Proc. Nat. Acad. Sci. USAVol. 72, No. 6, pp. 2242-2246, June 1975

Cloning, Isolation, and Characterization of Replication Regions of ComplexPlasmid Genomes

(DNA/restriction endonuclease/incompatibility/R-plasmids/heteroduplex analysis)

KENNETH TIMMIS, FELIPE CABELLO, AND STANLEY N. COHEN

Department of Medicine, Stanford University School of Medicine, Stanford, California 94305

Communicated by Allan Campbell, April 4, 1975

ABSTRACT EcoRI endonuclease-generated DNA frag-ments carrying replication regions of the F'lac and R6-5plasmids have been cloned and isolated, using as a selec-tion vehicle a nonreplicating ampicillin-resistance DNAfragment derived from a Staphylococcus aureus plasmid.Heteroduplex analysis of the constructed plasmid chimerasand the parent replicons has localized the cloned R6-5replication region to a DNA segment between kilobasepair coordinates 1.0 and 88.0 on the R6-5 map. Physicalproximity between the plasmid replication functions andthe locus governing plasmid incompatibility has beenshown for both parent replicons. The cloning methodreported appears to be generally applicable for the identi-fication and isolation of replication regions of a variety ofcomplex genomes.

Rapid progress has taken place recently in the study of DNAsynthesis involving small replicons such as simian virus 40 (1),the bacteriophages M13 and 4X174 (1, 2), and the colicin Elplasmid (Col El) of Escherichia coli (3, 4). In contrast, bio-chemical investigations of the replication of large repliconssuch as the E. coli chromosome, the sex plasmid F, and someantibiotic resistance (R) plasmids have been hindered by thegenetic complexity and structural fragility of these DNAmolecules. The recent demonstration that recombinant plas-mids are capable of containing and utilizing at least two dis-tinct sets of replication functions (5) has made evident stillother potential difficulties in investigating the replication oflarge genomes.

In this report, we describe the isolation and characteriza-tion of replication regions of the E. coli plasmids R6-5 andF'lac using an EcoRI-generated fragment of a Staphylococcusaureus plasmid as a selection vehicle. The staphylococcalplasmid DNA fragment, which carries genetic information forpenicillin-ampicillin (Ap) resistance, and which apparentlylacks an origin of replication in its original host (R. P. Novick,personal communication), is capable of propagation in E. colionly when linked to another EcoRI DNA fragment carryingfunctions required for replication in this bacterial host (ref.6 and A. C. Y. Chang and S. N. Cohen, unpublished data).Genetic and molecular investigations of the cloned E. coliplasmid replication region fragments demonstrate a physicalproximity between plasmid replication origins, replicationgenes, and incompatibility determinants.

MATERIALS AND METHODS

Escherichia coli K-12 strain C600 (7) and nalidixic-acid re-sistant (nalr) mutants of strains CRT46 (8) and CR34 (9)have been described. Plasmid R6-5 (10) expresses resistance

to chloramphenicol (Cm), kanamycin (Km), streptomycin(Sm), and sulfonamide (Su). The pSC102 plasmid was con-structed by in vivo ligation of EcoRI-treated R6-5 DNA, con-sists of three EcoRI-generated fragments of R6-5, and ex-presses Km and Su resistance (11). Plasmid pSC101 codes forresistance to tetracycline (Tc) (9). Plasmid pSC113 wasconstructed in vitro (6) and contains the entire pSC101 plas-mid plus two EcoRI-endonuclease-generated fragments ofthe S. aureus penicillinase plasmid pI258 (12); it codes forresistance to penicillin-ampicillin and Tc. Plasmids R100-1and R192-F7 (refs. 13 and 14, kindly provided by K. Hardy)are derepressed fertility mutants of R100 and R192, and ex-press resistance to Tc, Cm, Sm, and Su. F'lac is the classicParis F' plasmid and was obtained in strain DF109 (= bromo-deoxyuridine-resistant isolate of DF87, ref. 15) from D. Frei-felder via R. P. Silver.The procedures used for conjugal transfer (16) and trans-

formation of plasmid DNA (17), radioactive labeling and iso-lation of plasmids (18), sucrose and CsCl gradient centrifuga-tion (5), construction of hybrid plasmid DNA molecules bymeans of the EcoRI restriction endonuclease (11), and agarosegel electrophoresis (5) have been described. Plasmid hetero-duplex analysis procedures have been described by Sharp et al.(19). The EcoRI restriction endonuclease was purified fromE. coli strain RY-13 according to Greene et al. (20) throughthe phosphocellulose chromatography step. E. coli DNA ligasewas generously provided by S. Panasenko, P. Modrich, andI. R. Lehman.

RESULTSIsolation of the Selection Vehicle. Chang and Cohen re-

cently described the in vitro construction of a Tc and Apresistance plasmid chimera, pSC1 13, that contains the entirepSC101 plasmid replicon plus two EcoRI-generated frag-ments of the penicillin-ampicillin resistance staphylococcalplasmid pI258 (6). Cleavage of the pSC113 plasmid chimerawith the EcoRI endonuclease, ligation of the resulting frag-ments, and transformation of E. coli with the ligated mixtureyielded another plasmid (pSC122) that also expressed resis-tance to both Tc and Ap, but which contained only one of thetwo EcoRI-generated fragments of Staphylococcus plasmidDNA originally present in pSC113 (Fig. 1A). This Ap frag-ment, which is the larger of the two staphylococcal EcoRIfragments contained in pSC113, has a buoyant density in CsClof 1.692 g/cm3 (Table 1). Because of the substantial differencein the buoyant density of the Ap fragment and the buoyantdensity of the pSC101 plasmid (p = 1.710 g/cm3) the twoDNA species can be separated easily by preparative cen-

trifugation of EcoRI-cleaved pSC122 plasmid DNA in CsClgradients (Fig. 1B).

2242

Abbreviations: Ap, ampicillin-penicillin; Cm, chloramphenicol;Km, kanamycin; Sm, streptomycin; Su, sulfonamide; Tc, tetra-cycline; CCC, covalently closed circular; kb, kilobase; M.W.,molecular weight.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 24

, 202

0

Cloning of Replication Regions 2243

31

z

0-

0 10 20

FRACTION NUMBER30

wC

CX

0

z

mzCD)-i

co

FIG. 1. Isolation of staphylococcal Ap-resistance DNA. (A)Construction of pSC122 plasmid. One microgram of pSC113DNA in 50 ,u of a solution of 20 mM Tris.HCl pH 8.0, 1 mMNa2EDTA, 10 mM MgCl2 was digested for 30 min with 1 uA ofEcoRI endonuclease prepared as previously described (5). Thenuclease was then inactivated by incubation for 5 min at 600 andthe DNA mixture was ligated by addition of (NH4)2SO4 (to 10mM), NAD (to 100 uAM), bovine serum albumin (to 100 ug/ml),and 5 units of DNA ligase. After incubation at 140 overnight themixture was used to transform cells of E. coli K12 C600. Cova-lently closed circular (CCC) plasmid DNAs prepared fromtransformant clones that expressed Ap-resistance were analyzedby electrophoresis of their EcoRI digestion products through an

0.8% agarose gel in Tris-borate buffer (5). One Ap-resistanceplasmid, designation pSC122, contained only one of the twooriginal staphylococcal EcoRI endonuclease cleavage productsof the parent pSC113 plasmid. Fig. 1A is a photograph of an

ethidium bromide-stained agarose gel containing the followingEcoRI endonuclease-cleaved plasmids: 1, pSC101; 2, pSC113;3, pSC122; and 4, Ap DNA fragment isolated as described in1B below.

(B) Isolation of Ap fragment DNA by buoyant density cen-

trifugation. Thirty-three micrograms of [3H]thymidine-labeledpSC122 CCC-DNA (specific activity = 2.5 X 103 cpm/pg) were

cleaved with EcoRI enzyme as described above, dialyzed 4 hragainst a solution of 10 mM Tris HCl, pH 8.0, 1 mM EDTA(TE buffer) to remove NP40 detergent present in the enzyme

preparation, and then mixed with TE buffer to give a final volumeof 8 ml. Solid CsCl was added (refractive index = 1.3985) andthe solution was centrifuged in a bovine serum albumin-coatedcentrifuge tube in a 50 Ti rotor at 36,000 rpm for 60 hr at 200.Fractions were collected from a hole pierced in the bottom of thetube and the radioactivity (0) in an aliquot of each fraction was

measured and the refractive index was determined. Fractionsindicated by the shaded area were pooled and dialyzed ex-

haustively against TE buffer. This material was analyzed byagarose gel electrophoresis (Fig. 1A). About 10 jg of Ap fragmentDNA was recovered.

Isolation of Replication Regions of the R6-5 and F'lac Plas-mids. Aliquots (1.5,ug) of Ap fragment DNA purified from the

FIG. 2. Analysis of R6-5-Ap and F'lac-Ap hybrid plasmids byagarose gel electrophoresis. CCC plasmid DNA preparationswere cleaved to completion with EcoRI endonuclease and sub-jected to electrophoresis through 0.8% agarose slab gels asindicated in Fig. 1. Fragments of all plasmids are numberedfrom top to bottom in the gels, as was described previously (11).(A) 1, R6-5 + pSC135; 2, pSC135; 3, R6-5; 4, PSC136; 5, Apfragment; 6, pSC139; 7, pSC102; 8, R6-5 + pSC102. (B) 1,pSC138; 2, pSC137; 3, F'lac; 4, pSC140; 5, Ap fragment; 6,pSC141. The pSC140 and pSC141 plasmids contain one or moreother fragments of F'lac in addition to the replication regionfragments.

pSC122 plasmid were separately mixed with equal amounts ofEcoRI-cleaved R6-5 or F'lac plasmid DNA and ligation andtransformation were carried out as described in Materials andMethods and Fig. 1. Covalently closed circular (CCC) plas-mid DNA samples isolated from 10 separate Ap-resistantclones obtained by transformation with the R6-5 ligationmixture were treated with EcoRI endonuclease and examinedby agarose gel electrophoresis; several representative plas-mids are shown in Fig. 2A. All plasmid DNA preparationshad in common a single EcoRI fragment of the R6-5 plasmidin addition to the Ap fragment; one of the clones (pSC136)contained a second R6-5 DNA fragment (fragment XI, ref.11) which was not essential for replication of the chimera.Because fragments II and III of R6-5 have almost identical

mobilities in agarose gels, electrophoresis of a mixture ofEcoRI-cleaved pSC135 and R6-5 DNA (Fig. 2A-1) was car-ried out to identify the R6-5 DNA fragment that enablesreplication of the Ap fragment in E. coli. Mobility measure-ments, using the other R6-5 bands present in gel 1 of Fig. 2Aas internal standards, identified the band having increasedfluorescence intensity as EcoRI fragment II of R6-5. A similarco-electrophoresis experiment indicated that R6-5 fragmentII, and not the slightly smaller fragment III as was previouslybelieved (11), is contained also in the pSC102 plasmid (Fig.2A-8). Parallel experiments indicated that a single EcoRIfragment of F'lac (fragment VI of the parent plasmid) wascommon to all of the F'lac-Ap plasmid chimeras isolated (Fig.2B); several of the chimeras contained various other F'lac

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 24

, 202

0

2244 Biochemistry: Timmis et al.

TABLE 1. Properties of plasmids

Buoyant density (g/em3) MolecularMolecular weight" Copy

Plasmid complete EcoRI fragments 8e(S) lengthb (X 106) no.d Phenotype"

pSC122 1.703 1.710, 1.692 32 15.1 i 0.5 10.0 6-8 Tcr, Apr, Tra-, SUPd..A-, AOrF'lac 1.711 85 143.3 95 0.3-0.8 Lac+, Tra+(drd)SupdA+, AO", cllrpSC138 1.702 1.709, 1.692 31 J4.2 :1: 0.5 9.4 0.6-1.0 Apr, Tra-, SUpdRZA+, AOS, 410R6-5 1.711 75 98.5 65 0.7-3.0 Cmr, Smr, Sur, Kmr, Tra+, SUPdnA +, AOrpSC135 1.705 1.712,1.692 35 17.8 i 1.2 11.8 2-5 Apr, Tra , SUPd..A+, AOr

a Sedimentation coefficient, calculated by sucrose gradient-velocity centrifugation as previously described (5).b Molecular contour length in kilobase (kb) pairs. Double-stranded DNA of phage PM2 (8.92 kb) (27) was used as an internal reference.

F'lac and R6-5 values are from refs. 19 and 28. The length of the Ap DNA fragment was determined to be 6.4 kb.¢ F'lac and R6-5 from refs. 19 and 21.d Expressed per genome equivalent, assuming a molecular weight of 2.5 X 109 for the E. coli chromosome (30). Values obtained by

centrifugation of Sarkosyl "whole" lysates to equilibrium in CsCl density gradients in the presence of ethidium bromide (5) and by sedi-mentation of Brij 58 "cleared" lysates through neutral sucrose gradients (18) were in close agreement.

e Tra, ability of the plasmid to transfer itself to recipient bacteria by conjugation; drd, derepressed transfer; SUPdnaA, ability to suppressthe dnaA mutation in bacteria grown under non-permissive conditions, presumably by integrative suppression (31); AOs/r, sensitivity/resistance of plasmid replication to sub-lethal amounts of acridine orange (32), 4JIIr sensitivity resistance to the female-specific bacter-iophage 4II.

EcoRI fragments in addition to the one carrying the replica-tion functions for the plasmid.The R6-5-Ap plasmid, pSC135 [molecular weight (M.W.)

11.8 X .106] and the F'lac-Ap plasmid pSC138 (M.W. 9.4 X106), containing the replication region fragments of the re-

spective parent plasmids were selected for further study.Table 1 lists some molecular and biological properties of theseplasmids. The chimeric molecules, unlike the parent replicons,are nonconjugative plasmids and do not express either theantibiotic resistances of R6-5 or the lactose utilization of F'lac.However, both types of chimeras have replication propertiesthat are characteristic of the parent plasmids: (a) they sup-

press the dnaA mutation in bacteria grown under nonpermis-sive conditions; (b) replication of F'lac and pSC138 is sensi-tive to acridine orange, whereas that of R6-5 and pSC135 isnot; (c) although the chimeric plasmids are much smallerthan the parent replicons from which they were derived,they are maintained by host bacteria at the same copy num-

ber as the parent plasmids.

Heteroduplex Analysis of the pSC13S5, pSC138, and pSC102Plasmids. The location of the R6-5 DNA fragment containingthe cloned replication region of this plasmid was identified byexamination of the R6-5/pSC135 heteroduplex (Fig. 3). Twopreviously reported inverted repeats on the R6-5 plasmid (21)enabled precise localization of the region of homology. Thesegment of R6-5 contained in the R6-5-Ap constructed plas-mid chimera is 11.5 4 0.7 kilobase (kb) pairs in length, and islocated between coordinates 1.0 kb and 88.0 kb on the R6-5map (Fig. 3B). It thus is contained in the resistance transferunit (RTF) region of the plasmid. The pSC135/pSC102heteroduplex (Fig. 3C) confirms agarose gel electrophoresisdata (Fig. 2) indicating that these two plasmids contain thesame replication region fragment of R6-5; the entire 11.5 kbsegment of pSC135 derived from the R6-5 parent is presentalso in pSC102. Furthermore, the pSC135/pSC102 hetero-duplex demonstrates that the three EcoRI fragments com-

prising pSC102 are not contained contiguously in the R6-5plasmid, since the reverse repeat IR2 is located 20 kb and

TABLE 2. Compatibility of hybrid plasmids with related and unrelated replicons

C600 C600 C600 C600 C600 CR34 CR34DNA Selection C600 (pSC101) (F'lac) (pSC135) (pSC102) (R6-5) (R100-1) (R192-F7)

pSC122 I 2.9 X 105 - 1.4 X 106 - 2.2 X 105 3.1 X 106 1.2 X 105 2.1 X 10OI+R - 1.4 X 105 2.6 X 105 3.0 X 105 1.0 X 105 1.7 X 106

incompatibility index - - 1.0 0.85 1.0 1.2 1.2pSC138 I 1.2 X 105 1.9 X 105 7.6 X 10- 2.5 X 103 8.4 X 103 - -

I + R 1.9 X 105 1.9 X 101 2.5 X 103 8.3 X10- -incompatibility index - 1.0 42 - 1.0 1.0pSC135 I 1.9 X 106 2.4 X 105 1.1 X 106 - 2.4 X 104 5.3 X 104 2.4 X 102 1.5 X 103

I + R - 2.0 X 10 1.1 X 106 - 3.0 X 101 1.4 X 103 6 4.8 X 101incompatibility index 1.2 1.0 - 667 38 40 31pSC102 I 2.6 X 106 2.0 X 10 - 9.8 X 103 - 3.4 X 103 1.3 X 103

I + R - 2.4 X 105 1.2 X 101 - - 9.6 X 101 3.6 X 101incompatibility index 0.8 - 817 35 36

Transformation was carried out as described (17). Recipient bacteria were selected for the expression of antibiotic resistance deter-minants carried by either incoming (I) plasmid DNA, or by both incoming and resident (I + R) plasmid DNAs. The numbers of trans-formants per yg of DNA are given in the table. The incompatibility index (given in italics) is the ratio I/(I + R); it is about 1.0 forcompatible replicons and greater than 1.0 for incompatible plasmids.

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 24

, 202

0

Cloning of Replication Regions 2245

63.9 kb from the ends of EcoRI fragment II in the completeR6-5 plasmid and is situated 1.7 kb and 7 kb from the endsof this fragment in pSC102. Thus, formation of the pSC102plasmid necessarily occurred by intracellular ligation ofseparate DNA fragments, and not simply by in vivo recir-culation of a single DNA segment containing the three EcoRIfragments. Direct examination of the pSC102/R6-5 hetero-duplex (data not shown) confirmed this interpretation.The only region of homology seen in fifteen pSC135/pSC138

heteroduplexes (e.g., Fig. 3D) is the 6.4 kb Ap fragment,indicating that the EcoRI replication region fragments ofR6-5 and F'lac contain dissimilar base sequences. This isconsistent with the observed differences in replication-associ-ated properties shown by F'lac and R6-5 (copy number,acridine orange sensitivity, compatibility; Table 1 and ref.22). The F'lac/pSC138 heteroduplex (Fig. 3E) shows a regionof homology 7.8 kb in length, which corresponds to the sizeestimated by gel electrophoresis for the EcoRI replicationregion fragment cloned from the F'lac plasmid (M.W.5 X 106).

Compatibility Studies. Although incompatibility betweenbacterial plasmids has been considered to be a property ofplasmid replication (23), direct evidence for a structuralinterrelationship between these separate functions is lacking.Using constructed Ap resistance plasmid chimeras containingthe replication regions of R6-5 or F'lac, we have investigatedcompatibility of the plasmid chimeras with those plasmidsfrom which their replication functions have been derived(Table 2). Incompatibility was studied by transformationof the R6-5-Ap plasmid chimera pSC135, or the F'lac-Applasmid chimera pSC138 into bacteria containing the parentor a plasmid related to it, and was measured by the frequencyof expression of genes carried by the incoming plasmid in thepresence or in the absence of selection for determinants car-ried by the resident plasmid.The pSC122 plasmid, which provided the Ap fragment for

these plasmid chimeras, is seen to be compatible with F'lac,R6-5, R100-1, R192-F7, and pSC102 plasmids, and hence theAp fragment does not contribute to incompatibility. In con-trast, the R6-5-Ap plasmid pSC135 is incompatible withpSC102, R6-5, and other large plasmids (R100-1, R192-F7)(Table 2) of the same incompatibility group as R6-5 (FII, ref.22, and N. Datta, personal communication). The lowest incom-patibility ratios (about 40) were observed between pSC135 (orpSC102) and the R6-5, R100-1, and R192-F7 plasmids, whichappear to contain more than one set of replication functions(11, 24, 25). The greatest incompatibility ratio (about 800)was observed between pSC135 and pSC102. The pSC138plasmid, which contains a replication region from F'lac, isentirely compatible with plasmid R6-5 and its derivativepSC102, consistent with the absence of nucleotide sequencehomology in the replication regions of the R6-5-Ap and F'lac-Ap chimeras (Fig. 3D). As expected, pSC138 is incompatiblewith the F'lac plasmid.Measurements of incompatibility, which used a standard

conjugation method (16) and which employed nonconjugativeplasmids as resident replicons, yielded data which supportedthe interpretations derived from Table 2. In the absence ofcontinued selection for antibiotic resistance determinantscarried by both plasmids, rapid segregation occurred in in-stances where a high incompatibility index was observed.However, in both transformation and conjugation experi-ments, detectable segregation of or recombination between

('Sl1b REP jIR1 (ISl)a IIR2{Sl)bI= I /MEE \ /\/\ I

2.2 1.3 1.0 88.0 55.0 44.1 28.0 27.2 2.2 1.30/98.5 24.1 21.0__ RTF - r-DETERMINANT

FIG. 3. Heteroduplex analysis of R6-5-Ap and F'lac-Aphybrid plasmids. Standard procedures for plasmid DNA hetero-duplex analysis and electron microscopy were followed (19).OX174 single-stranded and PM2 duplex DNA were used asinternal standards for molecular length measurements (19, 27).The bar in each part of the figure represents 1 kb. Arrows indicatethe junctions of single-strand (SS) and double-stranded (D)regions of the heteroduplexes. (A) Heteroduplex between pSC135and R6-5 plasmids. The region of homology is 11.5 kb in lengthand is located between 1.0 and 88.0 kb on the standard R6-5physical map [shown in (B); see refs. 28 and 29]. A large single-stranded substitution loop (SS2) containing the two invertedrepeats IR1 and IR2 represents the R6-5 segment absent in thepSC135 plasmid. The smaller substitution loop (SS1) representsthe Ap-DNA fragment contribution to pSC135. (B) Map ofR6-5 showing the location of the cloned replication region. (C)The pSC135/pSC102 heteroduplex. The nonhomologous regionscomprising the Ap-DNA fragment (SS3) and the segment ofpSC102 (SS4) containing IR2 are indicated. (D) pSC135/pSC138heteroduplex. The duplex region represents the Ap DNA frag-ment common to both plasmids. The replication region frag-ments of R6-5 (SS5) and F/lac(SS6) are indicated. (E) pSC138/F'lac heteroduplex. The region of homology is 7.8 kb, whichcorresponds to the length of EcoRI fragment VI of F'lac. Thesingle-strand substitution loops SS7 and SS8 are Ap DNAfragments and the nonhomologous segment of F'lac, respec-tively.

plasmids, as determined by examination of CCC-DNA, didnot occur in cells carrying compatible plasmid replicons.

DISCUSSION

The Ap-resistance EcoRI staphylococcal plasmid DNA frag-ment used in these experiments has particular advantagesas a probe and vehicle for selection and cloning of replicationregions in E. coli. Because its buoyant density is substan-tially different from the buoyant density of pSC101 DNA, thefragment can be prepared in large quantities by CsCl den-sity gradient equilibrium centrifugation of the EcoRI-cleavedcomposite plasmid pSC122. The fragment can be separated

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 24

, 202

0

2246 Biochemistry: Timmis et al.

from cloned E. coli plasmid replication region fragments bythe same procedure, thus permitting the isolation of largeamounts of replication region DNA for physical character-ization and in vitro studies. While the level of ampicillinresistance expressed by the staphylococcal DNA fragment inE. coli (minimum inhibitory concentration, 200 ,g/ml) issubstantially lower than the levels achieved by commonpenicillin-ampicillin resistance plasmids indigenous to andwidely distributed among E. coli (Cabello and Cohen, inpreparation), it is sufficient for the selection of replication re-gions as described here. Although earlier results, and thosepresented in Table 2, suggest that at least two separatereplication regions are located on R6-5 and related plasmids(11, 24, 25), all of the plasmid chimeras cloned in these studiesby the use of the Ap fragment selection vehicle contain aunique EcoRI fragment. This finding suggests that essentialcomponents of other R6-5 replication region(s) may be dis-tributed on separate EcoRI fragments of the R6-5 plasmid.The use of other restriction endonucleases for cloning replica-tion regions of R6-5 should permit investigation of this pos-sibility.The demonstration that a 5.2 megadalton fragment of F'lac

DNA and a 7.6 megadalton fragment of the R6-5 plasmidcarry all of the functions required for replication of a DNAfragment in E. coli indicates that the replication origin andthe replication genes of the F'lac and R6-5 plasmids are clus-tered together in a small region of the genomes of these plas-mids; similar clustering of replication origin and replicationgenes has been observed in the genome of the bacteriophage X(26).The ability to clone specific segments of complex genomes

that carry genetic information for particular biological func-tions such as DNA replication appears to be highly useful forgenetic and biochemical investigations of such functions.The current experiments report the selective cloning andstudy of EcoRI-generated replication region fragments oftwo large plasmid genomes, R6-5 and F'lac. The methodsdescribed are potentially applicable for the isolation of DNAsegments containing the replication origin and/or genes ofany complex replicon capable of functioning in bacteria andmay be useful in the study of chromosome replication. Thesequestration of replication functions onto small plasmid DNAmolecules should facilitate in vivo and in vitro investiga-tions of gene products involved in replication. With appro-priate modification, the procedure describe1 may also permitthe isolation of particular DNA segments specifying functionsinvolved in the conjugal transfer of plasmids.These studies were supported by National Institute of Allergy

and Infectious Diseases Grant Al 08619, National ScienceFoundation Grant GB-30581, and American Cancer SocietyGrant VC-139. K.T. is the recipient of a postdoctoral fellowshipfrom the Helen Hay Whitney Foundation. We thank J. Zabielskifor technical assistance during part of this investigation.

1. Kasamatsu, H. & Vinograd, J. (1974) Annu. Rev. Biochem.43, 695-719.

2. Schekman, R., Weiner, A. & Kornberg, A. (1974) Science186, 987-993.

3. Sakakibara, Y. & Tomizawa, J.-I. (1974) Proc. Nat. Acad.Sci. USA 71, 802-806.

4. Lovett, M. A., Katz, L. & Helinski, D. R. (1974) Nature251, 337-340.

5. Timmis, K., Cabello, F. & Cohen, S. N. (1974) Proc. Nat.Acad. Sci. USA 71, 4556-4560.

6. Chang, A. C. Y. & Cohen, S. N. (1974) Proc. Nat. Acad. Sci.USA 71, 1030-1034.

7. Bachman, B. J. (1972) Bacteriol. Rev. 36, 525-557.8. Hirota, Y., Mordoh, J. & Jacob, F. (1970) J. Mol. Biol. 53,

369-387.9. Cohen, S. N. & Chang, A. C. Y. (1973) Proc. Nat. Acad.

Sci. USA 70, 1293-1297.10. Silver, R. P. & Cohen, S. N. (1972) J. Bacteriol. 110, 1082-

1088.11. Cohen, S. N., Chang, A. C. Y., Boyer, H. W. & Helling,

R. B. (1973) Proc. Nat. Acad. Sci. USA 70, 3240-3244.12. Peyru, G., Wexler, L. F. & Novick, R. P. (1969) J.

Bacteriol. 98, 215-221.13. Nishimura, Y., Ishibashi, M., Meynell, E. & Hirota, Y.

(1967) J. Gen. Microbiol. 49, 89-98.14. Meynell, E. & Cooke, M. (1969) Genet. Res. 14, 309-313.15. Freifelder, D. R. & Freifelder, D. (1968) J. Mol. Biol. 32,

15-23.16. Timmis, K. (1972) J. Bacteriol. 109, 12-20.17. Cohen, S. N., Chang, A. C. Y. & Hsu, L. (1972) Proc. Nat.

Acad. Sci. USA 69, 2110-2114.18. Clewell, D. B. & Helinski, D. R. (1969) Proc. Nat. Acad.

Sci. USA 62, 1159-1166.19. Sharp, P. A., Hsu, M.-T., Ohtsubo, E. & Davidson, N.

(1972) J. Mol. Biol. 71, 471-497.20. Greene, P. J., Betlach, M. D., Goodman, H. M. & Boyer, H.

W. (1974) in Methods in Molecular Biology, ed. Wickner,R. B. (Marcel Dekker, Inc. New York), Vol. 9, pp. 87-103.

21. Sharp, P. A., Cohen, S. N. & Davidson, N. (1973) J. Mol.Biol. 75, 235-255.

22. Willetts, N. (1972) Anau. Rev. Genet. 6, 257-268.23. Jacob, F., Brenner, S. & Cuzin, F. (1963) Cold Spring

Harbor Symp. Quant. Biol. 28, 329-348.24. Yoshikawa, M. (1974) J. Bacteriol. 118, 1123-1131.25. Rownd, R. H., Perlman, D. & Goto, N. (1974) in Micro-

biology-1974, ed. Schlessinger, D. (American Society forMicrobiology), pp. 76-94.

26. Stevens, W. F., Adhya, S. & Szybalski, W. (1970) in TheBacteriophage Lambda, ed. Hershey, A. D. (Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.), pp. 515-533.

27. Espejo, R. T., Canelo, E. S. & Sinsheimer, R. L. (1969)Proc. Nat. Acad. Sci. USA 63, 1164-1168.

28. Hu, S., Ohtsubo, E., Davidson, N. & Saedler, H. (1975) J.Bacteriol., 122, 764-775.

29. Ptashne, K. & Cohen, S. N. (1975) J. Bacteriol., 122, 776-781.

30. Cooper, S. & Helmstetter, C. (1968) J. Mol. Biol. 31, 519-540.

31. Nishimura, Y., Caro, L., Berg, C. M. & Hirota, Y. (1971)J. Mol. Biol. 55, 441-456.

32. Hirota, Y. (1960) Proc. Nat. Acad. Sci. USA 46, 57-60.

Proc. Nat. Acad. Sci. USA 72 (1975)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 24

, 202

0