Gene, 155 (1995) 51-59 ©1995 Elsevier Science B.V. All rights reserved. 0378-1119/95/$09.50 51

GENE 08704

Allele shuffling: conjugational transfer, plasmid shuffling and suppressor analysis in Saccharomyces cerevisiae

(Recombinant DNA; yeast; meiotic segregation; centromere vector; mutagenesis; extragenic suppression; trans-kingdom conjugation)

Robert S. Sikorski, William A. Michaud, Stuart Tugendreich and Philip Hieter

Department of Molecular Biology and Genetics, The Johns Hopkins University School of Medicine, Baltimore, MD 21205-2185, USA

Received by R.E. Yasbin: 20 July 1994; Revised/Accepted: 3 November 1994; Received at publishers: 13 December 1994

SUMMARY

Trans-acting suppressor analysis represents a powerful genetic technique capable of revealing interactions among biochemical pathways in vivo. Suppressor characterization in Saccharomyces cerevisiae has traditionally utilized meiotic segregation for the requisite manipulation of strain genotypes. Meiotic segregation is not compatible with all yeast genotypes and can be prohibitively labor intensive when examining large collections of suppressors. To facilitate rapid phenotypic analysis of suppressor mutations, we have devised a novel genetic strategy called 'allele shuffling'. This plasmid-based method should in principle identify allele-specific, allele-dependent and bypass suppressors. A centromere vector (YCp) was developed that can be directly transfrered from Escherichia coli to yeast via 'trans-kingdom' conjuga- tion. Suppressors ofa thermolabile cdc23 allele, cdc23-39, were isolated in the background of a yeast host strain harboring the mutant cdc23-39 gene positioned on a counterselectable plasmid. CDC23 or cdc23-39 genes cloned into a mobilizable YCp vector were then transferred directly from E. coli cultures to each suppressed yeast strain on the surfaces of agar plates. Plasmid shuffling of the cdc23-39 allele transconjugants segregated away the original cdc23-39 gene present during mutagenesis, allowing the intra- or extragenic nature of suppression to be determined. Phenotypes (if any) produced by suppressor mutations were revealed in those transconjugants receiving the wild-type CDC23-containing episome. The allele shuffling method should be generally applicable to the analysis of suppressors of any essential yeast gene.

INTRODUCTION

Genetic techniques for elucidating the biochemical function of a novel gene include the isolation and analysis of trans-acting suppressors of conditional mutations. Successfully applied initially to the study of bacterial phage morphogenesis, supressor analysis has now

Correspondence to: Dr. P. Hieter, 725 N. Wolfe Street, Baltimore, MD 21205, USA. Tel. (1-410) 955-3482; Fax (1-410) 614-2987; e-mail: phil hieter@qmail.bs.jhu.edu

Abbreviations: aa, amino acid(s); Ap, ampicillin; bp, base pair(s); cdc, cell-division cycle; CEN, centromere; Cyh, cycloheximide; E., Escherichia; kb, kilobase(s) or 1000 bp; LB, Luria-Bertani (broth); nt,

become widely used in the field of yeast genetics (for review, see Botstein and Maurer, 1982). Suppressor schemes are based on the assumption that compensatory mutations can be made in proteins that physically or biochemically interact with the protein product of the gene under study. Proteins whose functions are linked through biochemical pathways can produce complex

nucleotide(s); ori, origin(s) of replication; oriT, origin(s) of transfer; S., Saccharomyces; SD, synthetic dropout medium; ss, single strand(ed); Sz., Schizosaccharomyces; Tc, tetracycline; Ti, tumor inducing; TPR, tetratricopeptide repeat; tr, temperature resistant; ts, temperature sensi- tive; wt, wild type; YAC, yeast artificial chromosome; YPD, 2% glucose/l% yeast extract/2% peptone.

SSDI 0378-1119(94)00915-5

52

patterns of suppression. Careful analysis of these patterns can reveal mechanistic details not attainable by any other method (e.g., IRA and RAS interactions; Tanaka et al., 1989).

The isolation of suppressors of conditional lethal mut- ations in Saccharomyces cerevisiae is at first relatively straightforward. Large numbers of the conditional mutant can be incubated in the non-permissive environ- ment (high temperature for ts alleles), to select directly for suppressed clones which are capable of cell division and colony formation. The suppressor mutation may reside within the coding sequence of the original mutant allele (intragenic) or within another, unlinked locus (extragenic). In some cases, an extragenic mutation may itself produce a phenotype that can be useful for elucida- tion of the gene's function, or it may even display a condi- tional phenotype similar to the parent mutant. The suppressor may display specificity for certain mutant allele(s) in a pattern consistent with direct protein-protein interactions.

Distinguishing between intra or extragenic suppres- sors, determining allele specificity, and identifying the phenotype(s) associated with the suppressor mutation i t se l f - tasks central to meaningful suppressor analysis all require the creation and analysis of new genetic combi- nations. Traditionally, these new genotypes have been produced by passing appropriately constructed heterozy- gous diploid strains through meiosis and analyzing the phenotypes and genotypes of the resulting recombinant haploid progeny. The meiotic analysis of large collections of suppressors is cumbersome as each individual mutant necessitates the construction and meiotic analysis of sev- eral diploid strains. Also, certain genotypes are incapable of yielding viable spore progeny.

CDC23 is an essential gene in S. cerevisiae required for chromosome transmission and execution of a G2/M transition in the yeast cell cycle (for review, see Pringle and Hartwell, 1981). CDC23 encodes a 62-kDa nuclear protein with an unusual primary structure containing multiple, tandem copies a 34-aa unit (Sikorski et al., 1990a; 1993). Denoted TPR units, these repeats define a growing family of proteins with diverse biologic function (Sikorski et al., 1990a; 1993; Boguski et al., 1990; Goebl and Yanagida, 1991). TPR units may adopt amphipathic s-helical conformations which may promote protein- protein interactions (Sikorski et al., 1990a; Hirano et al., 1990; Lamb et al., 1994). In vitro mutagenesis of CDC23 has been used to reveal the essential nature of conserved TPR aa residues in executing CDC23 function in vivo and to produce a collection of mutant cdc23 alleles for suppressor studies (Sikorski et al., 1993; Lamb et al., 1994).

Standard suppressor analysis is difficult with cdc23

mutants because the homozygous diploid strain (cdc23tS/cdc23 ts) required for standard genetic crosses was found incapable of completing meiosis even at a per- missive temperature (Simchen, 1974; R.S.S and P.H., unpublished data). In order to pursue a detailed charac- terization of cdc23 revertants we set out to develop a scheme for genetic assortment based entirely on mitotic segregation of CEN-based plasmids. Termed allele shuffling, this new technique may find general applicabil- ity in a variety of S. cerevisiae genetic studies.

RESULTS AND DISCUSSION

(a) Allele shuffling strategy The allele shuffling technique as applied to a cdc23

mutant is diagramed in Fig. 1. The parent haploid strain (YRS201) was engineered for compatability with the 'plasmid shuffling' method of episome exchange (Boeke et al., 1987; Sikorski and Boeke, 1991) and the Supll- based visual assay of chromosome segregation. YRS201 also carries the ts cdc23-39 allele linked to the YCp- LEU2/CYH2 plasmid pRS318 (Sikorski and Boeke, 1991). The cdc23-39 mutation was produced through in vitro mutagenesis of CDC23 and has been shown to con- tain a single aa substitution (A404T) within the highly conserved G/A residue of the 'A' subdomain of TPR5 (Sikorski et al., 1990a; 1993). The coding sequence of the chromosomal CDC23 gene in YRS201 was deleted (and replaced with the yeast LYS2 gene), so that the sole source of the essential CDC23 gene product in this strain is pro- vided by the mutant allele on the plasmid, pRS318 was chosen to carry cdc23-39, so that counterselection could be performed by virtue of the CYH2 gene found on this vector. CYH2 confers dominant sensitivity to the drug Cyh (for review, see Sikorski and Boeke, 1991) and mitotic segregants that have lost pRS318 (and cdc23-39) can be selected directly on media containing Cyh. The yeast artificial chromosome YAC12 consisting of a 360-kb segment of human DNA linked to the yeast SUP11, URA3 and TRP1 genes (Pavan et al., 1990) was included in YRS201 to identify suppressors affecting chromosome stability by a colony color assay (Koshland and Hieter, 1988).

We planned the allele shuffling method to proceed as follows. First, a collection of tr revertant mutants of YRS201 would be isolated by selection at a temperature non-permissive for the cdc23-39 allele (3YC). A series of shuttle vectors carrying various CDC23 alleles would then be introduced into each individual mutant and 'exchanged' with the original cdc23-39 allele through Cyh-mediated plasmid shuffling. Depending on the CDC23 allele introduced, it would be possible to distin-

53

E. coli strains

Yeast strains I ;iiiiils ,il e] cis/trans

cdc23 revertants

-cdc23.':LYS2, Cyh R 1

.YAC12(SUPII) ~

YRS201

i i11111 11 hR ol allele specificity

el suppressor phenotype

iii:iiiiiiiiiii! ii!i!!i~i!!i!!i!!iii!:i!!i!!!ii!zlii!!ii!ii!i!ii:: :: II~iiii[liii~iiiii!~iiii~iiiliii!ilili cdc23::LYS2, Cyh R t ranskingdom plasmid conjugat ion shuffle bypass

mutant

[, YAC12(SUPll)

Fig. 1. Schematic diagram of allele shuffling utilizing ts cdc23 alleles. Details of the method can be found within the text. The numbered arrows represent the different experimental pathways that can be followed in order to achieve the desired phenotypic analysis. The cdc23-? plasmid could represent any of a number of different ts alleles of cdc23 (see Table I). Methods: YRS201 {MATa, ade2-101, lys2-801, his3-A200, trpl-A63, leu2-A1, ura3-52, cdc23::LYS2, cyh2 R, pRS237A2.1(LEU2/CYH2/cdc23-39), YAC12} was made in several steps as follows. The transplacement vector pRS187 was digested with Sinai and Nod, and the resulting fragments were used to transform the diploid YPH501 (Sikorski and Hieter, 1989) to Lys +, disrupting one copy of the chromosomal CDC23 loci. This heterozygous strain of the genotype CDC23/cdc23::LYS2 (confirmed by Southern analysis; data not shown) was then transformed to Ura + with the plasmid pRS20 (CDC23). The resulting strain was sporulated, tetrads dissected, and a Ura + clone from a four spore tetrad was isolated and designated YRS124. To introduce a YAC marked by the SUPll gene into this background, YRS124 was mated to YPH510, a strain carrying a 360-kb YAC (YAC12) consisting of anonymous human DNA linked to the yeast URA3 and TRP1 genes (Pavan et al., 1990). The resulting diploid, YRS223A, was sporulated and a YAC12-containing, cdc23::LYS2 (pRS20) haploid was isolated by microdissection. A cyh R derivative of YRS223A was selected directly in the presence of Cyh (see Sikorski and Boeke, 1991) to yield YRS140. The shuffle-plasmid pRS237A2.1 (CDC23/CYH2/LEU2) was introduced into YRS140 to make YRS191. Finally, the pRS20 plasmid was lost through outgrowth on S D - L e u + U r a medium to yield a ts segregant denoted YRS201. pRS20 contains an 8-kb SalI fragment of yeast genomic DNA (isolated from a ~ phage clone that spans the wt CDC23 locus) ligated into the unique SalI site of the URA3-based CEN vector YCp50 (Rose et al., 1987). This plasmid provided the original isolate of the CDC23 gene reported in Sikorski et al. (1990a). Plasmid pRS90 was made by cloning a 2.6-kb BamHI-XbaI fragment containing the wt CDC23 gene and flanking sequences (see Sikorski et al., 1990) into the BamHI and XbaI sites of pBLUESCRIPT-KS + (Stratagene, La Jolla, CA, USA). Plasmid pRS187 is a transplacement vector containing a deletion of CDC23 coding sequences with insertion of the yeast LYS2 gene. It was made by digesting pRS90 with SnaBI + NdeI, isolating the resulting large fragment (containing flanking CDC23 sequences), blunt-ending this fragment with PolIk, and ligating into it a blunt-ended (with PolIk) EcoRI-HindIII fragment containing the LYS2 gene as isolated from YIp601 (Barnes and Thorner, 1986). Plasmid pRS237A2.1 was made by cloning the cdc23-39 allele (Sikorski et al., 1993) as a 2.6-kb BamHI-XbaI fragment between the BamHI and XbaI sites of the YCp-LEU2-CYH2 vector pRS318 (Sikorski and Boeke, 1990). The nt sequence analysis of the cdc23-39 allele reveals a single nt alteration from wt (G 121° ~A) producing a GCG to ACG codon change and an Ala 4°4 to Thr change (Sikorski et al., 1993). Plasmid pRS229C4.2 (mutant donor) was made by cloning the cdc23-39 allele (Sikorski et al., 1993) as a 2.6-kb BamHI-XbaI fragment into the BamHI (partial digestion) and XbaI sites of pRS196. A clone was selected such the the BamHl site within the Tc R gene remained intact. Plasmid pRS229C5 (wt donor) was constructed by a scheme similar to that used for pRS229C4.2 but with insertion of the CDC23 allele instead of cdc23-39.

g u i s h b e t w e e n i n t r a a n d e x t r a g e n i c s u p p r e s s i o n , to e x a m -

ine t he a l le le spec i f ic i ty o f s u p p r e s s i o n , to f ind s u p p r e s s o r s

t h a t n o l o n g e r r e q u i r e a n y cdc23 f u n c t i o n (i.e., b y p a s s

m u t a t i o n s ) , a n d to r e v e a l t h e p h e n o t y p e ( if any ) of t he

s u p p r e s s o r i t se l f (see Fig. 1). To f ac i l i t a t e t he t r a n s f e r of

s eve ra l cdc23 al le les i n t o m a n y y e a s t m u t a n t s , a n d e l imi -

n a t e t h e n e e d to p e r f o r m l a rge n u m b e r s o f D N A - m e d i -

a t e d t r a n s f o r m a t i o n s , we d e c i d e d to e x p l o i t t he p h e n o -

m e n o n o f d i r e c t E. coli-to-S, cerevisiae trans-kingdom

c o n j u g a t i o n ( H e i n e m a n n a n d S p r a g u e , 1989).

54

Trans -k ingdom conjuga t ion , the transfer of p lasmid

D N A between d is tan t ly re la ted organisms, has been

d e m o n s t r a t e d in the l a b o r a t o r y with an E. coli d o n o r

and two yeast recipients: S. cerevisiae (He ineman and

Sprague, 1989) and to Schizosaccharomyces pombe

(Sikorski et al., 1990b). In the process of Sz. pombe

t r anscon juga t ion , p lasmids a p p e a r to suffer s t ruc tura l

r ea r rangemen t s at a high frequency (Sikorski et al.,

1990b). Before conjugal t ransfer could be cons idered a

general technique for in in t roduc ing genes on C E N plas-

mids into S. cerevisiae we needed to first explore several

technical issues.

(b) Construction of a mobilizable YCp vector A two-p lasmid system similar to tha t used by

H e i n e m a n n and Sprague (1989) was modi f ied for conju-

gal t ransfer a YCp vector from bacter ia to y e a s t . In this

system, a 'helper ' p lasmid ( p D P T 5 1 ) provides in trans all

of the requi red t ransfer (tra) and mobi l i za t ion (mob) func-

t ions necessary for con juga t ion of a second p lasmid con-

ta in ing the cis element, or iT (for review, see He inemann

and Sprague, 1991). A t t empt s to use an existing C E N -

based vector (YCp50) for t r anscon juga t ion yielded very

few t r anscon jugan t s in s imple plate ma t ing assays,

pe rhaps because c loning man ipu l a t i ons have al tered

sequences near the or iT site. A funct ional oriT site can

be found na tura l ly in the c o m m o n cloning vector pBR322

(nt 2212-2353), but it has been inac t iva ted by modif ica-

t ions to create subsequent der ivat ives such as p U C and

p B L U E S C R I P T . We, therefore, cons t ruc ted a new mobi -

l izable YCp shutt le vector, pRS196, inco rpora t ing the

or iT and Tc R sequences of pBR322 and the yeast-specific

e lements of pRS313 (Fig. 2). The bacter ia l Tc R gene on

pRS196 provides a means for g rowth selection dur ing

p r o p a g a t i o n of d o n o r bac te r ia (see M e thods in the legend

to Fig. 2). The yeast HIS3 gene on pRS196 allows yeast

t r anscon jugan t s to be selected direct ly on med ia lacking

his t idine by c o m p l e m e n t a t i o n of the his3-A200 m u t a t i o n

present in the b a c k g r o u n d of our yeast s trains (e.g.,

YRS201). Us ing a s t a n d a r d t r anscon juga t ion assay

Fig. 2. Construction of the mobilizable YCp vector pRS196. The oriT region of pBR322 (labelled T; nt 2212 2353) along with the Tc R gene were cloned into the backbone of the yeast shuttle vector pRS313 as described. For allele shuffling, CDC23 alleles were cloned as XbaI-BamHI fragments between the XbaI and BamHI sites of the polylinker. Note that the BamHI site is not unique in pRS196 (another site is present within the Tc R gene). The sequence labelled 'V' represents the vegetative ori for the plasmids in bacteria. Methods: Standard techniques were used for DNA cloning and restriction enzyme analysis (Maniatis et al., 1982) with a rapid boiling method used to characterize putative clones (Holmes and Quigley, 1981). E. coli strain DH5~ {F-, endA1, hsdl7(r~,m~ ), supE44, thi-1, recA, gyrA96, relA1, A(argF-lacZYA) U169 (q~80 lacZAM15)} from Bethesda Research Laboratories served as a bacterial host. Yeast transformations were performed by the Li-acetate method of Ito et al. (1983). Restriction enzymes were purchased from New England Biolabs and Bethesda Research Laboratories. The mobilizable YCp vector (pRS196) was constructed by ligating the 3.7-kb PvuI-EcoRI fragment of pBR322 to the 3.2-kb PvuI-EcoRI fragment of pRS313 (Sikorski and Hieter, 1989). The appropriate PvuI site in pRS313 was cleaved by partial digestion, as there are two PvuI sites within this vector. This cloning procedure should also allow the construction of mobilizable vectors containing alternate selectable markers using segments of the various pRS YCp (Sikorski and Hieter, 1989) and pRS YEp shuttle vectors (Christianson et al., 1992).

Fig. 3. A patch method for E. coli-to-S, cerevisiae episome transfer. Yeast transconjugants are shown forming colonies on this replica to S D - H i s medium after transfer of pRS196. See section c for detatils. Note that no colonies are formed if bacteria are omited prior to replica plating. The yeast colonies appear dark because of a pigment produced in ade2 strains grown medium containing limiting adenine. Methods: Yeast cultures were grown on YPD medium containing 2% agar. Yeast transformants were grown on SD medium plus the required auxotro- phic supplements (Rose et al., 1990). The permissive temperature for ts mutants was 25°C. The recipient yeast strains were patched to grids of approx. 1 cm 2 on YPD medium and incubated overnight at 25°C. The next day 10 pl of donor bacterial cells (see below for preparation and genotype) were pipetted directly onto each yeast patch, and the yeast/ bacteria mixes were cultured together overnight to allow conjugation. For selection of the transconjugants, the patches were then replica- plated to the appropriate selective SD medium (e.g., selecting for His + ). About 20-80 individual plasmid-bearing colonies appeared as papillae in each patch, depending on the concentration of bacterial suspension used. Because the donor E. coli strain used, JBl l7 , is a Thr - (as well as Leu ) auxotroph, background growth of bacteria on the yeast selec- tive plates is prevented if threonine is excluded from the SD supplement. Donor bacteria were prepared by growing the appropriate bacterial strain (containing the plasmid to be transferred) to saturation in selec- tive LB medium, washing the cells one time with TNB (50 mM Tris pH 7.6/0.05% NaCI), and concentrating the cells 10-1000-fold in TNB. Selection for strain JBl17 harboring pRS196 or its derivatives and pDPT51 is accomplished in LB plus trimethoprim (200 ~tg/ml) and Tc (12.5 pg/ml). Bacterial cells can be used immediately (for best results) or stored for future use as frozen aliquots in TNB/7.5% dimethylsulfate. The bacterial strain host routinely used to transfer plasmids by trans- kingdom conjugation is JB l l7 {supE44, thi-1, thr-1, leuB6, lacyl, tonA21, 2 , (r~, m~), (mcrA-, mcrB+)}, a derivative of C600 that contains the broad-host-range plasmid pDPT51 (Heinemann and Sprague, 1989; 1990). pDPT51 acts as a 'helper' by providing Mob and Tra functions necessary to mobilize (in trans) the transfer of plasmids

55

(Heinemann and Sprague, 1989) with donor bacteria car- rying pRS196 and recipient yeast strain YRS201 we have demonstrated the production of His + yeast ex-conjugants (data not shown).

(c) A simple method for E. coli-to-S, cerevisiae episome transfer

After exploring various methods for performing the transconjugation in order to minimize the amount of manipulations and time required we developed the following reproducible experimental protocol. The recipi- ent yeast strains were first streaked into square or X-shaped patches (see Fig. 3) onto solid rich medium (YPD) capable of supporting the growth of both bacteria and yeast. The next day, small samples of donor bacteria cultures were thawed from frozen stocks and pipetted directly onto each yeast patch, and transconjugation was allowed to proceed overnight. Replicas of the bacteria/yeast co-cultures were then printed to minimal medium for selection of His + yeast transconjugants. Yeast cultures interacting with donor bacteria via this method generated many His + colonies (Fig. 3).

(d) CEN plasmids are transferred to S. cerevisiae intact The mechanism by which trans-kingdom conjugation

occurs between bacteria and yeast is not known. However, many details of the mechanism behind the sym- biotic DNA transfer of Ti DNA from Agrobacterium to recipient plants have been elucidated (Zambryski et al., 1989). The circular Ti DNA plasmid is mobilized into the plant cell as a s s form where it traverses the cyto- plasm, enters the nucleus and integrates into the plant genome. By analogy to the Ti mechanism, a YCp vector may also be transferred through a ss intermediate. It is not known whether such species would be subject to a high rate of mutagenesis or rearrangement during the transfer or establishment in yeast. To examine the final structure of a YCp vector after transconjugation, we per- formed several independent transfers of plasmid pRS229C5 (pRS196 + CDC23) from E. coli to S. cerevis- iae as described above and isolated plasmid DNA from the resulting His + yeast clones. DNA preparations made from 12 independent yeast isolates all were capable of transforming E. coli to Ap resistance indicating that the plasmids had remained circular and that the Ap R gene had remained intact. To look for evidence of more subtle rearrangements, DNA samples from the E. coli trans-

containing the ColE1 oriT element, a DNA sequence found near the ori of pBR322 replicons. In addition to pRS196, a functional oriT can be found in the traditional yeast cloning vectors YEpl3 (LEU2) and YEp24 (URA3).

56

formants were digested with the restriction enzyme EcoRI to yield several diagnostic fragments (Fig. 4). N o gross

structural changes were detected in any of the plasmids

tested. Thus, it appears that the mechanism of transfer utilized during the bacteria-yeast t ransconjugat ion

process does not disrupt the structural integrity of the

mobilized YCp vector.

(e) Selection of a cdc23 allele for suppressor analysis We have isolated and characterized a large collection

of ts alleles of CDC23 produced by both in vitro and in

vivo mutagenesis techniques (Sikorski et al., 1993). To identify a cdc23 mutan t suitable for suppressor studies

we examined the growth temperature spectrum of ten alleles expressing the tightest ts phenotype. Each mutan t

allele was cloned into the pRS318 vector and introduced

into the yeast host YRS140 which contains a complete deletion of the CDC23 locus (see Methods in the legend

to Fig. 1). Cultures of the resultant strains were assayed

for their ability to form colonies at various temperatures

(Table I). The cdc23-1 allele appeared the most temper-

ature-labile member of the collection, but this allele

proved difficult to manipulate, since it grew very slowly at all temperatures. We chose cdc23-39 for suppressor

analysis based on its temperature spectrum, the fact that its muta t ion was well characterized, and the existence of

the muta t ion within one of the highly conserved residues

of the T P R domain (Sikorski et al., 1993). However, a caveat should be noted here, that the ability to isolate

extragenic suppressors of a part icular ts muta t ion is likely

to be both allele- and locus-dependent (and is not predict-

1 2 3 4 5 6 7 8 9 10 12 14

Fig. 4. Structural analysis of a YCp vector after mobilization by trans- kingdom conjugation by agarose-gel electrophoresis. DNA samples in lanes 2-14 were digested with EcoRI. Lane 2 contains pRS229C5 (pRS196+CDC23) DNA to serve as a reference for the migration of the various diagnostic fragments of the parent plasmid. Lanes 3 14 contain independent isolates of Ap R clones isolated from His + yeast tranconjugants of the strain YPH500 (Sikorski and Hieter, 1989), Lane 1 contains molecular weight markers of )~ phage DNA digested with BstEII.

TABLE I Plating efficiency of cdc23 ts strains

cdc23 allele Plating efficiency ~

Temperature (°C)

20 25 30 33 35 37

wt 354 339 335 313 336 329 -1 264 186 1 0 0 0 -2 120 148 172 134 92 0 -37 234 194 26 0 0 0 -39 268 246 73 0 0 0 -40 292 158 77 3 1 0 -49 463 448 175 8 0 0 -50 265 233 185 2 0 0 -51 316 339 208 32 2 0 -54 288 300 201 17 0 0 -56 224 298 237 114 11 0

" Plating efficiency was determined by spreading 500 cells (counted microscopically) onto YPD medium at the designated temperatures, and counting the colonies. The results are the average of two platings at each temperature.

able a priori). For example, it may not be possible to isolate extragenic supressors for some ts mutations.

(f) Analysis of cdc23 suppressors by allele shuffling Revertants of cdc23-39 were isolated after mutagenesis

and growth selection at 35°C and 64 strains were selected

for allele shuffling studies. Of these suppressor mutations, 23 were also capable of suppressing the cdc23 growth

defect at 37°C. 32 mutants in the collection were judged dominan t and 31 recessive in their ability to suppress the

35°C growth defect. Allele shuffling, as described above,

was performed on all 64 revertants to determine the intra

or extragenic nature of suppression (pathway 1 in Fig. 1) and the phenotype (if any) of the suppressor alone (path- way 3 in Fig. 1). Only one mutan t was found to be intra-

genic, that is, suppression being dependent on the presence of the original mutagenized pRS318-cdc23-39 episome. Complete nt sequence analysis of the cdc23 open

reading frame of the mutagenized plasmid isolated from this suppressor revealed a wt sequence, indicating that reversion was the result of an A 121° --*G change. All other

(63) suppressors were trans-acting. We next examined the suppressor collection for any

phenotypes produced by the mutated suppressor loci in the background of wt CDC23. Five suppressor mutat ions induced ch romosome instability as judged visually by the missegregation of a yeast artificial ch romosome (YAC). Six suppressors were themselves ts for growth at 37°C (Table II). We analyzed each ts suppressor strain micro- scopically to look for morphologies suggestive of a cell division cycle arrest similar that produced by cdc23-39.

T A B L E I I

Pheno types of ts cdc23 suppressors

57

Sup. Growth" C h r o m o s o m e loss b D o m / R e c c

35°C 37°C cdc23-39 CDC23

cdc23-39 CDC23 cdc23-39 CDC23

wt - + - + N A -- N A

3 + + - - - + Rec

9 + + - - + + + - Rec

11 + _+ - - + - Rec

35 + -- -- -- + + + Rec

39 + + -- - + - D o m

40 + + . . . . D o m

" The cdc23 geno type was m a n i p u l a t e d by allele shuffling as descr ibed in Fig. 1 such tha t the wt or cdc23-39 allele was present on the shut t le

vector pRS196.

b C h r o m o s o m e loss was de te rmined v isual ly a t 35°C us ing a co lony color assay (see K o s h l a n d and Hieter , 1987) and the SUP11-marked YAC,

YAC12. The results are presented relat ive to the wt hos t s t ra in YRS201.

c D o m i n a n c e and recessiveness was assayed by m a t i n g the or ig ina l suppressed s t ra in (in YRS201 b a c k g r o u n d ) to the tes ter s t ra in YRS232 (SUP/ cdc23-39). The resul t ing d ip lo ids were checked for g rowth at 35°C.

Methods: YRS201 was mutagen ized wi th ethyl me thane sulfonate as descr ibed (Spencer et al., 1988) a t a level sufficient to increase the frequency of

canavan ine - re s i s t an t m u t a n t s ( inac t iva t ion of the CAN1 locus) tenfold. This mutagen ized cul ture was d i s t r ibu ted over several Y P D pla tes and

incuba ted at 35°C for five days. The tr colonies were identif ied and co lony purif ied before fur ther analysis . The suppressors p roduced after mutagenes i s

are un l ike ly to be s ibl ings because no tr colonies were p roduced when a s imi lar n u m b e r of cells f rom an unmutagen ized cul ture was incuba ted at

35°C. D o m i n a n c e and recessiveness were checked by m a t i n g the suppressed s t ra ins to a der ivat ive of YRS201, YRS232 {MATer, ade2-101, lys2-801, his3-A200, trp1-A63, leu2-A1, ura3-52, cdc23::LYS2, cyh2 R, pRS229C4.2(HIS3/cdc23-39)}, select ing for His + and Leu + diploids. These d ip lo ids were

scored for g rowth at 35°C on Y P D medium. C h r o m o s o m e segregat ion defects in the suppressors were assayed qua l i t a t ive ly by visual inspect ion,

no t ing the relat ive frequency of red sectors ( loss of YAC12) p roduced in colonies formed on solid m e d i u m con ta in ing l imi t ing concen t ra t ions of

adenine. The basis for the SUPll-based sec tor ing assay is descr ibed in K o s h l a n d and Hieter (1987).

None of these mutants arrested with the typical large- bud phenotype of cdc mutants involved in nuclear division.

(g) Dependence of suppressor mutations on episomal location of cdc23-39

We expected to obtain a class of suppressors that would require a plasmid-borne cdc23-39 allele in order to express the suppression phenotype. To address this issue we selected a number of the suppressor clones and integrated a single copy of the cdc23-39 allele into each of their genomes by gene-conversion (see Methods in the legend to Table III). Analysis on the resultant stains showed that 50% (8/16) of the suppressors possessed the ability to fully suppress a chromosomal copy of cdc23-39,

19% (3/16) were markedly impaired in suppression and 31% (5/16) lost the ability to supress altogether. This last category may represent suppressors which allow the accumulation of YCp episomes to an elevated copy number.

(h) Episome-dependent suppressors of cdc23-39 We encountered a class of suppressors (episome-

dependent) that appear to function only in the context of an episome-borne copy of the cdc23-39 allele (Table III). It is possible that suppression of the ts pheno-

type in this subset of mutants is brought about through mutations that elevate the copy number of the YCp vector, and consequently elevate the intracellular levels of the thermolabile Cdc23 protein. Such mutants may be of interest to investigations of centromere function as the CEN element present on such YCp vectors normally functions to maintain 1-2 episome copies per cell. Although episome-dependent suppressors seem to be iso- lated frequently in the allele shuffling scheme, they can be easily identified and eliminated from further study after one-step gene conversion to generate a single chro- mosomal copy of the gene under study. Alternatively, one can envison modifications of the allele shuffling method to circumvent selection for episome-dependent suppres- sors. For instance, one may be able to utilize a lox-cre-

based system in which a chromosomal copy (flanked by lox sites) of the mutated gene under study is present during the suppressor selection phase. Induction of Cre recombinase in this strain can then be used to promote excision of the suppressed gene into an episome for shuffling manipulations.

(i) The allele shuffling method We have developed a series of genetic manipulations

termed 'allele shuffling' that are designed for rapid and extensive characterization of suppressor mutations in the

58

TABLE III

Effect of genome location on suppression a

Suppressor Suppression of chromosomal copy of cdc23-39

Sup2 + / - Sup3 Sup9 + Supl0 Supl6 Supl8 + Sup22 + Sup26 + / - Sup32 + Sup38 + Sup39 Sup40 Sup50 + Sup54 + Sup60 + Sup62 +/--

Methods: A double counter-selection was performed on suppressor strains in the YRS201 background to select for the gene conversion of cdc23::LYS2 to cdc23-39 by the plasmid-borne copy of cdc23-39.

Although virtually all of the CDC23 coding region is deleted in the cdc23::LYS2 locus, there remains approx. 300 bp of DNA on either side of the disruption that can recombine with the cdc23-39 fragment on the plasmid pRS237A2.1. The strain to be gene converted was first plated onto m-amino adipate plates to select against the growth of LYS2 +

strains (Sikorski and Boeke, 1991). Approx. 4 × 107 cells were plated onto m-amino adipate medium containing uracil, leucine, tryptophane, histidine, adenine, and lysine supplements (Rose et al., 1990) and incu- bated at 25°C. Colonies were visible after about six days, and were suitable for replica printing after approx. 8 days of incubation. Typically, 100 5000 ~x-amino-adipate-resistant colonies were obtained per plate. The colonies were then printed to YPD plates containing 10 gg Cyh per ml to select for loss of the CYH2-carrying vector. Generally about one-tenth of the ~x-amino-adipate-resistant colonies were able to grow on the Cyh plates. Cyh R colonies were then single-colony purified on YPD + Cyh medium and then tested for Leu + prototrophy. Those that grew on SD-Leu plates were discarded because the plasmid-linked LEU2 marker should have been lost from the strain during plasmid shuffling. A random selection of the resultant colonies were checked by Southern analysis to show that the precise gene conversion had taken place in all cases (data not shown).

only a small set of suppressors are being examined. Yeast strains grown on solid media can be made competent for DNA transformation by a simple procedure involving treatment with Li.isothiocyanate (Keszenman-Pereyra and Hieda, 1988) or by electroporation.

The strength of the allele shuffling method is that it allows one to process many candidate suppressors in a relatively short period of time with a few repetitive manipulations. The investigator can, therefore, assign stringent criteria for those phenotypes expected of a sup- pressor which acts directly in the genetic or biochemical pathway of interest. For example, one may consider studying only suppressors that by themselves exhibit a phenotype similar to the original conditional allele (for example, see Moir, 1982). Criteria such as allele specificity of suppression can also be factored into the choice of which suppressors to pursue in more detail.

Using an abbreviated version of allele shuffling (path- way 4, Fig. 1), one can also search for a class of mutants which 'bypass' the requirement for a function previously shown essential for viability. Bypass suppressors have been used successfully in the dissection of complex biochemical pathways, such as yeast cell cycle control networks (for example, see Russell and Nurse, 1986). Because bypass mutants are likely to be rare (or non- existent) a simple screening procedure is essential for the screening of a large number of candidate revertants. Using the host strain designed for allele shuffling, one need only transfer each revertant to the appropriate counterselection medium for the analysis (Cyh for shuffling pRS318). Growth indicates that the episome is no longer required and therefore the essential function it carried had been 'bypassed' by a trans-acting mutation.

The allele shuffle strategy and CEN-based E. coli/S. cerevisiae conjugation vectors described in this report should be uesful in a variety of applications in addition to suppressor isolation, and may provide general tools for novel yeast genetic studies.

yeast S. cerevisiae. Although our experiments have focused on suppressors of CDC23 function, allele shuffling should be applicable to any gene for which con- ditional-lethal alleles exists. The key to the allele shuffling method involves the positioning of a conditional allele on an episome compatible with plasmid shuffling (see Sikorski and Boeke, 1991). For our cdc23 studies we cloned the ts alleles directly into the shuffle vector; how- ever, one should be able to use double-strand break recombination ('allele rescue') to construct the appro- priate recombinants in vivo (Rothstein, 1991). We used transkingdom conjugation to introduce test alleles into the set of suppressors, but this feature is not essential if

A C K N O W L E D G E M E N T S

We thank Jack Heinemann for kindly providing E. coli host strains and Joachim Li for helpful discussions con- cerning transkingdom conjugation. In addition, we are grateful to Rob Jensen and Karen Chapman for their enthusiastic support during development of these meth- ods. R.S.S. was supported by a Medical Scientist Training Program grant from the National Institutes of Health. This work was supported by a grant from the Pew Memorial Trust to P.H. (T84-0200-025).

REFERENCES

Barnes, D.A. and Thorner, J.: Genetic manipulation of Saccharomyces cerevisiae by use of the LYS2 gene. Mol. Cell. Biol. 6 (1986) 2828-2838.

Boeke, J., Truehart, J., Natsoulis, G. and Fink, G.R.: 5-Fluoroorotic acid as a selective agent in yeast molecular genetics. Methods Enzymol. 154 (1987) 164 175.

Boguski, M., Sikorski, R.S., Hieter, P. and Goebl, M.:New functional roles for TPR genes. Nature 346 (1990) 114.

Botstein, D. and Maurer, R.: Genetic approaches to the analysis of microbial development. Annu. Rev. Genet. 16 (1982) 61-83.

Christianson, T.W., Sikorski, R.S., Dante, M., Shero, J.H. and Hieter, P.: Multifunctional yeast high-copy-number shuttle vectors. Gene 110(1991) 119 122.

Gerring, S.L., Spencer, F. and Hieter, P.: The CHLI(CTF1) gene pro- duct of Saccharomyces cerevisiae is important for chromosome transmission and normal cell cycle progression in G2/M. EMBO J. 9 (1991) 4347-4358.

Goebl, M. and Yanagida, M.: The TPR snap helix: a novel protein motif from mitosis to transcription. Trends Biochem. Sci. 16 (1991) 173-177.

Heinemann, J.A. and Sprague, G.F.: Bacterial conjugative plasmids mobilize DNA transfer between bacteria and yeast. Nature 340 (1989) 205 209.

Heinemann, J.A. and Sprague, G.F.: Transmission of plasmid DNA to yeast by conjugation with bacteria. Methods Enzymol. 194 (1990) 187-195.

Hirano, T., Hiraoka, Y. and Yanagida, M.: A temperature-sensitive mutation of the Schizosaccharomyces pombe gene nuc2+ that encodes a nuclear scaffold-like protein blocks spindle elongation in mitotic anaphase. J. Cell. Biol. 106 (1988) 1171-1183.

Hirano, T., Kinoshita, N., Morikawa, K. and Yanagida, M.: Snap helix with knob and hole: Essential repeats in S. pombe nuclear protein nuc2+. Cell 60 (1990) 319-328.

Holmes, D.S. and Quigley, M.: A rapid boiling method for the prepara- tion of bacterial plasmids. Anal. Biochem. 114 ( 1981 ) 193-197.

Icho, T. and Wickner, R.B.: Metal-binding, nucleic acid-binding finger sequences in the CDC16 gene of Saccharomyces cerevisiae. Nucleic Acids Res, 15 (1987) 8439 8450.

Ito, H., Fukuda, Y., Murata, K. and Kimura, A.: Transformation of intact yeast cells treated with alkali cations. J. Bacteriol. 153 (1983) 163 168.

Keszenman-Pereyra, D. and Kotaro, H.: A colony procedure for trans- formation of Saccharomyces cerevisiae. Curr. Genet. 13 (1988) 21-23.

Koshland, D. and Hieter, P.: Visual assay for chromosome ploidy. Methods Enzymol. 155 (1987) 351-372.

Maniatis, T., Fritsch, E.F. and Sambrook, J.: Molecular Cloning. A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1982.

Moir, D., Stewart, S.E., Osmond, B.C. and Botstein, D.: Cold-sensitive cell-division-cycle mutants of yeast: isolation, properties, and pseu- doreversion studies. Genetics 100 (1982) 547.

59

Pavan, W.J., Hieter, P. and Reeves, R.H.: Generation of deletion deriva- tives by targeted transformation of human-derived yeast artificial chromosomes. Proc. Natl. Acad. Sci. USA 87 (1990) 1300-1304.

Pringle, J.R. and Hartwell, L.H.: The Saccharomyces cerevisiae cell cycle. In: Strathern, J.N., Jones, E.W. and Broach, J.R. (Eds.), The Molecular Biology of the Yeast Saccharomyces, Life Cycle and Inheritance. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1981, pp. 97-142.

Rose, M.D., Novick, P., Thomas, J.H., Botstein, D. and Fink, G.R.: A Saccharomyces cerevisiae genomic plasmid based on a centromere containing shuttle vector. Gene 60 (1987) 237-243.

Rose, M.D., Winston, F. and Hieter, P.: Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1990.

Rothstein, R.: Targeting, disruption, replacement, and allele rescue: integrative transformation in yeast. Methods Enzymol. 194 (1991) 281-302.

Russell, P. and Nurse, P.: cdc25 + functions as an inducer in the mitotic control of fission yeast. Cell 45 (1986) 145-153.

Schultz, J., Marshall-Carlson, L. and Carlson, M.: The N-terminal TPR region is the functional domain of SSN6, a nuclear phosphoprotein of Saccharomyces cerevisiae. Mol. Cell. Biol. 10 (1990) 4744-4756.

Sikorski, R.S. and Hieter, P.: A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122 (1989) 19-27.

Sikorski, R.S., Boguski, M.S., Goebl, M. and Hieter, P.: A repeating amino acid motif in CDC23 defines a family of proteins and a new relationship among genes required for mitosis and RNA synthesis. Cell 60 (1990a) 307 317.

Sikorski, R.S., Michaud, W., Levin, H., Boeke, J., Hieter, P.: Transkingdom promiscuity. Nature 345 (1990b) 581 582.

Sikorski, R.S. and Boeke, J.B.: In vitro mutagenesis and plasmid shuffling: from cloned gene to mutant yeast. Methods Enzymol. 194 (1991) 302 318.

Sikorski, R.S., Michaud, W.A., Wootton, J., Boguski, M.S., Connelly, C. and Hieter, P.: TPR proteins as essential components of the yeast cell cycle. Cold Spring Harb. Symp. Quant. Biol. 56 (1991) 663-673.

Sikorski, R.S., Michaud, W.A. and Hieter, P.: p62 ca°z3 of S. cerevisiae: a nuclear tetratricopeptide repeat protein with two mutable domains. Mol. Cell. Biol. 13 (1993) 1212 1231.

Simchen, G.: Are mitotic functions required in meiosis? Genetics 76 (1974) 745-753.

Spencer, F., Connelly, C., Lee, S. and Hieter, P.: Isolation and cloning of conditionally lethal chromosome transmission fidelity genes in Saccharomyces cerevisiae. Cancer Cells 6 (1988) 441-452.

Tanaka, K., Matsumoto, K. and Toh-e, A.: IRAI, an inhibitory regula- tor of the RAS-cyclic AMP pathway in Saccharomyces cerevisiae. Mol. Cell. Biol. 9 (1989) 757-768.

Weinert, T.A. and Hartwell, L.: The RAD9 gene controls the response to DNA damage in Saccharomyces cerevisiae. Science 241 (1988) 317-322.

Zambryski, P., Tempe, J. and Schell, J.: Transfer and function of T-DNA genes from Agrobacterium Ti and Ri plasmids in plants. Cell 56 (1989) 193-201.