Proc. NatI. Acad. Sci. USAVol. 83, pp. 735-739, February 1986Genetics

Shuttle mutagenesis: A method of transposon mutagenesis forSaccharomyces cerevisiae

(Escherichia coLi/transposon Tn3/loxP/chromosomal transformation)

H. STEVEN SEIFERT*, EMILY Y. CHEN, MAGDALENE SO, AND FRED HEFFRONDepartment of Molecular Biology, Scripps Clinic and Research Foundation, La Jolla, CA 92037

Communicated by John Abelson, September 30, 1985

ABSTRACT We have extended the method of transposonmutagenesis to the eukaryote, Saccharomyces cerevisiae. Abacterial transposon containing a selectable yeast gene can betransposed into a cloned fragment of yeast DNA in Escherichiacoli, and the transposon insertion can be returned to the yeastgenome by homologous recombination. Initially, the clonedyeast DNA fragment to be mutagenized was transformed intoan E. coli strain containing an F factor derivative carrying thetransposable element. The culture was grown to allow trans-position and cointegrate formation and, upon conjugation,recipients were selected that contained yeast sequences withtransposon insertions. The yeast DNA was removed from thevector by restriction endonuclease digestion, and the trans-poson insertion was transformed into yeast. The procedurerequired a minimum number of manipulations, and eachtransconjugant colony contained an independent insertion. Wedescribe 12 transposon Tn3 derivatives for this procedure aswell as several cloning vectors to facilitate the method.

Transposon mutagenesis has many advantages over chemicalmutagenesis including a high mutation frequency withoutkilling the organism, single hit mutations, the ability torecover the mutated gene after mutagenesis, the ability tointroduce selectable markers in strain construction, and useas a portable region of homology for genetic manipulations.Most transposon mutagenesis protocols rely on a bacterio-phage or conjugative plasmid for delivery of the insertionelement into the organism, so they cannot be applied toorganisms that do not have phage or conjugative plasmids.The technique described here, shuttle mutagenesis, utilizescloned genes in Escherichia coli as targets for transpositionand requires only that the organism has a transformationsystem with the ability to recombine insertions into thechromosome. Efficient insertion and selection ofinsertions incloned genes within E. coli is achieved by conduction ofcointegrate forms of plasmids between strains (1). To allowselection in both yeast and E. coli we have constructedseveral mini-transposon derivatives containing the Tn3 (3-lactamase gene and one of five selectable yeast genes. Inaddition, we have constructed derivatives of each of thesetransposons that contain a copy of the E. coli lacZ codingsequence fused to an open reading frame in one Tn3 38-base-pair (bp) terminal repeat. Translational fusions of clonedyeast genes showing /3-galactosidase activity can be isolatedwhen in frame insertions into the genes are made.

Shuttle mutagenesis is designed to use derivatives ofbacterial transposon Tn3 as insertion sequences. Thistransposon confers ampicillin resistance to its host and hasseveral properties that make it efficient in producing inser-tions (2). Transposition of Tn3 normally occurs in two steps.In the first step the donor and target molecules becomejoined

with direct repeats ofTn3 at thejunction (a cointegrate). Thisstep requires transposase, an enzyme encoded by Tn3 andthe two 38-bp terminal repeats. The transposase can besupplied in trans and, thus Tn3-derived transposons as smallas 72 bp, encompassing only the ends, are functional forcointegrate formation (3). The second step in Tn3 transpo-sition is resolution ofthe cointegrate to give a simple insertionin the target DNA while reforming the donor. Resolution ofcointegrates occurs by site-specific recombination at aninternal resolution site (res or IRS) mediated by the Tn3resolvase (4). Bacteriophage P1 also encodes a site-specificrecombination system that can resolve cointegrates. Thebacteriophage produces the cre enzyme that catalyzes re-combination between lox sites (5). The major differencebetween the P1 and Tn3 systems is that cre can catalyze bothintra- and intermolecular recombination while the Tn3resolvase only acts intramolecularly. We have made use ofthe P1 lox/cre system to resolve cointegrates in shuttlemutagenesis.Tn3 shows transposition immunity to plasmids that contain

at least one Tn3 terminal repeat. Transposition immunity isonly active in cis leaving transposition to other plasmids inthe same cell unaffected (6). All common E, coli vectors thatconfer resistance to penicillin contain one Tn3 terminalrepeat and are immune to Tn3. Therefore, we constructedTn3-free vectors for use in this system. Furthermore, Tn3does not transpose to the E. coli chromosome, although thereason for the immunity of the E. coli chromosome isunknown. Transposition immunity provides a way of direct-ing transposition of Tn3 into a given target.To demonstrate the utility of shuttle mutagenesis, we have

mapped a yeast adenine biosynthesis gene, ADEI, withmultiple insertions of a HIS3 carrying mini-transposon,m-Tn3(HIS3).

MATERIALS AND METHODSBacterial and Yeast Strains. All E. coli and Saccharomyces

cerevisiae strains used in this study and their properties arelisted in Table 1.DNA Manipulations. Enzymes were purchased from

Boehringer Mannheim (except endonuclease Not I that wasobtained from New England Biolabs) and used under condi-tions recommended by the supplier. Details of the vector andmini-transposon constructions are described in the figurelegends. The polylinker used to construct pHSS4 was syn-thesized on an Applied Biosystems 380A DNA synthesizer,and the DNA was purified from a 20% polyacrylamide/7 Murea gel before cloning. Transfer of digested DNA to nitro-cellulose and hybridization of the filter with nick-translatedprobe, kinase treatment ofdephosphorylated ends, and filling

Abbreviations: bp, base pair(s); kb, kilobase(s).*To whom reprint requests should be addressed at: MB2, ScrippsClinic, 10666 N. Torrey Pines Rd., La Jolla, CA 92037.

735

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

Mar

ch 1

1, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986)

Table 1. Strains of E. coli and S. cerevisiae

Comments/source/Strain Relevant properties reference

Escherichia coliJM109 F'traD36 laclq-ZAM15

proAB+/recA1A(lac-pro)endAI gyrA96 hsdRl7 Ref. 7

DH1 F- recAl endAl gyrA96hsdRl7 Ref. 8

RDP145 F- recAl3 rpsL Derivative ofSK1592/Ref. 9

RDP146 F- recAlA(lac-pro) rpsE Ref. 10NS2114Sm F- recA X-cre rpsL Streptomycin-

resistantderivative ofNS2114/NatSternberg

W3110polA F- polA Derivative of W3110Saccharomyces cerevisiae

YP52ADE' a his3 ura3 lys2 trpl Spontaneous ADE2revertant ofYP52/Phil Hieter

in to make blunt ends with the Klenow fragment of DNApolymerase I were performed by standard techniques (11).

Biological Manipulations. E. coli cells were transformedaccording to the procedure of Hanahan (8) or Dagert andErhlich (12). S. cerevisiae cells were transformed with alithium acetate transformation procedure (13). E. coli conju-gations were as described (36). Donor and recipient cellswere grown at 37°C with good aeration to about 3 x 108 cellsper ml. Matings were done at a donor:recipient ratio of 1:2 for15 min. Transconjugants were selected on plates or in brothculture.

RESULTS

The Vectors. The Tn3-free vectors were designed to carrya minimal amount of nonessential information to favortransposition into clonedDNA sequences rather than into thevector. The important part of the vectors is the polylinkersurrounded by restriction endonuclease Not I sites. SinceNot I has an 8-bp recognition sequence (5'-GCGGCCGC) itwould be predicted to cut once every 16 kilobases (kb) in arandom DNA sequence. In an A-T-rich organism like yeast,the enzyme has a probability of cutting every 390 kb (assum-ing 60% AT). Therefore, DNA fragments that have beencloned between the Not I sites have a high probability ofbeing excised intact by Not I. The basic vectors used in theprocedure are derived from an amplifiable pMB8 replicon(pFH97; ref. 14), the kanamycin resistance determinant(Kmr) from Tn5 (15), and a synthetic double-stranded DNAmolecule containing recognition sites for the restrictionenzymes Not I, BamHI, and EcoRI (Fig. 1) The vectors alsocontain the polylinker from pLink322 (Fig. 2; ref. 11) thatcontains sites for Cla I, HindIII, Xba I, Bgl II, and Pst I. Thevectors are between 2.2 and 2.8 kb in size and contain from400 to 1000 bp of nonessential DNA sequence.

Vectors pHSS9 and pHSS10 (not shown) carry a DNAfragment that contains the M13 phage origin of replication(17). Single-stranded DNA is produced from these plasmidswhen an M13 helper phage is used. A gel purified 514-bp RsaI origin fragment from M13 replicative form I (RFI) DNA (18)was inserted into the Sma I site of pHSS6 to make pHSS9.The polylinker in pHSS9 was inverted between the Not I sitesto make pHSS10. The M13 origin vectors can be used to clonemutagenic insertions from a yeast cell into E. coli. Aftersingle stranded DNA is produced, oligonucleotide probesspecific for the ends of the mini-transposon can be used as

AH Bg Ap'N

EUC1Km rS ~~~~~~pFH97 R

on RI

|A. Xmal DigestionB. Ligation RI

dIll Digest A Ndel Digestion

B. Klenow Fill-inC. EcoRADigestionD. Ligation

H

Kmr p

FG1.Cntu ionetoni Synthetic Linker

into ~ AHXdm-n al IDigestedn pUN(16) Th smlN m

B. EcoRI Digestiction f AATT f K3'3'. -GGCC

A. LigationH

RIX N

FIG. 1. Construction of cloning vector pHSS4. The startingplasmid, pUC19Km, was made by inserting the HindIIIsSal Ikanamycin-resistance-encoding fragment from plasmid pRZ102 (15)into HindIII- and Sal I-digested pUC19 (16). The small Xma Ifragment was deleted from pUC19Km by cutting with Xma I andligating under dilute conditions to produce pUC19KmAX (notshown). The HindkeIthEcoRh restriction fragment from pUC19KmAXthat sencodes kanamycin resistance was ligated to the 700 bpEcoRmaNdeI origin fragment from pFH97 (14) after filling in theHindIII and Nde I sites with the Klenow fragment of DNA poly-merase I, to produce pHSS3. The HindIII site was retained inceplasmid. Two 33-base DNA oligomers were prepared on an AppliedBiosystemsr380A DNA synthesizer with the following sequences.A[5'-d(CCGGGCGGCCGCGGATCCGAATTCCGCGGCCGC)]and B [5 H'-d(AATTGCGGCCGCGGAATTCGGATCCGCGG-CCGC)]. The oligonucleotides were gel purified, annealed by slowcooling from100(Cto 22°C and treated with kinase. This producedthe synthetic linker with the restriction endonuclease sites shown.The synthetic linker was cloned at the EcoRI and Xma I sites ofpHSS3 to make pHSS4. Solid lines, sequencesdelrvedfuom pUC19;hatched lines, sequences derived fromTno; open lines, sequencesderived from pFH97. Representative symboys are as follows: ori, anorigin of replication; Apre ampicillin resistance; Kmr, kanamycinresistance;H( HindIII; BgP,BR 2Pst I; X, XmaI; RI, EcoRI;StSal I; N. Not I.

pnmers in dideoxy sequencing (19) to sequence the site ofthemini-transposon insertion directly. In addition, a series ofinsertions of a mini-Tn3 element into a cloned insert can beused to sequence a large insert without subcloning.The Mini-Transposons. The prototype mini-Tn3 derivative

was constructed as shown in Fig. 3, by introducing a cloned,synthetically prepared copy of the Tn3 terminal repeat frompIR38-8 (20) into pRH43, a pBR322 derivative containing twocloned loxP sites (21). This resulted in a plasmid (pTn) thathas a ,B-lactamase gene and a loxP site surrounded by two38-bp inverted repeats. Only one terminal repeat of Tn3 wasrequired since a second copy of the terminal repeat wasalready present at position 3148 ofpBR322 (27). Choice oftheloxP site was based on its smaller size as compared with theres site ofTn3 and for the convenience of flanking restrictionsites in the constructions. Also, many E. coli strains containthe transposon y8 (TnlO00) in their chromosome (28), and theresolvase of Tnl000 can complement that of Tn3. If trans-

736 Genetics: Seifert et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 737

ori

Bgl 11

Kmr Sma IPst I

- Not IEco RI

- Cla I- Hind IIIXba I

- Bgl IIPst IBarn HINot I

FIG. 2. Physical and genetic map of pHSS6. The vector shown isabout 2.3 kb in size. To make pHSS6, the 92-bp EcoRI/BamHIfragment from pLink322 (11) was cloned between the same sites ina pHSS4 (Fig. 1) derivative, in which the HindIII site had beenremoved by filling in with the Klenow fragment ofDNA polymeraseI.

position were done in a TnJO00-containing strain, mostcointegrates would resolve before transfer. Thus the use ofloxP alleviated the necessity of using TnlO00-free strains inthe procedure.The selectable yeast genes HIS3, URA3, TRPI, LEU2, and

SUPJI were inserted into this small Tn3 derivative byinserting the appropriate restriction fragment. Details of thecloning strategies are discussed in Fig. 3. To make deriva-tives of these transposons that can be used to producetranslational fusions, the (-galactosidase coding region, with-out the first eight amino acids, was cloned into the uniqueBamHI site in each of the six plasmids described above frompMC1871 (22). These plasmids have the P-galactosidasecoding region in frame with the one open reading frame of the38-bp terminal repeat. The structure of one of the lacmini-transposons is shown in Fig. 3b. When these mini-transposons transpose into a coding region in frame they willproduce translational fusion products.The mini-transposons shown in Fig. 3 are defective in that

b.

C.

d.

6.

f.

g.pTn Ps

Ir Xmori

they contain only the 38-bp repeats of Tn3. They requiretransposase to be supplied in trans for transposition to occur.We make use ofapACYC184 derivative, pLB101, as a sourceof transposase. This plasmid has had the right end of Tn3deleted leaving transposase as the only trans-acting Tn3function (3). This plasmid is immune to Tn3 transposition dueto the remaining 38-bp Tn3 repeat insuring that transpositionof an element will not occur into pLB101. Resolution ofcointegrate structures can not take place due to the absenceof a res site in the defective Tn3. Resolution of cointegratestructures can be obtained by site-specific recombinationbetween bacteriophage P1 loxP sites contained in the mini-transposons (5). This recombination event is catalyzed by thecre protein of bacteriophage P1 that is encoded for by abacteriophage X lysogen in E. coli strain NS2114. Although aP1 phage lysogen could also be used to resolve the cointe-grate structures, newly transferred DNA from a strain thatdoes not express the P1 modification activity would besubject to degradation by the phage restriction activity.We chose the F derivative pOX38 (29) to carry the

mini-transposons because it transfers at high frequency(100%o per hour) and does not contain any known insertionsequences that might form cointegrates with the vector. Eachof the 12 plasmids described above containing the mini-Tn3derivatives was transformed into a cell containing pLB101,the source oftransposase, and pOX38. Transconjugants wereselected in matings between these strains and a polA deriv-ative of strain W3110 carrying a X prophage that provides crefunction. Transconjugants were grown for several genera-tions to segregate out the pBR322 replicon. After transfer ofpOX38 carrying the mini-transposon from the polA strain toNS2114SM, the absence of the small multicopy plasmid wasascertained.

Mutagenesis of a Cloned Gene. The ADEI gene of S.cerevisiae produces phosphoribosyl-amino-imidazolesucci-nocarboximide synthetase, an enzyme involved in the ade-nine biosynthetic pathway. Yeast strains that contain an adel

P RI B RI B

lacZ'P RI B B

$ -Z-HIS3P RI H P H

t URA3P RI H RI SB

+ LEU2P RI HPSuX B

TRPl- fP RI RI

supli

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

FIG. 3. Construction of mini-transposon carrying plasmid pTn and genetic maps of the mini-transposon derivatives. (Left) The smallHindIII/Sma I fragment from pIR38-8 that contains a Tn3 terminal repeat (20) was inserted between the HindIII site and filled in Xho I siteof pRH43 (21) to make plasmid pTn. (Right) All of the mini-transposon derivatives were made from pTn. Solid lines, the Tn3 terminal repeats;hatched lines, the ,-lactamase coding sequence; cross with a vertical arrow under it, a loxP site; and open lines, additional sequences with theorientation of the selectable marker shown by a horizontal arrow, if known. Not all of the restriction endonuclease sites between the loxP siteand the terminal repeat have been shown for the derivatives b-d and g, but can be implied by following the cloning strategies below. (a) Theprototype mini-transposon m-Tn3 that is carried on plasmid pTn. (b) m-Tn3(lac): This mini-transposon was produced by subcloning the3galactosidase coding region as a BamHI fragment from plasmid pMC1871 (22). Analogous lacZ fusion derivatives of each transposon belowhave been made but are not shown. (c) m-Tn3(HIS3): Produced by subcloning the 1.8-kb BamHI HIS3 fragment from pSZ62 (23) into the BamHIsite of pTn. (d) m-Tn3(URA3): Made by subcloning the 1-kb HindIII URA3 fragment from pDURA (24) into the HindIII site of pTn. (e)m-Tn3(LEU2): Made by subcloning the 2.1-kb Sal I/HindIII LEU2 fragment from CV13 [YEpl3(25)] into these same sites in pTn. (f)m-Tn3(TRPI): Made by subcloning the 800-bp EcoRI/Bgl II TRPI fragment from pSE271 (26) into EcoRI/BamHI-digested pRH43. TheEcoRI/Xho I fragment containing loxP and TRPI from this plasmid was inserted between the EcoRI/Sal I sites in pTn. (g) m-Tn3(SUPII): Aderivative of pTn made by subcloning the 1.1-kb EcoRI SUPII fragment from p115 (26) into the EcoRI site of pTn. P, Pst I; RI, EcoRI; H,HindIII; S, Sal I; B, BamHI; Sm, Sma I; Su, Sau3AI; X, Xho I.

Genetics: Seifert et al.

HPSBP Rl.

a. 0

4 loxp

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986)

mutation require adenine for growth and have a red colorwhen plated on YPD media (11), making mutations in adeleasy to score. The ADEJ gene has been inserted into an E.coliplasmid, and a preliminary mapping ofthe gene by R-loopanalysis has been reported (30). We used the cloned ADEIDNA as a target for saturation mutagenesis with a mini-transposon to provide a more detailed mapping of the codingsequence of the gene.A 2-kb EcoRI fragment containing the ADEI gene was

subcloned from YEp13 (ADEI)1 (30) into the EcoRI site ofpHSS4 and used as a target for shuttle mutagenesis. Thecellular events that occur after the cloned ADEI gene istransformed into E. coli are schematically shown in Fig. 4.Fig. 4 is representative of the events occurring in a single cellof a culture regardless of the mini-transposon derivative orcloned gene used. After allowing transposition of m-Tn3(HIS3) from pOX38 into pHSS4(ADEJ) the culture wasmated with NS2114Sm for 30 min and transconjugantsselected with kanamycin, ampicillin, and streptomycin. Nor-mally, greater than 104 transconjugants per ml are found aftermating for 15 min. The plasmid DNA from 41 transconjugantcolonies was analyzed by digestion with restriction endonu-

A.

B.

clease Not I to distinguish insertions into the vector fromthose in the yeast DNA. Six insertions were in the vector andnot mapped. Three transconjugants showed multiple bands inrestriction digests and are thought to result from multiplematings of different donors with the same recipient. Thenumber of transconjugants showing multiple bands can bereduced by mating for 15 min (data not shown). All 32insertions in the cloned DNA were mapped by digestion withseveral restriction endonucleases. Fig. 5 shows a physicaland genetic map of pHSS4(ADEJ) and the insertions ofm-Tn3(HIS3) into theADEI gene. The insertions were spreadevenly throughout the cloned DNA with no two insertionsmapping as identical. Therefore, there are no hot spots form-Tn3 transposition in pHSS4(ADEJ).DNA from all 32 isolates was digested with Not I, and used

to transform S. cerevisiae strain YP52ADE' selecting forHIS' colonies. The 3 insertions in the pBR322 homology didnot transform, while the other 29 clones each gave about 200colonies with 5 jug of DNA. We assume that this low level oftransformation is due to the nonhomologous ends of the NotI-digested DNA. All m-Tn3(HIS3) insertions that resulted ina red phenotype on YPD media when transformed intoYP52ADE' were assumed to be in the ADEI gene, These 16insertions are shown by the thick line in Fig. 5. To show thatthe insertions were in the homologous region of the chromo-some, Southern blots were performed on six of the HIS'transformants using pHSS4(ADEI) DNA as a probe. DNAwas prepared from the HIS' Ye52 transformants (Fig. 5),completely digested with EcoRI, and transferred onto nitro-cellulose. The hybridization of this blot with labeled pHS-S4(ADEI) DNA is presented in the Inset of Fig. 5. Lane A ofthe Inset contains DNA from the HIS- parent strain. It canbe determined that all of the transformants contain anadditional 3.2 kb of DNA split by an EcoRI site. This isconsistent with the 3.2-kb m-Tn3(HIS3) having been insertedby one step transplacement (31) into the homologous regionof the chromosome to where the insertion mapped in theplasmid DNA. The faint bands in lanes C and D result frominsertions that map near the central EcoRI site in the ADE1

FIG. 4. Schematic representation of mutagenesis protocol. Thelarge boxes in the figure represent individual E. coli cells. Black filledlines, Tn3 derived sequences; hatched lines, the pHSS4 vectorsequences; the dot-filled line, the ADEI sequences; open lines, theyeast HIS3 gene. (A) Single E. coli cell containing three plasmids;pLI3101, pOX38: :m-Tn3(HIS3), and pHSS4(ADEI). (B)Transposase is constitutively expressed in the donor cell so thattransposition will occur forming a cointegrate between pOX38::m-Tn3(HIS3) and pHSS4(ADEI). The E. coli chromosome and pLB101are immune to Tn3 transposition. This cell is mated with NS2114Smand the cointegrate transferred. (C) The recipient strain (NS2114Sm)encodes cre that resolves the cointegrate structure resulting inpOX38::m-Tn3(HIS3) and pHSS4(ADEI)::m-Tn3(HIS3).

B C D F E G

Ih JTII. TITTII TI Ii

A 13 C D ES PGE

_-+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ d r-RI7Kmr f n

RI RI B B B R |N Bg N RI

. yeast insert - I pHSS4-IpBR322sequence

FIG. 5. Map of m-Tn3(HIS3) insertions in pHSS4(ADEI) andSouthern blot of selected insertions after transplacement into theyeast genome. Vertical lines show the sites of m-Tn3(HIS3) inser-tions in the plasmid. The arrows on the top of the vertical linesindicate the direction of the mini-transposon insertions with thearrow pointing in the same direction as the 3-lactamase gene. Thethick horizontal line shows those insertions that gave an adel (red)phenotype when transplaced into yeast. The symbols representingrestriction endonuclease sites are as indicated in the previous figures.(Inset) Autoradiogram of chromosomal DNA that was digested tocompletion with EcoRI and probed with pHSS4(ADEI) DNA thathad been 32p labeled by nick-translation. Lane A, parent strainYP52ADE'. The two hybridizing bands correspond to fragments of3.85 kb and 1.08 kb. Lanes B-G, HIS' transformants ofYP52ADE'that were transformed with Not I-digested pHSS4(ADEI)::m-Tn3(HIS3) plasmids as indicated above the line drawing.

738 Genetics: Seifert et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

020

Proc. Natl. Acad. Sci. USA 83 (1986) 739

locus and leave less than 200 bp of ADEI homology. Theseresults show that the mini-transposons can be used toinactivate a gene, map the coding region, and transplaceinsertions into homologous regions of the yeast genome.

DISCUSSIONWe have developed a shuttle mutagenesis system that ex-tends the method of transposon mutagenesis to the eukaryoteS. cerevisiae. The power of the new protocol lies in theefficient insertion of defective Tn3 derivatives into clonedyeast genes, and the ability to select the transplacement ofthetransposon(s) into the yeast genome. Sites within yeast genesthat have been mutated by using shuttle mutagenesis can beeasily identified using the ampicillin resistance marker car-ried by the mini-transposons. This method can be used togenerate a series of insertions in a single clone or a collectionof clones.

Shuttle mutagenesis has several advantages over conven-tional mutagenesis procedures. The use of Tn3 was based onseveral of its properties. (i) The transposition functions ofTn3 can be separated into cis- and trans-acting functions.During shuttle mutagenesis the trans-acting functions (e.g.,transposase) are only present in the cell with the mini-transposons for a short time. This reduces the number oftransposition induced adjacent deletions. (it) Tn3 showstransposition immunity to plasmids containing Tn3 or at leasta copy of the 38-bp repeat of Tn3, and to the E. colichromosome. Transposition immunity is used in shuttlemutagenesis to direct transposition to plasmids carryingcloned DNA sequences. (iii) Another property of Tn3 trans-position is its requirement for cointegrate formation with thetarget DNA. This is used in the mutagenesis to ensure thatevery clone with an insertion, and only those clones withinsertions, can be mated into the recipient cell. (iv) Finally,the efficiency of Tn3 transposition is independent of size.Therefore, all of the mini-transposon derivatives show a highfrequency of transposition.

Another useful feature of the system lies in the mini-transposon derivatives that carry the ,B-galactosidase codingsequence. These mini-transposons can be used to make lacZfusions with any cloned gene when transposition fuses the lacZsequence in frame with the cloned gene coding sequence. Thefusions can be used to study regulation of a gene (32, 33) or topurify a gene product using P-galactosidase antibody to isolatethe fusion product (34). To aid in the procedure, a series ofnewTn3-free cloning vectors was constructed. These plasmidscontain minimal nonessential information, leaving the largesttarget for insertions in the cloned sequences. In addition, the useof restriction endonuclease Not I sites to surround the cloningregion allows the excision ofintact fiagments ofyeastDNA thatcarry sufficient DNA homology surrounding the transposoninsertions to recombine into the yeast chromosome. The lowlevel of yeast transformants can be improved upon if the Not Iends of the DNA were removed by limited nuclease BAL-31digestion.We have demonstrated the utility of the technique by using

it to map the yeast ADEI gene. Insertions that produced anadel phenotype (red color) when returned to yeast spannedapproximately 900 bp. This is in close agreement with thepredicted size of the gene (30). Two insertions at one end ofthe gene produced a pink phenotype suggesting that they stillhad partial function. We believe that these insertions may bein the 3' end of the gene leaving a truncated protein withpartial enzymatic activity. However, from this data alone wecannot determine which end of the gene is 5'.

This method of shuttle mutagenesis allows rapid mappingand sequencing of cloned yeast genes. The mini-transposonthat carries URA3 can be used as a marker to move alleles

into yeast that have no selectable phenotype. The ability toselect for URA' cells on minimal medium allows cotrans-formation of the nonselectable mutation with URA'. Theability to select for URA- cells with 5-fluoroorotic acid (35)can then be used to select for those cells which have becomeURA- through a spontaneous loss of the URA3 mini-transposon. The selectable phenotype associated with thesemutations will aid in their cloning and characterization. Thebasic steps of this procedure can be adapted for use withother organisms besides yeast.

Helpful discussions by R. Rothstein, M. Jayaram, and J. Nickoloffwere greatly appreciated. Strains supplied by R. Hoess, D. Kaback,R. D. Porter, and N. Sternberg were essential for the completion ofthis project. This work was supported by National Institutes ofHealth Grants AI20978 (F.H.) and AI20845 (M.S.), National ScienceFoundation Grant DMB-8217002 (F.H.), and a Damon Run-yon-Walter Winchell Cancer Fund Fellowship, DGR-847 (H.S.S.).

1. Clark, A. J. & Warren, G. J. (1979) Annu. Rev. Genet. 13, 99-125.2. Heffron, F. (1983) in Mobile Genetic Elements, ed. Shapiro, J.

(Academic, New York), pp. 223-260.3. Huang, C.-J., Heffron, F., Twu, J.-S., Schloemer, R. H. & Lee,

C.-H. (1985) Gene, in press.4. Gill, R., Heffron, F., Douga, G. & Falkow, S. (1978) J. Bacteriol.

136, 742-756.5. Sternberg, N. & Hamilton, D. (1981) J. Mol. Biol. 150, 467-486.6. Robinson, M. K., Bennett, P. M. & Richmond, M. H. (1977) J.

Bacteriol. 129, 407-414.7. Yanisch-Perron, C., Vieira, J. & Messing, J. (1982) Gene 33,

103-119.8. Hanahan, D. (1983) J. Mol. Biol. 166, 557-580.9. Buzby, J. S., Porter, R. D. & Stevens, S. E. (1984) J. Bacteriol.

154, 1446-1456.10. Seifert, H. S. & Porter, R. D. (1984) Mol. Gen. Genet. 193, 269-274.11. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory,Cold Spring Harbor, NY), p. 7; pp. 107-146.

12. Dagert, M. & Erhlich, S. D. (1979) Gene 6, 23-28.13. Ito, H., Fukuda, Y., Murata, K. & Kimura, A. (1983) J. Bacteriol.

153, 163-168.14. Heffron, F., So, M. & McCarthy, B. J. (1978) Proc. Natl. Acad.

Sci. USA 75, 6012-6016.15. Jorgensen, R. A., Rothstein, S. J. & Reznikoff, W. S. (1979) Mol.

Gen. Genet. 177, 65-72.16. Norrander, J., Kempe, T. & Messing, J. (1983) Gene 26, 101-106.17. Johnston, S. & Ray, D. S. (1984) J. Mol. Biol. 177, 685-700.18. Zagursky, R. J. & Berman, M. L. (1984) Gene 27, 183-191.19. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad.

Sci. USA 74, 5463-5467.20. Lee, C.-H., Bahgwat, A., McCormick, M., Ohtsubo, E. & Heffron,

F. (1983) Mechanisms of DNA Replication and Recombination(Liss, New York).

21. Hoess, R. H., Ziese, M. & Sternberg, N. (1982) Proc. Natl. Acad.Sci. USA 79, 3398-3402.

22. Casadaban, M. J., Martinez-Arias, A., Shapira, S. K. & Chou, J.(1983) Methods Enzymol. 100, 293-308.

23. Orr-Weaver, T. L., Szostak, J. W. & Rothstein, R. J. (1981) Proc.NatI. Acad. Sci. USA 78, 6354-6358.

24. Bach, M.-L., LaCroute, F. & Botstein, D. (1979) Proc. Natl. Acad.Sci. USA 76, 386-390.

25. Broach, J. R., Strathern, J. N. & Hicks, J. B. (1979) Gene 8, 121-133.26. Heiter, P., Mann, C., Snyder, M. & Davis, R. W. (1985) Cell 40,

381-392.27. Sutcliff, J. G. (1978) Cold Spring Harbor Symp. Quant. Biol. 43,

77-90.28. Guyer, M. S., Reed, R. R., Steitz, J. A. & Low, K. B. (1977) Cold

Spring Harbor Symp. Quant. Biol. 45, 135-140.29. Guyer, M. S. (1978) J. Mol. Biol. 126, 347-365.30. Crowley, J. C. & Kaback, D. (1984) J. Bacteriol. 159, 413-417.31. Rothstein, R. J. (1983) Methods Enzymol. 101, 202-210.32. Guarente, L. & Ptashne, M. (1981) Proc. Natl. Acad. Sci. USA 78,

2199-2203.33. Rose, M., Casadaban, M. J. & Botstein, D. (1981) Proc. Natl.

Acad. Sci. USA 78, 2460-2464.34. Ullmann, A. (1984) Gene 29, 27-31.35. Boeke, J. D., LaCroute, F. & Fink, G. R. (1984) Mol. Gen. Genet.

197, 345-346.36. Seifert, H. S., So, M. & Heffron, F. (1986) Genetic Engineering,

Principles and Methods (Plenum, New York), Vol. 8, in press.

Genetics: Seifert et al.

Dow

nloa

ded

by g

uest

on

Mar

ch 1

1, 2

020