Replication in Escherichia co/i of Three Small Plasmids Derived from R Factor RI2
E. OHTSUBO, J. FEINGOLD, H. OHTSUBO, S. MICKEL, Department
of Microbiology, Health Sciences Center, State University at Stony Brook, Stony Brook, New York 11794
W. BAUER of New York
Replicating molecules of three small plasmids, pSM1, pSM2, and pSM3, were isolated from a CsCl density gradient containing ethidium bromide. These plasmids are all derived from R12, a mutant of NRf (same as RIOO). By means of pulse-labeling experiments, the replicating forms were located at buoyant densities intermediate between those of the closed circular and open circular DNA bands. These molecules were analyzed by electron microscopy following digestion with restriction endonucleases. Digestion of pSM2 with EcoRl and with Hind111 revealed the presence of a single origin of replication located 1.72 kilobases (kb) from the EroRl cutting site (2.04 kb from the Hind111 cutting site). These experiments also demonstrated that replication occurs in a unidirectional mode from the origin. Analysis of EcoRl-cleaved replicating molecules of pSMl and pSM3, which carry common sequences completely or partly homologous to pSM2. provides further evidence for the unidirectional replication of these plasmids from a common origin. The site of the origin of replication was fixed at 85.5 on the kilobase map of RlOO. This origin, which is located in the RTF region, probably corresponds to one of the replication origins of RlOO.
Bacterial sex factor RlOO (also called or 222) carries genes conferring resistance to several antibiotics, including tetracycline (ret), chloramphenicol (cat), streptomycin (srr), sulfonamide (~441) (Nakaya et al., 1960; Watanabe and Fukasawa, 1960), mercuric ions, and fusidic acid (Datta et al., 1974). The genome is apparently composed of two different replicons, the RTF (resistance transfer factor) and the r determinant, the latter of which contains all resistance genes except for tel (see review by Rownd er al., 1974).The physical structure of RlOO has been analyzed by the electron microscope heteroduplex method to localize resistance regions using a coordinate system with kilobases (kb) as a unit (Sharp et al., 1973; Hu et al., 1975; Ptashne and Cohen, 1975; Mickel et al., 1977). Perlman and Rownd (1976) have recently reported the replication mode of RlOO. They concluded from their denaturation mapping studies on the replicating moleNRl
cules isolated from a Proteus rnirabilis strain that replication proceeds either unidirectionally or bidirectionally from either of two replication origins which are located on the RTF and r-determinant regions, respectively. Mickel and Bauer (1976) have recently reported that R12, a mutant derivative of RlOO, gives rise to various small plasmids carrying no known drug resistances. These plasmids were found to contain a limited region of the RTF, including a replication origin and the genes necessary for replication of the RTF, as well as an IS1 (Insertion Sequence 1) region adjacent to the r determinant. In the present communication we report the results of a study of the replicating DNA molecules of three related small plasmids (pSM1, pSM2, and pSM3) harbored in an Escherichia coli K12 strain. Electron microscopic analysis is employed to ascertain the mode of replication and to map accurately the replication origin on
8 ISSN 0147-619X
the kilobase map of RlOO. We present evidence which indicates that replication occurs unidirectionally from a common origin for all three plasmids. This origin is located at the same distance from an EcoRl cutting site in all cases. The unidirectional mode of replication observed for these plasmids (which are derived from a sex factor) is similar to that observed in other nontransmissable plasmids such as ColEl (Tomizawa YI al., 1974: Inselburg, 1974; Lovett et al., 1974) and pSClO1 (Cabello et al., 1976), but is different from the bidirectional mechanism proposed for another sex factor, R6K (Lovett et al., 1975). MATERIALS
Bacterial plasmids. The plasmids examined were pSMI, pSM2, and pSM3, all derived from RI2 (Mickel and Bauer, 1976). The physical structures of these plasmids have been analyzed (Mickel et al., 1977) and are interpreted under Results. These plasmids are contained in an E. co/i strain, W677 (F- thr- leu- thi- lac- galmal- syl-). Thymine-requiring mutants of the strains were isolated using the technique described by Miller (1972) and were used for the isolation of replicating molecules of each plasmid. An E. co/i strain carrying pNT1, a derivative of ColE 1 plasmid (Tomizawa et al. , 1977), was kindly sent by Dr. J. Tomizawa. This plasmid DNA was used for the determination of molecular lengths of the fragments in a gel. Isolation qf closed circular plasmid DNA. Bacteria were grown overnight in penassay broth (Difco antibiotic medium 3) to stationary phase. Cleared lysate containing plasmid DNA was then prepared by the method outlined by Clewell and Helinski (1970). Closed circular DNA was then purified in a CsCl-ethidium bromide (EtdBr) density gradient. EtdBr was removed from the plasmid DNA by repeated extraction with isopropanol saturated with CsCl solution (1.50 g/ml) in TE buffer
OF pSM PLASMIDS
(pH 7.2) containing 0.01 M Tris, 0.001 M EDTA. The DNA solution so obtained was then dialyzed overnight against TE buffer at 4°C. Restriction enzyme treatment. The digestion with EcoRl (Miles Laboratories) was carried out in SO-p1solutions by the method of Greene et al. (1974). The reaction mixture contained, as final concentrations, 100 mM Tris-HCl, pH 7.6, 10 mM MgCl,, 50 mM NaCl, and appropriate volumes of enzyme and DNA. The samples were incubated at 37°C for 1 h. To stop digestion, a ~-PI aliquot of 0.4 M EDTA was added to each sample. Hind111 (Miles Laboratories) was assayed as recommended by Miles Labs. The reaction mixture contained, as final concentrations, 6 mM Tris-HCI, pH 7.5, 6 mM MgC&, 50 mM NaCI, 5 pg of BSA, and appropriate volumes of enzyme and DNA. The samples were incubated for 0.5 h at 37°C. Digestion was stopped by the addition of 5 ~1 of 0.4 M EDTA to the sample. HaelI was kindly supplied by Dr. H. Ohmori of the National Institutes of Health. For the digestion with HaeIII the reaction mixture contained, as final concentrations, 6 mM Tris-HCl, pH 7.9, 6 mM MgC&, 6 mM /3-mercaptoethanol, and 500 &ml of BSA. Gel electrophoresis. A slab gel (15 x 13 x 0.25 cm) was used for all analytical gel electrophoresis experiments. A suspension containing 1% agarose (w/v) was solubilized in electrophoresis (E) buffer (prepared from a 10x concentrated stock solution with a final concentration of 40 mM Tris-acetate, 50 mM Na-acetate, 5 mM EDTA, pH 7.9) by autoclaving for about 3 min. Gels were poured when the agarose solutions had cooled to 60°C. Acrylamide gels, 4% (w/v), were prepared in 75 ml of a solution containing 7.5 ml of the 10x concentrated E buffer, 10 ml of 3O:l acrylamide-bisacrylamide [30% (w/v) solution], 37.5 ~1 of N, N, N’, N’-tetramethylenediamine, 0.75 ml of 10% ammonium persulfate solution, and 56.75 ml of distilled water.
OHTSUBO ET AL.
Bromphenol blue and glycerol (to 20%, w/v) were added to give a final sample volume of 25 ~1. The sample was run into the gel at 100 V. After about 4 h at room temperature, the dye marker had moved approximately 80% down the gel. Gels were visualized by staining in electrophoresis buffer containing EtdBr (4 &ml). After about 30 min the stained bands were visualized by fluorescence produced by short-wave ultraviolet light illumination (CS1 Transilluminator, Ultraviolet Products, San Gabriel, California). Gels were photographed using a Tiffen Photar 23A filter, and measurements were taken from the prints (Polaroid 55 P/N).
lysozyme (in 0.05 M Tris-HCL), incubation for 5 min, addition of 0.4 ml of 0.25 M EDTA (pH 8.0), incubation for a further 5 min, addition of 1.6 ml of 10% Triton X-100 (in 0.05 M Tris-HCl, 0.063 M EDTA), and a final incubation for 10 min. The lysate was spun down in a Sorvall SS-34 rotor at 15 Krpm for 15 min and 2°C (volume preadjusted to 4.5 ml with TE buffer). The supernatant was centrifuged to equilibrium in a CsCl-EtdBr density gradient (density adjusted to 1.580 g/ml) in a type 40 rotor at 36 Krpm for 36-48 h at 20°C. Fifty fractions were collected from the bottom of the tube, approximately five drops per fraction. The fractions were prepared for scinDetection and isolation of replicative intermediates by a double-labeling experitillation counting by the addition of ment. The methods used for detection and 0.25 ml of 2 N NaOH and 0.1 ml of 500 isolation of replicating molecules of small pg/ml of carrier DNA and incubation at plasmids were similar to those described 37°C overnight or at 80°C for 2 h. Subby Lovett et al. (1975), Crosa et al. (1975), sequently, 1 ml of 1 N HCl and 0.3 ml of and Cabello et al. (1976). The thymine- 50% TCA were added and the fractions were requiring mutants of each strain were grown left at 0°C for 30 min. The fractions were with aeration at 37°C in 50 ml of M9 medium then filtered through a GF/C 2.4-mm containing 2 mg/ml of thiamine, 0.5% filter. The filter was washed two or three casamino acids, 4 pg/ml of thymine, and times with 5 ml of 1% TCA and then 0.5 $IJi/ml of [14C]thymine until the ASSO with 5 ml of 3:l EtOH-ether solution. The was 0.4 (43.2 Klett units). The cells were filters were dried in the oven for 30 min. harvested by filtration through a 0.45~pm Each filter was then placed into a vial Millipore filter. The filter was washed with containing 4 ml of scintillation fluid (toluene 50 ml of M9 medium, prewarmed to 37”C, and POPOP) and the fractions were counted complete except for thymine. The cells were for 5 min each in a Beckman scintillation resuspended in 50 ml of the thymine-free counter. Culture conditions for plasmids in elecmedium and were then starved for thymine for 30 min at 37°C with aeration. The tron microscope studies. Unlabeled M9 doubling time was approximately 40 min medium (500 ml) was prepared as just under these conditions. Cultures were next described for the labeled medium and cells cooled to 20°C and pulsed with [3H]thy- were cultured as described preceding DNA midine (50 &i/O.3 &ml) for 15 s at 2O”C, isolation. Electron microscope analysis. DNA molfollowed by plunging into a dry ice-ethanol bath and the addition of KCN to a final ecules digested with restriction enzymes concentration of 20 mM. The cells were were treated with phenol saturated with washed twice with TE-KCN (10 mM TE buffer. Residual phenol in the DNA Tris-HCl, pH 8.0, 1 mM EDTA, 20 mM solution was removed by ether extraction, KCN). Cell lysis was then carried out at following which the DNA was precipitated 0°C by the successive addition of 1 ml of with ethanol. The precipitation step was repeated twice to remove residual salts 25% sucrose solution (in 0.25 M Tris-HCl, 20 mM KCN) and 0.4 ml of 5 mg/ml of contained in the DNA solution. These DNA
OF pSM PLASMIDS
(1977). The sequence relationships among these plasmids and their parental plasmid, R12, have been determined by the electron microscope heteroduplex method (Mickel and Bauer, 1976; Mickel et al., 1977). R12 has the same structure as RlOO. All of the same plasmids are homologous to a portion of R12 which is localized in the RTF region. The small plasmids have a common EcoRl cutting site which has been mapped precisely at 87.2 in the RIO0 coordinate system as shown in Fig. 1 (Mickel et al., 1977). One of the Hind111 sites on RlOO has been mapped in the region covered by these pSM plasmids (H. Ohtsubo and E. Ohtsubo, unpublished results). This Hind111 site was fixed at 87.6 in the RlOO coordinate system (see Fig. 1) by gel electrophoresis analysis RESULTS using the following procedure. Closed circular DNA of pSM2 and pSM3 generates Structures of Plasmids pSM1, pSM2, and a single linear DNA fragment upon digestion pSM3 and Locations of EcoRl and with Hind111 (Fig. 2A, d and e). The pSM1 HindHI Cleavage Sites DNA, however, was not cleaved by Hind111 The physical structures of the small (Fig. 2A, a and b). suggesting that the plasmids used in these experiments are Hind111 cutting site is located in the shown schematically in Fig. 1. The plasmid 87.4-88.6 region of RIOO. As shown in coordinates in this study are expressed in Fig. 2A, f and g, sequential digestion with terms of RlOO coordinates, with the origin/ EcoRl and Hind111 of either pSM2 or terminus at the (ISl)[, region as shown in pSM3 generated two DNA bands, the Fig. 1 and as discussed by Mickel et al. smaller of which was of the same length
samples were dissolved in a solution containing a 0.5 M NH,COOCH, (pH 7.2) and 0.01 M EDTA and used for electron microscopy. DNA samples for electron microscope analysis were prepared by use of the aqueous spreading technique (Davis et al., 1971). EcoRl-cleaved molecules containing replication eyes and forks were photographed and measured and fractional lengths of replicated and unreplicated segments were calculated. Replicating molecules displaying whiskers (Delius et al., 1971), which were assumed to have arisen by the process of branch migration during the isolation period, were rarely observed. These molecules were not counted in the present experiments.
RTF RI2 = RIO0
IT1 87.6 , 88.6; 89.3/O
; 87.iia i:-
82.0: /\ EcoRl site
ISi 4 )20.0
deleted in pSM1
FIG. 1. Physical structures of pSM1, pSM2, and pSM3 plasmids derived from RI2 (=RlOO). The molecules are actually circular duplexes, but are displayed in a linear representation. RIO0 coordinates are distances in kilobases from an end point of an IS1 sequence as shown. However, only the relevant portion of RIO0 is shown. The coordinates of the pSM plasmids are expressed in terms of RlOO to show sequence relations among them. Note that pSM plasmids carry a segment of RlOO which is located in the RTF region. For locations of restriction enzyme cutting sites and origin of replication (on’), see the text.
OHTSUBO ET AL
A a bcde
FIG. 2. Electrophoresis of DNA in (A) 1% agarose and (B) 4% polyacrylamide gels. (b) Undigested pSM1 DNA. The dense band is closed circular DNA, whereas the faint band is open circular DNA. (a) pSM1 DNA treated with Hind111 (which does not cut pSM1 DNA). (c) pSM1 DNA digested with EcoRl. (d) pSM3 DNA digested with HindHI. (e and h) pSM2 DNA digested with HindIII. (f) pSM3 DNA digested with Hind111 and EcoRl. (g and i) pSM2 DNA digested with Hind111 and EcoRl. (j) mini ColEl DNA pNT1 digested with ffaeII1 (molecular length standard fragments in bases determined by H. Ohmori and J. Tomizawa, personal communication). Arrows indicate the position of a small DNA fragment of 380 bases in length, which can only be generated by double digestion of pSM2 and pSM3 with Hind111 and EcoRl.
from both plasmids. The size of this common fragment was determined in a 4% polyacrylamide gel to be 0.38 kb (Fig. 2B). The Hind111 cutting site on pSM plasmids was therefore located at 87.6 in the RlOO coordinate system. As will be described below, these restriction enzyme cutting sites are used to map the replication origin in the three small plasmids at 85.5 in the RlOO coordinate system. Isolation of Replicating of Plasmid pSM2
A thymine-requiring mutant of E. co/i W677 carrying pSM2 was prelabeled with [14C]thymine and pulse labeled with [3H]thymidine as detailed under Materials and Methods. The pulse-labeled DNA exhibited a broad density distribution between the positions of supercoiled (lower fractions)
and open circular (upper fractions) DNA (Fig. 3). The 3H/14Cratio is greatest in the fractions intermediate between those of closed and open circular DNA. It has been reported using plasmids ColEl, pSClO1, R6K, and RSF1040 that these pulse-labeled DNA molecules are replicating forms (Lovett et al., 1974, 1975; Cabello et al., 1976; Crosa et al., 1975). Electron microscope studies of this intermediate fraction, described below, show molecules with “eyes” and “forks.” indicating that replicating molecules are present in this fraction. Cleavage Analysis of Replicating Molecules of pSM2 by the EcoRl Hindttt Endonucleases
In this section we show that one of the plasmids, pSM2, replicates unidirectionally
FIG. 3. (A) Equilibrium CsCI-EtdBr density gradient centrifugation of pSM2 DNA prepared as described under Materials and Methods. Distributions of 14C radioactivity (uniformly prelabeled DNA) and 3H radioactivity (pulse-labeled DNA), as indicated by solid circles and by open circles, respectively. (B) The ratio of SH to IF. Note that the pulse-labeled material is found predominantly between the closed circular and open circular positions.
OF pSM PLASMIDS
molecules, as shown in Fig. 5, are represented in Figs. 6 and 7. We conclude that unidirectional replication of plasmid pSM2 from an origin, as represented in Fig. 3, occurs unequivocally. The average calculated from the measurement of 49 replicating molecules of pSM2 shows that the replication origin is located at 25.8 + 2.7% (1.72 2 0.18 kb) of the molecular length from the EcaRl cutting site (Fig. 6). The distance between the origin and the Hind111 cutting site is 30.7 ? 3.7% (2.04 + 0.25 kb) of the total length of the molecule (Fig. 7). The difference between these two numbers is 4.9 r 1.0% (0.32 + 0.07 kb). This value is consistent with the previously established distance, 5.8% (0.38 kb), between the Hind111 and EcoRl cutting sites (see Fig. 1). Figure 8 shows an alternative representation of replicating molecules of pSM2 digested with EcoRl. It may be seen that for molecules less than about 25% replicated, one unreplicated arm remains constant in size, whereas the other decreases progressively in length. The results presented in this section show that replication occurs in a unidirectional mode to the right (as depicted in our figures) of the genome.
from the origin. The evidence is based upon analysis of replicating molecules of pSM2 by cleavage with restriction endonucleases EcoRl and HindIII, each of which cleaves pSM2 once (see Fig. 1). Figure 4 presents a schematic representation of the expected results assuming that the plasmid replicates unidirectionally from a fixed origin. Digestion of replicating molecules would generate two types of molecules which can be observed in the electron microscope: molecules with a “bubble” lying between the two linear ends and molecules with two pairs of branches connected by a DNA segment, with members of each pair equal in length. For a plasmid which replicates unidirectionally from an origin, the distance from the fork, which remains at the origin, to the enzyme cutting site must always be the FIG. 4. Schematic diagram of the analysis of replicatsame. These results, obtained by measuring ing molecules of pSM2 digested with restriction electron micrographs of the replicating enzymes.
OHTSUBO ET AL.
FIG. 5. Electron micrographs of replicating DNA molecules treated with endonucleases. a, b. and c, pSM2 DNA treated with EcoRl: d and e, pSM2 DNA treated with HindHI; f and g, pSMl DNA treated with EcoRl: h, pSM3 DNA treated with EcoRl.
Cleavage Analysis of Replicating Molecules of pSM1 and pSM3 by EcoRl Endonuclease
Data similar to those presented above for pSM2 were obtained with replicating
molecules of pSM1 and pSM3. These experiments establish that replication in these two small plasmids is also unidirectional and proceeds from an origin that is common to both plasmids as well as to pSM2. Evidence was obtained from the
OF pSM PLASMIDS
were observed in the electron microscope for these two plasmids that are of the same type as for pSM2 (see Fig. 5). Figures 9 and 10 depict the representation of the replicating molecules of pSM1 and pSM3, respectively. The results indicate that these plasmids replicate unidirectionally from a corresponding origin. To compare the results with the three plasmids on a more direct basis, we have represented the molecules in Figs. 9 and 10 on a scale based on the ratio of the lengths of pSM1 and pSM3 to that of pSM2. The __--ratio of pSM1 to pSM2 is 0.82 and that of pSM3 to pSM2 is 1.10. Applying this normalization factor to our results for --pSM1, the distance from the EcoRl site is 27.2 k 2.2% (1.81 t-O.15 kb). For pSM3, the result after normalization is 26.4 4 2.5% (1.75 4 0.16 kb). These values are quite FIG. 6. Line diagrams showing the location of the consistent with the corresponding value, replicating portion (heavy lines) in the molecules of 25.8 k 2.7% (1.72 k 0.18 kb), for pSM2.
pSM2 which were treated with endonuclease EcoRl. All of the molecules were normalized to a scale of 100%. These data show that for all molecules one branch point may be aligned on a verticle line (25.8 2 2.7%. or 1.72 2 0.18 kb, from the end), whereas the location of the other branch point varies. At the top the map of pSM2 is shown (expressed in terms of RIO0 coordinates). The arrangement of features of the map including the origin of replication (on’) was based on the results above and obtained from the analysis of replicating molecules of pSM2 digested with Hind111 and of those of pSMl and pSM3 digested with EcoRl (see the text).
The evidence presented here shows that the small plasmids, which carry part of the RTF region of R12 (same as NRl or RlOO in their physical structure), replicate unidirectionally from a common origin. The mode of replication of these plasmids is similar to that observed for other small plasmids, such as ColEI (Tomizawa et al., 1974; Inselburg, 1974; Lovett cr al., 1974) analysis of replicating DNA molecules of and pSClO1 (Cabello et al., 1976), and pSM1 and pSM3 by cleavage with the differs from the bidirectional replication EcoRl endonuclease, which cuts each of observed for the rather large sex factor these plasmids once (Fig. 1). Molecules plasmid R6K (Lovett et al., 1975). 87.6 88.6 HhdLl 54te pSM2 1
1 ISI 7
OF LOOP P/o)
FIG. 7. Line diagrams of replicating pSM2 as in Fig. 6 except that digestion was with endonuclease HindIII. The common branch point was calculated to be at 30.7 f 3.7% (2.04 t 0.25 kb) from the right end.
OHTSUBO ET AL.
FIG. 8. A plot of the fractional lengths of replicated and unreplicated segments of EcoRl-cleaved molecules of pSM2 shown in Fig. 6. The graph indicates that one arm replicates completely before the second arm begins replicating. These results are indicative of unidirectional replication.
The unidirectional mode of replication in the pSM plasmids is different from that observed for NRZ (Perlman and Rownd, 1976), which has been reported to replicate either unidirectionally or bidirectionally. This difference may be due to several factors. (1) The denaturation mapping technique employed by Perlman and Rownd is rather imprecise compared with the use of restriction enzymes, since the reference point for the mapping of sites is a substitution loop of denatured single strands rather than a specific point generated by digestion with restriction enzymes. In this technique decisions involving the symmetry and alignment must be made, since mole-
cules are circular and the extent and regions of denaturation differ among the various molecules. (2) Since pSM plasmids are derived from R12, a replication mutant of NRl with a threefold increase in the number of DNA copies in exponentially growing E. cofi or P. mirabilis (Morris et al., 1974), it is possible that the pSM plasmids have a mutated origin of replication causing replication to proceed in a manner different from that of NRI. (3) NRl was studied in P. mirabilis, whereas the pSM plasmids were studied in E. coli K12. It is possible that host-specific genes which participate in the plasmid replication process can recognize different plasmid origin or termination
! I 82
FIG. 9. Line diagrams showing the location of the replicated portion (heavy lines) in the molecules of pSMl treated with EcoRl. The molecules were normalized to 82.8% to be directly comparable to the analysis of replicating pSM2 molecules (see Fig. 6). The common branch point for the molecules was calculated to be at 27.2 -t- 2.2% (1.81 c 0.15 kb) from the end. (The length of pSM2 represents 1004.) At the top the map oi pSM1 is shown.
07.2 ECORI site pSM3 1
88.6 89.3/O/82.0 I IS1 / / m /
87.2 I EC&l site 1
OF pSM PLASMIDS
FIG. 10. Line diagrams showing the location of the replicated portion (heavy lines) in the molecules of pSM3 treated with EcoRl. All of the molecules were normalized to 110% to be directly comparable with the analysis of replicating molecules of pSM2 (see Fig. 6). The common branch point (expressed in terms of pSM2) was calculated to be at 26.4 t 2.5% (1.75 ? 0.16 kb) from the end. The map of pSM3 is shown at the top.
regions. (4) pSM plasmids carry only a portion of R12, and it is possible that the mode of replication of the small plasmids is affected by the deletion of a large part of the entire R-factor genome. This interpretation is supported by the case of a deletion derivative, RSF1040, of the rather large sex factor R6K. The former plasmid replicates from one or the other or two origins either sequentially or simultaneously (Crosa er al., 1975). At present no data are available to permit a choice to be made among these alternatives. ACKNOWLEDGMENTS We thank Drs. H. Ohmori and J. Tomizawa for sending a bacterial strain and a restriction enzyme. This investigation was supported by U. S. Public Health Service Research Grant GM-22007 (to E. 0.) and GM-21176 (to W. B.). S. M. was a postdoctoral trainee of the USPHS, Grants Nos. CA-05243 and CA-09121. H. 0. received partial support from the USPHS on Training Grant CA-09176.
REFERENCES CABELLO, F., TIMMIS, K., AND COHEN, S. N. (1976).
Replication control in a composite plasmid constructed by in vitro linkage of two distinct replicons. Nature (London) 259, 285-290. CLEWELL, D. B., AND HELINSKI,
D. R. (1970). Properties of a deoxyribonucleic acid-protein relaxation complex and strand specificity of the relaxation event. Biochemist? 9, 4428-4440. CROSA, J. H., LLJTTROPP, L. K., HEFFRON, F., AND FALKOW, S. (1975). Two replication initiation sites on R-plasmid DNA. Mol. Gen. Genet. 140, 39-50. Datta, N., Hedges, R. W., Becker, D., AND Davies.
J. (1974). Plasmid determined fusidic acid resistance in the Enterobacteriaceae. J. Gen. Microbial. 83, 191-l%. DAVIS, R. W., SIMON, M., AND DAVIDSON, N. (1971). Electron microscope heteroduplex methods for mapping regions of base sequence homology in nucleic acids. In “Methods in Enzymology” (L. Grossman and K. Muldave, eds.), Vol. 21, p. 413. Academic Press, New York.
DErrus H, HoWE c., AND KozrNsKr A, w 9 %
(1971). Structure of the replicating DNA from bacteriophage T4. Proc. Nat. Acad. Sci. USA 68, 3049-3053.
GREENE, P. J., BETLACH, M. C., BOYER, H. W., AND GOODMAN, H. M. (1974). The EcoRl restric-
tion endonuclease. Itr “Methods in Molecular Biology” (R. B. Wickner. ed.), Vol. VII, pp. 87-111. Marcel Dekker, New York. Hu, S., OHTSLJBO, E., DAVIDSON, N., AND SAEDLER, H. (1975). Electron microscope heteroduplex studies of sequence relations among bacterial plasmids: Identification and mapping of the insertion sequences IS1 and IS2 in F and R plasmids. J. Bacterial. 122, 764-775. INSELBLJRG. J. (1974). Replication of ColEl plasmid
DNA in minicells from a unique replication initiation site. Proc. Nat. Acad. Sci. USA 71, 22562259.
M. A., KATZ, L., AND HELINSKI, D. (1974). Unidirectional replication of plasmid ColE 1 DNA. Nature (London) 251, 337-340. LOVETT, M. A., SPARKS, R. B., AND HELINSKI, D. R. (1975). Bidirectional replication of plasmid R6K DNA in Escherichia coli; correspondence between origin of replication and position of single-strand break in relaxed complex. Proc. Nar.
Acad. Sri. USA 72, 2905-2909. MICKEL, S., AND BAUER, W. (1976). Isolation by
tetracycline selection of small plasmids from the R factor R12 in Escherichia coli. J. Bacterial. 127,644-655.
MICKEL, S., OHTSLJBO, E., AND BAUER, W. (1977). Heteroduplex mapping of small plasmids derived from R-factor R12: Irr rpivo recombination occurs at IS1 insertion sequences. Gene, in press. MILLER, J. H. (1972). Selection of thy- strains with trimethoprim. In “Experiments in Molecular Genetics” (Miller, J. H.. ed.), pp. 218-220. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. MORRIS, C. F., HASHIMOTO, H., MICKEL, S., AND ROWND, R. (1974). Round of replication mutant of a drug resistance factor. .I. Bacterial. 118, 85S-
866. NAKAYA, (1960).
R., NAKAMURA. A., AND MURATA, Y. Resistance transfer agents in Shigella.
Biochem. Biophys. Res. Commun. 3. 654-659. PERLMAN, D.. AND ROWND, R. H. (1976). Two origins of replication in composite R plasmid DNA. Nature (London) 259, 281-284. PTASHNE, K., AND COHEN, S. N. (1975). Occurrence of insertion sequence (IS) regions on plasmid deoxyribonucleic acid as direct and inverted nucleotide duplications. J. Bacterial. 122, 776-781.
ET AL. ROWND, R. H., PERLMAN, D., AND GOTO, N. (1974). Structure and replication of R-factor deoxyribonucleic acid in Proteus mirabilis. In “First ASM Conference on Extrachromosomal Elements in Bacteria” (D. Schlessinger, ed.), pp. 76-94. American Society of Microbiology, Washington, D.C. SHARP, P. A.. COHEN, S. N., AND DAVIDSON. N. (1973). Electron microscope heteroduplex studies of sequence relations among plasmids of E. co/i. II. Structure of drug resistance (R) factors and F factors. J. Mol. Biol. 75, 235-255. TOMIZAWA, J., OHMORI, H., AND BIRD, R. E. (1977). The origin of replication or colicin El plasmid DNA. Proc. Nat. Acad. Sci. USA 74, 1865-1869. TohfIzAwA, J., SAKAKIBARA, Y., AND KAKEFUDA, T. (1974). Replication of colicin plasmid DNA in cell extracts. Origin and direction of replication. Proc. Nat. Acad. Sci. USA 71. 2260-2264. WATANABE, T., AND FUKASAWA, T. (1960). Resistance transfer factor, an episome in enterobacteriaceae.
Biochem. Biophys. Res. Commun. 3, 660-665.