Heterologous modules for efficient and versatile PCR-based gene targeting inSchizosaccharomyces pombe

Yeast 14, 943–951 (998)

Heterologous Modules for Efficient and VersatilePCR-based Gene Targeting in SchizosaccharomycespombeJU}RG BA}HLER1†‡, JIAN-QIU WU1†, MARK S. LONGTINE1, NIRAV G. SHAH1, AMOS MKENZIE III1,ALEXANDER B. STEEVER1, ACHIM WACH2§, PETER PHILIPPSEN2 AND JOHN R. PRINGLE1*1Department of Biology, University of North Carolina, Chapel Hill, NC 27599–3280, U.S.A.2Institut fur Angewandte Mikrobiologie, Biozentrum, Universitat Basel, CH-4056 Basel, Switzerland

We describe a straightforward PCR-based approach to the deletion, tagging, and overexpression of genes in theirnormal chromosomal locations in the fission yeast Schizosaccharomyces pombe. Using this approach and theS. pombe ura4+ gene as a marker, nine genes were deleted with efficiencies of homologous integration ranging from6 to 63%. We also constructed a series of plasmids containing the kanMX6 module, which allows selection ofG418-resistant cells and thus provides a new heterologous marker for use in S. pombe. The modular nature of theseconstructs allows a small number of PCR primers to be used for a wide variety of gene manipulations, includingdeletion, overexpression (using the regulatable nmt1 promoter), C- or N-terminal protein tagging (with HA, Myc,GST, or GFP), and partial C- or N-terminal deletions with or without tagging. Nine genes were manipulated usingthese kanMX6 constructs as templates for PCR. The PCR primers included 60 to 80 bp of flanking sequenceshomologous to target sequences in the genome. Transformants were screened for homologous integration by PCR.In most cases, the efficiency of homologous integration was §50%, and the lowest efficiency encountered was17%. The methodology and constructs described here should greatly facilitate analysis of gene function in S. pombe.? 1998 John Wiley & Sons, Ltd.

— fission yeast; gene deletions; gene truncations; overexpression studies; epitope tagging; polymerasechain reaction; gene expression; green fluorescent protein

*Correspondence to: John Pringle, Department of Biology, Uni-versity of North Carolina, Chapel Hill, NC 27599–3280, U.S.A.†Jurg Bahler and Jian-Qiu Wu contributed equally to this work.‡Present address: Imperial Cancer Research Fund, Cell CycleLaboratory, 44 Lincoln’s Inn Fields, London WC2A 3PX, U.K.§Present address: Bureco AG, Stadtweg 4, CH-4310Rheinfelden, Switzerland.Contract/grant sponsor: NIHContract/grant number: GM31006Contract/grant sponsor: RJEG TrustContract/grant sponsor: University of BaselContract/grant sponsor: Swiss Federal Office for Education andScience

INTRODUCTION

Deletion and overexpression studies are centralto the functional analysis of genes in yeast. In

Contract/grant number: 95.0191

CCC 0749–503X/98/100943–09 $17.50? 1998 John Wiley & Sons, Ltd.

addition, tagging of proteins can allow simple andefficient cell biological and biochemical analyses ofgene products (Smith and Johnson, 1988;Kolodziej and Young, 1991; Prasher, 1995). In thispaper, we describe a set of plasmid templates and aPCR-based method for the straightforwardmanipulation of genes directly in the genome ofSchizosaccharomyces pombe. The approach isbased on the one-step gene disruption method(Rothstein, 1983; Grimm and Kohli, 1988; Baudinet al., 1993; Grallert et al., 1993). In Saccharo-myces cerevisiae, it has been shown that transfor-mation with PCR products that terminate inshort stretches of homology (provided by theprimers) to genomic target sequences frequently
yields homologous integrants, thus replacing the
Received 24 December 1997Accepted 25 February 1998

944 . .

region between the target sequences with the PCRproduct (Baudin et al., 1993; Wach et al., 1997;and references cited therein). This methodologyallows direct manipulations of chromosomalgenes, such as deletion, overexpression, and tag-ging of gene products, without any cloning steps.Starting with plasmids described previously (Wachet al., 1994, 1997), we constructed several plasmidscontaining the heterologous selectable markerkanMX6. These plasmids serve as templates forPCR-based gene targeting in S. pombe.

Several different types of tags have proven to bewidely useful in localizing and isolating geneproducts. We constructed plasmids containingsequences encoding three copies of the influenzavirus hemagglutinin (HA) epitope (Field et al.,1988; Tyers et al., 1992), 13 copies of the humanc-myc (Myc) epitope (Evan et al., 1985; Munroand Pelham, 1987), glutathione S-transferase(GST) from Schistosoma japonicum (Smith et al.,1986), or the green fluorescent protein (GFP) fromthe jellyfish Aequorea victoria (reviewed byPrasher, 1995; Heim et al., 1995). Commercialantibodies to the HA and Myc epitopes are avail-able, and GST-tagged proteins can be purifiedeasily using commercially available glutathionebeads (Smith and Johnson, 1988). GFP fusionproteins can be detected in living (and sometimesin fixed) cells by fluorescence microscopy, andantibodies to GFP are also available. To facilitatetagging of proteins at their N-termini, as well asthe overexpression of both tagged and untaggedproteins, the plasmids constructed include onesincorporating the regulatable nmt1 promoter(Maundrell, 1990; Basi et al., 1993).

MATERIALS AND METHODS

Construction of plasmids containing PCR templatemodules

Standard recombinant-DNA methods were used(Sambrook et al., 1989). Plasmid DNA was pre-pared from bacteria and isolated from agarose gelsusing Qiagen kits. Plasmid KS-ura4 (Figure 1A; akind gift of S. Parisi and J. Kohli, University ofBern, Switzerland) contains the S. pombe ura4+

gene (coding region plus flanking sequences) on a1·8-kb HindIII fragment (Grimm et al., 1988) inthe HindIII site of pBluescript KS- (Stratagene).Construction of pFA6a-kanMX6 and pFA6a-GFP(S65T)-kanMX6 (Figure 1B) has beendescribed by Wach et al. (1997). Construction of

? 1998 John Wiley & Sons, Ltd.

the other plasmids shown in Figure 1B has beendescribed by Longtine et al. (1998).

Plasmids containing an nmt1 promoter togetherwith the kanMX6 marker (Figure 2) were con-structed as follows. The wild-type nmt1 promoter(Maundrell, 1990) and two attenuated versions ofthis promoter (Basi et al., 1993) were amplified byPCR (Expand system; see below) using thepREP3X, pREP41X, and pREP81X vectors (Basiet al., 1993; Maundrell, 1993; Forsburg, 1993)as templates and the primers 5*-TAACCTGAAGATCTCGCCATAAAAGACAGAATAAGTCATC-3* and 5*-TACATGACTTAATTAAAGACATGATTTAACAAAGCGACTATAAGTCAG-3* (restriction sites underlined), resulting in21·2-kb fragments (from position "1163 to +6of nmt1) with a BglII site near the upstream endand a PacI site near the downstream end. ThePCR products were digested with BglII and PacIand cloned into BglII/PacI-digested plasmidpFA6a-kanMX6-PGAL1 (Longtine et al., 1998),thus replacing the GAL1 promoter with one ofthe versions of the nmt1 promoter to create plas-mids pFA6a-kanMX6-P3nmt1, pFA6a-kanMX6-P41nmt1, and pFA6a-kanMX6-P81nmt1 (Figure2). About 200 bp of the downsteam end of thenmt1 promoter in each plasmid were sequenced bythe UNC-CH Automated Sequencing Facility ona Model 373A DNA Sequencer using the TaqDyeDeoxy Terminator Cycle Sequencing Kit(Applied Biosystems); each construct had theexpected sequence at the TATA box (Basi et al.,1993) and the correct junction at the PacI site.Sequences encoding a triple HA epitope, GST, orGFP (carrying the S65T mutation: Heim et al.,1995) were then cloned downstream of the nmt1promoters. To this end, plasmids pFA6a-3HA-kanMX6, pFA6a-GST-kanMX6, and pFA6a-GFP(S65T)-kanMX6 (Figure 1B) were digestedwith PacI and BglII, and the desired fragmentswere gel-purified and cloned into PacI/BamHI-digested plasmids pFA6a-kanMX6-P3nmt1,pFA6a-kanMX6-P41nmt1, and pFA6a-kanMX6-P81nmt1. Both junctions of the inserted tagsequences in the resulting plasmids were checkedby sequencing (as above), and the plasmids werenamed as indicated in Figure 2.

PCR amplification of fragments for transformationPCR primers were 80 to 101 nucleotides long

(see Tables 1 and 2); they were synthesized andPAGE purified by Integrated DNA Technologies.

Yeast 14, 943–951 (1998)

945-

Figure 1. Modules for use as PCR templates to generate fragments to be used for completegene deletion, partial deletion of C-terminal sequences, C-terminal tagging of proteins, or geneexpression studies. Arrows within the boxes show directions of transcription; arrows outsidethe boxes indicate PCR primers (not to scale; see Table 1). The approximate sizes of theexpected PCR products are indicated. (A) ura4+ module (including the S. pombe ura4+ codingregion and flanking sequences) cloned into the multiple cloning site of pBluescript KS-.Restriction sites used for cloning and for identifying insert direction are indicated. (B)kanMX6-based modules cloned into the multiple cloning site of pFA6a (Wach et al., 1994;Longtine et al., 1998). Restriction sites used for cloning are indicated; the AscI site at thejunction of 13Myc and TADH1 sequences was lost during the construction of plasmidpFA6a-13Myc-kanMX6 (Longtine et al., 1998). Gray boxes: kanMX6 module containing thepromoter and terminator sequences of the Ashbya gossypii translation elongation factor 1ágene together with the kanr gene from E. coli (Steiner and Philippsen, 1994; Wach et al., 1994,1997). Black boxes: protein tagging modules containing the terminator sequence of the ADH1gene from S. cerevisiae (Wach et al., 1994, 1997) together with the tags indicated.

DNA fragments were amplified using theExpand High Fidelity PCR system (BoehringerMannheim) and the plasmids shown in Figures 1and 2 as templates. PCR reactions were performedin HotStart 100 tubes (Molecular Bio-Products).The lower mix (total volume, 25 ìl) contained2·5 ìl of Expand buffer with 15 m MgCl2, 0·8 mof each dNTP, 10 ìg BSA, and 2 ì of eachprimer. The upper mix (total volume, 75 ìl) con-tained 7·5 ìl of Expand buffer, 2150 ng of DNAtemplate, and 0·75 ìl of Expand enzyme mixture.20 cycles of 94)C for 1 min, 55)C for 1 min, and68)C for 2 (for plasmids of Figure 1) or 3 (for
plasmids of Figure 2) min were executed. Some

products could not be amplified under these con-ditions (depending on the sequences of the 5*portions of the primers). In these cases, Taqpolymerase (Promega) or the TaqPlus PrecisionPCR system (Stratagene) were used together withthe buffers supplied. The amounts of enzymes andMgCl2were as recommended by the suppliers, andthe reaction mixtures were otherwise as describedabove. 35 cycles of 94)C for 1 min, 55)C for 1 min,and 72)C for 2 (for plasmids of Figure 1) or 3 (forplasmids of Figure 2) min were executed followedby an extension (72)C for 10 min).

The products from two to five PCR reactions
(210–20 ìg of DNA) were pooled, extracted with
Yeast 14, 943–951 (1998)

946 . .

an equal volume of phenol:chloroform:isoamylalcohol (25:24:1), precipitated with ethanol, anddissolved in 10 ìl of TE (pH 8) buffer (Sambrooket al., 1989). This concentrated DNA was useddirectly for transformation of S. pombe cells (seebelow).

Figure 2. Modules for use as PCR templates to generate fragments to be used for gene overexpression, partialdeletion of N-terminal sequences, and/or N-terminal tagging of proteins. Modules were cloned into the multiplecloning site of pFA6a (Wach et al., 1994; see Materials and Methods); restriction sites used for cloning areindicated. Arrows within the boxes show directions of transcription; arrows outside the boxes indicate PCRprimers (not to scale; see Table 1). Gray boxes: kanMX6 module (Wach et al., 1994, 1997; see Figure 1B). Whiteboxes: S. pombe nmt1 promoter (each module is available both with the wild-type promoter and with twoattenuated versions; see text). Black boxes: protein tagging modules (see Figure 1B). The approximate sizes of theexpected PCR products are indicated.

Transformation of S. pombe and selection ofG418-resistant or Ura+ transformants

S. pombe cells were transformed with the PCRfragments using a protocol based on the method ofKeeney and Boeke (1994). Wild-type (strain 972;Leupold, 1970) or ura4-D18 (Grimm and Kohli,1988) cells were grown at 30)C in YE medium(Moreno et al., 1991) to 2107 cells/ml (220 ml/transformation). Cells were washed once with anequal volume of water, and the cell pellet wasresuspended in 1 ml of water, transferred to anEppendorf tube, and washed once with 1 ml ofLiAc/TE made from 10#filter-sterilized stocks(10#LiAc: 1 lithium acetate, adjusted to pH 7·5with diluted acetic acid; 10#TE: 0·1 Tris–HCl,0·01 EDTA, pH 7·5). The cell pellet was thenresuspended in LiAc/TE at 2#109 cells/ml. 100 ìlof the concentrated cells were mixed with 2 ìlsheared herring testes DNA (10 mg/ml Yeast-maker carrier DNA; Clontech Laboratories) and10 ìl of the transforming DNA. After 10 min


incubation at room temperature, 260 ìl of 40%PEG/LiAc/TE (for 20 ml of solution: dissolve 8 gof PEG 4000 in 2 ml of 10#LiAc, 2 ml of10#TE, and 9·75 ml water, and filter sterilize; canbe stored up to 1 month) was added. The cellsuspension was mixed gently and incubated for30–60 min at 30)C. 43 ìl of DMSO were added,and the cells were heat shocked for 5 min at 42)C.Cells transformed with fragments carrying thekanMX6 marker were then washed once with 1 mlof water, resuspended in 0·5 ml of water, andplated onto two YE plates (250 ìl/plate). Theseplates were incubated for 218 h at 30)C, resultingin a lawn of cells. The cells were then replica platedonto YE plates containing 100 mg/l G418/Geneticin (Life Technologies; G418 was addedafter autoclaving the medium, and plates werestored at 4)C in the dark). The replica plateswere incubated for 2–3 days at 30)C, and largecolonies were restreaked onto fresh YE platescontaining G418. (The tiny colonies appearingon the initial G418 plates were not stable trans-formants.) Cells transformed with fragmentscarrying the ura4+ marker were washed withwater as described above, plated onto two EMMplates without uracil, grown at 30)C for 3 days,and restreaked onto fresh EMM plates withouturacil.

Yeast 14, 943–951 (1998)

947-

Table 1. PCR primers used to amplify the transformation modules.

Primer Sequence

Modules of Figure 1KS-ura4 (forward) 5*-(gene-specific sequence)-CGCCAGGGTTTTCCCAGTCACGAC-3*a

KS-ura4 (reverse) 5*-(gene-specific sequence)-AGCGGATAACAATTTCACACAGGA-3*a

pFA6a derivatives (forward) 5*-(gene-specific sequence)-CGG ATC CCC GGG TTA ATT AA-3*b

pFA6a derivatives (reverse) 5*-(gene-specific sequence)-GAATTCGAGCTCGTTTAAAC-3*c

Modules of Figure 2All modules (forward) 5*-(gene-specific sequence)-GAATTCGAGCTCGTTTAAAC-3*d

Modules without tag (reverse) 5*-(gene-specific sequence)-CATGATTTAACAAAGCGACTATA-3*e

Modules with 3HA (reverse) 5*-(gene-specific sequence)-GCA CTG AGC AGC GTA ATC TG-3*f

Modules with GST (reverse) 5*-(gene-specific sequence)-ACG CGG AAC CAG ATC CGA TT-3*f

Modules with GFP(S65T) (reverse) 5*-(gene-specific sequence)-TTT GTA TAG TTC ATC CAT GC-3*f

aThe sequences shown correspond to the M13 ‘forward’ and ‘reverse’ sequencing primers, as present in pBluescript on either sideof the multiple cloning site.bThe reading frame for the tag sequences is indicated. The primer sequence includes the BamHI and PacI sites (underlined) of themultiple cloning site of pFA6a (Wach et al., 1994; Figure 1B). For deletions, the gene-specific portion of the primer is typicallychosen to correspond to sequences immediately upstream of the start codon of the target gene. For C-terminal tagging offull-length proteins, the gene-specific portion of the primer corresponds to the C-terminal codons of the target gene, ending justupstream of the stop codon. For C-terminal deletions, the gene-specific portion of the primer corresponds to sequences in theregion where the truncation is desired. (If the gene fragment will not be tagged, a stop codon should be incorporated into theprimer; if the gene fragment will be tagged, the primer should preserve the reading frame.)cThe primer sequence includes the EcoRI and PmeI sites (underlined) of the multiple cloning site of pFA6a (Wach et al., 1994;Figure 1B). In the studies reported here, we typically used primers whose gene-specific portions corresponded to sequences80–200 bp downstream of the target gene stop codon (in order to leave a gap between the target sequences for C-terminal tagging).However, primers whose gene-specific portions correspond to sequences immediately downstream of the stop codon would beexpected to work well for deletions and probably also are effective for C-terminal tagging (Longtine et al., 1998); the use of suchprimers would reduce the risk of affecting the expression of neighboring genes.dThe sequence includes the EcoRI and PmeI sites (underlined) of the multiple cloning site of pFA6a (Wach et al., 1994; Figure 2).The gene-specific portion of the primer was typically chosen to correspond to sequences 90–200 bp upstream of the start codon ofthe target gene. However, primers whose gene-specific portions correspond to sequences immediately upstream of the start codonmay work as well (cf. note c).eThe complement of the start codon is underlined. For overexpression of full-length proteins, the gene-specific portion of the primercorresponds to the complement of the N-terminal codons of the target gene (without the start codon). For expression ofN-terminally deleted proteins, the gene-specific portion of the primer corresponds to sequences in the region where the truncationis desired (preserving the reading frame with respect to the start codon provided in the primer).fThe reading frames of the tag sequences are indicated. For N-terminal tagging of full-length proteins, the gene-specific portion ofthe primer corresponds to the complement of the N-terminal codons of the target gene (including the start codon). For tagging ofN-terminally deleted proteins, the gene-specific portion of the primer corresponds to sequences in the region where the truncationis desired (preserving the reading frame). The 3* portions of these primer sequences are specific to the tags and correspond to thecomplement of the C-terminal codons of the tags (without the stop codons).

Screening transformants for homologousintegration by PCR

We checked G418-resistant or Ura+ transform-ants by PCR for integration of the DNA fragmentby homologous recombination. Genomic DNA oftransformants was prepared using a procedurebased on that of Hoffman and Winston (1987). Areisolated transformant was grown to stationaryphase in 5 ml of YE medium at 30)C, centrifuged,resuspended in 0·5 ml of water, and transferred toan Eppendorf tube. After centrifugation, the cell
pellet was resuspended in 0·2 ml of 2% Triton

X-100, 1% SDS, 100 m NaCl, 10 m Tris, 1 mEDTA, pH 8·0. 0·2 ml of phenol:chloroform:isoamyl alcohol (25:24:1) and 0·3 g of acid-washedglass beads (425–600 ìm; Sigma) were added,and the tube was vortexed for 2·5 min. Aftercentrifugation for 5 min, 160 ìl of the aqueouslayer were transferred to a new tube, and 1 mlof ethanol was added. After mixing and centri-fugation for 2 min, the pellet was dried andthen dissolved in 50 ìl TE (pH 8·0). 1 ìl of thisDNA preparation was then used as template for a
PCR reaction using Taq polymerase (see above).
Yeast 14, 943–951 (1998)

948 . .

Table 2. Efficiency of homologous integration during PCR-based manipulations ofS. pombe genes.

Genea Manipulationb Marker Length of homologyc Positives/total checkedd

1 Deletion ura4+ 76 & 76 1/22 Deletion ura4+ 76 & 76 2/93 Deletion ura4+ 76 & 76 1/94 Deletion ura4+ 76 & 76 1/95 Deletion ura4+ 76 & 76 1/166 Deletion ura4+ 76 & 76 2/47 Deletion ura4+ 76 & 76 1/48 Deletion ura4+ 76 & 76 12/19

C tag kanr 80 & 80 4/12pom1 Deletion ura4+ 76 & 76 6/16

C tag kanr 80 & 80 26/30OE kanr 80 & 80 26/33

plo1 C tag kanr 80 & 80 8/20spn5 C tag kanr 80 & 80 4/14spn6 C tag kanr 80 & 80 10/11

9 Deletion kanr 80 & 80 4/10C tag kanr 80 & 80 15/18

10 Deletion kanr 70 & 70 2/2C tag kanr 70 & 70 15/15N tag kanr 80 & 80 16/17

11 Deletion kanr 81 & 80 2/4C tag kanr 79 & 80 12/12OE kanr 70 & 76 4/8N tag kanr 70 & 70 8/48

12 Deletion kanr 60 & 60 4/4C tag kanr 60 & 60 11/12OE kanr 60 & 60 7/8N tag kanr 60 & 60 20/24

aKnown genes are indicated by name: pom1 (Bahler and Pringle, 1998); plo1 (Ohkura et al., 1995); spn5and spn6 (septin-encoding genes: Longtine et al., 1996; O. Al-Awar et al., in preparation). Genesindicated by numbers do not yet have names; they were identified either in an overexpression screen(J. B. and J. R. P., unpublished results) or from genomic sequences available in the database.bC tag, C-terminal tagging; N tag, N-terminal tagging; OE, overexpression (nmt1 modules withouttags).cNumbers of nucleotides of homology to genomic target regions (forward & reverse primers).dNumber of transformants with homologous integration at target gene/total number of transformantschecked by PCR (see Materials and Methods). For a given gene, the data for experiments withdifferent C-terminal or N-terminal tags are pooled, as are the data from experiments using the differentversions of the nmt1 promoter.

One primer corresponded to sequences within thetransforming fragment. For modules containingkanMX6, we used primer 5*-GCTAGGATACAGTTCTCACATCACATCCG-3* (for the mod-ules in Figure 1B; corresponds to nucleotides "42to "14 with respect to the kanr gene start codon inthe kanMX6 module) or primer 5*-GCTACTGGATGGTTCAGTCAC-3* (for the modules inFigure 2; corresponds to nucleotides "247 to
"227 in the nmt1 promoter). To test deletions

made using ura4+, we used one of the primers thathad been used to generate the transformationfragment (Table 1). In all cases, the second primercorresponded to a region of the targeted geneoutside the sequences covered by the transformingfragment. A PCR product of the expected sizeshould be observed if the DNA had integrated byhomologous recombination at the targeted gene.In some cases, PCR checks were performed over
both junctions of chromosomal and inserted
Yeast 14, 943–951 (1998)

949-

DNA, using appropriate primers. To determine ifthe transforming fragment had integrated at asingle site in the genome, strains with a positivePCR reaction were crossed against a wild-typestrain, tetrads were dissected, and the resultingspore colonies were analysed for a 2:2 segregationof G418 resistance or the Ura4+ phenotype(together with the correlated positive PCR reac-tion). Some transformants were also checked bySouthern blotting (Sambrook et al., 1989) usingappropriate restriction fragments as probes.

Requests for plasmidsSend plasmid requests either to Jurg Bahler (fax:

(+44) 171 269 3258; e-mail: [email protected])or to Jian-Qiu Wu (fax: (+1) 919 962 0320; e-mail:[email protected]). Investigators who plan touse one or more of the plasmids for commercialpurposes should state this fact in their requests.For plasmids containing the GFP(S65T) allele, aHoward Hughes Medical Institute material trans-fer agreement must be signed. To obtain thisdocument, contact Roger Y. Tsien, HowardHughes Medical Institute, Cellular and MolecularMedicine, University of California at San Diego,9500 Gilman Drive, La Jolla, CA 92093-0647 (fax:(+1) 619 534 5270) and state that you will use thepFA plasmids with GFP(S65T) registered toA. Wach and P. Philippsen. A copy of the materialtransfer agreement must be received before theplasmids can be shipped.

RESULTS AND DISCUSSION

Gene deletion and tagging of protein C-terminiTo make gene deletions using the S. pombe

ura4+ marker, we used plasmid KS-ura4 (Figure1A) as a template for PCR. The primers had atotal length of 100 nucleotides; their 5* ends weretargeting sequences corresponding to 76 nucle-otides immediately upstream and downstream ofthe open reading frame to be deleted, and their3* ends were 24 nucleotides corresponding tosequences on either side of the pBluescript multiplecloning site (Table 1). Thus, the resulting PCRproducts contained the ura4+ gene on a 1·8-kbHindIII fragment, short flanking sequences(2130 bp) derived from the pBluescript KS- vec-tor, and the 76-bp tails homologous to the genomicsequences where integration was desired. Usingthis approach, we deleted nine genes in strainscontaining the ura4-D18 deletion, which removes


all sequences corresponding to those of the ura4+

marker (Grimm and Kohli, 1988). The efficienciesof homologous integration ranged from 6 to 63%(Table 2). This approach allows deletion of theentire open reading frame or of precisely specifiedsegments of C-terminal coding sequence, depend-ing on the target sequences in the forward primer.It should also be possible to perform targetedmutagenesis directly in the S. pombe genomeby a variation of the approach described for S.cerevisiae by Langle-Rouault and Jacobs (1995).

We then asked if modules containing thekanMX6 marker (Wach et al., 1997; Longtineet al., 1998) could be used in S. pombe. We foundthat the heterologous kanMX6 module doesindeed function in S. pombe, allowing selection ofG418-resistant transformants (see Materials andMethods). The kanMX6-based modules shown inFigure 1B allow gene deletion, tagging of proteinC-termini with any of several tags, and deletion ofC-terminal sequences with or without the additionof tags. Because the tagged genes are in theirnormal genomic locations and under their ownpromoters, such tagging should closely reflect thewild-type situation. It should also be possible touse these modules for gene expression studies: theopen reading frame of a target gene would bereplaced by a tag, whose expression would thenbe analysed. Because all of the modules shown canbe amplified by PCR using the same primersequences (Table 1), a wide variety of genemanipulations can be performed with a smallnumber of primers; in particular, the same forwardand reverse primers can be used for C-terminaltagging with each of the available tags, and thesame reverse primer can be used also for genedeletion. Using these modules and primers with 60to 81 nucleotides of homology to genomic targetsequences, we have deleted and/or tagged ninegenes with efficiencies of homologous integrationranging from 29 to 100% (Table 2).

Gene overexpression and tagging of proteinN-termini

To allow the regulated expression and/or over-expression of genes, tagging of protein N-terminiwith any of several tags, and deletion ofN-terminal coding sequences with or without theaddition of tags, we also constructed kanMX6-based modules containing the S. pombe nmt1 pro-moter (Figure 2). Each module was constructedboth with the wild-type promoter (Maundrell,

Yeast 14, 943–951 (1998)

950 . .

1990; designated P3nmt1) and with two weakerderivatives [designated P41nmt1 (mediumstrength) and P81nmt1 (low strength)] that containmutations that attenuate both repressed andinduced levels of expression (Basi et al., 1993;Forsburg, 1993; Maundrell, 1993). Each version ofthe promoter allows expression at a relatively lowlevel (with 5 mg/l thiamine in the medium) ora relatively high level (without thiamine in themedium), thus allowing a wide range of proteinexpression in the transformed cells. N-terminaltagging can sometimes yield functional proteins incases where C-terminal tagging does not, andoverexpression of tagged or normal proteins cansometimes facilitate localization or biochemicalstudies. All of the modules shown in Figure 2 canbe amplified using the same forward primer, butthey require different reverse primers (Table 1).Using these modules and primers with 60 to 80nucleotides of homology to genomic targetsequences, we have manipulated four genes withefficiencies of homologous integration rangingfrom 17 to 94% (Table 2).

ConclusionsThe methodology and modules described above

should facilitate the analysis of gene function in S.pombe. The possibility of using the heterologouskanMX6 marker and selecting for G418-resistantcells makes it unnecessary to target genes inspecific strain backgrounds, and the few otherselectable markers available for S. pombe can stillbe used for other purposes (e.g., selection of plas-mids) in the targeted strains. Various precise genemanipulations can be carried out directly in thegenome without any cloning steps. In most cases,we obtained §50% efficiency of homologous inte-gration of the PCR products (Table 2). Kaur et al.(1997) reported efficiencies of homologous integra-tion in S. pombe of 1–3% using PCR products with240 bp of sequence flanking the target genes. Thehigher efficiencies observed in our study mayreflect the longer (60–80 bp) flanking sequencesused. The efficiencies of homologous integrationvaried in different cases, but none of the 16 genesthat we manipulated by the approach describedhere presented special difficulties. The efficienciesof homologous integration observed here are com-parable to those observed in S. cerevisiae (Baudinet al., 1993; Wach et al., 1997; Longtine et al.,1998), although shorter flanking sequences seem tobe sufficient in S. cerevisiae. The efficiencies of


homologous integration at the desired target genewere high even with the modules containing thenmt1 promoter, which have 1·2 kb of sequencefrom the nmt1 locus itself.

The method described here presumably dependson a high quality of the long (80- to 100-nucleotide) primers used, but we have had very fewproblems with the PAGE-purified primers pro-vided by our supplier. The high cost of such longprimers is a drawback of the method, but themodular design of the templates allows a variety ofmanipulations to be performed with a few primers.Moreover, the high efficiencies of homologousintegration encourage the use of shorter primers(<80 nucleotides), especially because the screeningfor homologous integrants by PCR is straightfor-ward (see Materials and Methods). Our experiencesuggests that the specific transformation protocolused is also important for success with thismethod: although electroporation works well fortransformation with plasmids, the lithium acetate-based method described here (see Materials andMethods) seems to be more effective for inte-gration of linear PCR products into the S. pombegenome.

ACKNOWLEDGEMENTS

We thank Sandro Parisi, Jurg Kohli, SusanForsburg, Odile Mondesert, and Paul Russell forthe kind gift of plasmids and Primo Schar forinitial advice on PCR-based deletions. Work inJ.R.P.’s laboratory was supported by NationalInstitutes of Health grant GM31006 and by fundsfrom the RJEG Trust. Work in P.P.’s laboratorywas supported by a grant from the Universityof Basel and by grant 95.0191 from the SwissFederal Office for Education and Science. J.B.was supported by fellowships from the SwissNational Science Foundation and the Ciba-Geigy-Jubilaums-Stiftung. M.S.L. was supported in partby a postdoctoral fellowship from the NationalInstitutes of Health (GM15766).

REFERENCES

Bahler, J. and Pringle, J. R. (1998). Pom1p, a fissionyeast protein kinase that provides positional infor-mation for both polarized growth and cytokinesis.Genes Dev. 12, 1356–1370.

Basi, G., Schmid, E. and Maundrell, K. (1993). TATAbox mutations in the Schizosaccharomyces pombe
nmt1 promoter affect transcription efficiency but not
Yeast 14, 943–951 (1998)

951-

the transcription start point or thiamine repressibility.Gene 123, 131–136.

Baudin, A., Ozier-Kalogeropoulos, O., Denouel, A.,Lacroute, F. and Cullin, C. (1993). A simple andefficient method for direct gene deletion in Saccharo-myces cerevisiae. Nucl. Acids Res. 21, 3329–3330.

Evan, G. I., Lewis, G. K., Ramsay, G. and Bishop, J. M.(1985). Isolation of monoclonal antibodies specific forhuman c-myc proto-oncogene product. Mol. Cell.Biol. 5, 3610–3616.

Field, J., Nikawa, J.-I., Broek, D., et al. (1988). Purifi-cation of a RAS-responsive adenylyl cyclase complexfrom Saccharomyces cerevisiae by use of an epitopeaddition method. Mol. Cell. Biol. 8, 2159–2165.

Forsburg, S. L. (1993). Comparison of Schizosaccharo-myces pombe expression systems. Nucl. Acids Res. 21,2955–2956.

Grallert, B., Nurse, P. and Patterson, T. E. (1993). Astudy of integrative transformation in Schizosaccharo-myces pombe. Mol. Gen. Genet. 238, 26–32.

Grimm, C. and Kohli, J. (1988). Observations on inte-grative transformation in Schizosaccharomycespombe. Mol. Gen. Genet. 215, 87–93.

Grimm, C., Kohli, J., Murray, J. and Maundrell, K.(1988). Genetic engineering of Schizosaccharomycespombe: a system for gene disruption and replacementusing the ura4 gene as a selectable marker. Mol. Gen.Genet. 215, 81–86.

Heim, R., Cubitt, A. B. and Tsien, R. Y. (1995).Improved green fluorescence. Nature 373, 663–664.

Hoffman, C. S. and Winston, F. (1987). A ten-minuteDNA preparation from yeast efficiently releasesautonomous plasmids for transformation ofEscherichia coli. Gene 57, 267–272.

Kaur, R., Ingavale, S. S. and Bachhawat, A. K. (1997).PCR-mediated direct gene disruption in Schizosac-charomyces pombe. Nucl. Acids Res. 25, 1080–1081.

Keeney, J. B. and Boeke, J. D. (1994). Efficient targetedintegration at leu1–32 and ura4–294 in Schizosac-charomyces pombe. Genetics 136, 849–856.

Kolodziej, P. A. and Young, R. A. (1991). Epitopetagging and protein surveillance. Methods Enzymol.194, 508–519.

Langle-Rouault, F. and Jacobs, E. (1995). A method forperforming precise alterations in the yeast genomeusing a recyclable selectable marker. Nucl. Acids Res.23, 3079–3081.

Leupold, U. (1970). Genetical methods for Schizosac-charomyces pombe. Methods Cell Physiol. 4, 169–177.

Longtine, M. S., DeMarini, D. J., Valencik, M. L., et al.(1996). The septins: roles in cytokinesis and otherprocesses. Curr. Opin. Cell Biol. 8, 106–119.

Longtine, M. S., McKenzie III, A., DeMarini, D. J.,et al. (1998). Additional modules for versatile and


economical PCR-based gene deletion and modifi-cation in Saccharomyces cerevisiae. Yeast 13, 953–961.

Maundrell, K. (1990). nmt1 of fission yeast. A highlytranscribed gene completely repressed by thiamine.J. Biol. Chem. 265, 10857–10864.

Maundrell, K. (1993). Thiamine-repressible expressionvectors pREP and pRIP for fission yeast. Gene 123,127–130.

Moreno, S., Klar, A. and Nurse, P. (1991). Moleculargenetic analysis of fission yeast Schizosaccharomycespombe. Methods Enzymol. 194, 795–823.

Munro, S. and Pelham, H. R. B. (1987). A C-terminalsignal prevents secretion of luminal ER proteins. Cell48, 899–907.

Ohkura, H., Hagan, I. M. and Glover, D. M. (1995).The conserved Schizosaccharomyces pombe kinaseplo1, required to form a bipolar spindle, the actinring, and septum, can drive septum formation in G1

and G2 cells. Genes Dev. 9, 1059–1073.Prasher, D. C. (1995). Using GFP to see the light.

Trends Genet. 11, 320–323.Rothstein, R. J. (1983). One-step gene disruption in

yeast. Methods Enzymol. 101, 202–211.Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989).

Molecular Cloning: A Laboratory Manual, 2nd edn.Cold Spring Harbor Laboratory Press, Cold SpringHarbor, NY.

Smith, D. B., Davern, K. M., Board, P. G., Tiu, W. U.,Garcia, E. G. and Mitchell, G. F. (1986). Mr 26,000antigen of Schistosoma japonicum recognized byresistant WEHI 129/J mice is a parasite glutathioneS-transferase. Proc. Natl. Acad. Sci. USA 83, 8703–8707.

Smith, D. B. and Johnson, K. S. (1988). Single-steppurification of polypeptides expressed in Escherichiacoli as fusions with glutathione S-transferase. Gene 67,31–40.

Steiner, S. and Philippsen, P. (1994). Sequence andpromoter analysis of the highly expressed TEF gene ofthe filamentous fungus Ashbya gossypii. Mol. Gen.Genet. 242, 263–271.

Tyers, M., Tokiwa, G., Nash, R. and Futcher, B. (1992).The Cln3-Cdc28 kinase complex of S. cerevisiae isregulated by proteolysis and phosphorylation. EMBOJ. 11, 1773–1784.

Wach, A., Brachat, A., Pohlmann, R. and Philippsen, P.(1994). New heterologous modules for classicalor PCR-based gene disruptions in Saccharomycescerevisiae. Yeast 10, 1793–1808.

Wach, A., Brachat, A., Alberti-Segui, C., Rebischung,C. and Philippsen, P. (1997). Heterologous HIS3marker and GFP reporter modules for PCR-targetingin Saccharomyces cerevisiae. Yeast 13, 1065–1075.

Yeast 14, 943–951 (1998)

Documents

Heterologous modules for efficient and versatile PCR-based gene targeting inSchizosaccharomyces pombe