCl, 20 mM Tris-Cl, pH 8, 20 mM MgCl₂, 20% glycerol, and 0.01% bromophenol blue) in a total volume of 20 ml. Samples were electrophoresed in a 4% polyacrylamide gel at 20 mA (approximately 125 V) .

RNA analysis: Total RNA from BP1 and ADA mutant strains was prepared as described in Osborne and Guarente (1988) . RNA samples (20tg) were run in 1% agarose-MOPS-formaldehyde gels and blotted to nitrocellulose filters essentially as described in Sambrook et al. Cold Spring Harbor Laboratory. 1989. Filters were prehybridized and hybridized either in 5 x SSC, 5 x Denhardt's and 100 mg/ml denaturated calf thymus DNA at 45IC (oligonucleotide probes) or in 5x SSC, 5x Denhardt's, 50% formamide and 100 mg/ml denaturated calf thymus DNA at 42|C (ds-DNA probes) . Hybridization mixture for ds-DNA probes included 10% dextran sulfate. Prehybridization, hybridization and washings were performed as described in Sambrook et al.. Molecular Cloning; A Laboratory Manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory (1989) . Probes for ADH1 and GCN4 genes were 5'-labelled synthetic oligonucleotides

(5'-CCGTGGGATTCGTAGAAGATAACACC-3• and 5'-CCATCCAATGGTGAGAAACCCATTGG-3^* , respectively). The ACT1 probe was an EcoRI fragment from pYactl (Ng and Abelson,

Proc. Natl. Acad. Sci. USA 77:3912-3916_r 1980). The SPT15 (TFIID) probe was a BamHl/EcoRI fragment from pDE28-6 (Eisenmann et a ., Cell 58:1183-1191. 1989) . ds-DNA probes were labelled by random priming (United States Biochemical) . Results were analyzed on a Phosphorl ager (Molecular Dynamics) .

Western Blot: Yeast extracts from BP1 and ADA2-2 strains carrying the plasmid Lexgcn4-N125 (Hope and Struhl, Cell 4j5:885-894, 1986) were performed as described in Pfeifer et al. (1987) . Protein samples (lOO±g as determined by Bradford analysis, BioRad) were run on 12% SDS-polyacryla ide gels, electroblotted to Immobilon-P membranes (Millipore) and processed as described in Haldi and Guarente (1989) . Immunological detection of IexA-GCN4 was carried out using 1:500 dilution of rabbit Anti-lexA antiserum (Neosystems) as first antibody and 1:5000 dilution of goat anti-rabbit IgG (alkaline phosphatase conjugate, Sigma) as second antibody. The blot was developed with Lumi-Phos 530 (Boehringer Mannheim) and the result recorded with Kodak X-ray film.

Sequence analysis: Sequence searches in GenBank, EMBL data bank and PIR and TFDAA protein banks were performed using the Altschul et al. J. Mol. Biol. 215:403-410. (1990) algorithm at the National Center for Biotechnology Information (NCBI) using the BLAST network service.

In vitro transcription: Yeast nuclear extracts were prepared from BP1 and the ADA2 mutant derivative strain essentially as described by Lue and Kornberg PNAS, 84:8839-8843 1987), with modifications (Berger, et al.. Cell .61:1199-1208, 1990). The strains used were haploid and the final protein concentrations were approximately one-half (25-30 mg/ml) of that obtained previously using BJ296, a diploid strain.

In vitro transcription conditions using purified GAL4-VP16 and primer extension analyses were as described (Berger, et ah, Cell 61:1199-1208. 1990). The GCN4 protein was expressed in E. coli strain AR68 using plasmid pABlOO and was purified as described (Arndt and Fink, Proc. Natl. Acad. Sci. USA 83:8516-8520. 1986), with the additional step of a Bio-Gel P60 sizing column developed in 100 mM Tris-Cl (pH 7.5 at 4|C), 200 mM KC1, 10 mM 2-mercaptoethanol, 5% glycerol (v/v) . The GAL4-HAP4 fusion protein was prepared from crude E. coli lysates by centrifugation and ammonium sulfate precipitation at 40% saturation. This pellet was resuspended in Buffer A100 (Chasman et al.. Mol. Cell. Biol. £:4746-4749, 1989) and used as described below.

Activation by GAL4-VP16 and GCN4 in wild type and ADA2 mutant extracts. Reactions contained 55 ig of protein of wild-type or mutant nuclear extracts. The template contains the CYCl TATA box-mRNA initiation region and either a single GCN4-binding site positioned at - 178(HIS 14 X 1) (Hinnebusch et al.. PNAS_P 12.:498-502, 1985) or three GAL4 sites positioned at -65 (gal.3) (Berger et al.. 1990). The GCN4 sets contain 0.0, 0.32, and 0.8 ig of GCN4 purified from bacteria. The GAL4-VP16 sets contain 0.0, 0.06, and 0.3 ig of GAL4-VP16. Products of primer extension were run on 6% acylamide gels with 8 M urea at 1000 V. The gel was dried and exposed onto Kodak XR5 film. in vitro activation of transcription at the dA:dT UAS in wild-type and ADA2-2 mutant extracts. Reaction conditions were the same as stated in the previous paragraph, except 132 ig protein of wild-type and mutant extract was used. The higher extract concentration aids activation at the dA:dT UAS. The basal template was gal.3, and the dA:dT template was pAT2, which is activated in vitro by the dA:dT activator (Lue and Kornberg, 1987).

In vitro activation of GAL4-HAP4 and GAL4-VP16 in wild-type and ADA2-2 mutant extracts. Reaction conditions were the same as above, and the gal.3 template was employed throughout. GAL4-HAP4 was purified from E. coli and is about 10% pure as judged by analysis of a stained SDS-polyacrylamide gel. The GAL4-HAP4 sets contain 0.0, 1.33, and 4.0 ig of the preparation of GAL4-HAP4 and GAL4 VP16 set contains 0.0, 0.3, and 1.0 ig of GAL4-VP16.

RESULTS

Strategy for isolating mutations in a transcriptional adaptor

Our in vitro study (Berger et al. , Cell .61:1199-1208, 1990) and earlier in vivo studies (Gill and Ptashne, Nature 114.:721-724, 198.8) suggested that GAL4-VP16 could be used as a transcriptional inhibitor in yeast cells. We reasoned that inhibition of transcription in vivo might provide a genetic selection for isolating the VP16 target. Thus, a plasmid was constructed to express GAL4-VP16 from the constitutive ADH1 promoter in vivo. When this high copy plasmid (pSB201) was introduced into strain BP1 (bearing a deletion of the resident GAL4 gene) , cell growth was strongly inhibited. To assess whether growth arrest was specifically due to sequestration of the VP16 target, we examined the effects of mutations in the VP16 domain known to affect its activation potential (Cress and Triezenberg, Science 251:87-90, 1991). Three mutations change the critical Phe442 to Tyr, Ala, or Pro, while another changes a cluster of four acidic residues. These mutations were originally characterized in a truncated version of the VP16 activation domain (Cress and Triezenberg, Science 251:87-90, 1991). In yeast the fusion protein GAL4-VP16Δ456 is unstable. Therefore, we used the selected mutations in the context on the entire VP16 activation domain (residues 413 to 490) . The mutations relieved toxicity to a variable extent, with slight relief observed in DN2579 and FY, and greater relief observed in FA and FP, as shown in TABLE 1 below.

Table 1. Inhibition and Activation by GAL4-VP16 Expression Plasmids and Mutant Derivatives in vivo

Colony Size Growth on GAL

Activation

++ 1200 ND 1100 ND 610 ND 490 ND 420

5

Inhibition tests used the high copy expression plasmids, and growth and activation tests used single copy plasmids. Inhibition of growth by GAL4-VP16 is tabulated above. Growth on galactose plates reflects the relative ability of the expression plasmids to complement the GAL4 deletion in strain BP1. Activation assays test the ability of derivatives to trans-activate the GAL reporter, pLGSD5 (Guarente et al. , Proc. Natl. Acad. Sci., 79:741θ^'-7414, 1982) . Units of beta-galactosidase normalized to cell density are averages of triplicate assays that differed by less than 10%. WT is GAL4-VP16 constructed as described in Experimental Procedures above. DN2579 changes four Asp residues to Asn in the VP16 moiety at positions 429, 440, 443, and 449. F442Y, F442A, and F442P change Phe442 of VP16 to Tyr, Ala, and Pro, respectively. P42L, C14Y,

K17E, and L32P change residues in the GAL4 moiety, Pro42 to Leu, Cysl4 to Tyr, Lysl7 to Glu, and Leu32 to Pro, respectively.

We then determined the ability of the mutants to trans-activate transcription in yeast cells by expressing a non-toxic level of wild type (WT) or mutant proteins on a single copy plasmid (pSB202 and derivatives thereof) . The FA and FP mutants were substantially defective in activation of a CYCl-lacZ reporter driven by GAL4 binding sites, while FY or the DN2579 of acidic residues were only slightly defective. The ability of the mutants to inhibit growth paralleled their ability to activate transcription (TABLE 1) . Thus, we concluded that inhibition of growth was due to the interaction between VP16 and its target in vivo. To further define the action of GAL4-VP16 in vivo, we determined whether growth inhibition depended on DNA binding. For this purpose, mutants bearing alterations in the GAL4 DNA-binding domain (Johnston and Dover, Proc. Natl. Acad. Sci. USA 8_4:2401-2405, 1987) were examined for trans-activation and growth inhibition in yeast cells. Two of the mutants, Cysl4 to Tyr and Pro42 to Leu partially complemented the gal4 deletion when expressed at the non-toxic level, and two other mutants, Lysl7 to Glu and Leu32 to Pro did not complement at all (TABLE 1) .

This pattern followed the known effects of these mutations on DNA-binding as measured in vitro (Johnston and Dover, Proc. Natl. Acad. Sci. USA 84:2401-2405, 1987). The ability of the mutants to inhibit growth when expressed to a high level again correlated with their functional activity (TABLE 1) . The extent of inhibition by the mutants Pro42—Leu and Cys 14—Tyr more closely followed their ability to bind DNA (Johnston and Dover, Proc. Natl. Acad. Sci. USA 81:2401-2405, 1987) than did complementation of the gal4 mutation for growth. Thus, inhibition of yeast cell growth by GAL4-VP16 required the integrity of both the VP16 activation domain and the GAL4 DNA-binding domain.

A model to explain these observations is shown in Figure 3. A high level of GAL4-VP16 drives binding of the protein to many secondary or non-specific sites in the yeast genome, shown in Figure 3A. The binding of the protein in the proximity of sites that can bind TFIID will promote the assembly of transcription complexes including the general transcription factors and any bridging factors, i.e. adaptors. This assembly sequesters general factors, rendering them limiting and therefore inhibiting the growth of yeast cells. Mutations in either the VP16 activation surface or in the GAL4 DNA-binding domain prevent the assembly of these complexes and restore growth. Figure 3B shows the expectation that mutation of the adaptor will also disrupt the inhibition complex and restore growth.

We propose that inhibition is due to a process analogous to cis-inhibition in vitro (Berger et al. , Cell 61:1199-1208, 1990), in which GAL4-VP16 sequesters its target on the DNA. Because high levels of GAL4-VP16 are needed to cause inhibition, we infer that the inhibitor could bind to many secondary sites in the yeast genome. The complex trapped at these sites would include the putative adaptor, as well as one or more of the general factors. We further posit that growth inhibition is due to a depletion of general factors, since the non-DNA— binding forms of GAL4-VP16, presumably still capable of binding to the adaptor, are not toxic. By this interpretation, we hypothesized that mutations in the adaptor that weakened or abolished the interaction with VP16 or with general factors would free up general factors (Figure 3B) and restore growth. An assumption behind the prediction is that ^"normal yeast activators will function at a residual level in the mutant strain. This could occur if the mutation is leaky, leaving some residual activity; if the mutation were specific to VP16 or similar activators; or if some activation can occur in the absence of the adaptor. Our selection might also identify mutations in general factors that would be unable to bind adaptors or activators, and thereby avoid sequestration. Alternative interpretations of the growth inhibition data may also be considered. In one extreme, inhibition could be due to a direct interaction between GAL4-VP16 and one or more general transcription factor. In another, inhibition could be unrelated to sequestration of a transcription factor, but due to the inappropriate expression of cellular genes turned on by the presence of abnormally high levels of GAL4-VP16. Clearly, the properties of mutant strains resistant to GAL4-VP16 would distinguish among these alternatives.

Isolation of GAL4-VP16 resistant mutants

Strain BP1 was transformed with the high-expressing GAL4-VP16 construct, and mutants that gave rise to large colonies on glucose minimal agar plates were isolated (see Experimental Procedures) . The majority of strains that resulted from this procedure bore mutations in the GAL4-VP16 plasmid and were not studied further. To increase the chances of obtaining mutants of the host rather than GAL4-VP16, we also screened for secondary phenotypes, such as slow growth. Mutant strains were obtained which grew slowly compared to the wild type parent and retained their resistance to GAL4-VP16 when cured and retransformed with the expression plasmid. These strains, therefore, bore host mutations that resulted in resistance to toxicity.

When these mutants were mated with a wild type strain, sensitivity to GAL4-VP16 was restored. Thus, the mutants were recessive to wild type with respect to their ability to resist toxicity by GAL4-VP16. Because of the recessive nature of the mutations, they could be divided into complementation groups by crosses between mutant strains. We found that the mutants comprised three complementation groups termed ADAl, ADA2 and ADA3 (alteration/deficiency in activation) . These tests identified eight alleles of ADA 1, two alleles of ADA2, and one allele of ADA3. The other mutants were either unstable or had phenotypes too weak to score in complementation tests.

To test whether resistance to toxicity in the mutants was due simply to decreased synthesis of GAL4-VP16, the activity of the ADH1 promoter was determined by northern analysis. Endogenous ADH1 gene expression was probed in comparable amounts of total RNA from the isogenic wild type and mutant strains. This revealed that ADAl mutants had reduced levels of ADH1 RNA, while ADA2 and ADA3 mutants had levels comparable to the wild type. To further test whether resistance to GAL4-VP16 was due to effects on the ADH1 promoter, cells were transformed with a plasmid that expressed GAL4-VP16 from a different promoter (UAS2upl) . The ADA2 and ADA3 mutants were resistant to growth inhibition by this construct, while ADAl and wild type were partially sensitive. Thus, the ADA2 and ADA3 mutants were candidates for alterations in the VP16 target. This report describes the further characterization of ADA2 and the ADA2-2 mutant allele.

Cloning and seguencing^'of the ADA2 gene

When the ADA2-2 mutant was crossed to a wild type strain, slow growth segregated 2:2 in 7 complete tetrads and when tested cosegregated with resistance to GAL4-VP16. We therefore concluded that the two phenotypes were due to the same mutation, and employed the slow growth phenotype to isolate a complementing clone from a yeast genomic library in a single copy ARS CEN-containing vector (Rose et al. , Gene 60.:237-243, 1987). Two complementing clones contained overlapping inserts. In addition to restoring normal growth to the ADA2 mutant, the clones also restored sensitivity to GAL4-VP16. To determine whether the insert directed integration to the ADA2 region of the yeast genome, the ADA2 mutant was transformed with a plasmid missing an ARS and CEN but bearing the complementing insert and the URA3 marker. Integration was directed by cleaving the plasmid within the insert, and a Ura+ transformant was recovered which grew at the wild type rate. This transfor ant was crossed to a wild type strain, the diploid sporulated, and tetrads dissected. In 11 of 11 cases tested, tetrads contained 4 normal growing segregants and no slow growers. These data indicated that the cloned insert directed integration to the ADA2-2 locus, and therefore contained the ADA2 gene.

To locate the complementing region more precisely, 5* and 3' deletions were generated across the insert and the ability of deletion-bearing constructs to complement the ADA2-2 mutation was tested. The DNA corresponding to the complementing region was sequenced, revealing an open reading frame of 434 codons (Figure 1, SEQ ID NO. 1) . To prove that this open reading frame encoded ADA2, we employed PCR to subclone it into a yeast expression plasmid and found that the subclone complemented the ADA2 mutation. Data base searches indicated that the ADA2 sequence did not correspond to any known yeast gene, nor was it an obvious homolog of any gene from another organism.

There are several short sequence motifs of potential importance in the ADA2 sequence all of which occur in the amino-terminal quarter of the protein. First, the amino terminus (residues 7-34) contains 7 Cys residues. However, the spacing of the Cys residues does not align with any known.class of zinc finger DNA-binding proteins, such as the GAL4 family in yeast, or the steroid hormone receptors. Rather, an alignment is observed with the

Drosophila gene product, ref(2)P (Dezelee et al.. 1989) (Figure 4, SEQ ID No. 1) . Second, the Cys rich region is followed by a region (residues 65-95) with significant similarity to the yb motif, a repeat found in the myb family of oncogenes (Lane et al.. 1990) (Figure 4, SEQ ID No. 1) . There is a particularly good correspondence between the residues of ADA2 that correspond to the most conserved residues of the myb consensus. Curiously, while the myb sequence is found in 2 or 3 repeats in proteins which are members of the myb family, ADA2 contains a single copy. Third, region 89113 of ADA2 shows similarity to a region of the enzyme peptidylglycine monooxygenase (Stoffers et al. , 1989) .

The ADA2 mutation weakens trans-activation by several acidic activators

If the ADA2 mutant escaped toxicity by GAL4-VP16 because the VP16 target was altered, then trans-activation by VP16 should be reduced in the mutant strain. To test this surmise, we employed an ARS-CEN low copy plasmid bearing the GAL4-VP16 Phe442Ala mutant (GAL4-VP16FA) , and a reporter driven by the GAL-UAS, pLGSD5 (Guarente et al.. 1982) . Note that GAL4-VP16FA trans-activates with a minimum of toxicity in the wild type control strain. We found that levels of beta-galactosidase from the reporter plasmid were reduced about 6-fold in the ADA2 mutant strain, as shown in TABLE 2.

Table 2. Trans-Activation by GAL4-VP16 in the ADA2 Mutant Activity of GAL Reporter

WT Control 240

WT GAL4-VP16FA 7300

ADA2 GAL4-VP16FA 1200

An ARS-CEN plasmid containing GAL4-VP16 bearing the Phe442 to Ala mutation was transformed into the wild type strain BP1 and the ADA2 mutant. This plasmid confers minimal toxicity in BP1. The cells also bore the pLGSD5 GAL reporter. Levels of betagalactosidase were assayed in glass bead extracts. Results shown are the means from four replicate assays given as specific activity (units/mg total protein) . In two extracts each, the levels of GAL4-VP16FA were determined by a gel shift assay.

Preparation of nuclear extracts and transcription conditions are described in Experimental Procedures.

Reactions contained 55 ig protein of wild type or mutant nuclear extracts. The template contains the CYCl TATA box-mRNA initiation region, and either a single GCN4 binding site positioned at -178 (HIS 14 x 1) (Hinnebusch et al.. 1985) or three GAL4 sites positioned at -65

(gal.3) (Berger et al.. 1990). The GCN4 sets contain 0.0, 0.32, and 0.8 ϊg of GCN4 purified from bacteria. The GAL4-VP16 sets contain 0.0, 0.06, and 0.3 ig of GAL4-VP16. Products of primer extension were run on 6% acrylamide gels with 8 M urea at 1000 V. The gel was dried and exposed onto Kodak XR5 film.

In another gel shift whole cell extracts of BP1 (WT) with or without the GAL4-VP16 expression plasmid or the ADA2- 2 mutant with the plasmid were made in duplicate (see TABLE 2 and Experimental Procedures) . Extracts were mixed with a GAL4-specific oligonucleotide and electrophoresed. Purified GAL4-VP16 (Berger et al.. 1990) was mixed as a control. Complexes in the control not bearing the expression plasmid are found in all cases and correspond to cellular proteins that are not related to GAL4-VP16. The intact GAL4-VP16 complex is indicated by the heavy arrow, and complexes containing partially degraded GAL4-VP16FA are indicated by the light arrows. The amount of GAL4-VP16FA is comparable in the wild type and mutant strains. Gel shift analysis showed that the level of

GAL4-VP16FA was comparable in the wild type and mutant strains. Thus, the specific activity of the protein, that is its ability to activate transcription, is partially defective in the ADA2 mutant. If ADA2 were an adaptor for at least certain acidic activators, then the expression of other yeast genes might be affected in the mutant. Accordingly, we determined the activity of lacZ reporters that all bore the CYCl TATA box-mRNA initiation region, coupled to different UASs. The UASs were UASl or UAS2 of CYCl activated by HAPl or HAP2/3/4 respectively) (Forsburg and Guarente, Annu. Rev. Cell Biol. 5_:153-180, 1989), the HIS4 UAS (activated by GCN4 and BAS1/2) (Arndt et al.. Science 112:874-880, 1987) , or synthetic GCN4 binding sites (Hinnebusch et al. , Proc. Natl. Acad. Sci. USA 82:498-502, 1985). While the reporters activated by HAPl or HAP2/3/4 showed a small or no reduction in the ADA2 mutant, the reporters activated by GCN4 were reduced 5-10 fold in the mutant strain (TABLE 3).

Table 3. Activity of Several Yeast UASs in the ADA2 Mutant

Reporters were transformed into the wild type and ADA2 mutant strains, and beta-galactosidase levels assayed in permeabilized cells. The reporters were pLGΔAluXho (Guarente et al.. 1984), which contains UASl of CYCl, pLG265UPl (Forsburg and Guarente, 1989) , which contains UAS2 of CYCl, HIS(1)66 and HIS(2)14x2 (Hinnebusch et al.. 1985) which contain the HIS4 UAS or two synthetic GCN4-binding sites. All UASs lie upstream of the CYCl

TATA box-initiation region abutting CYCl-IacZ. LexA-GCN4 trans-activation was assayed in strains containing the lexA-GCN4-N125 expression plasmid (containing residues 1-87 of lexA and 2-125 of GCN4 expressed from the DED1 promoter) and the reporter plasmid YEp21-Sc3423 (Hope and Struhl, 1986) containing CYCl-IacZ under the control of a single lexA binding site. A western blot of shows that the levels of IexA-GCN4 in the wild type and mutant strains is the same. A similar experiment was attempted with a IexA-HAP4 expression plasmid, but it was not possible to detect the fusion protein by western blotting to normalize for levels in the wild type versus the mutant.

These differential effects were striking because all of these activators are of the acidic class. Transcription of several other yeast genes was monitored by northern blotting. This assay showed no significant reduction in RNA levels of ACT1 (Actin) , SPT15 (TFIID) , or GCN4 in the ADA2 mutant.

Twenty ig of total RNA from BP1, ADA1-1, ADA1-2, ADA2-1 and ADA3-1 strains were run on denaturating agarose gels and blotted to nitrocellulose. Sequential hybridizations with different gene specific probes (Experimental Procedures) were recorded on a Phosphorlmager (Molecular Dynamics) . Probes were specific for ACT1 (actin) , SPT15 (TFIID) encoding TATA-binding protein, GCN4, or ADH1. Quantitation of the different signals showed that TFIID and GCN4 transcription is not affected by ADA mutations, whereas ADH1 transcription is reduced 5-fold in ADA1-1 and ADA1-2 relative to BP1 and not affected in ADA2 and ADA3. ACT1 was used as standard for quantitation. To ascertain whether the reduction in activity of reporter constructs was due to a defect in the function of activation domains, we employed a lexA-GCN4 fusion in which the acidic activation domain of GCN4 (residues 12-125) is fused to the DNA-binding domain of the bacterial repressor, lexA (residues 1-87) (Hope and Struhl, Cell 16:885-894, 1986). The ability of the IexA-GCN4 fusion to activate a CYCl-IacZ reporter containing a single lexA binding site was reduced 12-fold in the ADA2 mutant compared to the wild type (TABLE 3) . There was no significant difference in the levels of the IexA-GCN4 protein in the wild type and mutant strains by Western blotting with lexA antibody. This finding suggests that the function of the GCN4 acidic activation domain is compromised in the ADA2 mutant strain.

The ADA2 mutation reduces transcriptional activation in vitro

The in vivo analysis indicated that the ADA2 mutation reduced the activity of the VP16 and GCN4 acidic activation domains. Models for the role of acidic activation domains in transcriptional activation fall into two broad categories; they counteract the tendency of histones to wrap up the promoter into chromatin (the chromatin pathway) , or they interact with the general factors by DNA looping (the interaction pathway) . Any mutation reducing the functional activity of an acidic activation domain in vivo, therefore, could affect either the chromatin pathway or the interaction pathway. If the ADA2 gene product were indeed an adaptor, then it would work by the interaction pathway in the absence of chromatin. To determine which pathway was affected by the ADA2 mutation, we turned to the yeast in vitro transcription assay, employing yeast nuclear extracts from wild type and mutant strains. In the absence of a chromatin assembly system, the in vitro reaction should proceed on DNA templates that are not wrapped into chromatin. An experiment was carried out. comparing the wild type and mutant extracts in their response to the addition of purified GAL4-VP16 or GCN4 proteins. In these reactions, the templates contained the CYCl-lacZ gene and CYCl TATA boxes. The GAL4-VP16 responsive template contained three GAL4-binding sites at -65, and the GCN4 responsive template bore a single GCN4 binding site at position -178. The ability of GAL4-VP16 to activate transcription was abolished in the ADA2 mutant extract compared to the wild type, while the ability of GCN4 to activate transcription was substantially*reduced. Basal transcription from both the GAL4-VP16 responsive template and the GCN4-responsive template was identical for equivalent amounts of wild type and mutant extracts.

Thus, the ADA2 mutation causes a defect in transcriptional activation by both GAL4-VP16 and GCN4 and does not affect basal transcription.

In previous experiments (Berger et al. , Cell 61:1199- 1208, 1990), we found that GAL4-VP16 could selectively inhibit activation of transcription by the dA:dT activator, without affecting basal transcription. From this and other findings, we inferred that the VP16 domain sequestered a factor required for the function of the dA:dT activator, which we termed an adaptor. Therefore, if ADA2 encoded this putative adaptor, we would expect activation by the dA:dT activator to be defective in the mutant. The dA:dT UAS does not function to activate transcription in the mutant extract. Thus, ADA2 appears to be required for the function of both GAL4-VP16 and the dA:dT activator, a property imputed to the putative adaptor (Berger et al.. 1990) .

The defect in the ADA2 mutant is activation domain-specific

Reporter assays indicated that the transcriptional defect in the ADA2 mutant was promoter-specific. For example, reporters activated by GAL4-VP16 or GCN4 were defective in the mutant, while reporters activated by HAPl or HAP2/3/4 were not. One explanation for this finding is that the activation domains of HAPl and HAP4 do not require ADA2. To test this possibility, we constructed a GAL4-HAP4 fusion gene in which codons for residues 1-147 of GAL4 (the DNA-binding domain) were fused to codons for the acidic activation domain of HAP4 (residues 330-554) . The fusion protein encoded by this gene was expressed in E. coli, partially purified, and used in the in vitro transcription reaction. The template employed in this experiment is the GAL4-responsive template used above. In The gal.3 template was employed throughout. GAL4-HAP4 was purified from E. coli and is about 10% pure as judged by analysis of a stained SDS polyacrylamide gel (Experimental Procedures). The GAL4-HAP4 sets contain 0.0, 1.33, and 4.0 ig of the preparation of GAL4-HAP4 and the GAL4-VP16 set contains 0.0, 0.3, and 1.0 ig of GAL4-VP16. GAL4-HAP4 activated transcription in the wild type extract almost as well as GAL4-VP16. Strikingly, activation by GAL4-HAP4 in the ADA2 mutant extract was as great or greater than in the wild type extract. GAL4-VP16, again, did not activate transcription in the ADA2 mutant extract. These results show that the requirement for ADA2 is specific; i.e. the protein is required for the function of the VP16 acidic activation domain, but not the HAP4^* acidic activation domain.

Finally, to verify that the requirement for ADA2 was also activation domain specific in vivo, we employed the GAL4-HAP4 fusion as an inhibitor of cell growth. The levels of expression of this protein driven by the UAS2UP1 promoter was in a sensitive range for this assay. GAL4-HAP4 was markedly toxic in the ADA2 mutant. Analyzed in this experiment were streaks of the ADA2-2 mutant transformed with the plasmids gLG265UPl (vector) and derivatives that express GAL4-VP16 or GAL4-HAP4 from the 265UP1 promoter. The plates contain minimal medium supplemented with 2 % glucose, 0.004% adenine, and 0.003% casamino acids. Growth inhibition by GAL4-HAP4 is clearly evident in the ADA2-2 mutant whereas inhibition by

GAL4-VP16 is not observed.As expected, GAL4-VP16 driven by UAS2UP1 was not toxic in the ADA2 mutant. The fact that GAL4-HAP4 is toxic in the. ADA2 mutant strain suggests that the HAP4 target has not been eliminated by the ADA2 mutation.

The ADA2 gene is not essential

To determine whether the ADA2 gene was essential, we constructed a deletion mutation in the gene. The coding sequence of the ADA2 clone, from residues 33 to the stop codon, was replaced.with the hisG-URA3-hisG cassette (Alani et al. , Genetics 116:541-545, 1987) (Experimental Procedures) . This construct was used to transform haploid and diploid strains to Ura+. Both strains gave Ura+ transformants that were large colonies, and the haploid strain also gave a predominant class of small colonies. These small colonies were purified and found to be resistant to GAL4-VP16. Further, the slow growth of these transformants was complemented when mated to an ADA2+ strain, but not an ADA2- strain. This complementation test suggested that the hisG-UFlA3-hisG cassette had disrupted the ADA2 gene. A final test of this suggestion was whether the slow growth of these transformants was complemented by the ADA2 clone. The gene was jettisoned from the hisG-URA3-hisG cassette by homologous recombination, leaving the hisG gene inserted in ADA2.

When this strain was transformed with the ADA2 clone (on a URA3 plasmid) , the Ura+ transformants regained a wild type growth rate, and also regained sensitivity to GAL4-VP16. These findings show that the ADA2 gene can be disrupted in haploids, indicating that the gene is not essential. The phenotypes of the strain bearing the disruption are very similar to the ADA2-2 strain, showing slow growth and resistance to GAL4-VP16, indicating that ADA2-2 is a null mutation.

Example 2

ISOLATION OF TRANSCRIPTIONAL ADAPTOR ADA3 EXPERIMENTAL PROCEDURES

Strains and plasmids: Assays were carried out in strain BP1 (MATA, gal4::HIS4, ura3-52, leu2-2,2-112, his4-519, adel; Berger et al.. Cell in Press., 1992) and its ADA3 mutant derivative. ADA3 disruption was performed in PSY316 (MATA, ade2-101, his3del.200, leu2-3,2-112, lys2, ura3-53. Plasmids expressing GAL4-VP16 and GAL4-VP16_FA in yeast (2i origin, ADH1 promoter) have been described elsewhere (Berger et al.. Cell in Press, 1992). The plasmid YCp88-GCN4, expressing IexA-GCN4 (DED1 promoter) and the lexA reporter YEp21-Sc3423 (Hope and Struhl, 1986) were generously given by K.Struhl. Transformation were performed by the lithium acetate method (Sherman et al. , Cold Spring Harbor Laboratory. Cold Spring Harbor, New York, 1986) . β-galactosidase assays were carried out in glass bead-generated extracts as described (Rose and Botstein, J. Mol. Biol. 170:883-904. 1983). General yeast manipulation were performed as described in Guthrie and Fink (1991) with slight modifications.

Gel retardation assays: Whole cell extracts and gel shift assays were performed as described (Berger et al. _f Cell in press. 1992) .

Western blotting: 100 ig of protein from whole cell extracts were and analyzed in a 12% polyacrylamide SDS gel and electroblotted to Immobilon-P (Millipore) . The filter was blocked, washed and incubated with anti-lexA antiserum

(first antibody, Neosystem) and alkaline phosphatase-conjugated anti-rabbit IgG (from goat, Sigma) as described (Haldi and Guarente, J. Biol. Chem.

161:17107-17112, 1989). The blot was developed with

Lumi-Phos (BioRad) and the results recorded on Kodak

X-Omat AR film. Cloning and sequencing ADA3: The single ADA3 mutant strain was transformed with a yeast genomic library on a ARS-CEN vector and faster growing colonies isolated on minimal medium with 2% glucose. Two different clones with overlapping inserts were isolated and checked for their ability to restore growth and sensitivity to GAL4-VP16 when retransformed into ADA3.1. The smaller of the two clones (2112.3, 6.5 Kb), was used for further characterization. A Hindlll-EcoRV 2.5 Kb fragment was subcloned into pRS31 6 (Sirkoski and Hieter, Genetics 122:19-27. 1989) and used for nested deletions (Erase-a-Base, Promega) and sequencing the gene (Sequenase, United States Biochemical) . A clone sufficient for complementation was obtained by subcloning a 2.2 Kb Hindlll fragment into the Hindlll site of the above subclone. Sequences upstream of the Hindlll site in ADA3 were sequenced with customized oligonucleotides. The DNA sequence identified an ORF of 702 codons. This ORF was isolated on a DNA fragment generated by polymerase chain reaction and subcloned into the yeast expression plasmid pDB20LB (Berger e£ al.. Cell in press, 1992; made by B. Turcotte) . The subclone fully complemented the ADA3 mutation, verifying that it encodes the ADA3 function.

ADA3 gene disruption: BamHI-Bglll fragment from pNYK51 (Alani et al.. 1987) , containing the URA3 gene inserted between two copies of E. coli hisG gene was introduced into a Hindlll-EcoRV fragment from the genomic ADA3 clone (subcloned in Bluescript KS+)in place of the Bglll fragment that spans codons 273 to 307. The resultant construct, pAA3-i, was integrated into the pSY316 genome as a Sall-SacI fragment. URA3^* transformants were screened for resistance to GAL4-VP16 and slow growth on minimal medium. Such transformants were recovered and were treated with fluoroacetic acid (FOA, Alani et al.. 1987) to select for deletion of Ura3, to give yBPA91 f.

Primer extension analysis: RNA was isolated from yByPl and ADA3.1 grown in YEP-lactate (Osborne and Guarente, Genes Dev. 1:766-772, 1988). 50 ϊg of RNA was hybridized with 5 x 10⁴ c.p.m. of 5•-(³²P)labeled oligonucleotide (5'-GTGTAGCACCIIICTTAGCAGAACCGGCC-3*) , complementary to CYCl mRNA. The primer was extended with MLV-reverse transcriptase as described (Ausubel et al.. Molecular Biology, ,l:4.8.1-ff Wiley Interscience, 1989) and analyzed in a 8% polyacrylamide sequencing gel.

In vitro transcription: In vitro transcription reactions were carried out in nuclear extracts as described previously (Berger et al.. Cell in press, 1992). GAL4-VP16, GAL4-HAP4, and GCN4 proteins were prepared as described (Berger gt al.. 1992) .

Sequence analysis: DNA and predicted protein sequence analysis was performed at the Massachusetts Institute of Technology Computer Center at the Whitaker College, using the GCG package. Sequence searches in GenBank, EMBL data bank and PIR and TFDAA protein banks were performed using the Altschul et al. J. Mol. Biol. 215:403-410, (1990) algorithm at the National Center for Biotechnology Information (NCBI) using the BLAST network service. Sequence Information: Yeast genes sequences (Figure 5) : KEX1 (Dmochowska et al.. Cell, 5_0:573-584, 1987) VPS15 (Herman et aJ , Cell. 61:425-437, 1991), HAPl (Pfeifer et al.. Cell, .56:291-301, 1989), and LAC9 (Salmeron and Johnston, Nucleic Acids Res.. 14:7767-7781, 1986). Gene sequences from higher eukaryotes: LpSl (from sea urchin, Xiang et al.. J. Biol. Chem.. 263:17173-17180. 1988) and Smg p25A ( Bovine, Matsui et al.. Mol. Cell. Biol.. iO:4116-4122, 1990). Gene sequences from eukaryotic viruses: FIPV-22 (Feline infectious peritonitis virus, DoGroot et al.. Virology. 167:370-376. 1988), SIVgag (Simmian immunodeficiency virus, Chakrabarti et al.. Nature, 328:543-547. 1987) and HIVgag (Human immunodeficiency virus Kumar et al.. J. Virol.. 64:890- 901, 1990). Bacterial genes: gtIC (from Bacillus subtillis, Bohannon and Sonenshein, J. Bacteriol.. 171:4718-4727. 1989) and bacteriophage PRDl DNA polymerase (UDNApol; Jung et al.. PNAS. 84:8287-8291. 1987). HIVgag (Human immunodeficiency virus I, Kumar et. al., 1990) (Figure 5B) .

RESULTS

Properties of the ADA3 mutant

We found a single allele of ADA3 among mutations that resisted growth inhibition by GAL4-VP16. In these experiments, the gene encoding GAL4-VP16 was borne on a multi-copy 2-i plasmid and under control of the ADH1 promoter (Berger et al.. 1992) . Two findings indicated that the ADA3 mutant did not escape toxicity by reducing synthesis of GAL4-VP16. First, the activity of the ADH1 promoter as assayed by northern blotting of the chromosomal ADH1 gene was not affected in the mutant. _Second, when challenged with GAL4-VP16 under the control of a different promoter, UAS2UP1, the mutant was resistant to toxicity. We surmised that the ADA3 mutation reduced the functional activity of the VP16 activation domain. In testing whether the ADA3 mutant supported a reduced level of trans-activation by GAL4-VP16, we took two measures to prevent toxicity by the activator in the parental control strain. First, we moved GAL4-VP16 to a single copy plasmid to reduce the levels of its synthesis. Second, we employed a single amino acid change of Phe442Ala (FA)in the full length VP16 domain (Cress and Triezenberg, Science 251:87-90. 1991. This mutant protein will trans-activate to a significant degree in yeast cells, but is not nearly as toxic as the wild type (Cress and Triezenberg, 1991; Berger et al.. 1992) . The activity from a reporter driven by this activator protein was reduced about three-fold in the ADA3 mutant compared to the parent as shown.in TABLE 4.

Table 4. Transactivation by GAL4-VP16FA.

Activity of Gal Reporter

mg of A-galactosidase

WT CONTROL 240 WT GAL4-VPi6FA 7,300

ADA3 GAL4-VPI6FA 2,700 Isogenic strains BP1 and the ADA3 mutant were transformed with the reporter pLGSD5, bearing the GAL UAS and CYCl TATA boxes driving expression of CYCl-lacZ, as well as the ARS-CEN plasmid in which synthesis of GAL4-VP16FA is driven by the ADH1 promoter (Berger et al.. 1992) . β— galactosidase was assayed by glass bead disruption of cells and units of activity are expressed per mg of protein.

A gel shift of GAL4-VP16 in wild type and ADA3 strains was conducted. Whole cell extracts of BP1 (WT) transform with pDB20L (-) of with the GAL4-VP16 expression plasmid and of the ADA3 mutant were made in duplicate and mixed with the GAL4-specific nucleotide (Berger et al.. , 1982). Purified GAL4-VP16 (left lane, Berger et al... 1990) was used to show the mobility of the correct complex. Complexes in the control not bearing the expression plasmid (-) are found in all cases and correspond to cellular proteins not related to GAL4-VP16. Arrows indicate GAL4VP16-specific complexes, the slowest one corresponding to the intact GAL4-VP16. Two faster complexes correspond to partially degraded GAL4-VP16 molecules. . No differences either in the overall amount of complexes or the proportion of degraded species are evident between wt and ADA3. The levels of GAL4-VP16FA present in the two strains was quantitated by gel shift analysis and very similar levels of GAL4-VP16 and several degradation products of the fusion protein were found in the two strains. We concluded, therefore, that the ability of GAL4-VP16 to activate transcription was substantially reduced by the ADA3 mutation. Interestingly, the ADA3 mutant strain bore another property similar to that of the ADA2 mutant. Cell growth was severely limited on minimal medium, but was only modestly affected on rich medium. On the basis of the resistance to GAL4-VP16 and this growth phenotype, it seemed possible that ADA2 and ADA3 perform related functions in yeast cells. To further investigate the similarity between the ADA3 and ADA2 mutants, we determined the effects of the ADA3 mutation on several yeast promoters containing the CYCl TATA box-mRNA region, and different UAS elements. In these reporter assays (TABLE 5) strong defects were observed using two promoters that were activated by the yeast activator, GCN4.

Table 5. UAS Activator WT ADA3

ADA3/WT

The various reporters all contain the CYCl TATA boxes and CYCl-lacZ, and the indicated UAS inserted at the Xho site at -178 (Berger et al. , 1992) . The reporters used for these assays were the following: UASl: pLGΔAluXho (Guarente et al.. 1984); UAS2:pLG255upl (Forsburg and Guarente, 1989) ; HIS4 fragment: HIS66 and HiS4 oligo HIS 14 X 2 (Hinnebusch et al. , 1985) . The activator known to act at each UAS is indicated, β-galactosidase in this case was assayed in permeabilized cells and units are expressed per OD₆₀₀ of cells.

One promoter bore a fragment of HIS4 DNA as UAS and the other bore two synthetic GCN4 binding sites. Promoters that were driven by HAPl or HAP2/3/4 were also affected by the ADA3 mutation, but to a lesser degree. The range of yeast activators affected in the ADA3 mutant is thus broader than was found in the ADA2 mutant. The ADA2 and ADA3 mutants are similar, however, in that the strongest effects were observed with reporters driven by GCN4.

To delineate further which domain of transcriptional activators required ADA3, the activity of a IexA-GCN4 fusion protein.was assayed in the mutant. In this fusion protein, the acidic activation domain of GCN4 is fused to the DNA-binding domain of the bacterial repressor lexA. Using a reporter with a single lexA site, we found that trans-activation by IexA-GCN4 was reduced about three-fold in the ADA3 mutant compared to the wild type. Activation of a lexA reporter is reduced to about 30-40% in the ADA3 mutant relative to wt (left) , whereas the level of the expressed protein, as revealed by Western blotting, is the same in both strains (right) . The expected size of the fusion protein is about 24,000 daltons. "-" denotates an extract of BP1 not bearing the IexA-GCN4 expression plasmid. The levels of IexA-GCN4, as quantitated by western blotting with lexA antibody, were comparable in the wild type and mutant strains. We conclude that the ADA3 product is required for the normal function of the acidic activation domains of GCN4, VP16, and possibly the HAP proteins. Cloning and sequencing of the ADA3 gene

We used the slow growth phenotype of the ADA3 mutant on minimal medium to clone the gene from ARS-CEN libraries. From over 30,000 ADA3 yeast colonies transformed with the genomic yeast library, two independent clones were selected for their ability to grow in minimal medium. The transforming plasmids were isolated and 7 independent bacterial clones originating from them were retransformed into the ADA3 mutant strain with (B) or without (A) the toxic construct GAL4-VP16.

Plate A shows growth in minimal medium, where ADA3 mutant cells should grow slower that wt cells. On plate B (rich medium) , ADA3 mutants should be able to overcome GAL4-VP16 toxicity, whereas rescued strains (i.e., with a wt phenotype) should be sensitive. When introduced into ADA3 mutant cells, clones 1, 2, 3 and 7 restore the wild type phenotype: the cells grow well in minimal medium (plate A) and are sensitive to GAL4-VP16 (plate B) . Clones 4,5, and 6 do not alter the mutant phenotype, and therefore were not considered. In both plates, BP1 (wt) and ADA3

(ADA3+vector) are included to show wild type and mutant growth in each medium; note that these two controls do not bear the toxic plasmid. ADA3 complementing DNA clones 1, 2, 3, and 7 contained identical or overlapping inserts; number 1 (2112.1) was chosen for further analysis. Clones were isolated that not only reversed the slow growth of the mutant but also restored sensitivity to the toxicity of GAL4-VP16, verifying that these two phenotypes were due to the same mutation. Two of these clones were analyzed and found to bear inserts that overlapped by about six kb. By deleting portions of the insert, we narrowed the complementing region to a three kb fragment. Exonuclease-generated deletions further localized the complementing gene and DNA sequencing identified an open reading frame (ORF) of 702 codons (Figure 2, SEQ ID NO. 3) .

We verified that the 702 codon ORF encoded ADA3 by a complementation experiment involving a gone disruption generated by the clone. To generate a gene disruption, we constructed a vector containing the hisG:URA3-hisG cassette (Alani et al.. 1987) inserted between the Bgl II sites at codons 273 and 307 in the ORF. This cassette generated haploid Ura+ transformants which were viable but grew slowly on minimal medium compared to the wild type parental strain. These transformants were temperature sensitive, failing to grow at 37"C. Further, the disrupted strain was resistant to GAL4-VP16. The disrupted strain was crossed to a strain bearing the original ADA3 mutation, or to the isogenic ADA3 parent. The diploid of genotype Disruption/ADA3 grew very slowly on minimal medium, while the diploid Disruption/+ grew at the same rate as +/+. Therefore, the cloned gene corresponds to the same complementation group as the original ADA3 mutation. We note that the viability of the disrupted strain suggests that ADA3 is not required for growth.

A data base search indicated that the 702 codon ORF did not correspond to any known yeast gene. Further, it was not an obvious homolog of any gene from another source. Analysis of the ADA3 sequence revealed three regions of possible interest. The first is a 15 amino acid hydrophobic region (292-306) containing five Phe residues which bears significant similarity to numerous proteins in the data base, including HIV-gag (Figure 5A) . The second is a 16 amino acid region of ADA3 (487-502) that is 75% identical to residues 293-307 of HIV gag. A third ADA3 region (157-170) shows also high homology with HIV-gag (Figure 5B, SEQ ID NO. 3) . All these three regions are homologous to the p24 product from HIV-gag, that constitutes the major core protein of the HIV virus (Veronese et al.. J_-- Virol. 62:795-801. 1988). Why ADA3 bears three regions with similarity to p24 from HIV is a question to frame further experimentation of the function of these proteins. The genomic insert also contained two known yeast genes, the UBC1 gene (Seufert et al.. EMBO J. 9.:4535-4541, 1990) to the downstream side and the ARG82 gene (ARGRIII) (Dubois et al.. Mol. Gen. Genet. 207:142-148. 1987) to the upstream side. ARG82, and therefore ADA3, maps to chromosome 4R (Guthrie and Fink, Methods in Enzymology 194:835. 1991 ) .

Effects of the ADA3 mutation on transcriptional activation in vitro

There are several possible explanations for the defect of GCN4 or GAL4-VP16 in the ADA3 mutant. By a model invoking effects on chromatin, the mutation would prevent these acidic domains from removing histones from the promoter. By an interactional model, the mutation would weaken protein-protein interaction between these activators at the UASs and the general factors at the TATA box. In order to distinguish between these models, we tested the response of an ADA3 deletion mutant extract to activators in nuclear extracts programmed by DNA templates.

In one experiment, we assayed the activation potential of GAL4-VP16 in nuclear extracts. The template used in this experiment was gal.3, which contains three GAL4 binding sites positioned at -65. Activation by GAL4-VP16 was reduced in the mutant extract compared to the wild type (lanes 3 and 4) . A reduction in activation by GCN4 was also observed in the mutant extract using a template with a GCN4 binding site at -178 (not shown) . Because the ADA3 mutant exerted only modest effects on the activity of the HAP2/3/4 activated UAS2UP1 reporter in vivo, we also tested the activity of a GAL4-HAP4 fusion in vitro. GAL4-HAP4 contained the acidic activation domain of HAP4 in place of the VP16 moiety. Activation by GAL4-HAP4 was also defective in the ADA3 mutant compared to the wild type. This finding contrasts with the case of the ADA2 mutant which did not affect activation by GAL4- HAP4 (Berger et al.. 1992) . Thus the ADA3 mutant is more general than the ADA2 mutant because activation of a broader spectrum of acidic activators is affected.

Effects of the ADA3 mutation on basal transcription If ADA3 were a component of the general transcriptional machinery, effects on basal transcription might be observed in the mutant extract. In fact, in the activation experiment above basal transcription in the ADA3 mutant extract was reduced compared to the wild type. The gal.3 template (Chasman et al.. Mol. Cell. Biol. 9_:4746-4749, 1989) employed in this experiment contains one strong TATA box, the CYC1-52 TATA box, which drives initiation primarily at the +26 and +43 starts in vitro (Hahn et al.. 1985) .

We wished to examine the effects of the ADA3 deletion mutation using a template with several TATA boxes. pLGSD5 (Guarente et al.. Proc. Natl. Acad. Sci. USA 79:7410-7414, 1982) contains a larger region of CYCl upstream DNA than does gal.3 including the TATA box at -106. The 106 TATA box drives initiation at an additional start site in vitro, the +10 site (Hahn et al, Proc. Natl. Acad. Sci. USA 82:8562-8566. 1985). We compared basal transcription in the wild type and mutant extracts using the pLGSD5 template. Strikingly, while the reduction in initiation at the +25 and +43 sites observed above with the gal.3 template was also seen with the pLGSD5 template, initiation at the +10 site was slightly elevated in the mutant extract. In addition, new initiation sites were observed in the vicinity of -10 in the mutant extract. These findings show that basal transcription is qualitatively altered in the ADA3 mutant, indicating that the mutation affects the general transcriptional machinery. This alteration displays a clear pattern in that transcripts directed by the -52 TATA box are under represented in the ADA3 mutant, while transcripts directed by the 106 TATA box are not.

We next examined whether any alteration in mRNA initiation sites occurred in vivo in the ADA3 mutant. We noted previously that the in vitro reaction does not accurately represent the spectrum of CYCl initiation sites seen in vivo (Berger et al.. Cell 61:1199-1208, 1990). Most strikingly, the +43 transcript, which is the major RNA in vitro, is a relatively minor species in vivo. Conversely, the +1, +16, and +25 RNAs are under represented in vitro compared to in vivo. Primer extension of in vivo RNA from the chromosomal CYCl gene showed similar levels of the +1, +10, +16, and +25 transcripts in the wild type and mutant strain. However, the ADA3-1 mutant displayed a clear reduction in the +43 and +50 transcripts (filled triangles) . The levels of the +34 transcripts were also reduced in the ADA3 mutant, although to a smaller degree. This reduction in utilization of proximal initiation sites is similar to what was observed in vitro, although the effects were smaller, perhaps because the ADA3-1 mutation does not cause a total loss of function. (A novel initiation site whose significance is not clear was also observed in the mutant (open triangle) ) . Taken together with the in vitro data, our results suggest that the shorter CYCl transcripts are under represented in the ADA3 mutant while the longer transcripts are not affected.

EQUIVALENTS

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. These and all other such equivalents are intended to be encompassed by the following claims. SEQUENCE LISTING

(1) GENERAL INFORMATION:

(i) APPLICANT: Guarente, Leonard Berger, Shelley PIna, Benjamin

(ii) TITLE OF INVENTION: Transcription Adapters in Eukaryotes

(iii) NUMBER OF SEQUENCES: 4

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE: Hamilton, Brook, Smith & Reynolds, P.C.

(B) STREET: Two Militia Drive

(C) CITY: Lexington

(D) STATE: MA (F) ZIP: 02173

(v) COMPUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: FC-DOS/MS-DOS

(D) SOFTWARE: Patentin Release #1.0, Version #1.25

(vi) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER: US 07/794,732

(B) FILING DATE: 18-NOV-1991

(C) CLASSIFICATION:

(viil) ATTORNEY/AGENT INFORMATION:

(A) NAME: Granahan, Patricia

(B) REGISTRATION NUMBER: 32,227

(C) REFERENCE/DOCKET NUMBER: MIT-5739

(ix) TELECOMMUNICATION INFORMATION:

(A) TELEPHONE: 617-861-6240

(B) TELEFAX: 617-861-9540

(2) INFORMATION FOR SEQ ID NO:l:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 3798 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: DNA (genomic)

(ix) FEATURE:

(A) NAME KEY: CDS -

(B) LOCATION: 308..1609 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:l:

ATCAATCAAT GCGCGTCTTC CAGATGAAAT TCTTCGGAAT GCCCATAAAT ACGAAAAACT 60

AACGATGGAA AAAATACTAT AAATGTATGG GTGTAATTTT AACCGAATGT ATGGATGTAA 120

CAGGGAGCTT TATATATGTA GAAAATTCCA GACGTTCCAA ACAAATAAGT GAAAATCAAC 180

ACGATTGGGA AAGTACATGC CAGAGTTCGA TGTTTAGGTT AATGATAAAA ATAATTTTTC 240

TCTAATCACC CTCCATTTTC GATAAAATAT CAGCGTAGTC TGAAAATATA TACATTAAGC 300

AAAAAGA ATG TCA AAC AAG TTT CAC TGT GAC GTT TGT TCA GCT GAT TGC 349 Met Ser Asn Lys Phe His Cys Asp Val Cys Ser Ala Asp Cys 1 5 10

ACG AAC AGA GTG AGG GTT TCA TGT GCT ATT TGT CCC GAA TAC GAC TTA 397 Thr Asn Arg Val Arg Val Ser Cys Ala lie Cys Pro Glu Tyr Asp Leu 15 20 25 30

TGC GTG CCC TGC TTT TCA CAA GGC TCG TAT ACA GGA AAG CAT CGC CCT 445 Cys Val Pro Cys Phe Ser Gin Gly Ser Tyr Thr Gly Lys His Arg Pro 35 40 45

TAC CAT GAT TAT AGG ATA ATA GAG ACC AAT TCA TAC CCC ATT CTC TGT 493 Tyr His Asp Tyr Arg lie lie Glu Thr Asn Ser Tyr Pro lie Leu Cys 50 55 60

CCT GAT TGG GGT GCC GAT GAA GAA TTG CAG TTG ATT AAA GGC GCA CAG 541 Pro Asp Trp Gly Ala Asp Glu Glu Leu Gin Leu lie Lys Gly Ala Gin 65 70 75

ACT TTA GGG CTC GGT AAT TGG CAG GAT ATT GCT GAC CAT ATA GGC AGC 589 Thr Leu Gly Leu Gly Asn Trp Gin Asp He Ala Asp His He Gly Ser 80 85 90

AGA GGC AAA GAA GAA GTT AAG GAA CAT TAC CTA AAA TAT TAT CTG GAA 637 Arg Gly Lys Glu Glu Val Lys Glu His Tyr Leu Lys Tyr Tyr Leu Glu 95 100 105 110

AGC AAA TAC TAT CCA ATA CCT GAT ATT ACC CAA AAT ATA CAT GTC CCA 685 Ser Lys Tyr Tyr Pro He Pro Asp He Thr Gin Asn He His Val Pro 115 120 125

CAA GAT GAA TTT TTG GAA CAG CGA AGG CAT AGA ATC GAG TCC TTC CGG 733 Gin Asp Glu Phe Leu Glu Gin Arg Arg His Arg He Glu Ser Phe Arg 130 135 140

GAG AGG CCG CTA GAG CCT CCA AGA AAG CCC ATG GCA TCG GTT CCT AGC 781 Glu Arg Pro Leu Glu Pro Pro Arg Lys Pro Met Ala Ser Val Pro Ser 145 150 155 TGC CAT GAA GTA CAG GGG TTT ATG CCG GGC AGA TTA GAA TTT GAA ACA 829 Cys His Glu Val Gin Gly Phe Met Pro Gly Arg Leu Glu Phe Glu Thr 160 165 170

GAA TTC GAA AAT GAA GCT GAA GGT CCC GTA AAG GAT ATG GTT TTT GAA 877 Glu Phe Glu Asn Glu Ala Glu Gly Pro Val Lys Asp Met Val Phe Glu 175 180 185 190

CCT GAT GAC CAG CCC TTA GAC ATT GAA TTG AAG TTT GCC ATC TTG GAC 925 Pro Asp Asp Gin Pro Leu Asp He Glu Leu Lys Phe Ala He Leu Asp 195 200 205

ATT TAC AAT TCT AGA TTG ACA ACC AGA GCA GAA AAG AAA AGG CTA TTA 973 He Tyr Asn Ser Arg Leu Thr Thr Arg Ala Glu Lys Lys Arg Leu Leu 210 215 220

TTT GAA AAC CAT TTA ATG GAT TAC AGA AAG TTG CAA GCT ATT GAC AAA 1021 Phe Glu Asn His Leu Met Asp Tyr Arg Lys Leu Gin Ala He Asp Lys 225 230 235

AAG AGG AGC AAA GAA GCC AAA GAA TTG TAT AAC CGC ATA AAA CCT TTT 1069 Lys Arg Ser Lys Glu Ala Lys Glu Leu Tyr Asn Arg He Lys Pro Phe 245 245 255

GCT CGT GTT ATG ACT GCA CAG GAT TTT GAG GAA TTT AGC AAG GAT ATA 1117 Ala Arg Val Met Thr Ala Gin Asp Phe Glu Glu Phe Ser Lys Asp He 255 260 265 270

TTA GAA GAA TTG CAT TGC AGG GCA AGA ATA CAG CAA TTA CAA GAA TGG 1165 Leu Glu Glu Leu His Cys Arg Ala Arg He Gin Gin Leu Gin Glu Trp 275 280 285

AGA AGC AAT GGG TTA ACC ACG TTA GAA GCA GGT CTT AAG TAT GAA CGA 1213 Arg Ser Asn Gly Leu Thr Thr Leu Glu Ala Gly Leu Lys Tyr Glu Arg 290 295 300

GAT AAG CAA GCT AGA ATT AGC AGT TTC GAG AAA TTT GGC GCT TCT ACG 1261 Asp Lys Gin Ala Arg He Ser Ser Phe Glu Lys Phe Gly Ala Ser Thr 305 310 315

GCG GCG TCA CTC AGT GAG GGC AAT AGC CGT TAT AGG TCG AAT TCT GCA 1309 Ala Ala Ser Leu Ser Glu Gly Asn Ser Arg Tyr Arg Ser Asn Ser Ala 320 325 330

CAT AGG TCG AAC GCA GAA TAC AGC CAA AAC TAC AGC GAA AAT GGC GGT 1357 His Arg Ser Asn Ala Glu Tyr Ser Gin Asn Tyr Ser Glu Asn Gly Gly 335 ^■ 340 345 350

AGA AAG AAA AAT ATG ACC ATT AGC GAT ATA CAG CAT GCA CCT GAC TAT 1405 Arg Lys Lys Asn Met Thr He Ser Asp He Gin His Ala Pro Asp Tyr 355 360 365 GCA CTG TTG TCC AAC GAC GAG CAG CAA CTT TGT ATT CAA CTG AAA ATT 1453 Ala Leu Leu Ser Asn Asp Glu Gin Gin Leu Cys He Gin Leu Lys He 370 375 380

TTA CCT AAA CCG TAT CTC GTT CTA AAA GAG GTA ATG TTC AGA GAA CTG 1501 Leu Pro Lys Pro Tyr Leu Val Leu Lys Glu Val Met Phe Arg Glu Leu 385 390 395

TTG AAG ACT GGG GGT AAT CTA AGC AAA AGC GCA TGT AGG GAA CTA CTT 1549 Leu Lys Thr Gly Gly Asn Leu Ser Lys Ser Ala Cys Arg Glu Leu Leu 400 405 410

AAT ATA GAC CCC ATA AAA GCG AAT AGA ATA TAC GAT TTT TTC CAG AGC 1597 Asn He Asp Pro He Lys Ala Asn Arg He Tyr Asp Phe Phe Gin Ser 415 420 425 430

CAG AAT TGG ATG TAACAAAAAA AAAATTGAAA AGTAACTACA ATTGTCACTA 1649

Gin Asn Trp Met

CTTAGAAGCT GATTATAAAT GTTGTGAATT TTTTTGTATG ATGTTTGGTT ACCTCTCGCT 2789

TTCATGTAGT TCAAATACAT TGCCCAAAAC TTTAAATCTT GTGGAAATTT GTTCGTTCCT 2849

CTTTGATAAA TAAAACCTAT TCTTTGTTGA ATAGACCAAT CAGATAGACT GTTCGTCTTT 2909

TTAACCTGTA AGATTCTCTT ACACCGCTTG GCACGTAATT TATTGACGTT ACTTTCATAG 2969

TTAATGTATT TTATATAGTC ATTTATGCTA GACCCTCTGG AATTTAATCT ATGCTCAAAA 3029

TCAGTTCTTT TCTTCATAAT TAATGAAACT TCATTTTTAG TGAATAGTCC TTTCTCCACC 3089

AAGTCATCCA TTTCAGGAAT ACATTGTTCC AAATAGTATC TTGTCTTCGA CATTATTAGA 3149

AATCGGTGAT GTGTGTATCT CTGGTCAATT TCACAATGTT ATGCCCTTAG ATATGTGTTT 3209

ATTCACGTTC ACTGTGTGTT CCTGATATTT AGCTCATCGC CCAGATCTTT CTAAAAAATT 3269

TTCAAGAGAG AATACCCGGA TGAAAAAGCG TGTTCGGATT CCCATCACTT AATGAAGCTT 3329

ATATCTGACA GCGCAGGAAA AGCACAAAAT AGAACCTAGC CACTTCGCGT AAAAGCTCAT 3389

AATGTTAACG ATATAGTAAT ATTTTGACGC ATCGAACTTT ATCACCCGCA GCATAAAGTA 3449

CTGATACTAC AAGTTTTTTT CGTTATTCTT ACCGGTATTG CGCTAAAGCA GTTTTGCACC 3509

TGAAAACAAG TAAAAATAGG ACTTGAAAAA ATTAAATAGC ATACATGTAT AATATAATCT 3569

TGATCACGGC AAGTGTTTCA ATAGAGATGT GGTTATAATT TGTCACTTGT GCAGTCAATA 3629

CCTTAAATTA ACTTTTACAA AGTATTTTTT TTAAACGGCC CATGAGCATA AATTGAGAAC 3689

GATATTATTA ATTCTAGTAT ATTTTTAGGA ACAGCAGGCC ATTATATCCA AAACTTTTTT 3749

AAAATAGCAT CCACATCACT TACTGGCGTA CAGTTTCGTG TCTGGCAAG 3798

(2) INFORMATION FOR SEQ ID NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 434 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:2:

Met Ser Asn Lys Phe His Cys Asp Val Cys Ser Ala Asp Cys Thr Asn 1 5 10 15

Arg Val Arg Val Ser Cys Ala He Cys Pro Glu Tyr Asp Leu Cys Val 20 25 30

Pro Cys Phe Ser Gin Gly Ser Tyr Thr Gly Lys His Arg Pro Tyr His 35 40 45

Asp Tyr Arg He He Glu Thr Asn Ser Tyr Pro He Leu Cys Pro Asp 50 55 60

Trp Gly Ala Asp Glu Glu Leu Gin Leu He Lys Gly Ala Gin Thr Leu 65 70 75 80

Gly Leu Gly Asn Trp Gin Asp He Ala Asp His He Gly Ser Arg Gly 85 90 95

Lys Glu Glu Val Lys Glu His Tyr Leu Lys Tyr Tyr Leu Glu Ser Lys 100 105 110

Tyr Tyr Pro He Pro Asp He Thr Gin Asn He His Val Pro Gin Asp 115 120 125

Glu Phe Leu Glu Gin Arg Arg His Arg He Glu Ser Phe Arg Glu Arg 130 135 140

Pro Leu Glu Pro Pro Arg Lys Pro Met Ala Ser Val Pro Ser Cys His 145 150 155 160

Glu Val Gin Gly Phe Met Pro Gly Arg Leu Glu Phe Glu Thr Glu Phe 165 170 175

Glu Asn Glu Ala Glu Gly Pro Val Lys Asp Met Val Phe Glu Pro Asp 180 185 190

Asp Gin Pro Leu Asp He Glu Leu Lys Phe Ala He Leu Asp He Tyr 195 200 205

Asn Ser Arg Leu Thr Thr Arg Ala Glu Lys Lys Arg Leu Leu Phe Glu 210 215 220 Asn His Leu Met Asp Tyr Arg Lys Leu Gin Ala He Asp Lys Lys Arg 225 230 235 240

Ser Lys Glu Ala Lys Glu Leu Tyr Asn Arg He Lys Pro Phe Ala Arg 245 250 255

Val Met Thr Ala Gin Asp Phe Glu Glu Phe Ser Lys Asp He Leu Glu 260 265 270

Glu Leu His Cys Arg Ala Arg lie Gin Gin Leu Gin Glu Trp Arg Ser 275 280 285

Asn Gly Leu Thr Thr Leu Glu Ala Gly Leu Lys Tyr Glu Arg Asp Lys 290 295 300

Gin Ala Arg He Ser Ser Phe Glu Lys Phe Gly Ala Ser Thr Ala Ala 305 310 315 320

Ser Leu Ser Glu Gly Asn Ser Arg Tyr Arg Ser Asn Ser Ala His Arg 325 ' 330 335

Ser Asn Ala Glu Tyr Ser Gin Asn Tyr Ser Glu Asn Gly Gly Arg Lys 340 345 350

Lvs Asn Met Thr He Ser Asp He Gin His Ala Pro Asp Tyr Ala Leu 355 360 365

Leu Ser Asn Asp Glu Gin Gin Leu Cys He Gin Leu Lys He Leu Pro 370 375 380

Lys Pro Tyr Leu Val Leu Lys Glu Val Met Phe Arg Glu Leu Leu Lys 385 390 395 400

Thr Gly Gly Asn Leu Ser Lys Ser Ala Cys Arg Glu Leu Leu Asn He 405 410 415

Asp Pro He Lys Ala Asn Arg He Tyr Asp Phe Phe Gin Ser Gin Asn 420 425 430

Trp Met

(2) INFORMATION FOR SEQ ID NO:3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 3126 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(ii) MOLECULE TYPE: DNA (genomic)

(ix) FEATURE:

(A) NAME/KEY: CDS

(B) LOCATION: 605..2710

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:

TTTTTAATAT CGTTAACATT TGAGTACTCA GAAGTGGTTA ATAAGGCATT CAGTTCGTCT 60

TTGTTAGGTT TATACTCTGA ACCCAATCTA GCCATTCTCT CCCAATATAA CTCTAAAATT 120

CTTTGAGGAG GTAAACCCTT CCATTTTTCG GGATGCATAT AGAGATCTGC TGAAAAAGCT 180

GGTCCTTCTT CACAGTTTTA GCTTCATTAC CAGCACTGTT CAAACATCTA GACATTGTCA 240

AAACCCTTTT CTGGGTGCTC CACAAATTAC CCCGAGAGAC TTTCCATAAC CCCAATGGTA 300

CTTTCATCTT CCTAAGAATA GTATTTTAGT AGTGTCAAAT GTTTCGGTAA CTCTACTGCC 360

TTGTTCACAG ATCGAAATAT ACTCAGAATC AGTTACAGAA AAGGTCCCAT TTAGCGTAAT 420

TTCTTAGCCA ACTCATTAAA GAGCTCCGAC GTGCAACGCG ATAAAGGTTC GCCGACGACA 480

ACAAAACAAT GGCAGGCAGT CATTAAATGA ACGCCATCAA ATGAAATATT CGGGTAGCGA 540

TTACAATAAG GAAGTTGGAA AAACATAACA AAGACGGAGC GACGAGAAGT ATTGGACAGG 600

ACAT ATG CCT AGA CAT GGA AGA AGA GGA AAA CTG CCA AAA GGC GAG AAA 649 Met Pro Arg His Gly Arg Arg Gly Lys Leu Pro Lys Gly Glu Lys 1 5 10 15

CTT CCT AAA AAG GAG GGA GGC GAC AAT ACT CCG TCA AAG CTT TTA TCG 697 Leu Pro Lys Lys Glu Gly Gly Asp Asn Thr Pro Ser Lys Leu Leu Ser 20 25 30

TCA ATG CTT AAA ACT TTA GAT TTG ACC TTC GAG CGT GAT ATT GGA ATG 745 Ser Met Leu Lys Thr Leu Asp Leu Thr Phe Glu Arg Asp He Gly Met 35 40 45

CTC AAT GGC AAA AGT GTT AGG TCC ATA CCC AAT AAA AAA ACT TTA CTT 793 Leu Asn Gly Lys Ser Val Arg Ser He Pro Asn Lys Lys Thr Leu Leu 50 55 60 GAA TTG CAA AGT CAA TTG GAT AGT CTC AAT GAA ATA CTG GGT ACA ATT 841 Glu Leu Gin Ser Gin Leu Asp Ser Leu Asn Glu He Leu Gly Thr He 65 70 75

GCC CGA GGT GAT CAA GAG ACC ATT GAA GCG CTT CGG AAA ATA AGA GAC 889 Ala Arg Gly Asp Gin Glu Thr He Glu Ala Leu Arg Lys He Arg Asp 80 85 90 95

AGT AAA AAT GAG AAG CAA GCA AAC GAT GAA AAG CAG GAA ACT TCA AAT 937 Ser Lys Asn Glu Lys Gin Ala Asn Asp Glu Lys Gin Glu Thr Ser Asn 100 105 110

GCA GAC GGG CAG CAC GAA AGT TCT ACT GCT ACT GAA GAA ACG AAT ATA 985 Ala Asp Gly Gin His Glu Ser Ser Thr Ala Thr Glu Glu Thr Asn He 115 120 125

ATT GAT AAA GGG GTA CAA AGC CCA CCA AAA CCT CCT CCT TCT AAT GAA 1033 He Asp Lys Gly Val Gin Ser Pro Pro Lys Pro Pro Pro Ser Asn Glu 130 135 140

ATC TCC GGC ACA ATT GAG AAC GAT GTA GAA AGC ATC AAA CAA GCG GCT 1081 He Ser Gly Thr He Glu Asn Asp Val Glu Ser He Lys Gin Ala Ala 145 150 155

GAT AAT ATG GCT AAA GAA GAA ATT AAT GAA GAC AAA GAT TTA CAA GTT 1129 Asp Asn Met Ala Lys Glu Glu He Asn Glu Asp Lys Asp Leu Gin Val 160 165 170 175

CAT AGA GAT CAA CCC AGA GAA AAG AGG CCG TTC GAT TCT GAA ACT GAA 1177 His Arg Asp Gin Pro Arg Glu Lys Arg Pro Phe Asp Ser Glu Thr Glu 180 185 190

AAT AGG GCC ACT GAA AAT GAA AAC ACA CAG AGG CCA GAC AAT AAG AAG 1225 Asn Arg Ala Thr Glu Asn Glu Asn Thr Gin Arg Pro Asp Asn Lys Lys 195 200 205

CAA AAA ATT GAC GTT GAC AAA ATG GAA AAC GAT CCA ACT GTT AAA AAT 1273 Gin Lys He Asp Val Asp Lys Met Glu Asn Asp Pro Thr Val Lys Asn 210 215 . 220

CCA AAA TCA GAA TTT GTA GTA TCG CAA ACG TTA CCT CGA GCA GCA GCA 1321 Pro Lys Ser Glu Phe Val Val Ser Gin Thr Leu Pro Arg Ala Ala Ala 225 230 235

GCA CTC GGC TTA TTC AAT GAA GAG GGA CTA GAA AGT ACA GGA GAA GAC 1369 Ala Leu Gly Leu Phe Asn Glu Glu Gly Leu Glu Ser Thr Gly Glu Asp 240 245 250 255

TTC CTA AAA AAA AAA TAC AAT GTA GCT AGT TAT CCA ACA AAT GAT TTG 1417 Phe Leu Lys Lys Lys Tyr Asn Val Ala Ser Tyr Pro Thr Asn Asp Leu 260 265 270

AAA GAT CTG TTA CCA GGT GAA TTA CCG GAT ATG GAC TTT TCG CAT CCT 1465 Lys Asp Leu Leu Pro Gly Glu Leu Pro Asp Met Asp Phe Ser His Pro 275 280 285 AAA CCA ACC AAC CAA ATT CAA TTC AAT ACC TTT CTA GCA TTT GTG GAA 1513 Lys Pro Thr Asn Gin He Gin Phe Asn Thr Phe Leu Ala Phe Val Glu 290 295 300

AAT TTT TTC AAA GAT CTT AGT GAT GAC AAC TTA AAG TTT TTG AAG ATG 1561 Asn Phe Phe Lys Asp Leu Ser Asp Asp Asn Leu Lys Phe Leu Lys Met 305 310 315

AAA TAT ATT ATA CCC GAT AGC TTG CAA TTT GAC AAA ACT TAT GAC CCT 1609 Lys Tyr He He Pro Asp Ser Leu Gin Phe Asp Lys Thr Tyr Asp Pro 320 325 330 335

GAG GTA AAC CCG TTT ATT ATA CCG AAA CTA GGC CCT TTA TAC ACA GAT 1657 Glu Val Asn Pro Phe He He Pro Lys Leu Gly Pro Leu Tyr Thr Asp 340 345 350

GTT TGG TTC AAA GAT GAA AAC GAC AAA AAT TCT GCC TAT AAA AAA CCT 1705 Val Trp Phe Lys Asp Glu Asn Asp Lys Asn Ser Ala Tyr Lys Lys Pro 355 360 365

TCA CCA TAT TCA AAC GAT GCA TCT ACT ATA CTA CCA AAA AAA AGT GCC 1753 Ser Pro Tyr Ser Asn Asp Ala Ser Thr He Leu Pro Lys Lys Ser Ala 370 375 380

AAC GAA CTT GAT GAT AAT GCT TTG GAA TCG GGC AGT ATA TCG TGT GGG 1801 Asn Glu Leu Asp Asp Asn Ala Leu Glu Ser Gly Ser He Ser Cys Gly 385 390 395

CCC TTA TTA TCT AGG CTG TTG AGT GCT GTA TTA AAA GAT GAC AAT GAC 1849 Pro Leu Leu Ser Arg Leu Leu Ser Ala Val Leu Lys Asp Asp Asn Asp 400 405 410 415

AAA TCA GAA TTG CAA TCT TCT AAA ATA ATA CGG GAT GGC GGA CTA CCG 1897 Lys Ser Glu Leu Gin Ser Ser Lys He He Arg Asp Gly Gly Leu Pro 420 425 430

AGA ACC GGA GGG GAA GAC GAT ATA CAA TCG TTT AGA AAC AAC AAC AAT 1945 Arg Thr Gly Gly Glu Asp Asp He Gin Ser Phe Arg Asn Asn Asn Asn 435 440 445

GAT ACT GTA GAC ATG ACA TTA TCT CAG GAG AAC GGG CCC AGT GTT CAA 1993 Asp Thr Val Asp Met Thr Leu Ser Gin Glu Asn Gly Pro Ser Val Gin 450 455 460

ACA CCT GAC AAT GAC ATT GAT GAG GAA GCA TCT TTC CAA GCA AAA CTG 2041 Thr Pro Asp Asn Asp He Asp Glu Glu Ala Ser Phe Gin Ala Lys Leu 465 470 475

GCA GAA AAT AAA GGC AGT AAT GGC GGT ACT ACA AGT ACG TTA CCC CAA 2089 Ala Glu Asn Lys Gly Ser Asn Gly Gly Thr Thr Ser Thr Leu Pro Gin 480 485 490 495

CAA ATT GGG TGG ATA ACA AAT GGA ATT AAC CTG GAC TAT CCA ACA TTC 2137 Gin He Gly Trp He Thr Asn Gly He Asn Leu Asp Tyr Pro Thr Phe 500 505 510 GAA GAA CGA TTG AAA AGA GAA CTA AAG TAT GTA GGG ATA TAC ATG AAT 2185 Glu Glu Arg Leu Lys Arg Glu Leu Lys Tyr Val Gly He Tyr Met Asn 515 520 525

TTG CCT AAG GAT GAG AAC AAC CCT AAC TCA GAT GAC CCC GAT TGG GTC 2233 Leu Pro Lys Asp Glu Asn Asn Pro Asn Ser Asp Asp Pro Asp Trp Val 530 535 540

ACC GGT AGA GAA GAC GAC GAA ATA AGC GCA GAG TTA AGA GAA TTG CAA 2281 Thr Gly Arg Glu Asp Asp Glu He Ser Ala Glu Leu Arg Glu Leu Gin 545 550 555

GGC ACT TTA AAA CAA GTG ACC AAG AAA AAC CAA AAG AGG AAA GCC CAA 2329 Gly Thr Leu Lys Gin Val Thr Lys Lys Asn Gin Lys Arg Lys Ala Gin 560 565 570 575

TTG ATT CCA CTA GTG GAA AGA CAA CTA GCA TGG CAG GAA TAC TCA TCT 2377 Leu He Pro Leu Val Glu Arg Gin Leu Ala Trp Gin Glu Tyr Ser Ser 580 585 590

ATT TTA GAA GAT TTA GAC AAA CAA ATC GAC CAG GCT TAT GTC AAA CGT 2425 He Leu Glu Asp Leu Asp Lys Gin He Asp Gin Ala Tyr Val Lys Arg 595 600 605

ATT CGC GTA CCC AAG AAA AGA AAG AAG CAT CAT ACA GCA GCT TCA AAT 2473 He Arg Val Pro Lys Lys Arg Lys Lys His His Thr Ala Ala Ser Asn 610 615 620

AAT GTG AAC ACA GGA ACC ACG TCT CAA ATA GCA CAA CAA AAG GCT GCA 2521 Asn Val Asn Thr Gly Thr Thr Ser Gin He Ala Gin Gin Lys Ala Ala 625 630 635

AAC TCA AGT TTA AAA TCC CTT CTG GAT AAA AGG CAA AGG TGG ATT AAT 2569 Asn Ser Ser Leu Lys Ser Leu Leu Asp Lys Arg Gin Arg Trp He Asn 640 645 650 655

AAG ATT GGT CCA TTG TTT GAT AAA CCT GAA ATT ATG AAG AGA ATC CCC 2617 Lys He Gly Pro Leu Phe Asp Lys Pro Glu He Met Lys Arg He Pro 660 665 670

AAT GAA AGC GTA TTC AAG GAC ATG GAC CAA GAA GAA GAT GAG GAT GAA 2665 Asn Glu Ser Val Phe Lys Asp Met Asp Gin Glu Glu Asp Glu Asp Glu 675 680 685

GCC GAT GTA TTT GCA CAA AAC ACT AAT AAG GAC GTG GAA CTA AAT 2710

Ala Asp Val Phe Ala Gin Asn Thr Asn Lys Asp Val Glu Leu Asn 690 695 700

(2) INFORMATION FOR SEQ ID NO:4:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 702 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: protein

(xi) SEQUENCE DESCRIPTION: SEQ ID N0:4:

Met Pro Arg His Gly Arg Arg Gly Lys Leu Pro Lys Gly Glu Lys Leu 1 5 10 15

Pro Lys Lys Glu Gly Gly Asp Asn Thr Pro Ser Lys Leu Leu Ser Ser 20 25 30

Met Leu Lys Thr Leu Asp Leu Thr Phe Glu Arg Asp He Gly Met Leu 35 40 45

Asn Gly Lys Ser Val Arg Ser He Pro Asn Lys Lys Thr Leu Leu Glu 50 55 60

Leu Gin Ser Gin Leu Asp Ser Leu Asn Glu He Leu Gly Thr He Ala 65 70 75 80

Arg Gly Asp Gin Glu Thr He Glu Ala Leu Arg Lys He Arg Asp Ser 85 90 95

Lys Asn Glu Lys Gin Ala Asn Asp Glu Lys Gin Glu Thr Ser Asn Ala 100 105 110

Asp Gly Gin His Glu Ser Ser Thr Ala Thr Glu Glu Thr Asn He He 115 120 125

Asp Lys Gly Val Gin Ser Pro Pro Lys Pro Pro Pro Ser Asn Glu He 130 135 140

Ser Gly Thr He Glu Asn Asp Val Glu Ser He Lys Gin Ala Ala Asp 145 150 155 160

Asn Met Ala Lys Glu Glu He Asn Glu Asp Lys Asp Leu Gin Val His 165 . 170 175

Arg Asp Gin Pro Arg Glu Lys Arg Pro Phe Asp Ser Glu Thr Glu Asn 180 185 190

Arg Ala Thr Glu Asn Glu Asn Thr Gin Arg Pro Asp Asn Lys Lys Gin 195 200 205

Lys He Asp Val Asp Lys Met Glu Asn Asp Pro Thr Val Lys Asn Pro 210 215 220 Lys Ser Glu Phe Val Val Ser Gin Thr Leu Pro Arg Ala Ala Ala Ala 225 230 235 240

Leu Gly Leu Phe Asn Glu Glu Gly Leu Glu Ser Thr Gly Glu Asp Phe 245 250 255

Leu Lys Lys Lys Tyr Asn Val Ala Ser Tyr Pro Thr Asn Asp Leu Lys 260 265 270

Asp Leu Leu Pro Gly Glu Leu Pro Asp Met Asp Phe Ser His Pro Lys 275 280 285

Pro Thr Asn Gin He Gin Phe Asn Thr Phe Leu Ala Phe Val Glu Asn 290 295 300

Phe Phe Lys Asp Leu Ser Asp Asp Asn Leu Lys Phe Leu Lys Met Lys 305 310 315 320

Tyr He He Pro Asp Ser Leu Gin Phe Asp Lys Thr Tyr Asp Pro Glu 325 330 335

Val Asn Pro Phe He He Pro Lys Leu Gly Pro Leu Tyr Thr Asp Val 340 345 350

Trp Phe Lys Asp Glu Asn Asp Lys Asn Ser Ala Tyr Lys Lys Pro Ser 355 360 365

Pro Tyr Ser Asn Asp Ala Ser Thr He Leu Pro Lys Lys Ser Ala Asn 370 375 380

Glu Leu Asp Asp Asn Ala Leu Glu Ser Gly Ser He Ser Cys Gly Pro 385 390 395 400

Leu Leu Ser Arg Leu Leu Ser Ala Val Leu Lys Asp Asp Asn Asp Lys 405 410 415

Ser Glu Leu Gin Ser Ser Lys He He Arg Asp Gly Gly Leu Pro Arg 420 425 430

Thr Gly Gly Glu Asp Asp He Gin Ser Phe Arg Asn Asn Asn Asn Asp 435 440 445

Thr Val Asp Met Thr Leu Ser Gin Glu Asn Gly Pro Ser Val Gin Thr 450 455 460

Pro Asp Asn Asp He Asp Glu Glu Ala Ser Phe Gin Ala Lys Leu Ala 465 470 475 480

Glu Asn Lys Gly Ser Asn Gly Gly Thr Thr Ser Thr Leu Pro Gin Gin 485 490 495

He Gly Trp He Thr Asn Gly He Asn Leu Asp Tyr Pro Thr Phe Glu 500 505 510 Glu Arg Leu Lys Arg Glu Leu Lys Tyr Val Gly He Tyr Met Asn Leu 515 520 525

Pro Lys Asp Glu Asn Asn Pro Asn Ser Asp Asp Pro Asp Trp Val Thr 530 535 540

Gly Arg Glu Asp Asp Glu He Ser Ala Glu Leu Arg Glu Leu Gin Gly 545 550 555 560

Thr Leu Lys Gin Val Thr Lys Lys Asn Gin Lys Arg Lys Ala Gin Leu 565 570 575

He Pro Leu Val Glu Arg Gin Leu Ala Trp Gin Glu Tyr Ser Ser He 580 585 590

Leu Glu Asp Leu Asp Lys Gin He Asp Gin Ala Tyr Val Lys Arg He 595 600 605

Arg Val Pro Lys Lys Arg Lys Lys His His Thr Ala Ala Ser Asn Asn 610 615 620

Val Asn Thr Gly Thr Thr Ser Gin He Ala Gin Gin Lys Ala Ala Asn 625 630 635 640

Ser Ser Leu Lys Ser Leu Leu Asp Lys Arg Gin Arg Trp He Asn Lys 645 650 655

He Gly Pro Leu Phe Asp Lys Pro Glu He Met Lys Arg He Pro Asn 660 665 670

Glu Ser Val Phe Lys Asp Met Asp Gin Glu Glu Asp Glu Asp Glu Ala 675 680 685

Asp Val Phe Ala Gin Asn Thr Asn Lys Asp Val Glu Leu Asn 690 695 700

Claims

1. An isolated transcription adaptor which stimulates activated eukaryotic DNA transcription on a region of a DNA strand, wherein the adaptor interacts with two factors: a) the first factor being capable of binding a first region on the DNA strand; and b) the second factor being capable of binding a second region on the DNA strand located downstream from the first region, thereby providing a transcriptional complex required for normal activated transcription.

2. An isolated transcriptional adaptor of Claim 1 wherein the adaptor interacts with the first factor by binding an acidic activation domain on the first factor.

3. An isolated transcription adaptor which stimulates activated eukaryotic DNA transcription on a region of a DNA strand, wherein the adaptor interacts with an acidic activation domain of an activator and further interacts with a general transcription factor, thereby providing a transcriptional complex required for activated transcription.

4. DNA encoding a transcriptional adaptor which stimulates activated DNA transcription in a eukaryotic cell.

5. DNA of Claim 4, wherein the DNA is selected from the group consisting of: all or a portion of the ADA2 gene having the sequence of Figure 1 (SEQ ID NO.:l) ; all or a portion of the ADA3 gene having the sequence of Figure 2 (SEQ ID NO.:3); DNA which hybridizes to all or a portion of the ADA2 gene having the sequence of Figure 1; and DNA which hybridizes to all or a portion of the ADA3 gene having the sequence of Figure 2.

6. DNA which hybridizes to DNA encoding a transcriptional adaptor which stimulates activated DNA transcription in a eukaryotic cell.

7. DNA of Claim 6, wherein the DNA hybridizes to all or a portion of a DNA sequence selected from the group consisting of: the ADA2 gene sequence of

Figure 1 (SEQ ID N0.:1) and the ADA3 gene having the sequence of Figure 2 (SEQ ID NO.:3).

8. A method for producing a recombinant transcriptional adaptor, comprising transforming a cell with DNA encoding the transcriptional adaptor in expressible form.

9. A method of Claim 8, wherein the DNA is DNA of Claim 5.

10. A transcriptional adaptor produced by transforming a cell with a DNA encoding a transcriptional adaptor in expressible form.

11. The transcriptional adaptor of Claim 10, wherein the DNA is DNA of CLaim 5.

12. A transformed host cell expressing a transcriptional adaptor, wherein the transcriptional adaptor stimulates activated eukaryotic DNA transcription.

13. A transformed host cell of Claim 12, wherein the expressed transcriptional adaptor is ADA2.

14. A transformed host cell of Claim 12, wherein the expressed transcriptional adaptor is ADA3.

15. A method of stimulating activated transcription in a eukaryotic cell, comprising introducing into the cell a transcriptional adaptor in sufficient quantity and under conditions appropriate for the transcriptional adaptor to associate with two factors, wherein the first factor binds to a first region on the DNA and the second factor binds to a second region on the DNA sequence located downstream from first region, thereby stimulating activated transcription.

16. The method of Claim 15, wherein the first factor is a transcriptional activator and the second factor is a general transcription factor.

17. The method of Claim 16, wherein the general transcription factor is TATA box-binding protein

18. A method of determining the ability of a substance to inhibit transcription by interfering with a transcriptional adaptor which stimulates activated DNA transcription in a eukaryotic cell, comprising the steps of: a) combining the substance with the transcriptional adaptor, under conditions appropriate for stimulation of activated transcription in a eukaryotic cell; and b) determining whether stimulation of activated transcription in the eukaryotic cell results, wherein absence of activated transcription the eukaryotic cell is an indication that the substance interferes with the transcriptional adaptor.

19. A method of inhibiting activation of transcription in a eukaryotic cell, comprising introducing into the cell a substance to inhibit transcription by interfering with a transcriptional adaptor which stimulates activated DNA transcription in a eukaryotic cell, comprising the steps of: a) combining the substance with the transcriptional adaptor, under conditions appropriate for stimulation of activated transcription in a eukaryotic cell; and b) determining whether stimulation of activated transcription in the eukaryotic cell results, wherein absence of activated transcription in the eukaryotic cell is an indication that the substance interferes with the transcriptional adaptor.