Proc. Natl. Acad. Sci. USAVol. 77, No. 12, pp. 7333-7337, December 1980Genetics

Replication of plasmids from Staphylococcus aureus inEscherichia coli

(heterospecific gene exchange/recombinant DNA/Bacillus subtilis)

A. GOZE AND S. D. EHRLICHGroupe d'Etude des Acides Nucl6iques, DMpartement de Microbiologie, Institut de Recherche en Biologie Moleculaire, 2, pl. Jussieu, Paris 75005, France

Communicated by Joshua Lederberg, August 5, 1980

ABSTRACT Plasmid pBR322 derives from plasmid ColE1and does not replicate in Escherichia coli strains lacking DNApolymerase I. Hybrids between pBR322 and a plasmid isolatedfrom Staphylococcus aureus, pC194, replicate in such E. colistrains, provided that the pC194 replication region is intact.Inactivation of the pBR322 replication region does not interferewith the replication of hybrids in E. coli. Hybrids betweenpBR322 and. two other plasmids from S. aureus, pT127 andpUB112, replicate at the restrictive temperature in E. coli hav-ing thermosensitive DNA polymerase I. Similar hybrids in-volving pC221 and pHV400, plasmids from S. aureus and Ba-cillus subtilis, res ectively, do not replicate under such condi-tions. These resu ts show that some plasmids from a Gram-positive bacterium, S. aureus, can replicate in a Gram-negativeone, E. coli.

The study of plasmids during the past 20 years has broughtforth, among others, the realization that the same replicons canfunction in many bacterial species. This ability represents afoundation for genetic exchanges among diverse species. Thescope of such heterospecific genetic exchanges is revealed bydetermining the diversity of organisms in which particularplasmids may replicate.

Organisms belonging to numerous and phylogenetically verydistant species of Gram-negative bacteria can serve as hosts forthe same plasmids (1). The evidence is so abundant that we feeljustified in generalizing this ability to almost any bacteria be-longing to the Gram-negative group. Not so many data areavailable for Gram-positive bacteria, but a similar generaliza-tion is likely to hold true within this group as well, because someplasmids replicate in species as distant as Staphylococcus aureusand Bacillus subtilis (2). Genetic exchanges between organismsbelonging to the same group, Gram-positive or Gram-negative,thus seem possible. Plasmids that could replicate in bothGram-positive and Gram-negative bacteria would link thesegroups into a unique network of potential exchangers of geneticinformation, covering most of the known prokaryotes. In thisreport we present evidence that plasmids originally isolatedfrom a Gram-positive bacterium, S. aureus, can replicate in aGram-negative bacterium, Escherichia coil.

MATERIALS AND METHODSBacterial Strains and Plasmids. The B. subtilis strain used

was HVS49 trpC2 his-2 tyr-1 aro-2 (our laboratory). All E. colistrains used were K-12 derivatives: HVC45, thrAl leu-6 thi-ltonA21 supE44 hsdR str (3); TS214, leu thy his argG nmtB lacYxyl polA214(ts) strA (4); JG1 12, thyA lacZY14 rha polAl str (J.Gross via R. Devoret); HVC293, leu thi AtrpE5 lacYl hsdR strrecAl (our laboratory); and HVC337, leu thi AtrpE5 lacYlhsdR recB21 (our laboratory). Phage X i434 int-102 red-3 was

The publication costs of this article were defrayed in part by pagecharge payment. This article must therefore be hereby marked "ad-vertisement" in accordance with 18 U. S. C. §1734 solely to indicatethis fact.

7333

obtained from R. Devoret. Plasmids used are listed in Table1.Media. All bacteria were grown in L liquid or solid media.

Resistant E. coli bacteria were selected and grown by usingampicillin (Ap) at 10 or 100 ,g/ml, tetracycline at 15 ,g/ml,or chloramphenicol (Cm) at 25 Atg/ml. Chloramphenicol-re-sistant (CmR) B. subtilis were selected and grown by usingchloramphenicol at 3 ,ug/ml and 25 ug/ml, respectively.Enzymes. EcoRI restriction endonuclease and phage T4 li-

gase were prepared and used as described (11, 12). All otherenzymes were purchased from Bethesda Research Laboratoriesand used according to instructions supplied by the manufac-turer.Gel Electrophoresis. Agarose and acrylamide gel electro-

phoresis were performed by standard methods, using Tris/EDTA/borate or acetate buffer. DNA was extracted from gelsas described (13).

Plasmid DNA Preparation. Plasmid DNA was extracted andpurified by hydroxyapatite column chromatography and ce-sium chloride density gradients in the presence of ethidiumbromide (3, 14, 15).Competence Induction and Transformation. B. subtilis and

E. coli competent cells were prepared and transformed as de-scribed (3, 16, 17).

Biohazard Considerations. The potential biohazards asso-ciated with the described experiments were examined by theFrench National Control committee; the appropriate experi-ments were performed at LI containment level.

RESULTSPlasmid pC194 was isolated from S. aureus (6) and shown toreplicate and express its chloramphenicol resistance gene in B.subtils (2). It has, in our hands, consistently failed to transformE. coil to chloramphenicol resistance. This result suggests eitherthat pC194 cannot replicate in E. coli or that the plasmid ismaintained in this host, but that its chloramphenicol resistancegene is not adequately expressed. To distinguish these possi-blities we linked pC194 to E. coli plasmid pBR322 (5), whichcarries a f3-lactamase gene conferring upon E. coli a discerniblephenotype even when present in very few copies per cell (18).The hybrids between pC194 and pBR322 were introduced inE. coli cells and their replication properties were examined.

Hybrids Between pBR322 and pC194 Replicate in E. coliDeficient in DNA Polymerase. pBR322 contains a repliconessentially identical to that of ColEl, which does not functionin cells having a low level of DNA polymerase 1 (19). We haveintroduced pBR322 and the hybrids between pBR322 andpC194, named pHV14 and pHV15 [differing in the orientationsof the two parental plasmids (Fig. 1; ref. 10)] into the E. coli

Abbreviations: Ap, ampicillin; Cm, chloramphenicol; R, resistant;MDal, megadaltons.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

7334 Genetics: Goze and Ehrlich

Table 1. Plasmids usedMolecu-

lar weightPlasmid Description X 10-6 Ref.

pBR322 Chimeric cloning vector 2.8 5pC194 Natural isolate 1.8 6pC221 Natural isolate 3.0 7pC223 Natural isolate 3.0 8pUB112 Natural isolate 3.0 8pT127 Natural isolate 2.8 9pHV400 Natural isolate 5.1 3pHV14 pBR322 + pC194* 4.6 10pHV1S pBR322 + pC194* 4.6 10pHV32 pHV14 with in vivo deletion 3.7 3pHV50 Hae II segment A (3.0 MDal) of 3.0 This work

pHV14pHV51 Hae II segment A (3.0 MDal) of 3.0 This work

pHV15pHV211 pBR322 + pC221* 5.8 This workpHV213 pBR322 + pC223* 5.8 This workpHV215 pBR322 + pUB112* 5.8 This workpHV433 Sau3A segments B (0.8 MDal) 4.7 B. Niaudet,

and D (0.2 MDal) of pT127 unpub-inserted in BamHI site of lishedpHV32

pHV426 HindIII segment B (1.8 MDal) of 5.5 3pHV400 + pHV32

pHV427 HindIII segment B (1.8 MDal) of 5.5 3pHV400 + pHV32

MDal, megadaltons.* Plasmids were joined to pBR322 at their unique HindIII sites.

polA214(ts) bacteria. At the permissive temperature (300C) thestrain carrying pBR322 was resistant to ampicillin and tetra-cycline, while the strains carrying pHV14 and pHV15 were.resistant to ampicillin and chloramphenicol, the tetracyclineresistance encoded by pBR322 being inactivated by insertion.At the restrictive temperature (420C) all the three strains ap-peared sensitive to chloramphenicol. They could not be clearlydifferentiated by their resistance to ampicillin, probably be-cause of the instability of the antibiotic at that temperature. Wecould not ascertain, therefore, that the hybrids were maintainedin E. coli cells having a low DNA polymerase I level. For this

H nd III

Ap Cm

pC194

Haell DpHV.1,2Rep

pBR322 2

IHnd IIIHoa- 11

,Hond III

Hind III

Ap 2 Rep

3 PC 194

Hoell1 . pHVIS

pBR322 2

Hind IIII Ha* 11

Ap pBR322 4 Cm pBR32T2A 4h Rep

3 pHV32A

pHVSO pHVSI2

HoiJelnl 3 PC 194 3 pC 194Hoe~ ~2 Hind III Rep HindIII Cm

FIG. 1. Representation of some of the hybrid plasmids. pBR322is represented as a thin line, pC194 as a double line. Kilobase pairsfrom the EcoRI site for pBR322 and HindIII site for pC194 are in-dicated. ini stands for the site of initiation of DNA replication ofpBR322 (20), Rep for the replication region of pC194, and pBR322Aand pC194A for the plasmids with deletions. Direction ofDNA rep-lication in pBR322 is indicated by an arrow. Only the two extreme HaeII sites are shown for pHV14 and pHV15, the nine others (situatedbetween positions 416 and 2351 of pBR322) are omitted for the sakeof clarity of the drawing.

reason we tested plasmid maintenance by culturing the cellsat the restrictive temperature without selective pressure anddetermining at the permissive temperature the proportion ofcells still resistant to antibiotics. A representative experimentis shown in Fig. 2. After 15 generations at 42°C the fraction ofampicillin-resistant (ApR) cells in the culture containing pBR322was 10-4. This is close to 3-10-5, the value expected if plasmidreplication were arrested when the culture was shifted to therestrictive temperature. When the cells carried pHV14, 1000times more ApR cells were present; when they carried pHV15,100 times more (Table 2).Two representative ApR clones from each of the three cul-

tures grown at 42°C, carrying pBR322, pHV14, and pHV15,respectively, were selected for further analysis. They containedplasmid DNA indistinguishable from the parental one by size,HindIII restriction pattern, and genetic markers. This indicatesthat the resistance to ampicillin is due to the presence of plas-mids in the resistant cells. The ApR clones were found to bepolA(ts), as judged by their UV and methyl methanesulfonatesensitivity (21) and their ability to support growth of Xred phage(22); these properties were all identical to those of the parentalstrain devoid of plasmids. The higher maintenance of pHV14and pHV15 could not thus be attributed to reversion or com-plementation of the polA214 mutation. Moreover, when therepresentative ApR clones from the three cultures were sub-mitted to a second cycle of growth at 420C, plasmid mainte-nance was identical to that observed in the first cycle, whichindicates that no enrichment of more stable mutants, eitherplasmid or chromosome-borne, had occurred during the growthat 420C.A possible explanation of higher maintenance of hybrid

plasmid than of pBR322 is that pC194 encodes a diffusibleproduct that can replace DNA polymerase I in the replicationof pBR322 and that was not detected by the tests for the poly-merase-positive phenotype that we performed. This was ruled

if)-J-JwuIxa.

I-0

Uz

a

crLaJ

5 10 15GENERATIONS AT 420C

FIG. 2. Maintenance of plasmids in the E. coli polA214(ts) bac-teria at restrictive temperature. 0, pHV14; *, pBR322.

Proc. Natl. Acad. Sci. USA 77 (1980)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

Proc. Natl. Acad. Sci. USA 77 (1980) 7335

Table 2. Maintenance of hybrids between pBR322 and pC194 inE. coli polA214(ts) at restrictive temperature

Fraction ofcells* carrying

Plasmid plasmid

pBR322 10-4pHV14 10-1pHV15 10-2pBR322 ex pHV14t 10-4pBR322 ex pHV15t 10-4pHV32 10-3

* Fifteen generations ofgrowth at 420C were allowed. An average fromseveral experiments is shown.

t See text.

out because, at 420C, the pBR322 was as unstable when intro-duced by transformation in the polA214(ts) cells carryingpHV14 as when introduced into the plasmid-free cells. Thisinstability was not due to incompatibility between pHV14 andpBR322, because in a similar experiment performed at 30'Csegregation of the two plasmids was negligible.

Another possible explanation of higher maintenance of hy-brids is that the pBR322 associated with pC194 has undergonea mutation that has rendered its replication independent of theDNA polymerase I. To test this hypothesis we have recon-structed pBR322 by excising it out of pHV14 and pHV15 thatsurvived growth at 420C for 15 generations. The reconstructedplasmids, labeled pBR322 ex pHV14 and pBR322 ex pHV15,were lost as rapidly as the authentic pBR322 from polA214(ts)cells at 420C (Table 2). This shows that the higher stability ofpHV14 and pHV15 depends on the presence of pC194 in themolecule.A further indication that the maintenance of pHV14 in the

polA214(ts) E. coli bacteria at 42°C depends on the pC194replicon is obtained from experiments with the plasmid pHV32(Fig. 1). This plasmid was obtained by deleting from pHV14the replication region of pC194 (ref. 3; unpublished). At therestrictive temperature 100 times less pHV32 than pHV14 ismaintained in polA214(ts) cells (Table 2), whereas both plas-mids are stable at 300C. We do not know the reasons for asomewhat higher maintenance of pHV32, compared topBR322, under the restrictive conditions (Table 2).

E. coli polA214 retains 16% of the wild-type DNA poly-merase I activity at nonpermissive temperature (4), whereasthe polAl strain retains less than 2% (21). If the hybrids betweenpC194 and pBR322 indeed do not require DNA polymeraseI for replication, they should be maintained not only in theformer but also in the latter bacteria. To test this point, E. colipolAl competent cells were treated with 1 ,ug of pHV14,pBR322, or pHV32 DNA. Some 100 transformants resistant toampicillin at 10 ,g/ml were obtained with pHV14 DNA andnone with the other two DNAs. Under the same conditions withpolymerase-positive competent cells the three DNAs yieldedabout 106 transformants per ,ug. Two representative ApR clonesobtained with the polAl strain were tested for UV sensitivityand Xred growth, and they were found to be indistinguishablefrom the polAl parental strain. Plasmid DNA isolated fromthese clones was identical to the parental pHV14 in size, Hin-dIII restriction pattern, and genetic markers. The yield of theplasmid DNA was consistent with the presence of one to threecopies per host cell. These results show that a pC194-pBR322hybrid can replicate in E. coli containing very low amounts ofthe DNA polymerase I.The polAl strain carrying pHV14 was resistant to ampicillin

(10 Mg/ml) but could not be clearly distinguished from theplasmid-free strain by its chloramphenicol resistance. This maybe due to the low copy number of pHV14 in this strain, and may

explain our failure to obtain CmR transformants when treatingcompetent E. coli cells with pC194 DNA (see above).pHV14 is maintained in the absence of the selective pressure

in the polAl bacteria with a stability comparable to that ob-served in the polA(ts) bacteria at 42°C. It is worth noting thatpHV14 also segregates from B. subtilis, 30-70% of the cellslosing pHV14 when grown for 15 generations without selectivepressure. The parental plasmid pC194 is stable in this host underthe same conditions. These results suggest that the insertion ofpBR322 into the HindIII site of pC194 interferes somewhatwith the maintenance of the latter, and may explain, at leastin part, the instability of pHV14 in E. coli strains containinglow levels of DNA polymerase I.Hybrids Between pBR322 and pC194 Lacking the DNA

Initiation Site of pBR322 Replicate in E. coli. The resultspresented in the preceding section indicate that the pC194replication functions are expressed in E. coli, because the hy-brids between pC194 and pBR322 are maintained in the strainsthat are unable to replicate pBR322 due to their lack of DNApolymerase I. In order to test even more stringently the repli-cation of pC194 in E. coli, we decided to delete from the hy-brids the site of initiation of pBR322 DNA replication. For thispurpose we took advantage of the fact that the Hae II sites areconveniently placed in pHV14 and pHV15. The first of the 11sites maps at the nucleotide 235 of pBR322, and the last at thenucleotide 2721 (23), whereas none is found in pC194 or in thefl-lactamase gene. Because the site of initiation of pBR322 DNAreplication maps at the nucleotide 2534 (20, 23, 24) it can beeasily eliminated by deleting from pHV14 and pHV15 theregion spanned by the two extreme Hae II sites (Fig. 1).

Plasmids with deletions, named pHV50 and pHV51 (Fig. 1),deriving from pHV14 and pHV15, respectively, were obtainedby cleaving the corresponding DNAs with Hae II, ligatingthem, and transforming B. subtilis to chloramphenicol resis-tance. The plasmids isolated from the representative B. subtilistransformants were of the expected size and released expectedsegments upon cleavage with Hae II, HindIII, and Hae IIIrestriction endonucleases (Figs. 3 and 4).

u e'jjh iroq

FIG. 3. Agarose gel electrophoresis of Hae II- and HindIII-cleaved pHV50 and pHV51 plasmid DNA. Lanes b-f, HindIII-cleavedplasmid DNAs; g-l, Hae II-cleaved plasmid DNAs; a and m-q, un-cleaved plasmid DNAs. The order of plasmid DNAs used is the fol-lowing: a and b, pC194; c, h, and m, pHV50 extracted from E. coli; d,i, and n, pHV50 extracted from B. subtilis; e, j, and o, pHV51 ex-tracted from E. coli; f, k, and p, pHV51 extracted from B. subtilis; g,1, and q, pHV14.

Genetics: Goze and Ehrlich

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

7336 Genetics: Goze and Ehrlich

a b c d e F q

Fi;. 4. AXrxvlmide -el elec-t rphoresis of Hae 111-cleavedH\O-)() and pHV-1 pla.>mid DNA.

a, pHVII 1)h. pH\ Nl extractedfrom B. sdhbtlis; c pHV50 ex-

t ra(Itted frorin E. o 1i; d. pBR3 22: e,

pHVl-.1 extracted from E. coli: t,

pH\-;l extracted from B. subtilis:VI 5.

Both pHV50 and pHV51 DNAs transformed E. coli to am-picillin resistance, with an efficiency of 103 transformants per

jug, which is about 1/1000th of that of the parental plasmids.The transformants selected on plates supplemented with am-picillin at 10 jg/ml were resistant to the drug at 100 ,ug/ml. Wecould not reproducibly distinguish their chloramphenicol re-

sistance from that of the plasmid-free recipient E. coli strain.Plasmids were extracted from representative E. coli ApR

transformants. They were indistinguishable from the pHV50and pHV51 DNAs extracted from B. subtilis by size, Hae II,

HindIII, and Hae III restriction patterns (Fig. 3 and 4), andgenetic markers in E. coli and B. subtilis. The yield of DNA was

compatible with the presence of one to three copies of theplasmid per bacterial cell. These data show that the initiationsite of pBR322 DNA replication is not necessary for replicationof hybrid plasmids in E. coil. To ascertain that the Hae IT-in-troduced deletions inactivated replication of pBR322, pHV50DNA was cleaved with the HindIII nuclease, which separatedpC194 from the deleted pBR322 (Fig. 1). The mixture was li-gated and fractionated on an agarose gel. The gel was sliced,and DNA was extracted from the slices and used to transformE. coli to ampicillin resistance. No transformants were obtainedwith the ligated segment corresponding to the deleted pBR322,while about a hundred were obtained with this segment re-

joined to pC194. Plasmid DNAs extracted from representativetransformants contained the two segments (pBR322 with a

deletion and pC194) in both possible orientations, as expectedif the cleavage and the ligation were successful. This result in-dicates that the deletion introduced in pBR322 interferes withits replication. It is consistent with the data obtained by otherauthors, showing that the region that we have deleted is indis-pensible for the replication of ColEl (20, 25-27).

Maintenance of pHV50 and pHV51 in E. coil could possiblybe due to the presence of homologous sequences on pC194 andon the E. coil chromosome. The plasmid would then be repli-cated solely when integrated in the chromosome, but wouldrecombine out of it at a frequency that would allow detectionof the plasmid DNA in the cells. This is unlikely, because wecould transform E. coil recAl (HVC293) and recB21 (HVC337)competent cells with pHV50 and pHV51 DNAs. The trans-formants which were recombination-deficient, as indicated bytheir UV sensitivity, contained plasmids indistinguishable bysize and HindIII restriction pattern from the parental pHV50and pHV51.pHV50 and pHV51 segregated from E. coil cells grown

without selective pressure. In different clones between 10-1 and0-3 cells still carried pHV50 after 15 generations. A typical

value was 3.10-2, not very different from that observed with

pHV14 and pHV15 maintained in the strains lacking DNApolymerase I.pHV14 was maintained more efficiently than pHV15 in E.

coli polA(ts) cells at the restrictive temperature (see precedingsection). No such difference was found in the experiments withplasmids pHV50 and pHV51, which were obtained frompHV14 and pHV15. A possible explanation of these results isthat a function (or functions) that interfered in some way withthe maintenance of pHV15 under restrictive conditions was

eliminated by the deletion giving rise to pHV51.Maintenance in E. coli of Hybrids Between pBR322 and

Several Plasmids Originating from S. aureus or B. subtilis.We constructed hybrids by inserting pC221, pC223, andpUB112 (7, 8) into the HindIII site of pBR322 (Table 1). Thethree former plasmids were isolated from S. aureus; they rep-licate in B. subtilis and encode resistance to chloramphenicolin both hosts (2). Our hybrid plasmids replicate in, and conferchloramphenicol resistance upon, E. coli and B. subtilis (S.aureus was not tested). This indicates that they preserve intactboth the resistance gene and the replication region active in B.subtilts. The hybrid plasmids were introduced into E. colipoiA214(ts) cells, and their maintenance at 42°C was deter-mined. A representative experiment is shown in Table 3. Thehybrid containing pUB112 was maintained at the level similarto that of pHV14, while that containing pC221 was lost as

rapidly as the pBR322 alone. An intermediate efficiency wasobserved for the hybrid containing pC223.

Similar data were obtained with another class of hybridplasmids. pHV400 is a cryptic B. subtils plasmid (3) and pT127is a tetracycline resistance-encoding plasmid originating fromS. aureus (9). No convenient restriction sites are known that canbe used to insert these two plasmids in their entirety intopBR322. In addition, because pHV400 carries no knownmarker, a hybrid between this plasmid and pBR322 could notbe easily introduced into B. subtilis to ascertain that its repli-cation region is still intact. For these reasons we have used hy-brids constructed from HindIII segments of pHV400 or Sau3Asegments of pT127 and pHV32 (Table 1), which could replicatein B. subtilis (ref. 3; B. Niaudet, personal communication).These hybrids were introduced into E. coli polA214(ts) cells,and their maintenance at 42°C was measured (Table 3).pHV426 and pHV427, which contain the replication region ofpHV400 inserted in the HindIll site of pHV32 in either of thetwo possible orientations, are lost as rapidly as pBR322. On theother hand, the hybrid pHV433, which contains the replicationregion of pT127, is maintained as efficiently as pHV14.

DISCUSSIONTwo research groups have previously reported the transfer ofplasmids between Gram-negative bacteria and B. subtilis. The

Table 3. Maintenance of hybrids between pBR322 and differentplasmids in E. coli polA214(ts) at restrictive temperaturePlasmidjoined to Hybrid Fraction ofpBR322 plasmid cells carryingor pHV32 name plasmid*

None pBR322 3.8.10-5pC194 pHV14 4.6-10-1pT127 pHV433 4.0-10-1pUB112 pHV215 1.2.10-1pC223 pHV213 6.0-10-3pC221 pHV211 1.2-10-5pHV400 pHV426 1.4-10-5pHV400 pHV427 4.7-10-6

* Fifteen generations of growth at 420C were allowed.

Proc. Natl. Acad. Sci. USA 77 (1980)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0

Proc. Natl. Acad. Sci. USA 77 (1980) 7337

evidence for heterospecific replication of these plasmids was,however, not entirely convincing. In one report the putativeB. subtilis transformants did not have the auxotrophic markerscarried by the recipient cells (28). In another, plasmids isolatedfrom the transformants differed in size from the one used totransform, and the capacity of the new plasmids to replicatein E. coli was not verified (29). The most recent work reportsphenotype change of B. subtilis recipient but lacks evidencefor the presence of the plasmid in the recipient cells (30).Two lines of evidence show that the plasmid pC194, which

was isolated from S. aureus (6) and subsequently introducedinto B. subtilis by DNA transformation, (2) can replicate in E.coli. (i) Plasmid pBR322, which needs a high level of DNApolymerase I for its replication, is maintained in E. coli strainscontaining low levels of the enzyme, provided that it is linkedto pC194. The maintenance is not due to the synthesis of apC194-encoded analogue of the DNA polymerase I and re-quires that the replicon of pC194 be intact. Similar tests werepreviously used to analyze plasmid replication regions (i.e., ref.31). (ii) pC194 linked to the pBR322 whose replication regionhas been inactivated by a deletion is maintained in E. coli.Recombination functions of the host are not involved in thismaintenance.Two other plasmids isolated from S. aureus, pT127 and

pUB1 12 (8, 9), that replicate in B. subtilis (2) can also replicatein E. coli, as judged by the ability of the hybrids between theseand pBR322 to be maintained at the restrictive temperaturein E. coli cells containing thermosensitive DNA polymerase I.Another such plasmid, pC221 (7), and the cryptic plasmidpHV400 isolated from B. subtilis (3) cannot replicate in E. coli,as indicated by a similar test. Plasmid pC223, from S. aureus(8), seems to be able to replicate to a certain extent in E. coli.pC194, pT127, and pUB112, which replicate in S. aureus,

B. subtilis, and E. coli, are likely to use at least two, and perhapsthree, different replicons, because (i) they belong to differentincompatibility groups; (ii) there is less than 10% homologybetween pT127 and pC194 whereas there is 60% homologybetween pT127 and pUB112 (32); (iii) their restriction patternsare different (refs. 2, 32, 33; unpublished data).pC194 does not replicate in E. coli as efficiently as it does in

S. aureus or B. subtilis, as indicated by the low copy number,and possibly by the segregation of the hybrid plasmids that wehave examined. Further work is needed to explain this behavior,and, more generally, the mechanism of replication of pC194,pT127, and pUB112 in E. coli. pT127 is likely to encode aprotein necessary for its replication as shown by a study of asimilar plasmid, pT181 (34). We have no evidence at presentthat would imply involvement of a plasmid-encoded proteinin the replication of pC194. If this plasmid relies solely on hostfunctions for its replication, as is the case for ColE1 (35, 36), itwould be interesting to determine the sequence of the repli-cation region recognized by three hosts as different as S. aureus,B. subtilis, and E. coli.

In conclusion, our results show that some plasmids can rep-licate in both Gram-positive and Gram-negative prokaryotes.These plasmids represent a link that joins the organisms be-longing to these two groups into a network of potential ex-changers of genetic information. Whether this network islimited to prokaryotes remains an open question.

We are grateful to M. Meunier, B. Niaudet, and the participants ofthe European Molecular Biology Organization course on GeneticEngineering (Paris, 1979) for carrying out some of the experiments,and to A. Dedieu and E. Pierre for their technical help. The authorsare on the Institut National de la Sante et de la Recherche Medicalestaff. This work was supported by grants from the D6legation GCn6ralea la Recherche Scientifique et Technique (78.7.0346), Institut National

de la Sant6 et de la Recherche M6dicale (72-79-104), and the Fondationde la Recherche M6dicale Franqaise.1. Reanney, D. (1976) Bacteriol. Rev. 40, 552-590.2. Ehrlich, S. D. (1977) Proc. Natl. Acad. Sci. USA 74, 1680-

1682.3. Niaudet, B. & Ehrlich, S. D. (1979) Plasmid 2, 48-58.4. Kingsbury, D. T. & Helinski, D. R. (1973) J. Bacteriol. 114,

1116-1124.5. Bolivar, F., Rodriguez, R. L., Greene, P. J., Betlach, M. C.,

Heyneker, H. L., Boyer, H. W., Crosa, J. H. & Falkow, S. (1977)Gene 2, 95-113.

6. Iordanescu, S. (1975) J. Bacteriol. 124, 597-601.7. Chabbert, Y. A., Baudons, J. G. & Gerbaud, G. R. (1964) Ann.

Inst. Pasteur (Paris) 107, 678-690.8. Novick, R. (1976) J. Bacteriol. 127, 1177-1187.9. Ruby, C. & Novick, R. P. (1975) Proc. Natl. Acad. Sci. USA 72,

5031-5035.10. Ehrlich, S. D. (1978) Proc. Natl. Acad. Sci. USA 75, 1433-

1436.11. Ehrlich, S. D., Bursztyn-Pettegrew, H., Stroynowski, I. &

Lederberg, J. (1976) Proc. Natl. Acad. Sci. USA 73, 4145-4149.

12. Sgaramella, V. & Ehrlich, S. D. (1978) Eur. J. Biochem. 86,531-537.

13. Harris-Warrick, R. M., Elkana, Y., Ehrlich, S. D. & Lederberg,J. (1975) Proc. Natl. Acad. Sci. USA 72,2207-2211.

14. Clewell, D. B. & Helinski, D. R. (1969) Proc. Natl. Acad. Sci. USA62, 1159-1166.

15. Colman, A., Byers, M. J., Primrose, S. B. & Lyons, A. (1978) Eur.J. Biochem. 91,303-310.

16. Anagnostopoulos, C. & Spizizen, J. (1961) J. Bacteriol. 81,741-746.

17. Dagert, M. & Ehrlich, S. D. (1979) Gene 6,23-28.18. Yasuda, S. & Hirota, Y. (1977) Proc. Natl. Acad. Sci. USA 74,

5458-5462.19. Kingsbury, D. T. & Helinski, D. R. (1970) Biochem. Biophys. Res.

Commun. 41, 1538-1544.20. Oka, A., Nomura, N., Morita, M., Sugisaki, H., Sugimoto, K. &

Takanami, M. (1979) Mol. Gen. Genet. 172, 151-159.21. De Lucia, P. & Cairns, J. (1969) Nature (London) 224, 1164-

1166.22. Zissler, J., Signer, E. & Schaeffer, F. (1971) in The Bacteriophage

lambda, ed. Hershey, A. D. (Cold Spring Harbor Laboratories,Cold Spring Harbor, NY), pp. 469-475.

23. Sutcliffe, J. G. (1979) Cold Spring Harbor Symp. Quant. Biol.43, 77-90.

24. Tomizawa, J., Ohmori, H. & Bird, R. E. (1977) Proc. Natl. Acad.Sci. USA 74, 1865-1869.

25. Backman, K., Betlach, M., Boyer, H. W. & Yanofsky, S. (1979)Cold Spring Harbor Symp. Quant. Biol. 43, 69-76.

26. Kahn, M. L., Figurski, D., Ito, L. & Helinski, D. R. (1978) ColdSpring Harbor Symp. Quant. Biol. 43,99-103.

27. Ohmori, H. & Tomizawa, J. I. (1979) Mol. Gen. Genet. 176,161-170.

28. Domaradsky, I. V., Levadnaja, T. B., Sitnikov, B. S., Rassadin,A. S. & Denisova, T. S. (1976) Dokl. Akad. Nauk. SSSR 226,1443-1445.

29. Naumov, L. S., Numerov, V. K., Bozkko, A. N., Lushnikov, I. V.,Tyurin, V. S. & Domaradsky, I. V. (1977) J. Microbiol. Epi-demiol. Immunobiol. 5,33-37.

30. Buheitel, G. & Klingmuller, W. (1979) Eur. J. Appl. Microbiol.Biotechnol. 37, 79-84.

31. Taylor, D. P. & Cohen, S. N. (1979) J. Bacteriol. 137, 92-104.32. Iordanescu, S., Surdeanu, M., Della Latta, P. & Novick, R. (1978)

Plasmid 1, 468-479.33. Wilson, C. R. & Baldwin, J. N. (1978) J. Bacteriol. 136,402-413.34. Iordanescu, S. (1979) Plasmid 2, 207-215.35. Donoghue, D. J. & Sharp, P. A. (1978) J. Bacteriol. 133, 1287-

1294.36. Kahn, M. & Helinski, D. R. (1978) Proc. Natl. Acad. Sci. USA 75,

2200-2204.

Genetics: Goze and Ehrlich

Dow

nloa

ded

by g

uest

on

Nov

embe

r 16

, 202

0