Vol. 139, No. 3

JOURNAL OF BACTERIOLOGY, Sept. 1979, p. 817-823 002 1-9193/79/09-0817/07$02.00/0

Effect of Protein Synthesis on Plasmid Maintenance in Streptococcus faecalis MICHAEL MURPHEY-CORB AND MICHAEL L. MURRAY* Department of Microbiology, Louisiana State University Medical Center, New Orleans, Louisiana 70119 Received for publication 20 March 1979

Plasmid-to-chromosome ratios in Enterobacteriaceae, upon interruption of protein synthesis by chloramphenicol, are either conserved or increased when measured by dye buoyant density centrifugation. We have found, on the other hand, that the effect of inhibition of protein synthesis on the amount of covalently closed circular deoxyribonucleic acid visualized by this method in two strains of Streptococcus faecalis appears to differ from these established systems. A threeto sixfold decrease in covalently closed circular deoxyribonucleic acid was observed when lysates of chloramphenicol-treated cultures were submitted to dye buoyant density centrifugation. A loss of covalently closed circular deoxyribonucleic acid was also evident from electrophoretic profiles of these lysates. Several conditions which could account for the apparent loss of covalently closed circular deoxyribonucleic acid upon inhibition of protein synthesis are discussed. Control of plasmid replication in the family Enterobacteriaceae has been characterized as either stringent or relaxed, depending upon the amount of replication which occurs in the absence of protein synthesis (1). Little is known about control of replication of plasmids in streptococci. The experiments described in this paper demonstrate significant differences between the behavior of replicating plasmids in the Enterobacteriaceae and in streptococci. MATERIALS AND METHODS

Bacterial strains and growth conditions. The bacterial strains chosen for these experiments were obtained from the Diagnostic Laboratory of Charity Hospital, New Orleans, La. They have been identified as Streptococcus faecalis on the basis of their reaction on Enterococcosel agar (BBL Microbiology Systems, Cockeysville, Md.), their resistance to 6.5% NaCl, and growth at 45°C. Strain CS92 was found to be resistant to erythromycin (minimum inhibitory concentration [MIC], >1,000 jig/ml), lincomycin (MIC, 250 jig/ml), ampicil-

lin/penicillin (MIC, 31 jig/ml), tetracycline (MIC, 62

jig/ml), and cephalothin (MIC, 125 jig/ml). Strain

CS93 was isolated as a derivative of CS92 by treatment of CS92 with ethidium bromide. The antibiotic resistance pattern of CS93 follows that of CS92 with the exception that it is sensitive to erythromycin and lincomycin (MIC, 5 jig/ml). CS40 is resistant to erythromycin (>1,000 jig/ml) and lincomycin (>1,000 jg/ ml). Brain heart infusion broth (BBL Microbiology Systems) supplemented with 0.1 M Tris-hydrochloride, pH 7.0, was used with further supplements as indicated. Unsupplemented brain heart infusion agar was 817

used for solid medium. When labeled DNA was desired, either 1 or 10 juCi of [methyl-3H]thymidine (0.2 Ci/mmol; Research Products International) per ml was added to the culture. A mixture of "C-labeled amino acids (no. 3130-08, Schwarz/Mann, Orangeburg, N.Y.) was added at 1 jiCi/ml when labeled protein was desired. Chloramphenicol (CAP) (Sigma Chemical Co., St. Louis, Mo.) was added at a concentration of 300 jig/ml, except where otherwise noted. Cell lysis. Cellular lysis was obtained by the method of Clewell et al. (3), with the following modification. The cells were harvested by centrifugation for 10 min at 4,000 x g and 4°C and suspended in 0.06 M disodium EDTA-50 mM Tris, pH 8.0. After incubation in this buffer for 15 min at 37°C, the cells were washed twice by alternate centrifugation and resuspension in 50 mM Tris, pH 8.0. The washed cells were resuspended in 25% sucrose in Tris, pH 8.0, and the procedure for lysis was continued as described by Clewell et al. (3). Pretreatment of the cells with EDTA permitted complete lysis of all cultures examined, including those exposed to CAP. When enrichment of the lysate for low-molecularweight DNA was desired, a high-salt extract of the streptococcal lysate was prepared, using the method of Hirt (7). Further concentration of the supernatant was obtained by precipitation of the DNA with the addition of 2 volumes of cold (-20°C) 95% ethanol. The DNA precipitate was then redissolved in 50 mM Tris, pH 7.5, 50 mM NaCI, and 5 mM EDTA (TES). If subsequent electrophoresis was to be performed, the plasmid-containing samples were dialyzed against TES overnight at 4°C. Dialyzed preparations were stored at -20°C until use. Dye buoyant density centrifugation. Dye buoyant density centrifugation was performed using the method of Grossman et al. (5). CsCl, TES, and lysate were combined to yield a solution with a final volume

818

MURPHEY-CORB AND MURRAY

of 10 ml and a density of 1.546 g/cm '. Propidiulm diiodide (Calbiochem-Behring, La Jolla, Calif.) was then added to a final concentration of 300 jig/ml, and all subsequent manipulations were performed in reduced light to prevent nicking of the DNA. The solution was centrifuged for a minimum of 48 h at 44,000 rpm at 25°C, using a Ti6O rotor in a Beckman L2-65B or L3-50 preparative ultracentrifuge. The bottom of the polyallomer tube was punctured, and equal fractions were collected. Aliquots of each fractioni were placed on 1.25-cm filter paper disks (Schleicher and Schuell, 740E) and precipitated with I N HCl at room temperature (4), and the acid-precipitable counts were measured with an Intertechnique scintillation counter and 5 ml of scintillation fluid composed of 1.3 g of 2,5diphenyloxazole per liter of toluene. When further analysis of CsCl-purified plasmid DNA was desired, fractions containing the covalently closed circular (CCC) DNA were pooled, and propidium diiodide was removed by extraction with CsCI-satturated isopropatnol.

Soome lysates were analyzed directly by dye buoyant density centrifugation. T'hese lysates were prepared as described above, except that Sarkosyl was substituted for sodium dodecyl sulfate. The viscous lysate was passed 20 times through a 22-gauge needle to shear high-molecular-weight DNA before centrifugation. Quantitation of CCC DNA. The quantity of CCC D)NA was determined, as a function of time after the simultaneous removal of label and addition of CAP, from the radioactivity found in the plasmid peak obtained upon dye buoyant density centrifugation. Lowmolecular-weight D)NA was purified from cellular lysates by high-salt extraction and further purified by precipitation with ethanol before dye buoyant density analysis. This procedure gave more reproducible results than could be obtained with unextracted DNA. The yield at each successive manipulation was determined, and the resulting values were normalized to correct for varying degrees of recovery. The average recovery was

60

±

15% standard deviation.

Agarose gel electrophoresis. Agarose gel electrophoresis of high-salt extracts or of dye buoyant density-purified DNA was accomplished as described by Harkess and Murray (6). RESULTS Our initial studies on the maintenance of

streptococcal plasmid DNA involved primarily the effect of the inhibition of protein synthesis on the plasmids in strains CS92 and CS40. These strains were chosen for our studies because both bear several plasmids, each of which might respond differently to the inhibition of protein synthesis. Preliminary experiments were performed to observe the effect of CAP on host macromolecular synthesis in strain CS92. CAP was added to a portion of a culture growing in the presence of [3H]thymidine and '4C-amino acids. Turbidity, [3H]thymidine incorporation, and '4C-amino acid incorporation were monitored during further incubation; the measurements are reported

,J.

BACTERIOL.

in Fig. 1. Immediate cessation of protein synthesis upon addition of CAP is evident; DNA synthesis and turbidity are subsequently inhibited. CAP was remnoved from an aliquot of the culture after 120 min; DNA synthesis resumed in 45 min, and protein synthesis and increase in turbidity resumed in 100 min. Cell viability was also measured (data not shown); no change in the number of viable cells was observed during incubation in the presence of CAP.

Dye buoyant density analysis. To determine the effect of inhibition of protein synthesis on plasmid maintenance, the amount of CCC DNA was measured after incubation of CS92 in

the presence of CAP. A [3H]thymidine-labeled culture was split into three aliquots. The first was placed at -20°C immediately and reserved as the zero-time control. The other two aliquots were incubated for an additional 2 h, one in the presence of CAP and the other in the absence of any inhibitor. At this time, all three cultures were harvested, and lysates were prepared and analyzed by dye buoyant density centrifugation. Figure 2 represents the dye buoyant density profiles, from left to right, of the zero-time control, the 2-h CAP, and the 2-h control samples, respectively. The denser peak represents label incorporated into CCC DNA; the lighter peak is comprised of sheared chromosomal fragments along with the linear and open circular forms of plasmid DNA. A distinct loss of counts incorporated into the CAP-treated sample as CCC DNA relative to that found in the controls is evident. Continuous label experiments identical to the one described above were performed repeatedly; all demonstrated depressed levels of CCC DNA upon treatment of the culture with CAP. Quantitation of CCC DNA. To monitor the fate of the plasmid DNA after the introduction of CAP, a culture of CS92 was treated in the following manner. A midlog-phase culture of CS92 grown for several generations in the presence of ['H]thymidine was washed with chilled unlabeled medium to remove any traces of radioactive precursor. The culture was resuspended in fresh medium containing no isotope and split into two portions. One portion was maintained as a control, and CAP was added to the other. Both cultures were incubated at 37°C, and aliquots were removed from each at timied intervals and prepared for dye buoyant density analysis. The amount of CCC DNA from each culture is plotted as a function of time (Fig. 3). The amount of label present in the chromosomal pellet of each sample was also measured to assess the effectiveness of the chase. It is apparent from these data that by 30 min after the addition

PLASMII) MAINTENANCE IN S. FAECALIS

VOL. 139, 1979

-

819

1

-. 7 -. 4 -. 3 -. 2

101

.0

:.i

--L

CL

-.1

30

90

60

120

150

180

210

240

270

300

Minutes After CAP Addition

FIG. 1. Analysis of the effects of CAP on growth, DNA synthesis, and protein synthesis of CS92. A culture of CS92 was grown in the presence of radiolabeled precursors of DNA and protein for approximately four generations before the addition of CAP. Thereafter, control (0, El) and CAP-treated (0, *) cultures were monitored as a function of time after the addition of CAP for changes in turbidity (0--- 0, @--- 0) (0.1 -0), and optical density unit = 2 x 108 cells per ml), incorporation of [3H]thymidine (0-OC, incorporation of '4C-amino acids (L}---I, *E- ). At 120 min (T) a portion of the CAP-treated culture was washed and resuspended in fresh medium without CAP, with continued monitoring (A). 400-

200-

8a-

4-

2-

20

40

60

80

100

20

40

60

80

100

'70 Gradient

FIG. 2. Dye buoyant density analysis of a continuously labeled culture of CS92 in the presence or absence of CAP. The dye buoyant density profiles from the total lysates of 20-ml cultures are shown, from left to right, at zero time, after 2 h of incubation in the presence of CAP, and after 2 h of incubation without inhibitor.

of CAP, the amount of CCC DNA found in the CAP-treated samples has declined approximately sixfold relative to that present in the control sample. The decrease in CCC DNA levels observed in the control culture after 90 min is reproducible and may be associated with aging

of the cells upon entering stationary phase. After 2 h of incubation in the presence of CAP, a portion of the CAP-treated culture was washed and resuspended in an equal volume of unlabeled medium without CAP, and the measurement of CCC DNA at timed intervals was

820

MIJRPHEY-CORB AND MURRAY

PJ. BACTERIOLI.

continued. These data are also shown in Fig. 3. A return of radiolabel in the forim of CCC DNA was observed upon removal of CAP from the culture. Partial inhibition of protein symnthesis is sufficient to cause decreases in thee level of CCC DNA, as shown by continuous la]bel experiments similar to those of Fig. 2, perfoirmed with concentrations of CAP close to the MIC for strain CS92 (10 ,g/ml). The effects of 5, 10, and 20 ytg of CAP per ml were determine d. A decreased level of CCC DNA was observe d even when a concentration as low as 5 ,ug of C"AP per ml was used, although increase in turb idity was only partially suppressed (data not shiown). Agarose gel electrophoresiis. The electrophoretic profiles of high-salt ext,racts of strains CS92 and CS40 denmonstrate tihe presence of four plasmids in CS92 (Fig. 4A-C') and two plasmids in CS40 (Fig. 4K). The ery thromycin-sensitive segregant of CS92 (CS93) ,isolated after

growth of CS92 in the presence of 0.5 jig of ethidium bromide per ml, demonstrates a loss of one plasmid (pCS12) (see Fig. 4F-H). The bands observed as a result of electrophoresis of highsalt extracts were comprised of CCC DNA representing distinct plasmid species, since the same profile was obtained consistently after electrophoresis of dye buoyant density-purified CCC DNA (data not shown). Plasmid designations and molecular weights, calculated from the distance traveled relative to that of standard DNAs, are listed in Table 1. To determine if all plasmids present in strains CS92, CS93, and CS40 respond to the same degree to the effects of CAP, the intensity of each plasmid band demonstrated in photographs of agarose gels containing plasmid DNA fronm CAP-treated cultures was compared to the band intensity in photographs of uninhibited control cultures. A representative gel is shown in Fig. 4. Figure 4A, B, and C shows the electrophoretic patterns of plasmids from lysates of control samples of strain CS92 obtained at zero time, 2 h, and 24 h, respectively. Figure 4D and E illustrates the electrophoretic patterns of plasmids from lysates of strain CS92 after incubation for 2 and 24 h, respectively, in the presence of CAP. Figure 4F through J and Fig. 4K and L depict - parallel studies of strains CS93 and CS40, respectively. A loss of CCC DNA is evident upon incubation of all three strains with CAP. Moreover, loss of CCC DNA within a strain was uniform among all plasmids harbored by the organisms. To determine if the loss of CCC DNA observed by dye buoyant density centrifugation resulted in a concomitant loss of plasmid-borne genetic information, the frequency of loss of the FI(;. :3. Effect of a chose in(ubati!on period in the erythromycin resistance determinant borne on presence of CAP on the amouniot of laobele( CC DNA pCS12 was measured for both control and CAPin strain CS92. After growth for sec eral generati(ons treated samples of strain CS92. The curing frein the pr-esence of 10 0Ci of [f(H/thviy idine per ml, the quency was determined after 2 h in the presence was washed and resuspendled in fiesh me- of CAP and again after five serial passages by cultur'-e diumi without isotope to remrove rua in brain heart infusion in the sor. At the time of removtl of the racliouctueprecur- overnight growth of absence CAP to remove anv error in measor- (zer-o time,), CAP was added to a Port thc por tion ofo f surement due to the tendency of streptococci to culture. U'pon conttiniie d in ciiubation, both control ((X) aind CAP-treated (4*) cultures wvere chain. The curing frequencies obtained by these remo(ed at timed intervials and assa yed by dye buoy- experiments are presented in Table 2. No inant density centrifugation for the ,wresence of CCC crease in segregation events was observed, even DNA. I)ashed lines indicate acid-preecipitable counts after multiple passages in the absence of CAP. pr-esent inl chromosomal DNA of 'ontrol (OI) and It is possible that plasmid transfer within the CAP- treated (-) culltUrI'es thr-oughoiilt the period of culture obscured any segregation which did octhe chase. At 120 mnin, a portion of Ithe CAP-treated cur. However, repeated attempts to demonstrate saniple ia5s iwashed and resuspend(ed in fiesh ediuni to remolve CAP, and the incuba tlon wxas contin- transfer of pCS12 to several recipients have been unsuccessful (data not shown). ued with further monitoring (A). AcUztj"acil[CItt ctn "/ tillC apparent number of counts present ir7 CCC DNA was DISCUSSION performed as described to correct foir loss of label as a result of experimental man ipulatio n. The effect of the inhibition of protein synthe-

Iloa(tiue pre(cr20o sioples

tiomnt was

m'

PLASMID MAINTENANCE IN S. FAECALIS

VOL. 139, 1979

821

FIG. 4. Agarose gel analysis of the effects of CAP on plasmids contained in strains CS92, CS93, and CS40. High-salt extracts of designated samples were dialyzed against TES overnight at 4°C and layered onto 0.7%o vertical agarose gels. Electrophoresis was performed at room temperature for 4 h at 5 V/cm. The contents of each lane are: (A, B, C) 0-, 2-, and 24-h control samples of CS92; (D, E) 2- and 24-h incubation of CS92 in the presence of CAP; (F, G, H) 0-, 2-, and 24-h control samples of CS93; (I, J) 2- and 24-h incubation of CS93 with CAP; (K) zero-time control of CS40; (L) 2-h incubation of CS40 with CAP. The band designated by the arrow is comprised of the linear chromosomal fragments which contaminate Hirt extracts. TABLE 1. Plasmids contained in strains CS92 and CS40 with their corresponding molecular masses Molecular massa

(megadaltons) CS92

pCS10 pCSll pCS12 pCS13

30 17 11 9

2.8 0.9 Molecular mass determinations were obtained by comparing the distance traveled by plasmid unknowns to that of pSC101 and ColEl standard DNA (10).

CS40

pCS14

pCS15

sis on plasmid copy number in two strains of streptococci seems to differ from the gram-negative systems previously studied. Upon inhibition of protein synthesis, the plasmid-to-chromosome ratio has been shown to be conserved ("stringent control") in both an F plasmid (9) and an R plasmid (11) of Escherichia coli. In these systems, the protein synthesis inhibitor

TABLE 2. Effects of CAP treatment on the frequency of curing of erythromycin resistance in CS92 No. of Noof % SegresegreSerial colonies olof passage tested gants iso- gation" lated 2 1 0.04 4,400 Control 18 0.5 5 3,400 Control 2 0.2 1 880 CAP treated 3 0.3 980 5 CAP treated " The number of erythromycin-sensitive segregants was scored after 2 h of incubation with or without CAP, as indicated, and again after subculturing for 5 successive days in fresh brain heart infusion without CAP.

Sample

CAP interrupts DNA synthesis of both plasmid and chromosome, indicating not only coupling of DNA synthesis with protein synthesis, but also coupling of plasmid replication to chromosomal replication. For example, label shift experiments of plasmid NRl in E. coli demon-

822

MURPHEY-CORB AND MtURRAY

strate an interruption of plasmid synthesis upon the introduction of CAP to the medium. Plasmid DNA replicated before CAP treatment, however, remains as a stable component of the culture (11). It has also been shown that an increase in plasmid DNA may occur in the absence of protein synthesis. One studied example of this "relaxed" mode of replication is that of the plasmid ColE1, where DNA synthesis can proceed for many rounds after the cessation of protein synthesis (2). These modes of plasmid replication are further established by similar studies with the plasmids R6K and R28K by Arai and Clowes

,J. BACTF,RIOL,.

was not lost from the cell, but rather was altered in a manner which rendered it undetectable by

dye buoyant density centrifugation. It should be noted here that these results do not rule out the possibility that the observed return of radiolabel during the postwash incubation was due to reutilization of radioactive DNA degradation products produced as a consequence of CAP treatment. However, we feel this is unlikely since we can find no published precedence for this phenomenon.

(iv) Preferential binding of plasmid DNA to another substance. Inhibition of protein

synthesis could conceivably alter the binding properties of either the plasmid DNA or some Our streptococcal strains, on the other hand, unknown substance. Binding of this substance demonstrate a three- to sixfold decrease in the to plasmid DNA could have resulted in a complasmid-to-chromosome ratio due to a reduction plex which was undetectable in our measurein plasmid copy number when measured by dye ments. Formation of the complex after the adbuoyant density centrifugation. A loss of CCC dition of CAP would account for failure to reDNA was also evident from electrophoretic pro- cover the DNA at the buoyant density expected files of high-salt extracts of lysates. Several con- of the CCC form and the absence of electropho(1).

ditions which would account for the apparent loss of CCC DNA upon inhibition of protein synthesis are discussed below. (i) Integration of the plasmid genome into the host chromosome. Integration seems improbable in that both strains CS92 and CS40 are multiplasmid organisms; simultaneous integration of all plasmids within a strain is unlikely. (ii) Alteration of the supercoiled circular structure of the plasmid to another form. The introduction of a single-stranded nick into the closed duplex (producing a relaxed open circle) or a double-stranded break (producing a linear molecule) would alter the plasmid buoyant density to that of the chromosomal linear fragments. Consequently, the plasmid DNA would be indistinguishable from the chromosome in dye buoyant density analysis. However, the linear and relaxed circular forms of lowmolecular-weight plasmids are clearly resolved from the supercoiled circular duplex forms after electrophoretic migration through 0.7% agarose (8). In this regard, gel analysis of the effect of CAP on the plasmids in CS40 is of interest since any alteration in form of these plasmids would be evident. Gel studies of CS40 demonstrated no conversion of the supercoiled duplex of either plasmid to another form (Fig. 4L). (iii) Loss of plasmid DNA from the cell. Loss of plasmid DNA from the cell is unlikely since an increase in erythromycin-sensitive variants of strain CS92 would be expected in the presence of CAP. This did not occur. This result, coupled with the observation of a return of radiolabel in the form of CCC DNA upon removal of CAP (Fig. 3), suggests that the plasmid DNA

retic bands identified with nicked circular or linear forms of plasmid molecules. Binding of DNA to cellular components has also recently been suggested by Womble and Rownd to explain their inability to recover putative plasmid replicative intermediates in the form of CCC DNA (11). We have demonstrated a reduction in the CCC form of plasmid DNA in S. faecalis as a result of the interrruption of protein synthesis. Although some discussion concerning the nature of the mechanism responsible is presented, further study is necessary to define the governing molecular interactions involved. ACKNOWLEDGMENTS This work was supp)orted hy a Hibernia National Banik of New Orleans-Amiierican Heart Association-Louisiana, Inc. Research Award antd hb P'tublic Health Service Biomedical Research Support froni the National InstitUtes of Health to the I-S.1i. School of Dentistry. LITERATURE CITED

Arai, T., and R. Clowes. 1974. Replication of stringent and relaxed plasmids, p. 141-155. In D. Schlessinger (ed.), Microhiology-1974. Aniericain SocietY for Microbiology, Washington, D.C. 2. Clewell, D. B. 1972. Nature of Col El plasmid replication 1.

in Escherichia coli in the presence of chloramphenicol. J. Bacteriol. 110:667-676. :3. Clewell, D. B., Y. Yagi, G. M. Dunny, and S. K. Schultz. 1974. Characterization of three ptlasmid deoxyribonucleic acid molecules in a strain of Streptococcus faecalis: identification of a plasmid determining erythromycin resistance. J. Bacteriol. 117:283-289. 4. Freifelder, D., and A. G. Braun. 1976. A mtultiple sample precipitator. Anal. Biochem. 74:154-159. 5. Grossman, L. T., R. Watson, and J. Vinograd. 1974. Restricted uptake of ethidium bromide and propidium diiodide by denatured closed circular DNA in buoyant cesium chloride. J. Mol. Biol. 86:271-283.

VOL. 139, 1979

PLASMID MAINTENANCE IN S. FAECALIS

6. Harkess, N. K., and M. L. Murray. 1978. Restriction

plasmids from Haemophilus influAgents Chemother. 13:802-808. 7. Hirt, B. 1967. Selective extraction of polyoma DNA from infected mouse cell cultures. J. Mol. Biol. 26:365-369. 8. Johnson, P. H., and L. I. Grossman. 1977. Electrophoresis of DNA in agarose gels. Optimizing separations of conformational isomers of double and single-stranded DNAs. Biochemistry 16:4217-4225. 9. Kline, B. C. 1974. Mechanisms and biosynthetic requireenzyme analysis of enzae. Antimicrob.

823

for F plasmid replication in E. coli. Biochemistry 13:139-146. 10. Meyers, J. A., D. Sanchez, L. P. Elwell, and S. Falkow. 1976. Simple agarose gel electrophoresis method for identification and characterization of plasmid deoxyribonucleic acid. J. Bacteriol. 127:1529-1537. 11. Womble, D. D., and R. H. Rownd. 1979. Effects of chloramphenicol and rifampicin on the replication of R plasmid NR1 deoxyribonucleic acid in E. coli. Plasmid 2:79-94. ments