Proc. Natl. Acad. Sci. USAVol. 89, pp. 11589-11593, December 1992Genetics

Dominant genetics using a yeast genomic library under the controlof a strong inducible promoter

(AYES/expression library/truncated proteins/mating response pathway/STEHJ)

SANDRA W. RAMER, STEPHEN J. ELLEDGE*, AND RONALD W. DAVISDepartment of Biochemistry, Stanford University School of Medicine, Stanford, CA 94305

Contributed by Ronald W. Davis, September 1, 1992

ABSTRACT In Saccharomyces cerevisiae, numerous geneshave been identified by selection from high-copy-number li-braries based on "multicopy suppression" or other phenotypicconsequences of overexpression. Although fruitful, this ap-proach suffers from two major drawbacks. First, high copynumber alone may not permit high-level expression of tightlyregulated genes. Conversely, other genes expressed in propor-tion to dosage cannot be identified if their products are toxic atelevated levels. This work reports construction of a genomicDNA expression library for S. cerevisiae that circumvents bothlimitations by fusing randomly sheared genomic DNA to thestrong, inducible yeast GAL] promoter, which can be regulatedby carbon source. The library obtained contains 5 X 107independent recombinants, representing a breakpoint at everybase in the yeast genome. This library was used to emineaberrant gene expression in S. cerevisiae. A screen for domi-nant activators of yeast mating response identified eight genesthat activate the pathway in the absence of exogenous matingpheromone, including one previously unidentified gene. Oneactivator was a truncated STEII gene lacking 41000 base pairsof amino-terminal coding sequence. In two different clones, thesame GALI promoter-proximal ATG is in-frame with thecoding sequence ofSTEII, suggesting that internal initiation oftranslation there results in production of a biologically active,truncated STEII protein. Thus this library allows isolationbased on dominant phenotypes of genes that might have beendifficult or impossible to isolate from high-copy-number librar-ies.

A common genetic approach employed to analyze the cellularfunction of a protein is an examination of the phenotypesassociated with perturbing its expression level. The effects ofa protein's absence can be studied readily in Saccharomycescerevisiae by using homologous recombination to deletegenes (1). In addition, many proteins show novel phenotypeswhen overproduced or inappropriately produced. Increasedgene dosage is sometimes sufficient to produce elevatedlevels of a given protein (2). High-copy-number libraries areused frequently in yeast to isolate genes based on phenotypesdue to increased expression levels. However, these librariessuffer from two major drawbacks due to their dependence onsimple gene dosage to elevate protein production. First, highlevels of protein production cannot always be achieved byhigh copy number alone (3). Classes of genes such as thoseexpressed at specific points in the cell cycle or duringparticular growth or differentiation stages, for example, maybe tightly regulated and thus not subject to overexpression byincreased gene dosage. Thus many genes that might showinteresting phenotypes ifoverexpressed cannot be isolated inthis manner. Second, because these libraries lack a mecha-nism for controlling protein production, they are deficient in

genes whose products are toxic at high concentration and ingenes tightly linked to these toxic genes.To circumvent these limitations, we have constructed a

yeast genomic DNA expression library consisting of randomgenomic fragments cloned downstream of the regulatableGAL] promoter (PGL,). Thus a portion of the library con-tains coding sequences that have been separated from theirendogenous promoter regions and joined instead to PG4LI.PGALI is a very strong yeast promoter that is tightly repressedin cells utilizing glucose as a carbon source and is inducedover 1000-fold when the carbon source is changed to galac-tose (4). Thus clones from the library that contain PGALIfusions proximal to coding sequences can achieve high levelexpression of these genes in an experimentally regulatablemanner. Because genes can be isolated based on the domi-nant phenotypes displayed when they are overexpressed, thismethod is termed "gene isolation by conditional overexpres-sion," and the regulatable library is referred to as a "condi-tional expression library."The library was tested by a variety of standard criteria, and

the ability of clones in the library to complement amino acidauxotrophies in both a galactose-dependent and -independentfashion was confirmed. Two additional screens were under-taken to examine the ability of the features of this library topermit the isolation ofgenes with interesting phenotypes. Thefirst screen identified a genet whose protein product isdeleterious at high concentration.The second screen was designed to identify genes that

show novel phenotypes when overexpressed. Two suchgenes that have been identified previously in yeast, STE4 andSTEI2, are involved in pheromone-induced mating response.Overexpressing either gene from PGAS can activate themating-response pathway in the absence of exogenous mat-ing pheromone (5-8). The conditional expression library wasscreened for other genes that could activate the pathwaywhen overproduced. The identification of several such genes,taken together with the other uses ofthe library, suggests thatthis regulatable genomic DNA expression library will be avaluable alternative to standard high-copy-number librariesin S. cerevisiae.

MATERIALS AND METHODSPlasmids, Strains, and Media. Plasmids and strains used for

library construction have been described elsewhere in detail(9). The features of the pYES-R vector relevant to this workare diagrammed in Fig. 1. Genomic yeast DNA was preparedfrom strain SNY243 (MATa leu2-04 adel ade6 circ°; from M.Snyder, Yale University). Other yeast strains used were YC3

Abbreviation: PGALI, GAL) promoter.*Present address: Department of Biochemistry and Institute forMolecular Genetics, Baylor College of Medicine, Houston, TX77030.tThe sequence reported in this paper has been deposited in theGenBank data base (accession no. L02617).

11589

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

June

3, 2

020

11590 Genetics: Ramer et al.F igare

EcoRI-Xhol-EcoRI

OOP-GAL1 P-lac

Ad11 Apn:1.4

pYES-R

orn

ARS1CEN4

FIG. 1. Features of the pYES-R vector used for genomic libraryconstruction. Details of the vector have been described (9).

(MATa ade6 canl-100 his3-11,15 leu2-3,112 trpl-1 ura3-1;from R. Esposito, University of Chicago), CRY1 (MATacani-100 ade2-1 his3-11,15 leu2-3,112 trpl-1 ura3-1; from R.Fuller, Stanford University), YNN490 (MATa/a ade2-1/++/ade6 can]-100/can) -100 his3-11,15/his3-11,15 leu2-3,112/leu2-3,112 trpl-1/trpl-J ura3-1/ura3-1), YNN217(MATa lys2-801 ura3-52 ade2-101 his3A200), and NNY19(MATa FUSJ-lacZ::LEU2 ura3 leu2 his3 trpl lys2; ref. 6).Standard media for yeast and Escherichia coli were used as

described (10, 11). Uracil-deficient media were used to selectfor the pYES library plasmids.

Preparation of a S. cerevisiae Genomic DNA Library. YeastgenomicDNA from strain SNY243 was prepared and purifiedas described (12). The DNA was sheared by passage througha 27-gauge needle until the average size was :6000 base pairs(bp). The sheared DNA was made blunt by using T4 DNApolymerase. Phosphorylated Xho I adaptors (5'-CGAGC-TACGTCAGGG-3' and 5'-CCCTGACGTAGC-3') were li-

gated to the repaired genomic DNA at a 10-fold molar excess

over genomic DNA. Fragments of 3600-8400 bp were gelpurified, ligated to "T-filled" AYES-R arms, and packaged as

described (9). The library was amplified by one passagethrough E. coli (11).The phage library was converted to a plasmid library

suitable for transformation into yeast by using cre-lox re-

combination essentially as described (9). Fifty million plaque-forming units, equal to the number of independent clones inthe initial unamplified library, yielded 4 mg of cesium chlo-ride-purified plasmid DNA.Complementation of his3 Auxotrophy. Transformants

were selected on SDC-Ura (10) plates and replica-plated toSGC-His (10) plates. Colonies that grew on the SGC-Hismedium were tested for growth on SDC-His (10) plates.

Filter Replica Assay for (-Galactosidase Activity. Nitrocel-lulose filter replicas were made from plates containing freshlygrown yeast colonies, and the yeast cells were permeabilizedby immersing the filter in liquid nitrogen for 5-10 sec. Filterswere thawed and then placed in a Petri dish colony-side-up onWhatman 3MM paper saturated with Z buffer (60 mMNa2PO4/40mM NaH2PO4/10mM KCl/1 mM MgSO4/50mM2-mercaptoethanol) containing 5-bromo-4-chloro-3-indolyl-j3-D-galactoside (300 ,ug/ml). Filters were incubated at 300Cfor between 30 min and 18 h and then scored for blue colonies.For determining galactose-dependent f3-galactosidase activ-ity, filters were incubated colony-side-up on plates contain-ing galactose as the sole carbon source for 12-18 h prior toassay.

Quantitative P-Galactosidase Assays. A crude protein ex-tract was prepared from 109 late-logarithmic-phase cells in 1

Proc. NatL. Acad. Sci. USA 89 (1992)

ml of ice-cold Z buffer containing 1 mM EDTA, 1 mMphenylmethylsulfonyl fluoride, 2 mM pepstatin A, 0.6 mMleupeptin, and 0.5 mM benzamidine by using glass beads (13),and 8-galactosidase activity was determined by standardtechniques (14). For assays requiring galactose induction,logarithmic-phase cells in glucose-containing media werepelleted by centrifugation, washed with sterile water, andthen resuspended in galactose-containing media for 17 hbefore extracts were prepared.DNA Sequencing. Double-stranded plasmid DNA was se-

quenced by the method of Sanger et al. (15) by usingSequenase 2.0 as recommended by the manufacturer (UnitedStates Biochemical). The oligonucleotide 5'-CAACAAAA-AATTGTTAATATACCTC-3', corresponding to a region ofthe yeast PGALI was used as a primer to obtain sequenceinformation from genomic insert DNA proximal to PG,4LI.Sequence analysis and homology searching were performedusing the Intelligenetics programs.

RESULTSA yeast genomic DNA library was prepared in a vectordesigned to allow conditional, high-level protein productionin yeast resulting from the fusion of coding sequences to astrong, vector-borne promoter. The resulting library waslarge and contained a high proportion of DNA inserts. Theability of the library to allow the identification of clones thatwould not be found in a traditional high-copy-number librarywas examined. Several such clones were isolated, indicatingthat this conditional expression library can be a usefulalternative to such high-copy-number libraries.

Preparation of a Yeast Genonic DNA Library in pYES-R. Alibrary of 5 x 107 recombinant phage was constructed inAYES-R (9). An aliquot of the amplified library representing5 X 107 phage was converted to the plasmid form (pYES-R)by using the "automatic subcloning" feature of the AYESsystem (9). A library of this complexity contains enoughclones to represent fusion of PGALJ at every base in the yeastgenome (14 million bp). Nine of 10 randomly selected,independent library clones analyzed contained genomicDNA inserts, with an average insert size of 3800 bp.Complementation of Amino Acid Auxotrophy. To verify

that plasmids in the library contain genes under the control ofPG4L,, the library was screened for clones able to comple-ment the histidine auxotrophy of a his3A200 strain, YC3, ina galactose-dependent fashion. Out of 100,000 YC3 trans-formants, 42 clones were capable of growing on SGC-Hismedium. Four of these clones could not grow on SDC-Hismedium, indicating that the HIS3 gene was under the controlof PGALI. Plasmid DNA was rescued from one galactose-dependent HIS+ clone. The PGALI proximal end of thegenomic insert DNA from this plasmid was sequenced. Thesite of the fusion was 59 bp upstream ofthe translational startcodon of the HIS3 gene (Fig. 2A).

Isolation ofa Gene Conferring Galactose-Dependent GrowthArrest. The ability of PG,4LI to be regulated experimentally bychoice of carbon source allows plasmids from the library tobe maintained under conditions in which there is no PGAI-driven transcription. This feature should permit the isolationof genes whose products are toxic at high concentration(Toxic on High Expression, or THE genes). Such genes couldnot be represented in a standard high-copy-number library.To identify such THE genes, 40,000 transformants inYNN490 were plated at low density (<500 per plate) andreplica-plated to glucose plates and to galactose plates.Visual comparison of the two sets of plates revealed severalcolonies whose growth on galactose plates was much poorer.When streaked for single colonies, all but one of thesetransformants proved capable of growing on galactose, al-though much more slowly than the majority of the library

Dow

nloa

ded

by g

uest

on

June

3, 2

020

Figure 2

Genetics: Ramer et al.

A GAL! Promoter

Proc. NatL. Acad. Sci. USA 89 (1992) 11591

HtS3 Upstream Region HIS3 StartXhol Adaptor

ACTITIAACGTCAAGGAGAAAAAACCTCTAGCC GAA CCTCG AGCAACGICAGGGATGTGATTC7CGAAGAATATACTAAAAAAATGG CAG'cAAG ATAAAC f A A G C: AAAGATG

B5'-GAAAACT5CCAAAGACCCAATTGGAGAATAGCGACAAITTTTIGAGGAAC-AGAAGGAAACACT'ACGAAAATAGCGAA- CTAGAAAC i,-AsCCC- i:C- :Si I

CAGAGCACTATCTGAAACAATTAAAAATAAAI GAAAGAACT7GAATCTCAAAGGGAAGAATTAAAiAAAGAGAd.CAC TGAA.A GAA CA'-'AA A

FIG. 2. (A) Structure and sequence of the PGALI-HIS3 gene fusion junction. The sequence of the yeast PGALI immediately proximal to theEcoRI-Xho I cloning site of pYES-R is represented by the open box with the arrow; the stippled box delimits the Xho I adaptor. The HIS3upstream region is represented by an open box. Two upstream, out-of-frame ATGs are underlined. The HIS3 translational initiation codon ATGis in boldfaced type. (B) Partial sequence of the galactose-dependent growth arrest gene THE). The sequence shown begins at the point offusionwith the GAL) promoter. The 5'-most ATG is underlined.

transformants. The final transformant produced only occa-sional papillae when streaked onto galactose plates but grewwell on glucose medium. The library plasmid was rescuedfrom this clone and was used to transform several otherstrains (YNN490, YNN217, and NNY19). In each case, theplasmid conferred galactose-dependent growth arrest, indi-cating that the effect is neither strain nor cell-type specific.Galactose-induced cells of all three strains containing theplasmid were mononucleate, showed no gross morphologicalabnormalities, and did not arrest with aCDC phenotype (datanot shown).Sequence data from the end of the genomic insert DNA

proximal to PGAL, revealed an open reading frame (THE])with no similarity to other sequences presently in the DNAdata bases (Fig. 2B). Sequence data from the other end of thegenomic insert DNA, nearest the lac promoter, indicated thata portion of the ADE2 gene was present on the insertfragment, thus mapping THE) to the long arm of chromo-some XV, within 3000 bp of ADE2.Screen for Activators of the Mating Response Pathway.

Activation of the pheromone response pathway in the ab-sence of pheromone is a phenotype exhibited by the STE4and STEJ2 genes, two of the genes involved in the pathway,when overproduced under PG,4LI control (5-8). To test for thepresence in the library of other such genes, transformants inNNY19 were screened for the ability to induce transcriptionof FUSI, a gene that is induced over 40-fold soon after cellsare exposed to mating pheromone. FUSI transcription hasbeen used often as a measure of primary intracellular re-

sponse to pheromone exposure (16). A chromosomally inte-grated copy ofa FUSJ-IacZ fusion was used as a reporter forFUS) transcriptional activation (6). Thirteen of 150,000transformants screened conferred plasmid-dependent, galac-tose-dependent activation of the FUS] fusion not requiringexogenously supplied mating pheromone. The PGALI proxi-mal ends of the genomic inserts in these plasmids were

sequenced, and the sequence data were used to identify 12 ofthese clones. One clone was isolated that did not matchsequences available in the DNA data bases. This clone isbeing analyzed further, and its characterization will be de-scribed elsewhere. The remaining clones corresponded topreviously identified genes (Table 1).The positions of the vector-insert fusion junctions for the

identified genes are listed in Table 1. Three of the genes,OBFI, PHO2, and STE)H, were isolated twice. The site offusion was unique for each of these clones, indicating thatthey were independently derived. Three MATal clones wereisolated, each with a unique fusion junction ranging from 34to 84 bases upstream of the initiation codon. PGAL,-genefusions occurred as close as 18 bases upstream of theinitiation codon to as far as 256 bases upstream of the ATG.The average distance from the start codon for all the knowngenes obtained is 93 bp.

In the case of two of the MATal clones and the MSNJisolate, the fusion points each occurred upstream of twoATGs, neither of which was in-frame with the coding se-

quence. The leader sequences of the remaining genes con-tained no intervening ATGs.The two STE)I isolates were particularly interesting. Nei-

ther clone contained the full-length STE)) gene. The insertfragment in each clone contained only the carboxyl-terminalcoding region for STE]) plus sequences further downstreamof the gene. In each case, approximately half of the STE))coding region was absent. In one clone, the site offusion withPGALI was located 934 bp into the coding sequence ofSTE)),and in the other it was 953 bp into the gene. In both cases, thePGAL,-gene fusion occurred such that the first ATG after thejunction was in-frame with the STEH) gene coding sequence.With leader sequences of 68 and 87 bp, the two clonesputatively initiate translation from the same internal, in-frame ATG. One of the truncation alleles has been charac-terized further in a study described elsewhere (17).

Quantitatlon of Transcriponal Activation of FUSIacZ.To assess the sensitivity ofthe visual screen and to determinethe relative levels of FUS1)-acZ induction elicited by over-expression ofeach clone, quantitative P-galactosidase assayswere performed. Table 1 lists the levels of activity producedby cells containing each of the clones after 17 h of growth ingalactose-containing media. As little as a 3-fold increase inf3-galactosidase activity was perceptible in the visual screen.

DISCUSSIONThe development of an experimentally regulatable yeastgenomic DNA expression library was prompted by a con-sideration of the advantages such a library would have overstandard high-copy-number libraries. The ability to experi-mentally control protein production enables studies of inap-propriate gene expression that cannot be undertaken usinghigh-copy-number libraries. Such studies could include, for

Table 1. Results of screen for activators of matingresponse pathway

Fusion junction point, bp f-Galactosidase,Gene upstream of initiation codon fold induction*OBF1 105 3 + 0.2OBFl 20 7 ± 0.3MSNI 151 4 ± 2MATal 68 13 ± 0.8MATal 36 27 ± 1MATal 84 49 + 3PHO2 135 36 ± 3PHO2 18 52 ± 3MCMI 256 44 5STE4 84 1600 ± 100STEH) 68t 2800 ± 300STE)) 87t 3000 ± 200*Fold induction relative to the parent strain without a plasmid, grownon galactose. Values are the average (± SD) of two trials.

tPosition of fusion junction point upstream of the internal ATG(corresponding to amino acid 382 of the full-length protein) puta-tively used as start codon for truncation protein.

Dow

nloa

ded

by g

uest

on

June

3, 2

020

Proc. Nadl. Acad. Sci. USA 89 (1992)

example, investigating the effects of constitutive expressionof genes that are normally temporally regulated with the cellcycle, expression in all cell types of genes that are normallyexpressed only in certain cell types (e.g., haploid- or diploid-specific genes), or expression of genes in inappropriategrowth states of the cell (e.g., mating, meiosis, or sporula-tion-specific genes during vegetative growth). The condi-tional expression library should be able to increase theprotein level of virtually any yeast gene above its wild-typelevel, because PGAL1 is one of the strongest yeast promotersidentified (4). Because cells containing the library can bemaintained under conditions where transcription from PGAL,is repressed, the conditional expression library should alsopermit identification of genes whose protein products aretoxic at high levels. Such clones should be absent from ahigh-copy-number library because their unregulated overex-pression should be lethal. Finally, because the conditionaloverexpression described here is plasmid-based, the librarycan be used to study dominant effects of overexpressionwithout the attendant genetic difficulties often associatedwith studying dominant phenotypes. Thus, the potentialadvantages of a conditional expression library over a tradi-tional high-copy-number library are manifold.To achieve regulatable, high-level expression of a given

gene under the strict control of PGALI, two criteria must bemet. First, PGALI should be fused to a region between thecoding sequence and its endogenous promoter and upstreamregulatory regions. Second, there should not be fortuitousATG start codons between PGALI and the bona fide start ofthe gene, because translation in yeast usually begins at thefirst ATG in a given message (18). These theoretical consid-erations and an examination of the upstream regions ofseveral sequenced yeast genes suggested that a PGALI-genefusion occurring up to 100 bp upstream of most codingsequences should result in regulatable control by PGALI. Fora haploid genome of 14 million bp, a minimal library size ofat least 140,000 clones, representing a breakpoint every 100bases, would be necessary to achieve a functional PGALI-gene fusion for each gene in the yeast genome.The library obtained in this study is complex enough to

contain clones representing a breakpoint at every base in thegenome, assuming sequence-independent shearing. Thus thelibrary should contain every possible PGALJ-gene fusion.However, only those fusions that meet the criteria describedabove will result in PGALI-regulated protein production. Themajority of the clones in the library contain genes that retaintheir endogenous regulatory sequences, and thus the libraryfunctions independently of carbon source as a standardsingle-copy library as well. Additionally, because the librarywas made in the pYES vector and the genomic DNA wasinserted nondirectionally, the library should also containgenes that can be expressed in E. coli by using the lacpromoter. In its phage form, it should be suitable for screen-ing with antibodies as well.The isolation of clones capable of complementing amino

acid auxotrophies in both a galactose-independent and ga-lactose-dependent fashion demonstrated that the library con-tains both endogenously regulated genes and coding se-quences under PGALI control. Endogenously regulated HIS3genes were isolated at approximately the frequency expectedfor a standard genomic DNA library (19). In addition, fourclones were isolated that could complement a his3 mutationin a strict galactose-dependent manner. To achieve strictgalactose-dependent transcription, the PGA,4 fusion pointmust usually occur close enough to the coding sequence startto remove the endogenous regulatory signals. In the galac-tose-dependent HIS3 clone that was sequenced, the fusionjunction occurred immediately downstream of the HIS3TATA boxes (20), thus putatively eliminating endogenousHIS3 regulation. Interestingly, however, it occurred up-

stream of two out-of-frame ATGs, which might have beenpredicted to interfere with PGAL-driven HIS3 expression.Three clones identified in the mating-response screen con-tained intervening out-of-frame ATGs as well. Thus in at leastsome cases, the presence ofinterveningATGs between PGLIand the coding sequence initiation codon does not precludeexpression of gene products under galactose control.The set of genes whose RNA or protein products are

deleterious at high dosage represents an important class ofgenes that cannot readily be isolated by using existing tech-niques. Although there have been reports of specific genesthat were lethal when cloned into an overexpression system(21), these genes cannot generally be found by using ahigh-copy-number library. Thus it has not been feasible toidentify more than isolated examples of such genes. Geneisolation by conditional overexpression using this libraryshould permit the identification of a whole class of thesegenes, and, indeed, screening a small fraction of the libraryfor genes that are toxic on overexpression readily identifiedone such gene, THE).A conditional overexpression library also makes possible

the study of dominant overexpression phenotypes. Twogenes in the yeast pheromone response pathway, STE4 andSTE)2, cause dominant, pheromone-independent activationofFUS) transcription when overproduced from PGAI (5, 6,8). In the current study, 150,000 library transformants werescreened for dominant activation of the mating responsepathway. Analysis of the positive clones was greatly simpli-fied by their galactose dependence, which implied that thegenes causing the phenotypes were PGA1 proximal, thuseffectively mapping their positions on the insert fragments. Asingle sequencing primer based on sequences within PG,4L,allowed identification of each of these clones and will be ofgeneral use for the identification of other galactose-dependent clones. From 13 independent positive clones,seven known genes and one previously unknown gene wereisolated.Of the seven known genes, one of the strongest activators

isolated was STE4, which confirms previous work in otherlabs. Genes were isolated that displayed a thousandfold rangeof levels of activation. Among the genes identified in thescreen were three transcription factors, OBFD, PHO2, andMCMI. MCMJ has been shown previously to be involved intranscription of pheromone-inducible genes (22). OBFD andPHO2 are involved in a variety of cellular processes (23, 24)but had not previously been implicated in mating response. Itis possible that the transcriptional activation of FUSI is anindirect result of increased levels of a general transcriptionfactor. MSN1 is a multicopy suppressor of a defect in theSNF protein kinase (25). The fact that the mating responsepathway involves at least four protein kinases suggests thatMSN1 may be performing a function in the mating responsescreen that is analogous to its role in suppressing the defec-tive SNF1 kinase.MATal is an example of a gene that was isolated based on

the phenotype caused by its expression in an inappropriatecell type. MATal is normally transcribed only in cells ofmating type a. Its ectopic expression in an a strain (such asNNY19, in which the screen was performed) results insimultaneous production of a- and a-specific gene products(26) and presumably autocrine activation of the pheromoneresponse pathway.The final known gene identified in the screen was STE)),

one of the protein kinases in the mating pathway. The genewas isolated twice, and neitherclone contained the full-lengthgene. Both plasmids contained inserts lacking z40o of theamino terminus of the coding region. Interestingly, thesetruncated STE] I clones were potent activators of the matingresponse pathway, showing the highest levels of FUSI in-duction of any of the genes isolated in the screen. Thus the

11592 Genetics: Ramer et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

020

Proc. Natl. Acad. Sci. USA 89 (1992) 11593

mating response activation screen revealed yet another in-teresting feature of the regulatable overexpression library,the possibility of identifying dominant truncation alleles. Thegalactose-dependent growth arrest gene may be anotherexample of such a dominant truncation allele, because thereading frame of that clone is open up to the fusion break-point. Such truncation alleles are not represented in a high-copy-number library because they lack functional promoterelements.The results of the library characterization, then, suggest

that this regulatable yeast genomic DNA library will allowaccess to a number of different classes of genes that havephenotypic consequences when overexpressed but havebeen difficult or impossible to isolate to date using traditionalhigh-copy-number libraries.

We thank P. Foreman and R. Fuller for critical comments on themanuscript. We would also like to thank P. Foreman, R. Fuller, andM. Schena for helpful discussions and suggestions and R. Fuller, R.Rugieri, R. Esposito, P. Webster, and B. Cairns for gifts of strainsand oligonucleotides. This work was supported by National Insti-tutes of Health Grants GM28428 and R37HG00198 (to R.W.D.).S.W.R. is a predoctoral trainee supported in part by a Training Grant2TRGM07599 from the National Institutes of Health. S.J.E. was aSenior Fellow of the American Cancer Society.

1. Rothstein, R. J. (1983) Methods Enzymol. 101, 202-211.2. Rine, J., Hansen, W., Hardeman, E. & Davis, R. W. (1983)

Proc. Natl. Acad. Sci. USA 80, 6750-6754.3. Valdivieso, M. H., Mol, P. C., Shaw, J. A., Cabib, E. &

Duran, A. (1991) J. Cell Biol. 114, 101-109.4. Johnston, M. & Davis, R. W. (1984) Mol. Cell. Biol. 4, 1440-

1448.5. Cole, G. M., Stone, D. E. & Reed, S. I. (1990) Mol. Cell. Biol.

10, 510-517.6. Nomoto, S., Nakayama, N., Arai, K. & Matsumoto, K. (1990)

EMBO J. 9, 691-6%.

7. Whiteway, M., Hougan, L. & Thomas, D. Y. (1990) Mol. Cell.Biol. 10, 217-222.

8. Dolan, J. W. & Fields, S. (1990) Genes Dev. 4, 492-502.9. Elledge, S. J., Mulligan, J. T., Ramer, S. W., Spottswood, M.

& Davis, R. W. (1991) Proc. Natl. Acad. Sci. USA 88, 1731-1735.

10. Sherman, F., Fink, G. R. & Hicks, J. B. (1986) Methods inYeast Genetics (Cold Spring Harbor Lab., Cold Spring Harbor,NY).

11. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) MolecularCloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

12. Philippsen, P., Stotz, A. & Scherf, C. (1991) Methods Enzymol.194, 169-182.

13. Jazwinski, S. M. (1990) Methods Enzymol. 182, 154-174.14. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold

Spring Harbor Lab., Cold Spring Harbor, NY).15. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.

Acad. Sci. USA 74, 5463-5467.16. Trueheart, J., Boeke, J. D. & Fink, G. R. (1987) Mol. Cell.

Biol. 7, 2316-2328.17. Cairns, B. R., Ramer, S. W. & Kornberg, R. D. (1992) Genes

Dev. 6, 1305-1318.18. Cigan, A. M. & Donahue, T. F. (1987) Gene 59, 1-18.19. Clarke, L. & Carbon, J. (1976) Cell 9, 91-99.20. Struhl, K. (1983) Cold Spring Harbor Symp. Quant. Biol. 2,

901-910.21. Burke, D., Gasdaska, P. & Hartwell, L. (1989) Mol. Cell. Biol.

9, 1049-1059.22. Jarvis, E. E., Clark, K. L. & Sprague, G. J. (1989) Genes Dev.

3, 936-945.23. Rhode, P. R., Elsasser, S. & Campbell, J. L. (1992) Mol. Cell.

Biol. 12, 1064-1077.24. Berben, G., Legrain, M. & Hilger, F. (1988) Gene 66, 307-312.25. Estruch, F. & Carlson, M. (1990) Nucleic Acids Res. 18,

6959-6964.26. Ammerer, G., Sprague, G. J. & Bender, A. (1985) Proc. Natl.

Acad. Sci. USA 82, 5855-5859.

Genetics: Ramer et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

020