JOURNAL OF BACTrOLOGY, May 1972, p. 667-676Copyright 0 1972 American Society for Microbiology

Vol. 110, No. 2Printed in U.S.A.

Nature of Col E1 Plasmid Replication inEscherichia coli in the Presence of

ChloramphenicolDON B. CLEWELL

Departments of Oral Biology and Microbiology, The University of Michigan, Ann Arbor, Michigan 48104

Received for publication 10 February 1972

The colicinogenic factor E1 (Col E1) in Escherichia coli continues to replicateby a semiconservative mechanism in the presence of chloramphenicol (CAP)for 10 to 15 hr, long after chromosomal deoxyribonucleic acid (DNA) synthesishas terminated. Following CAP addition, the rate of synthesis of plasmid DNAgradually increases to an extent dependent on the medium employed. Within 2to 4 hr after the addition of CAP, replication in a glucose-Casamino Acidsmedium approaches a maximum rate representing approximately eight timesan average rate which would be required for a net doubling of DNA per cell inone generation. The number of copies of Col El DNA molecules that accumu-

late under these conditions approaches about 3,000 copies per cell, represent-ing a 125-fold increase over the normal level of 24 copies per cell. The system isparticularly convenient for studying the mechanism of DNA replication.

The colincinogenic factor E, (Col E ,) inEscherichia coli has certain features whichmake it an attractive system for studies on thenature and control of deoxyribonucleic acid(DNA) replication. It is a relatively smallplasmid (4.2 x 106 daltons) that can be iso-lated as a covalently closed supercoiled cir-cular duplex molecule (2, 3, 20). Recently ithas been found that when protein synthesis isinhibited, by either amino acid starvation ortreatment with 'chloramphenicol (CAP), the'plasmid continues to replicate while the chro-mosomal DNA replicates only to a point con-sistent with completion of rounds of replica-tion (4; Clewell and Helinski, J. Bacteriol., inpress). In this communication, results are pre-sented dealing with a more detailed analysis ofthe nature of Col E, replication during theexposure of cells to CAP. Included are studieson the rate and extent of replication as well asa determination of the semiconservative na-ture of the duplication process.

MATEIALS AND METHODSMaterials. Reagents and sources were as follows:

Brij 58, from Emulsion Engineering, Inc.; sodiumdeoxycholate (DOC) from Difco Laboratories; Sar-kosyl NL30 (sodium dodecyl sarcosinate) from theGeigy Chemical Corp.; egg white lysozyme, Pronase,bromouracil, and ethidium bromide from Calbi-ochem; CAP from Parke, Davis and Co.; CsCl (opti-

cal grade) from Schwartz Bioresearch, Inc.; thy-mine-methyl-3H (11.7 Ci/mmole), thymidine-methyl-3H (6.7 Ci/mmole), and thymine-2-l"C (55.8mCi/mmole) from New England Nuclear Corp.

Bacteria and media. E. coli strains JC411thy-(Col EJ) (6) (auxotrophic for methionine, leucine,histidine, arginine, and thymine obtained from D.Kingsbury) and CR34(Col E1) (auxotrophic for threo-nine, leucine, vitamin B1, and thymine) (17) wereused in this study. Each strain harbors the Col E,plasmid DNA which was transferred by conjugationfrom E. coli K-30. An M9 medium (as described inreference 9) containing thymine (1-4 Ag/ml), B1 (2.5jg/ml), and required amino acids (50 jg/ml) or 0.5%Casamino Acids (Difco) was employed for most ofthis work. In one case, Difco antibiotic medium no. 3(Penassay Broth) was utilized (thymidine ratherthan thymine isotope was used in this case in orderto get better incorporation of isotope). Cell growthwas at 37 C and was followed by measuring turbidityin a Klett-Summerson colorimeter.

Preparation of cleared lysate. The lysing proce-dure has been described in detail previously (7, 8). Itessentially involves the lysis of ethylenediamine-tetraacetic acid (EDTA)-lysozyme spheroplastswith a detergent mixture of Brij 58 and DOC. Centrif-ugation of the crude lysate for 25 min at 48,000 x gpellets about 95% of the total DNA, leaving theplasmid DNA in the supematant fluid. This super-natant fluid is referred to as cleared lysate. Thelatter can be frozen (-70 C) until ready for furtherexamination.

Preparation of Sarkosyl lysates. Cells were har-vested and resuspended in 1 ml of 25% sucrose [0.05

667

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

M tris(hydroxymethyl)aminomethane (Tris), pH 8.0].Lysozyme was added [0.2 ml of a 5 mg/ml solutionin TES (0.03 M Tris-0.005 M Na2EDTA-0.05 M NaCI) ],and after 5 min at 25 C, 0.4 ml of 0.25 M EDTAwas added. After another 5 min at 25 C, Pronase wasadded (0.2 ml of a 5 mg/ml solution in TES, predi-gested at 37 C for 30 min). After 5 min, lysis wasbrought about by the addition of 1.8 ml of 2% Sar-kosyl (in TES). At this point the crude lysate can befrozen (-70 C).

Preparative buoyant density-equilibrium cen-trifugation. Centrifugation was performed on aBeckman model L3-50 ultracentrifuge in a type Ti-50 fixed-angle rotor at 44,000 rev/min at 15 C for 60hr. Each polyallomer centrifuge tube contained 7.5 gof CsCl and 6 ml of sample in TES. When dye wasincluded, 8 ml of sample containing 1.5 mg of ethi-dium bromide was mixed with 7.5 g of CsCl. Theremaining space in the centrifuge tube was filledwith mineral oil. At the end of the run, the bottomof the tube was punctured and 0.2- to 0.3-ml frac-tions were collected.

Sucrose density gradients. Sucrose density gra-dient centrifugations, dropwise fractionation of thegradients, and the counting of radioisotopes werecarried out as described in detail previously (8).When cleared lysates were sedimented, the gradientscontained 0.55 M NaCl, 0.005 M Na2 EDTA, and 0.02M Tris, pH 8.0. Samples pooled from buoyant den-sity gradients were sedimented on gradients con-taining similar levels of Na2 EDTA and Tris, pH 8.0,but with 0.05 M NaCl.Medium shifts. Medium shifts were carried out

by one of two methods involving either centrifuga-tion or filtration of cells. In the former case, cellswere: (i) pelleted by centrifugation in an Interna-tional model PR-6 centrifuge at 5,000 rev/min for 5min at 25 C; (ii) washed by resuspending and centri-fuging in twice the original volume of medium salts;(iii) resuspended in new fresh M9 medium (pre-warmed to 37 C). The entire procedure required 12to 15 min. Cell growth during the shift was minimal,as evidenced by less than a 10% increase in turbidityimmediately following the shift.The other method of medium shift involved col-

lecting the cells on a membrane filter (Millipore,HA, 0.45 Am), washing with 20 ml of medium salts,and resuspending the cells in the new medium. Thelatter procedure required about 5 min.

Counting of radioisotopes. Samples were pre-pared and counted as previously described (8) in aBeckman LS250 liquid scintillation system.

RESULTSA culture of E. coli JC411thy-(Col El) was

grown for several generations in glucose min-imal medium containing "4C-thymine to labeluniformly all cellular DNA. The cells werethen pelleted, washed, and resuspended in asimilar medium containing 3H-thymine inplace of '4C-thymine. A portion of this suspen-sion served as a control and was allowed toundergo one division before being rapidly

chilled in an ice bath. CAP was added to theremainder, after which samples were removed1, 2, and 4 hr later and chilled on ice. All sam-ples were then lysed by the Brij 58-DOC pro-cedure (see above). Samples of the crude ly-sates were spotted on filter papers andcounted to determine the ratio of 3H to 14C.Since Col E, DNA is usually a very minorcomponent of the total DNA in crude lysates[plasmid normally represents less than 2% ofthe total DNA (3, 8)], the ratio of 3H to "4Chere reflects the extent of replication of chro-mosomal DNA. The lysates were then centri-fuged to obtain "cleared lysates" (see Mate-rials and Methods), which were then analyzedon sucrose density gradients as shown in Fig. 1to determine the ratio of 3H to 14C in Col ElDNA. The plasmid DNA is centered in frac-tions 14 to 17, typical of supercoiled Col ElDNA under these conditions (Clewell and Hel-inski, J. Bacteriol., in press). The rather broadlow-molecular-weight peak is of unknown or-igin but apparently represents labeled thymineincorporated during the most recent round ofreplication. This is suggested by the fact thatthere is mainly 3H in the peaks and very little,if any, '4C. This material may represent degra-dation products of recently synthesized DNA.The extent of replication of both plasmid

and chromosomal DNA is depicted in Fig. 2. Itis clear that, although the addition of CAP al-lowed chromosomal DNA to continue to anextent representing slightly more than half ofthat represented by the control culture (whichwas allowed to double its DNA content), theplasmid not only continued to replicate butgradually increased its rate of replication.From the slopes of the lines in Fig. 2, it is evi-dent that the average rate of Col El replicationbetween the second and fourth hours was al-most three times that observed for the firsthour after chloramphenicol addition, andabout twice the average rate that would benecessary to produce a normal doubling ofplasmid DNA during a normal division cycle(i.e., as judged by comparison to the 3H to 14Cratio of the one-generation control cells). It isseen also that for about two generation equiva-lents (doubling time in this medium is 75 min),Col El accumulation paralleled the theoret-ical logarithmic increase that would be ex-pected if cells were growing logarithmically.

Figure 3 shows the results of an experimentidentical to that of Fig. 2 except that Casa-mino Acids (0.5%) were present in the me-dium. In this case Col El replication increasedwith time to a greater extent than that seen inminimal medium; and here the replication

668 CLEWELL

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COL E1 PLASMID REPLICATION

SE)

In

a-

9

Fraction NumberFIG. 1. Sedimentation analyses of cleared lysates prepared from cells treated with chloramphenicol (CAP).

A culture (50 ml) of E. coli JC411thy- (Col E1) was grown in glucose minimal medium containing "4C-thy-mine (1.6 1.Ci per 3.6 ug per ml) at 37 C for several generations in log phase. The cells were then shifted to asimilar medium (83 ml) containing 3H-thymine (2.4 gCi per 5 ,ug per ml) in place of "4C-thymine. Immedi-ately the cells were divided into two portions. One portion (20 ml) was allowed to continue until the turbidityhad doubled, at which point growth was terminated by rapid chilling in an ice bath. To the remaining por-tion, CAP was added to a concentration of 170 ,g/ml. While incubating at 37 C, samples (20 ml) of the latterculture were removed and quickly chilled on ice at 1, 2, and 4 hr after the medium shift. Cleared lysates wereprepared from cells, diluted twofold with TES, and 0.2-ml samples were centrifuged (from right to left)through 5 to 20% sucrose density gradients in an SW50.1 rotor (15 C) at 48,000 rev/min for 135 min. (A) Par-allel control in which cells were shifted into medium identical (with the exception of the isotope) to the pre-shift medium for one generation (75 min). (B), (C), and (D) represent cells shifted to a medium containingCAP for 1, 2, and 4 hr, respectively.

rate between hours 2 and 4 exceed the "nor-mal" doubling time (50 min in this case) byapproximately eightfold. The rate of accumu-lation of Col E1 parallels the theoretical loga-rithmic increase for about four generationequivalents.

It is important to point out that the phrase"normal doubling time," as used here, refers toa value corresponding to the synthesis in-volved in a net doubling of the number of ColE, DNA molecules per cell (i.e., 24 copies) pergeneration. Although the overall rate of syn-thesis may increase exponentially during onegeneration, it can, for the sake of making com-parisons, be represented by a value corre-sponding to a linear rate of synthesis which isconstant throughout one division cycle. Tofollow plasmid replication for longer periods oftime, CAP (150 ug/ml) was added to a log cul-ture (100 ml) growing in M9 Casamino Acidsmedium containing "4C-thymine (1.6 MCi per3.6 Mg per ml), and subsequently samples (20

ml) were removed (and put on ice) at 0, 4, 9,20, and 26 hr. Lysates were prepared by theSarkosyl procedure (see Materials and Meth-ods) and centrifuged to equilibrium in CsCldensity gradients containing ethidium bro-mide. The gradients were photographed withan ultraviolet (UV) lamp to promote ethidiumbromide-mediated fluorescence of the DNAbands (Fig. 4). The higher-density DNA band(lower band) represents covalently closed cir-cular DNA (18). The gradients were then frac-tionated, and radioactivity was determined insamples of the fractions (Fig. 5).The UV photographs (Fig. 4) of the glucose-

Casamino Acids-grown cells, which comparethe DNA in cells that have not been exposedto CAP with that of cells that have been ex-posed for various lengths of time, clearly indi-cated that Col El DNA (the lower band),which is not readily apparent at zero time,gradually approaches a level which is compa-rable to that of the chromosomal DNA (the

669VOL. 110, 1972

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

20/

i 5

3H o1/14c "

Plosmid (CON),

Chrom(CON Chrom.(CAP)

01 2 3 4

Hours after Medium ShiftFIG. 2. Extent of replication of plasmid and chro-

mosomal DNA after treatment of cells in glucose-minimal medium with chloramphenicol. The counts-per-minute ratios of 'H to 14C represent the valuesobtained in the experiment depicted in Fig. 1.Corrections have been made regarding the base linelevel of counts within the peaks. The values obtainedfor the chromosome were based on the ratio of 3H to14C counts in the crude lysate. CAP, cells treatedwith chloramphenicol; CON, control without chlor-amphenicol for one generation. Dashed line repre-sents the theoretical increase in DNA that wouldarise during normal logarithmic cell growth.

A

100y

3H4c /

50-II

Chrom. (CON) /

Plosmid (CON)Chrom.(CAP)

0 -2 3 4

Hours after Medium ShiftFIG. 3. Extent of replication of plasmid and chro-

mosomal DNA after treatment of cells in glucose-Casamino Acids medium with chloramphenicol. Thisexperiment was carried out in essentially the samemanner as the experiment in Fig. 1 and 2, with theexception of the medium employed. Normal dou-bling time in this medium, however, is 50 min. CAP,cells treated with chloramphenicol; CON, controlwithout chloramphenicol for one generation. Dashedline represents the theoretical increase in DNA thatwould arise during normal logarithmic cell growth.

C D

FIG. 4. Photographs of UV-illuminated dye-buoyant density gradient of Sarkosyl lysates prepared fromcells treated for varying lengths of time with chloramphenicol. (A), (B), (C) and (D) represent cells treated for0, 4, 9, and 20 hr, respectively. Upper band represents chromosomal DNA; lower band represents covalentlyclosed circular DNA.

670 CLEWELL

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COL E1 PLASMID REPLICATION

upper band). The positions of the UV-ab-sorbing bands correspond well with peaks ofradioactivity in the fractionated gradients (Fig.5), indicating that incorporation of radioactivethymine into Col E1 DNA relates to a net in-crease in mass and not simply a repair process.

to

04 A

50.

48

n

0.

a.,

1)

l0 20Fraction Number

The Col E1 DNA band, when analyzed (afterpooling and dialyzing fractions) on neutralsucrose density gradients behaves as normal23S material and also contains a small about(5%) of a 31S substance (Fig. 6) presumablyrepresenting dimer molecules (either circularor catenated dimers, or both). Dimers are gen-erally not found to any extent in E. coli (3)except in certain abnormal situations (11), al-though oligomers are present in relatively highpercentages when the plasmid is harbored inProteus mirabilis (2, 12). The small amount of17S DNA probably represents some circularCol E, DNA that has acquired a nick (see ref-erence 2).Beyond 4 hr in CAP it is reasonable to as-

sume that the level of chromosomal DNA re-mains constant (see Fig. 2 and 3), in whichcase the amount of plasmid DNA can be meas-ured relative to the level of chromosomalDNA. Figure 7 portrays these values for thisexperiment as well as for two similarly per-formed experiments carried out in a minimalmedium (M9 glucose plus required aminoacids) and a rich broth. These data indicatethat after 3 to 4 hr of exposure to CAP, repli-cation of Col E, DNA occurs linearly for sev-

5

4

toI00x

a-

I

FIG. 5. Dye-buoyant density centrifugation ofSarkosyl lysates from cells treated with chloram-phenicol (CAP) for increasing lengths of time. Thefractionation profiles shown correspond to those gra-dients pictured in Fig. 4 (density increases from rightto left). In addition, the fractionation profile of a 26-hr treatment (from the same experiment) is shown.(A), (B), (C), (D), and (E) correspond to CAP treat-ments of 0, 4, 9, 20, and 26 hr, respectively. Frac-tions of equal volume were collected only in the re-gion of the gradients where it was visually observedthat the DNA was present. A portion (0.05 ml) ofeach fraction was counted. In the case of A, thecounts were plotted also on a scale different from theothers so that the higher-density satellite peak couldbe resolved.

3

2

010 20 30Fraction Number

FIG. 6. Sedimentation analysis of pooled fractionscorresponding to covalently closed circular DNA iso-lated by dye-buoyant density centrifugation. Thesample corresponds to combined fractions obtainedfrom the experiment of Fig. 5. A 0.2-ml sample wascentrifuged (from right to left) through 5 to 20%o su-crose density gradients in an SW50.1 rotor (15 C) at48,000 rev/min for 90 min.

671VOL. 110, 1972

I

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

0.9

0.81-

0.7 -

0.6 FE|a .m

0.51-

0.4 -

0.3 F0.2 1

0.1

05 10 15 20 25Hours of CAP Treatment

FIG. 7. Extent of Col E1 replication during treat-ment with chloramphenicol (CAP) in three differentmedia. (A), (B), and (C) correspond to cells grown inglucose-Casamino Acids, Difco Penassay Broth (anti-biotic medium no. 3), and glucose-minimal medium,respectively. After 3 to 4 hr in CAP, chromosomalDNA no longer replicates, and it is assumed here toremain at a constant level beyond this time. Thusthe chromosomal DNA is used as a normalizingfactor in estimating the amount of plasmid DNApresent. The experiment corresponding to A repre-sents the experiment of Fig. 5. The ratio of plasmidto chromosomal DNA was estimated on the basis ofthe areas enclosed in the two peaks of the dye-buoyant density gradients (see Fig. 5). Relative areas

were determined by weighing pieces of paper cut tothe dimension of the areas enclosed by the peaks.The data of B and C were calculated in the samemanner.

eral more hours and then levels off. In the case

of glucose-Casamino Acids-grown cells,leveling off occurs at a point representingabout 82.5% of the level of chromosomal DNA.Assuming the size of a chromosomal genomeequivalent is 2.5 x 109 daltons (10), we can

estimate that there are 2.1 x 109 daltons ofplasmid DNA per chromosomal genome equiv-alent. Since Col El DNA has a molecularweight of 4.2 x 106, this corresponds to 500copies per chromosomal genome equivalent. Itwas determined on the basis of cell counts(using the Petroff-Hauser bacteria counter)and DNA analyses [by the method of Burton(5)] that log cells growing in this medium con-tain approximately four (actually 3.75 0.30)chromosomal genome equivalents of DNA percell. Since normally there are about 24 copies

of Col El DNA per cell independent of thenumber of chromosomal equivalents (Clewelland Helinski, J. Bacteriol., in press), there arein this case six copies of plasmid DNA perchromosomal equivalent. Under conditions ofprotein synthesis inhibition, cell mass in-creases only slightly (10-15%), and the amountof chromosomal DNA increases by about 50%(15; and see Fig. 2 and 3). Thus, if we assumethat after extended CAP treatment there is anaverage of six chromosomal genome equiva-lents per cell, this would correspond to 6 x500 = 3,000 copies of Col E1 DNA per cell (seeTable 1), representing about a 125-fold in-crease per cell over the normal level.

In the case of cells grown in Difco PenassayBroth, the relative amount of plasmid DNA tochromosomal DNA leveled off at a point repre-senting about 47.5% that of the chromosomalDNA (also see Table 1). Cells in this mediumhave a normal doubling time of 30 min (ascompared to 50 min for glucose-CasaminoAcids-grown cells) and might be expected, onthe basis of experiments performed on otherstrains of E. coli and Salmonella, to have a

TABLE 1. Rate of synthesis and extent ofaccumulation of plasmid DNA in the presence of

chloramphenicol

Avg no. ofCol E, DNA

molecules syn- Final no. ofthesized per min- copies of Col E,ute during max- DNAimum rate of

Medium synthesisaPer Per

chromo- chromo-somal Per somal Pergenome cellb genome cellbequiva- equiva-lent lent

A. Glucose-Cas- 0.68 4.1 500 3000amino Acids

B. Broth 0.55 3.3 283 1698C.Glucose mini- 0.20 0.7 119 440malIa "Maximum" rate of synthesis refers to that rate

of synthesis which is achieved after the first 3 to 4 hrof exposure to chloramphenicol and corresponds tothe time related to the linear regions of the curvesshown in Fig. 7. Rates were calculated by deter-mining the difference between the number of Col E,copies at two arbitrarily chosen time points and di-viding by the time elapsed between the two points.

°Based on presence of six chromosomal genomeequivalents per cell for A and B and 3.7 chromo-somal genome equivalents for C (or 1.5 times thenumber of chromosomal genome equivalents per cellduring normal log-phase growth).

.A

B

CIDsoI

,,,,

672 CLEWELL

L

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COL El PLASMID REPLICATION

greater cell mass and more chromosomal DNAper cell (10, 16). Efforts to confirm this in ourK-12 strain indicated (reproducibly) that thisis not the case for these cells and that the cellmass and amount of DNA per cell in PenassayBroth is indistinguishable from that of the glu-cose-Casamino Acids-grown cells. The reasonfor this is not clear but could simply reflectthe nature of this particular strain. Thus, thefact that the leveling off of plasmid replicationin broth-grown cells at a level approximatelyhalf that of Casamino Acids-grown cells isapparently not related to a difference in theamount of chromosomal DNA per cell. Morelikely, Col El replication in these (broth-grown) cells had become somewhat dependenton one or more nutrients in the broth, presentin very low amounts, which were exhaustedrelatively quickly. In the presence of CAP, thecell could not compensate with new enzymesynthesis, and thus replication terminated.

In the case of cells in minimal medium (nor-mal doubling time is 75 min), replication ofplasmid DNA continued for about 10 hr butleveled off at a point representing only 20% ofthe level of chromosomal DNA. This result isprobably simply a reflection of the relativelypoor medium. The number of chromosomalgenome equivalents per logarithmicallygrowing cell was found to be 2.45 i 0.35 (thecorresponding number of copies of Col El isshown in Table 1).As indicated in Table 1, the replication rates

during the time of linear increase in plasmidDNA (i.e., after 3 to 4 hr in CAP) indicateagain that Col El DNA replicates at rates sig-nificantly faster than that which occurs undernormal cell growth conditions. In the case ofglucose-Casamino Acids-grown cells, a normaldoubling of Col El DNA corresponds to thesynthesis of 24 copies in 50 min and may berepresented as occurring at an average rate of0.48 copies per min. Beyond 4 hr in CAP therate is 4.1 per min per cell or slightly morethan an eightfold increase in the average rate.This is approximately the same relative ratethat we obtained between hours 2 and 4 in theshorter-term experiment of Fig. 3. It is thusapparent that upon addition of CAP to cells inthis medium, replication of Col E1 occurs loga-rithmically for a time period representingabout four generation equivalents before be-coming linear.The following experiment was performed to

determine whether Col El replication in thepresence of CAP occurs by a semiconservativeprocess. A 45-ml culture of CR34(Col E1) cellswas grown for several generations (in log

phase) in a glucose-Casamino Acids mediumcontaining 14C-thymine (0.67 ,Ci per 1.5 ,ug perml). CAP (150 Ag/ml) was added, and after4 hr the cells were shifted to 48 ml of mediumcontaining bromouracil (BU) (200 ,ug/ml) and3H-thymine (50 uCi per 1.2 ,ug per ml) in amolar ratio of 110 to 1 (BU to thymine). Thenew medium also contained chloramphenicolat the same concentration as was present inthe preshift medium. Samples (20 ml) wereremoved and rapidly chilled on ice after 2 and4 hr. Crude lysates were prepared by the Sar-kosyl procedure and centrifuged to equilibriumin CsCl density gradients. The results areshown in Fig. 8. The sample representing cellsincubated for 2 hr in BU exhibits one ratherbroad 3H peak, which contains also some "IC-labeled DNA. The material in this peak has ahigher density than unreplicated "4C-DNA

I,

I ~~10 20 30 404

4 A -8

3-6

2~~~~~~~~~* )

10 20 30 40 L1)2345 6

FoI

-4 I-I B 8

3 6I

4

10 20 30 40Fraction Number

FIG. 8. CsCl-buoyant density gradient centrifuga-tion of Sarkosyl lysates prepared from cells exposedto chloramphenicol and bromouracil. Cells prela-beled with l4C-thymine and exposed to chloram-phenicol for 4 hr were shifted to a medium con-taining 9H-thymine (in place of 14C-thymine), bromo-uracil, and chloramphenicol (as described in text).(A) and (B) represent cells exposed to bromouracilfor 2 and 4 hr, respectively (density increases fromright to left). Arrows indicate fractions which werepooled and analyzed on sucrose gradients (see Fig. 9).

VOL. 110, 1972 673

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CLEWELL

(devoid of 3H) and apparently represents hy-brid (heavy:light) DNA. After 4 hr, a moredense DNA appears which is practically de-void of 14C label and must therefore representDNA with BU in both strands. Such a patternis what one would expect as a result of semi-conservative replication. The 3H-labeled DNAis Col El DNA, as was demonstrated by pool-ing fractions (as indicated) through regions ofthe gradient, dialyzing, and analyzing by sedi-mentation on sucrose density gradients (Fig.9). The typical supercoiled 23S DNA is themajor species in all pools except that repre-senting unreplicated 14C-labeled DNA, inwhich case larger chromosomal DNA is thepredominant species. The amount of 17S opencircular DNA seems greater than that normallyseen. Nicking of covalently closed circularDNA could have resulted from nonspecificnuclease(s) activity before or during the lysisprocedure. That this is a likely possibility issupported by the fact that in these experi-ments lysis of cells after exposure to BU waspremature. That is, lysis occurred to some de-gree even before lysozyme was added to thecell suspension. Since no EDTA was presentprior to this time, nucleases (activated byendogenous divalent cations) could have pro-duced nicks in the DNA.

Material representing Col El DNA that hasundergone one round of replication is centeredin pool 4 (of Fig. 8B). When the DNA in thispool was denatured by boiling for 20 min (thisprocedure nicks as well as promotes denatura-tion) and then was fractionated after alkalineCsCl centrifugation, the 3H counts all ap-peared in a band of higher density than the14C-labeled band. When a similar analysis wasperformed on material in pool 2, representingtwice-replicated Col El DNA, there was onlyone radioactive peak and it contained almostentirely 3H. These data indicate that Col Elreplication in the presence of CAP is semicon-servative.

DISCUSSIONThe results presented above clearly indicate

that, in the presence of a high level of CAP, E.coli (Col E,) cells continue to replicate Col E,DNA by a semiconservative process long afterchromosomal DNA synthesis has terminated.The rate of Col E1 replication under theseconditions depends somewhat on the mediumused. Accumulation of plasmid DNA occurslogarithmically during the first few hours.After 2 to 4 hr of CAP treatment, cells in aglucose-Casamino Acids medium allow replica-tion at a rate approximately eight times the

10 20 30 8B6 9/,l

4 4l' 5000

*g S

10 20 30

Fraction NumberFIG. 9. Sedimentation analyses of the pooled frac-

tions indicated in Fig. 8. Al through A5 representpools 1 to 5, respectively, corresponding to cells ex-posed to bromouracil for 2 hr. Bl through B6 repre-sent pools 1 to 6, respectively, corresponding to cellsexposed to bromouracil for 4 hr. Samples were sedi-mented through 5 to 20%o sucrose density gradients(from right to left) in an SW50.1 rotor (15 C) at48,000 rev/min for 90 min.

normal rate. Replication continues for up to 10to 15 hr following the addition of CAP, duringwhich time the number of Col E, moleculesper cell increases approximately 125-fold.

674 J. BACTERIOL.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

COL E, PLASMID REPLICATION

The reason for the increased rate of replica-tion is not clear. One possibility is that eachnew plasmid DNA molecule is, under theseconditions, immediately used as a template byreplicating enzymes present in excess

amounts, thus yielding a logarithmic increasein DNA synthesis until achievement of a max-

imum rate corresponding to saturation of thereplicating enzymes. It is still not clear, how-ever, why such replicating machinery, if itwere present in such excess, should be more

accessible to Col El DNA during inhibition ofprotein synthesis. Perhaps the indirect inhibi-tion of chromosomal DNA synthesis somehowfrees replicating enzyme(s). The KornbergDNA polymerase (polymerase I), an enzyme

upon which the Col El plasmid has been pre-viously shown to be dependent (13), is nor-

mally present to the extent of about 400 mole-cules per cell (19) and may be involved in thisphenomenon.Another possibility is that there may always

be an excess of available replicase, but that thelevel of certain other rate-limiting factors hasbecome altered. For example, suppose that theoverall rate of Col El synthesis is regulated bythe frequency of initiation. If initiation were

negatively controlled through the interactionof the appropriate protein subunit(s) (e.g., a

"repressor" substance which binds and in-hibits an initiator substance), then conceivablya situation could arise (e.g., after extendedCAP treatment) resulting in a breakdown ofthe normal maintenance of the repressed statedue to abnormal dissociation of the regulatorycomponents. Alternatively, an increased fre-quency of initiation could result from the accu-mulation of higher levels of "signal" activitywhich normally "triggers" initiation. In anycase, a breakdown of the control of initiationwould then make the maximum rate of syn-thesis dependent on other factors, such as,perhaps, the replicase level.

Lastly, there is always the unattractive pos-

sibility that a protein involved in Col El repli-cation continues to be synthesized in the pres-ence of high levels of CAP and gives rise toaccelerated Col El synthesis. Such a possi-bility seems unlikely but should not be com-pletely overlooked.

It has been reported that under conditions ofamino acid starvation Col El DNA immedi-ately begins to replicate at a rate of two tothree times the normal rate (4; Clewell andHelinski, J. Bacteriol., in press). Thus an

acceleration of Col El DNA synthesis in theabsence of protein synthesis is brought onmuch more rapidly by amino acid starvation

675

than by CAP treatment, although the mecha-nism by which acceleration occurs may be thesame in both cases.

In certain bacteriophage systems, replicationof phage DNA in the presence of high levels ofCAP (100 jg/ml) has been shown to occur tosome extent, but usually at rates which de-crease rapidly with time (1, 14). Thus, in thesesystems, unlike the case of Col E1, inhibitionof protein synthesis by CAP impedes continua-tion of DNA replication.The replication of Col E1 DNA in the pres-

ence of CAP provides an extremely convenientand simple in vivo system for studying themechanism of DNA replication. The Col E,DNA molecule can be isolated with ease and,if desired, in relatively large quantities. Thesystem is currently being used in our labora-tory to study the effects of various chemicalcompounds (antibiotics, etc.) on DNA replica-tion as well as the mechanism by which thesecompounds act.

ACKNOWLEDGMENTS

I wish to thank Sandra White for her excellent technicalassistance.

This work was supported by Public Health researchgrants AI-10318-91 (National Institute of Allergy and Infec-tious Diseases) and DE-02731 (National Institute of DentalResearch), and an institutional research grant (IN-40K) fromthe American Cancer Society to The University of Mich-igan.

LITERATURE CITED

1. Amati, P. 1970. Chloramphenicol: effect on DNA syn-

thesis during phage development in Escherichia coli.Science 168:1226-1228.

2. Bazaral, M., and D. R. Helinski. 1968. Characterizationof multiple circular DNA forms of colicinogenic factorE, from Proteus mirabilis. Biochemistry 7:3513-3520.

3. Bazaral, M., and D. R. Helinski. 1968. Circular DNA ofcolicinogenic factors E,, E, and E, from Escherichiacoli. J. Mol. Biol. 36:185-194.

4. Bazaral, M., and D. R. Helinski. 1970. Replication of a

bacterial plasmid and an eipsome in Escherichia coli.Biochemistry 9:399-406.

5. Burton, K. 1956. A study of conditions and mechanismsof the diphenylamine reaction for the colorimetric es-timation of DNA. Biochem. J. 62:315-323.

6. Clark, A. J., and A. D. Margulies. 1965. Isolation andcharacterization of recombination-deficient mutantsof Escherichia coli K-12. Proc. Nat. Acad. Sci. U.S.A.53:451-459.

7. Clewell, D. B., and D. R. Helinski. 1969. Supercoiledcircular DNA-protein complex in Escherichia coli:purification and induced conversion to an open cir-cular DNA form. Proc. Nat. Acad. Sci. U.S.A. 62:1159-1166.

8. Clewell, D. B., and D. R. Helinski. 1970. Properties of a

deoxyribonucleic acid-protein relaxation complex andstrand specificity of the relaxation event. Biochem-istry 9:4428-4440.

9. Clowes, R. C., and W. Hayes. 1968. Experiments in mi-crobial genetics, p. 187. John Wiley & Sons Inc.,New York.

VOL. 110, 1972

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

676 CLEWELL

10. Cooper, S., and C. Helmstetter. 1968. Chromosome rep-lication and the division cycle of Escherichia coli B/r.J. Mol. Biol. 31:519-540.

11, Goebel. W. 1970. Studies on extrachromosomal DNAelements: replication of the colicinogenic factor Col E,in two temperature sensitive mutants of Escherichiacoli defective in DNA replication. Eur. J. Biochem.15:311-320.

12. Goebel, W., and D. R Helinski. 1968. Generation ofhigher multiple circular DNA forms in bacteria. Proc.Nat. Acad. Sci. U.S.A. 61:1406-1413.

13. Kingsbury, D. T., and D. R. Helinski. 1970. DNA po-

lymerase as a requirement for the maintenance of thebacterial plasmid colicinogenic factor E,. Biochem.Biophys. Res. Commun. 41:1538-1544.

14. Levine, A. J., and R. L. Sinsheimer. 1969. The process

of infection with bacteriophage OX174. XXV. Studieswith bacteriophage 4X174 mutants blocked inprogeny replicative form DNA synthesis. J. Mol. Biol.39:619-639.

15. Maaloe, O., and P. C. Hanawalt. 1961. Thymine defi-

J. BACTERIOL.

ciency and the normal DNA replication cycle. J. Mol.Biol. 3:144-155.

16. Maaloe, O., and N. 0. Kjeldgaard. 1966. Control of mac-romolecular synthesis. W. A. Benjamin, Inc., NewYork.

17. Okada, T., K. Yanagisawa, and F. J. Ryan. 1960. Elec-tive production of thymine-less mutants. Nature(London) 188:340-341.

18. Radloff, R., W. Bauer, and J. Vinograd. 1967. A dye-buoyant-density method for the detection and isola-tion of closed circular duplex DNA: the closed cir-cular DNA in HeLa cells. Proc. Nat. Acad. Sci.U.S.A. 57:1514-1521.

19. Richardson, C. C., C. L. Schildkraut, H. V. Aposhian,and A. Komberg. 1964. Enzymatic synthesis of deoxy-ribonucleic acid. XIV. Further purification and prop-erties of deoxyribonucleic acid polymerase of Esche-richia coli. J. Biol. Chem. 239:222-232.

20. Roth, T. F., and D. R. Helinski. 1967. Evidence for cir-cular DNA forms of a bacterial plasmid. Proc. Nat.Acad. Sci. U.S.A. 58:650-657.

on March 17, 2020 by guest

http://jb.asm.org/

Dow

nloaded from