Molec. gen. Genet. 143, 311-318 (1976) © by Springer-Verlag 1976

Effects of Different Alleles of the E. coli K12 polA Gene on the Replication of Non-transferring Plasmids Nigel D.F. Grindley 1 and William S. Kelley 2 Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 06510 USA 2 Department of Biological Sciences, Carnegie-Mellon University, Pittsburgh, Pa. 15213, USA

Summary. The effects of eight different polA- alleles

on the replication of six different non-transferring enterobacterial plasmids have been tested. Using phage P1CM transduction, different allelic poIAmutations were introduced into E. coli K12 strains carrying one of several antibiotic resistance plasmids. Plasmid stability in the transductants was examined by testing clones for drug resistance after growth under various conditions. From the results, the R factors may be divided into three different classes. One plasmid is only affected by PolA- conditions which inhibit host cell growth, three plasmids (from the same compatibility group) are unstable under conditions in which the cells are severely deficient in DNA polymerase I and two other plasmids (compatible with each other and with the other four) are immediately lost from such transductants and are unstable in a number of others. Furthermore, the plasmids which are most dependent on D N A polymerase I have been shown to replicate in the presence of chloramphenicol and therefore typify a class ofplasmids which includes bacteriocinogenic factors such as Cole 1 and CloD F 13, resistance determinant RSF1030 and the E. coli 15 minicircular plasmid.

Introduction

Enterobacterial plasmids may be divided into two functional groups based on their conjugational transfer properties: a) Transfer factors or conjugative plasmids-factors which mediate their own transfer from cell to cell. b) Non-transferring or non-conjugative plasmids-plasmids which can only be transferred from strains harbouring a transfer factor. In general, members of the second group are rela-

tively small, having weights of less than 10 7 daltons and are found in multiple copies per chromosome equivalent in the host cell (Clowes, 1972). A number of previous reports have indicated that DNA polymerase I is necessary for the replication of several different non-transferring plasmids. Kingsbury and Helinski (1970) have shown that a colicinogenic plasmid, ColE1, is not replicated in E. coli K12 carrying the polA1 allele. E. coli strains with this mutation make about one percent of the DNA polymerase I found in the corresponding wild type cells (de Lucia and Cairns, 1969; Lehman and Chien, 1973). Using this property of ColE1, Kingsbury and Helinski (1973a, b) were able to isolate a further (temperature sensitive) polA- mutant of E. coli, poIA ts214, in which the plasmid replicates at permissive temperatures b u t from which it is lost at non-permissive temperatures. Similar dependency on DNA polymerase I has been reported by Goebel and Schrempf (1972a) in experiments with the minicircular plasmid of E. coli 15 and by Veltkamp and Nijkamp (1973) with the cloacinogenic factor CloDF13. The latter authors have also shown that the CloDF13 plasmid continues to replicate in the presence of the polA'107 mutant allele (Glickman et al., 1973) which has been shown to be deficient in the Y--, 3' exonuclease activity of DNA polymeraseI (Heijneker etal., 1973). Although a number of investigators have examined the replication of conjugative plasmids in polA- strains (e.g., Kingsbury and Helinski, 1970) in these cases no requirement for DNA polymerase I has been demonstrated. We have investigated the maintenance and replication of a number of nontransferring drug resistance plasmids in a series of E. coli strains differing in their mutantpolA alleles. Our results suggest that these plasmids fall into three groups: a) Those with no demonstrated requirement for DNA polymerase I, except under conditions when host replication is also inhibited.

N.D.F. Grindley and W.S. Kelley: Plasmid Replication inpolA- Mutants

312

b) Those which require DNA polymerase I for their replication. c) Those which do not show a requirement for this enzyme for replication but which are subject to aberrant segregation in its absence. In addition we have demonstrated that plasmids of the second type (requiring DNA polymerase I) continue to replicate in pop E. coli strains when chromosomal D N A replication is inhibited due to the blockage of protein synthesis by chloramphenicol. Materials and Methods Non-transferring Plasmids The plasmids used in this study are listed in Table 1. All were obtained from Dr. E.S. Anderson and have been extensively characTable 1. Non-transferring enterobacterial plasmids Plasmid No.

Drug resistance

References

NTP1

A

Anderson and Lewis (1965 a, b)

NTP2

SSu

Anderson and Lewis (1965 a, b)

NTP3

ASu

Anderson et al. (1968)

NTP5

T

Smith, Humphreys and Anderson (1974)

NTP7

ASSu

Smith, Itumphreys and Anderson (1974)

NTP1 i

K

H.R. Smith and E.S. Anderson, unpublished observations

Symbols for plasmid mediated drug resistances: A=ampicillin, K = kanamycin, S = streptomycin, Su = sulphonamides, T = tetracycline.

terized in his laboratory (see Smith, Humphreys and Anderson, 1974). All are small (5 10 x 106 daltons), exist as multiple copies per host chromosome and are unable to mediate their own transfer but can be transferred from donor strains carrying a suitable transfer factor (conjugative plasmid). NTP2, NTP3 and NTP7 are members of the same compatibility group; the remainder are compatible with these three and with each other and with the colicinogenic plasmids ColE1, ColE2 and ColE3 (Smith et al., 1974; H.R. Smith and E.S. Anderson, unpublished observations).

Bacterial Strains and Phage The bacterial strains are listed in Table 2. All the studies of plasmid stability and replication were performed using the W3110 derivatives CM1062 or CM3643 as host strains. The plasmids were introduced into these pop hosts by transduction with phage P1CM (Rosner, 1972) or by conjugation mediated by the transfer factor T-Adrpl (Grindley, Grindley and Anderson, 1972). The polA- strains were obtained from the individual sources listed in Table 2. Phenotypic properties of these mutants and the enzymatic defects of DNA polymerase I produced by them are summarized in Table 3. It should be noted that the conditional lethal polA mutants polAexl and polABT4113 do grow at the non-permissive temperatures in liquid culture, although more slowly than thepolA + control. This behavior is fully compatible with the conditional lethal phenotype as reported by Konrad and Lehman (1974) and most survivors of growth at the non-permissive temperature still display the representative MMS sensitive phenotype characteristic of these mutant alleles. In the course of our experiments we also have found that polAts214 strains grown at 32° have a low viability when plated on nutrient medium at high temperatures (plating efficiency is about 0.1% at 42° C). This behavior has also been observed by D.J. Sherratt (personal communication) and appears to be a standard phenotypic trait of such strains. When a broth-grown culture of such cells (grown at 32° C) is diluted 1 : 2 x l0 s in the same medium and incubated at 42 ° C, growth to the original turbidity takes about twenty-four hours compared with eight hours for a polA + control. Thus, the phenotype of the polAts214 strains is very similar to that ofpolABT4113 andpolAexl strains. The latter twopolA - alleles

Table 2. Escherichia coli strains Reference No.

Description"

Source

(a) Host strains CM1062

K12F- thyA

metE- rha- lac- Str~

CM3643

K12F- thyA metE- rha- lac- Spcr Str S

(b) Pol- strains DS694 CM3634 JG111 MM383 DS624 RS5064 BT4113 KMBL1789

K12FK12FK12FK12F K12FK12FK12FK12F-

thyA- rha lac- StffpolA1 thyA- metE- lac- StrrpolA5 thyA metE- cys- polA6 thyA- rha- lae- Strr polA12 thyA- rha lac- Strr polAts214 trpA- polAexl thyA gua polABT4113 thyA- argA bio- pheA- endA- polA'107

" Gene symbols are from Taylor and Trotter (1972).

Derivative of W3110 from J. Wechsler Transduction of CM 1062 to resistance to spectinomycin

D.J. Sherratt M. Peacey J.D. Gross M. Peacey D.J. Sherratt I.R. Lehman B.M. Olivera H.L. Heijneker

N.D.F. Grindley and W.S. Kelley: Plasmid Replication inpolA

313

Mutants

Table 3. Enzymatic defects in D N A polymerase I from polA- mutants Allele

5' ~ 3' exonuclease

Polymerase

Phenotype

Reference

polAl

Approximately normal levels

Premature termination results in _< 1% of wild type levels

MMS ~, UVL viable at all temperatures

Lehman and Chien (1973)

polA5

Greater than normal levels

Altered pH optimum, altered template specificity, lower overall activity

MMS~,UV ~, viable at all temperatures

Lehman and Chien (1973)

polA6

Normal or greater than normal

Altered pH optimum, altered template specificity, thermolabile in vitro

MMS ~, UV ~, viable at alI temperatures

Kelley and Whitfield (1971); W.S. Kelley and N.D.F. Grindley, unpublished

polA12

Approximately normal levels

Altered pH optimum, sensitive to low salt conditions, thermolabile in vitro

M M S S - m o r e sensitive at 42 ° C than at 30 ° C, viable at all temperatures, conditional lethal in combination with recA or recB

Monk and Kinross (1972); Lehman and Chien (1973)

polAts214

No data reported

Thermolabile in in vitro crude extract assays

MMS r at 32 ° C MMS s at 42 ° C, low viability at 42 ° C (see text)

Kingsbury and Helinski (1973b); D.J. Sherratt, personal communication

polAexl

Low levels

Lowered levels, thermolabile in vitro

MMS ~, low viability when plated on rich medium at 43 ° C

Konrad and Lehman (1974)

polABT4113 Thermolabile

Lowered levels, thermolabile in vitro

MMS s, low viability when plated on rich medium at 45°C

Olivera and Bonhoeffer (1974)

polA'107

Normal levels

MMS ~, viable at all temperatures

Heijneker et al. (1973)

Low Ievels

produce D N A polymerase I molecules defective in 5' ~ 3' exonucleolytic activity; at present the intragenic site of the polAts214 mutation is not known.

Meda Bacterial strains were generally grown in L broth (Lennox, 1955). Solid medium was prepared by adding 13 g/liter of Difco Bacto Agar to L broth. Transductants were selected on M9 minimal medium containing glucose or rhamnose (0.2%) as carbon source and supplemented with thymine (50 gg/ml) and L-amino acids (20 ~tg/ml) as required. Liquid minimal medium used for investigating plasmid replication during chloramphenicol treatment was as described by Neidhardt, Bloch and Smith (1974).

Stability of Plasmids in polA- Strains All transductants were examined for the presence of the particular plasmid by replica plating for plasmid-conferred resistance to suitable antibiotics. IfpolA- transductants carried the plasmid, its stable inheritance during growth in drug-free liquid medium (L broth) was examined. This was done at 37 ° C for the polA1, polA5, polA6 and polA'107 mutations. Cultures of the temperature-conditional mutants were initially grown at the permissive temperature to a turbidity equivalent to 5 x 108 cells/ml, then diluted 1:2 x 105 for polA12 and polAts214 or 1 : 500 for polAexl and polABT4113 alleles, and grown at the non-permissive temperature to the same turbidity. PolA + controls were always grown in parallel. After growth the liquid cultures were plated for single colonies (at the permissive temperatures) and these were tested for plasmid-encoded drug resistance.

Preparation of Cell Lysates for Density Gradient Centrifugation Transduction of polA Alleles The heat-inducible lysogenic phage P1CM was obtained from Dr. J.L. Rosner and was used for transduction of the mutant alleles from their original genetic background into the plasmid bearing derivatives of CM1062 or CM3643. Transductions were performed at 32 ° in 0.01 M MgSO 4, 0.005 M CaC12 for thirty minutes using a multiplicity of infection of approximately 1. Met + or Rha + transductants were selected by plating on suitably supplemented M9 minimal medium. Nonlysogenic transductants were purified by streaking on L-agar or the selective medium and scored for co-transduction of the polA allele by replicating patches onto L-agar plates containing 0.02% MMS. Replicas were incubated at 37 ° C, 42°C or 45°C depending on the polA mutation being tested.

Cells labelled with thymidine-methyl-3H (Amersham/Searle Corporation) were lysed with Sarkosyl as described by Bazaral and Helinski (1968). Solutions for density gradients were prepared by dissolving 5.2 g of cesium chloride in 0.9 ml of lysate, 2.9 mI of water and 1.6 ml of ethidium bromide solution(700 ~tg/ml). Centrifugation was performed in nitrocellulose tubes in a Type 50 titanium rotor in a Spinco L3-50 ultracentrifuge at 36,000 rpm for 48 to 60 hours. Gradients were fractionated from the bottom and radioactivity of each fraction was assayed by sampling onto Whatman 3MM filter paper discs, washing the discs with 5% trichloroacetic acid, drying and counting in a liquid scintillation counter. Cleared lysates for sucrose gradient analysis were prepared by the method of Clewell and Helinski (1969), substituting five percent Triton X100 for the Brij58-deoxycholate detergent system.

314

N.D.F. Grindley and W.S. Kelley: Plasmid Replication

Results

Effect of polA Alleles on Stability of Non-transferring Plasmids The results of introducing the different poIA mutations into strains carrying the various plasmids are summarized in Table 4. Of the 6 plasmids, NTP1 and NTP11 are the most severely affected, being undetectable following transduction of polA1, polA5 or poIA6, and rapidly lost from strains carrying the temperature sensitive mutations polA12, polAts214 or polABT4113 when grown at non-permissive temperatures. It appears that these two plasmids require substantial amounts of DNA polymerase I for their replication. This is demonstrated by their instability (particularly noticeable with N T P l l ) in polA12 strains at 28 ° Cconditions under which poIA12 recA double mutants have sufficient DNA polymerase activity to remain viable (Monk and Kinross, 1972). Strains carrying polA'107 which are therefore defective in the 5'--+ 3' exonucleolytic activity but retain the polymerizing activity of DNA polymerase I, did not lose NTP1, NTP11 or any of the other plasmids tested, suggesting that normal levels of exonucleolytic activity are not required for replication of these plasraids. In view of this result we suggest that the loss of N T P l l (but not of NTP1) from polAexl strains

T a b l e 4. Stability o f p l a s m i d s

PolA Allele

inpolA-

Mutants

is due to altered polymerizing activity or template specificity rather than the defective 5 ' ~ 3' exonucleolytic activity of these strains. The three plasmids from the SSu compatibility group, NTP2, 3 and 7 (Anderson et al., 1968; Smith et al., 1974), all appear to have some requirement for DNA polymerase I as they are not stably maintained by poIA1 strains. The rate of segregation of plasmid negative lines appears to be about 5% per generation. These plasmids are also lost at a low and rather variable rate frompolA12 strains grown at 45 ° C. No loss of plasmids from this group was observed from other polA mutant strains. The plasmid NTP5 was stably maintained by all the poIA- strains except those grown under conditions that strongly inhibit host cell growth, namely polAts214, polABT4113 and polAexl grown at 45 ° C. We checked that the polA1 (NTP5) lines contained normal amounts of NTP5 DNA by labelling the DNA with 3H-thymidine, preparing cleared lysates (Clewell and Helinski, 1969) and sedimenting the DNA through 15-50% sucrose gradients. The acid precipitable label recoveredinthe cleared lysates of the polA ÷ and polA1 strains was respectively 1.4% and 1.3% of the label in the whole lysate (compared with a mean recovery of 1.5% obtained by Smith et al., 1974). The sucrose gradients both gave single DNA bands of indistinguishable sedimentation coefficients (data not shown).

strains

Replication of NTP1 and NTP2 in Strains Producing a ThermolabiIeDNA Polymerase

Ternperature

NTP1

NTP2

(°c)

(A)

(SSu) (ASu) (ASSu)(T)

(K)

polA1 polA6

37

-

u

u

u

+

-

37

-

+

+

+

+

-

polA5

37

-

+

nt

nt

+

-

polA12

28 45

u u

+ u

+ u

+ nt

+ +

u nt

polAts214

32 42

+ u

+ +

+ +

+ +

+ +~

+ u

polABT4113

32 45

+ u

+ +

nt

nt

+ u

+ u

polAexl

32 45

+ +

+ +

nt

+ +

+ u

-

polA'107

37

+

+

+

nt

+

+

-=polA-

inpolA

Plasmid NTP3

NTP7

NTP5

NTPll

t r a n s d u c t a n t s w e r e all p l a s m i d n e g a t i v e ; + =polAt r a n s d u c t a n t s w e r e all p l a s m i d positive a n d n o loss o f t h e p l a s m i d w a s o b s e r v e d a f t e r g r o w t h in l i q u i d c u l t u r e a t the i n d i c a t e d t e m p e r a t u r e ; u (=unstable)=mostpolA- t r a n s d u c t a n t s w e r e p l a s m l d positive b u t g a v e rise to p l a s m i d n e g a t i v e cells d u r i n g g r o w t h in liquid c u l t u r e at the i n d i c a t e d t e m p e r a t u r e ; nt = n o t tested. a U n s t a b l e at 45 ° C.

The instability of NTP1 (and NTP11) in the various

polA strains and of the SSu group of plasmids in polA1 lines could be caused by a reduction in plasmid DNA replication in cells with limiting amounts of DNA polymerase I, or the enzyme could be involved in a step subsequent to polymerization but essential for normal plasmid segregation at cell division. To distinguish between these possibilities we examined the replication of NTP1 and 2 in polAts214 lines. Dye buoyant density gradients of DNA from these strains labelled with 3H-thymidine at 32° C or after a shift to 45 ° C are shown in Fig. 1. The yields of plasmid DNA obtained from these gradients and from polA + controls are shown in Table 5. Formation of covalently closed circular NTP1 DNA (the band of greater density) is greatly inhibited in the polAts214 strain within 10 min of the shift to 45 ° C while the yield of plasmid DNA at 32° C is normal (5.5% of total 3H-label recovered-see Table 5). Moreover analysis by neutral sucrose gradient sedimentation of cleared lysates of the poIAts214

N.D.F. Grindley and W.S, Kelley: Plasmid Replication inpolA- M u t a n t s

315

a (.~

40

<~ J Q. Or)

B

A NTP1 32°C

NTP1 45°C

)_o

Z CI

15

2O

J

IO

1.0

S 10

0.5

I-z

C, -4 r-

C N T P P- 5Z°C

g 1.O(

•

410

0.5 50TOP 50TOP 80"11 I 0 20 BOTTOM 10 20 FRACTION NUMBER FRACTION NUMBER

Fig. 1. Replication of NTP1 and N T P 2 in K12 polAts214 at 32 ° C and 45 ° C. ThepolAts214 derivatives o f C M 1 0 6 2 (NTP1) or CM3643 (NTP2) were grown in L broth at 32 ° C. At a turbidity of about 108 cells/ml, 5 ml portions were shifted to 45 ° C or maintained at 32 ° C. After one generation (or 10 min for NTP1 at 45 ° C), 3Hthymidine (2 pCi/ml) was added to each culture. Incubation was continued for a further 20 min, then the cells were harvested, lysed with sarkosyl and run in CsCl-ethidium bromide gradients. The resulting profiles are of polAts214 (NTP1) labelled at (A) 32 ° C, (B) 45 ° C and ofpolAts214 (NTP2) labelled at (C) 32 ° C, (D) 45 ° C. (Note the change of scale at about fraction 20)

Table 5. Recovery of plasmid D N A from E. coli polA +,polAts214, and polA12 strains at 32 ° C a n d 45 ° C Plasmid

PolA Allele

% total label recovered in plasmid peak 32 ° C

NTP 1 NTP1 NTP2 NTP2 NTP2

polA + polAts214 polA + polAts214 polA12

45 ° C

4.1

7.6

5.5 9.3 7.2 9.8

0.6" 6.9 7.3 7.1

Growth, labelling and density gradients were performed as described in the legend to Fig. 1. This estimate is rendered inaccurate by the small size of the plasmid peak compared to the " b a c k g r o u n d " counts recovered in the region below the chromosomal peak (see Fig. 1).

(NTP1) strain labelled at non-permissive conditions showed that there was no accumulation of open circular plasmid molecules. By contrast, the yield of closed circular NTP2 DNA from the polAts214 host strain was unaffected after one generation at 45 ° C. Similarly NTP2 continued to replicate normally in a polA12

I-Z bJ (.> n," bJ 13..

O

.

~

0 5 10 15 HOURS A FTER ADDITION OF CHLORAMPHENICOL

Fig. 2. Replication of plasmid D N A in the presence of chtoramphenicol. E. coli thyA polA ÷ strains containing NTP1 o--Q;NTP2oo;NTP5×--×;orNTPll---Iweregrown at 37 ° C in 40 ml of minimal m e d i u m containing 3H-thymine (20 pg/ ml, 0.4 pCi/~tg). W h e n the culture reached a turbidity of about l0 s cells/ml, a 5 ml sample was removed, lysed and prepared for cesium chloride-ethidium bromide density gradient centrifugation. At the same time, chloramphenicol was added to the remainder to a final concentration of 100 pg/ml and incubation was continued at 37 ° C. Further samples were removed and prepared for centrifugation at various times after addition of chloramphenicoI. The a m o u n t of closed circular plasmid D N A was determined from the dye buoyant density gradients

host after one generation at non-permissive temperature (see Table 5).

Plasmid Replication in the Absence of Protein Synthesis The amounts of closed circular plasmid DNA recovered from cells at various times after addition of chloramphenicol to exponentially growing cultures are shown in Fig. 2. It should be noted that chromosome replication continues at a decreasing rate for about three hours after addition of chloramphenicol giving an increase in DNA of about 60% to 70%. NTP1 and 11 clearly continue to be replicated in the presence of chloramphenicol for several hours and long after the cessation of chromosome replication. With the plasmids NTP2 and 5, however, there is little or no detectable increase in the proportion of plasmid DNA. Discussion

When discussing the role of DNA polymerase I in plasmid replication one must consider both the distinct enzymatic functions of the polymerase and the complex series of steps by which covalently closed circular plasmid molecules must be replicated. By determining the effects of different mutant DNA polymerase I species on plasmid stability we hoped to distinguish between different possible modes of plasmid replication.

316

N.D.F. Grindley and W.S. Kelley: Plasmid Replication inpolA Mutants

The E. colipolA mutants which we have used may be broadly classified into three groups (see Table 3 for references): 1. Mutants defective only in polymerizing ability. This group includes strains with the mutations polA1, polA5 and poIA6 and the temperature conditional polA12, all of which produce normal levels of 5 ' ~ 3' exonuclease and either drastically reduced amounts ofpolymerizing activity or polymerase molecules with altered polymerizing activity. A second conditional mutant, polAts214, may also belong to this category, but no data describing its exonucleolytic activities are available. 2. Mutants defective only in 5 ' ~ 3' exonuclease activity. This group is defined by the mutant allele polA'107. DNA polymerase I produced by E. coli bearing this mutation is functionally normal when assayed for polymerizing activity in vitro. • 3. Mutants defective in both polymerizing and 5<-+ 3' exonycleolytic activities. Strains with the conditional-lethal mutations polAexl and polABT4113 are of this group. These two mutations presumably lie within the portion of the polA gene encoding the amino-terminal end of the DNA polymerase I molecule which includes the active site of the 5 ' ~ 3' exonuclease. In addition to causing loss of the exonucleolytic activity, the mutations cause further alterations of configuration in the enzyme resulting in altered polymerizing activity. The combination of these effects apparently causes the conditional lethal phenotype because of defects in chromosomal DNA replication. More rigorous definition of the differences between the mutant enzymes is difficult due to a paucity of comparable biochemical data and the acknowledged hazards of correlating in vivo response with in vitro activity. In discussing the mechanism of replication of small non-transferring R factors a number of analogies may be drawn between them and the bacteriocinogenic factors ColE1 and CloDF13. Both of these plasmids are of the same approximate size as the non-transferring R factors, occur in multiple copies per chromosome and are dependent on DNA polymerase I for their replication (Kingsbury and Helinski, 1970; Veltkamp and Nijkamp, 1973 respectively). The replication of both plasmids has been examined in polC- (DNA polymerase III deficient) strains of E. coli. When growing cultures of plasmid-containing polC strains are shifted to non-permissive temperatures, plasmid replication continues for some time but gradually diminishes, and ceases within three hours (Goebel, 1972; Collins, Williams and Helinski, 1975; Veltkamp and Nijkamp, 1973). From these results Collins et al.

(1975) suggested that DNA polymerase III is required for some stage of normal ColE 1 replication. However, neither ColE1 nor CloDF13 can be propagated in polA1 strains ofE. coli and synthesis of these plasmids stops abruptly in polAts214 strains following a shift to non-permissive temperature (Durkacz and Sheratt, 1973; Kingsbury and Helinski, 1973; Veltkamp and Nijkamp, 1974). Since DNA polymerase I has multiple enzymatic functions, different roles in chromosomal DNA synthesis have been ascribed to it. Because of the low plating efficiency of mutant strains defective in both 5<--,3, exonuclease and polymerase functions, DNA polymerase I's role in chromosome replication has been postulated to be one of excising the RNA primers and concomittantly extending the DNA chain of Okazaki fragments (Sugino, Hirose and Okazaki, 1972). Clearly the role of this enzyme is more extensive in ColE1 and CloDF13 synthesis since mutant alleles such aspoIA1 produce normal levels of 5 ' ~ 3' exonuclease but do not have the polymerizing capacity to replicate such plasmids. The data indicating the effects of the various polA alleles on plasmid replication are summarized in Table 4. From these results we can derive a number of conclusions about the role of DNA polymerase I in plasmid replication. As presented here, the plasmids fall into three distinct classes: NTP1 and NTPll. As with ColE1 and CloDF13, these plasmids fail to replicate in strains lacking substantial amounts of DNA polymerase I. The most dramatic effects are found with polA1, polA5 and polA6 strains in which the plasmids cannot be propagated at all. The thermolabile polymerase produced by polA12 replicates the plasmids poorly even at a "permissive" temperature and they are lost from polAts214 strains at non-permissive temperatures. 3Hthymidine labelling experiments indicate that plasmid synthesis is virtually nil under the latter conditions although replication is normal in pol + strains at these temperatures. We were unable to demonstrate any accumulation of newly synthesized open circles or catenated forms, supporting the conclusion that DNA polymerase I is the polymerizing enzyme for this class of plasmids. This is consistent with the experiments of others (Goebel, 1972; Collins etal., 1975) which show no accumulation of intermediate forms of ColE 1 in polAts214 strains at non-permissive temperatures. Their experiments do show the accumulation of multimeric forms of ColE1 inpolC strains at non-permissive temperatures, suggesting that DNA polymerase III may be involved in a maturation step subsequent to polymerization. The stability of NTP1 in polA'107 and polAexl strains suggests that the 5'-+3' exonucleolytic activity of polymerase I is not essential for replication of this plasmid, or may be supplied

N.D.F. Grindley and W.S. Kelley: Plasmid Replication inpolA- Mutants

by other cellular enzymes such as DNA polymerase III (Livingston and Richardson, 1975). NTP11 is also stably maintained in poIA'107 strains but is not replicated in E. colipoIAexl. This could be due to a greater sensitivity of N T P l l to the altered polymerizing activity ofpolAexl, or a low level of the 5 ' ~ 3 ' exonuclease activity might also be required. Both NTP1 and 11 are lost from polABT4113 strains at non-permissive temperatures but, as such strains are greatly deficient in polymerizing activity at high temperatures, this result may reflect only the requirement for this activity. NTP2, NTP3 and NTP7. The replication of NTP2 is apparently normal in both polA12 and poIAts214 strains at the non-permissive temperature, indicating that these strains contain adequate levels of polymerizing enzymes for replication of the plasmid. Although our results suggest that DNA polymerase I is not the polymerizing enzyme for the plasmids of this group, we do observe a gradual loss from polA1 strains, and from E. colipolA12 at elevated temperatures. This indicates that a low level of DNA polymerase I activity (other than 5'--+ 3' exonucleolytic activity) is required for normal plasmid maintenance although the role of the enzyme is not clear. NTP5. The replication of NTP5 appears to be very similar to chromosomal replication and only requires levels of DNA polymerase I activity necessary to ensure cell viability. The only other non-transferring plasmid known to have no requirement for DNA polymerase I is pSC101 (Timmis, Cabello and Cohen, 1974) which also codes for tetracycline resistance. The latter plasmid was artificially synthesized from the sheared D N A of the conjugative plasmid R6-5 (Cohen and Chang, 1973). Thus, NTP5 remains a unique naturally occurring plasmid. Supporting the division of plasmids into groups based on their polymerase requirements is the observation of NTP1 and NTP11 DNA replication in the presence of chloramphenicol. Again, this type of behavior is observed with the bacteriocinogenic plasmids ColE1 (Clewell and Helinski, 1972) and CloDF 13 (Veltkamp, Barendsen and Nijkamp, 1974) and also with the resistance determinant RSF 1030 (Crosa, Luttrop and Falkow, 1975) and the E. coli 15 minicircular plasmid (Messing, Staudenbauer and Hofschneider, 1972) all of which require DNA polymerase I for replication. By contrast, there is no increase in the plasmid/chromosome ratio with NTP2 or NTP5, the representative species of the other non-transferring plasmid groups, after addition of chloramphenicol. Obviously, the replication of the DNA of NTP1 and N T P l l requires only proteins which pre-exist in the cell and have a long half-life. This finding also indicates that the continued replication of these plasmids is not coupled to chromosomal DNA replication or cell division un-

317

der these experimental circumstances. The failure of NTP2 and NTP5 to continue replication in the presence ofchloramphenicol suggests that they are dependent on some labile protein(s) for their replication and/ or their replication is coupled to the cell cycle. Acknowledgements. This work was supported by U.S.P.H.S. Grant GM- 19374 to W. S.K. and by Allegheny County (Pa.) Health Research and Services Foundation Grants 0-89 to W.S.K. and R-85 to N.D.F.G.; the latter is a postdoctoral research fellow of the Science Research Council of Great Britain. We are grateful for gifts of strains from the individuals mentioned in the text. Numerous colleagues have given us helpful suggestions in the preparation of our manuscript. In particular, we wish to thank Drs. P.M.A. Broda, June N. Grindley, G.O. Humphries, D.J. Sherratt, H.R. Smith and N.S. Willetts. We also wish to acknowledge the assistance of Mr. Y.B. Mitchel in analyzing the gradient data.

References Anderson, E.S., Kelemen, M.V., Jones, C.M., Pitton, J.-S. : Study of the association of resistance to two drugs in a transferable determinant in Salmonella typhimurium. Genet. Res. 11, 119-124 (1968) Anderson, E.S., Lewis, M.J.: Drug resistance and its transfer in Salmonella typhimurium. Nature (Lond.) 206, 579 583 (1965a) Anderson, E.S., Lewis, M.J. : Characterization of a transfer factor associated with drug resistance in Salmonella typhimurium. Nature (Lond.) 208, 843 849 (1965b) Bazaral, M., Helinski, D.R. : Circular DNA forms of colicinogenic factors El, E2, and E3 from Escherichia coil J. molec. Biol. 36, 185-194 (1968) Clewell, D.B., Helinski, D.R.: Supercoiled circular DNA-protein complex in Eseherichia coli: Purification and induced conversion to an open circular form. Proc. nat. Acad. Sci. (Wash.) 62, 1159-1166 (1969) Clewell, D.B., Helinski, D.R.: Effect of growth conditions on the formation of supercoiled ColE1, deoxyribonucleic acid, and protein in Eseherichia coil J. Bact. 110, 1135-1146 (1972) Clowes, R.C. : Molecular structure of bacterial plasmids. Bact. Rev. 36, 361 405 (1972) Cohen, S.N., Chang, A.C.Y.: Recircularization and autonomous replication of a sheared R-factor DNA segment in Escherichia coli transformants. Proc. nat. Acad. Sci. (Wash.) 70, 1293 1297 (1973) Collins, J., Williams, P., Helinski, D.R. : Plasmid ColE1 DNA replication in Escherichia cob strains temperature-sensitive for DNA replication. Molec. gen. Genet. 136, 273-289 (1975) Crosa, J.H., Luttropp, L.K., Falkow, S. : Nature of R-factor replication in the presence of chloramphenicol. Proc. nat. Acad. Sci. (Wash.) 72, 654-658 (1975) de Lucia, P., Cairns, J. : Isolation of an E. coli strain with a mutation affecting DNA polymerase. Nature (Lond.) 224, 1164-1166 (1969) Durkacz, B.W., Sherratt, D.J. : Segregation kinetics of Colicinogenic Factor ColE1 from a bacterial population temperature sensitive for DNA polymerase I. Molec. gen. Genet. 121, 71 75 (1973) Glickman, B.W., van Sluis, C.A., Heijneker, H.L., R6rsch, A. : A mutant of Escherichia coli K 12 deficient in the 5 ' - 3 ' exonucteolytic activity of DNA polymerase I. I. General characterization. Molec. gen. Genet. 124, 69 82 (1973) Goebel, W. : Replication of the DNA of the Colicinogenic Factor E1 (ColE1) at the restrictive temperature in a DNA replication

318

N.D.F. Grindley and W.S. Kelley: Plasmid Replication inpolA- Mutants

mutant thermosensitive for DNA polymeraseIII. Nature (Loud.) New Biol. 237, 6%70 (1972) Goebel, W., Schrempf, H. : Replication of the minicircular DNA orE. coil 15 is dependent on DNA polymerase I but independent of DNA polymerase III. Biochem. biophys. Res. Commun. 49, 591--600 (1972a) Goebel, W., Schrempf, H. : Replication ofplasmid DNA in temperature-sensitive DNA replication mutants of Escherichia coli. Biochim. biophys. Acta (Amst.) 262, 32 41 (1972b) Grindley, N.D.F., Grindley, J.N., Anderson, E.S. : R-factor compatibility groups. Molec. gen. Genet. 119, 287-297 (1972) Heijneker, H.L., Ellens, D.J., Tjeerde, R.H., Glickman, B.W., van Dorp, B., Pouwels, P.H, : A mutant ofEscherichia coli K12 deficient in the 5 ' ~ 3' exonucleolytic activity of DNA polymerase I, II Purification and properties of the mutant enzyme. Molec. gen. Genet. 124, 83 96 (1973) Kelley, W.S., Whitfield, H.J. : Purification of an altered DNA polymerase from an E. coli strain with apol mutation. Nature (Lond.) 230, 33-36 (1971) Kingsbury, D.T., Helinski, D.R.: DNA polymerase as a requirement for the maintenance of the bacterial plasmid Colicinogenic factor El. Biochem. biophys. Res. Commnn. 41, 1538 1544 (1970) Kingsbury, D.T., Helinski, D.R.: Temperature sensitive mutants for the replication of plasmids in Escherichia coli I. Isolation and specificity of host and plasmid mutations. Genetics 74, 17-31 (1973a) Kingsbury, D.T., Helinski, D.R.: Temperature-sensitive mutants for the replication of plasmids in Escherichia coli: Requirement for deoxyribonucleic acid polymerase I in the replication of the plasmid ColE1. J. Bact. 114, 1116-1124 (1973b) Konrad, E.B., Lehman, I.R. : A conditional lethal mutant of Escherichia coli K 12 defective in the 5<+ 31 exonuclease associated with DNA polymerase I. Proc. nat. Acad. Sci. (Wash.) 71, 2048-2051 (1974) Lehman, I.R., Chien, J.R.: Persistence of deoxyribonucleic acid polymerase I and its 5 ' ~ 3' exonuclease activity in PotA mutants of Escherichia coil K12. J. biol. Chem. 248, 7717-7723 (1973) Lennox, E.S. : Transduction of linked genetic characters of the host by bacteriophage P1. Virology 1, 190-206 (1955) Livingston, D.M., Richardson, C.C. : DNA polymerase III of E. coli Characterization of associated exonuclease activities. J. biol. Chem. 250, 470-478 (1975)

Messing, J., Staudenbauer, W.L., Hofschneider, P.H.: Inhibition of minicircular DNA replication in E. coli 15 by rifampicin. Nature (Lond.) New Biol. 238, 202-203 (1972) Monk, M., Kinross, J.: Conditional lethality of recA and recB derivatives of a strain ofEseherichia coli K 12 with a temperaturesensitive deoxyribonucleic acid polymerase I. J. Bact. 109, 971978 (1972) Neidhardt, F.C., Bloch, P.L., Smith, D.F.: Culture medium for enterobacteria. J. Bact. 119, 736 747 (1974) Olivera, B.M., Bonhoeffer, F.: Replication of Escherichia coli requires DNA polymerase I. Nature (Loud.) 250, 513-514 (1974) Rosner, J.L.: Formation, induction, and curing of bacteriophage P1 lysogens. Virology 48, 679 689 (1972) Smith, H.R., Humphreys, G.O., Anderson, E.S. : Genetic and molecular characterization of some non-transferring plasmids. Molec. gen. Genet. 124, 224-242 (1974) Sugino, A., Hirose, S., Okazaki, R.: RNA-linked nascent DNA fragments in Escherichia coli. Proc. nat. Acad. Sci. (Wash.) 69, 1863-1867 (1972) Taylor, A i . , Trotter, D.C.: Linkage map of Escheriehia eoli strain K12. Bact. Rev. 36, 504-524 (1972) Timmis, K., Cabello, F., Cohen, S.N.: Utilization of two distinct modes of replication by a hybrid plasmid constructed in vitro from separate replicons. Proc. nat. Acad. Sci. (Wash.) 72, 22422246 (1975) Veltkamp, E., Barendsen, W., Nijkamp, H.J.: Influence of protein and ribonucleic acid synthesis on the replication of the bacteriocinogenic factor CloDF 13 in Escheriehia coli cells and minicells. J. Bact. 118, 165-174 (1974) Veltkamp, E., Nijkamp, H.J.J.: The role of DNA polymerase I, II and III in the replication of the bacteriocinogenic factor CloDF13. Molec. gen. Genet. 125, 329 340 (1973) Veltkamp, E., Nijkamp, H.J.J. : Replication of the bacteriocinogenic plasmid CIoDF13 in thermosensitive Escherichia coli mutants defective in initiation or elongation of deoxyribonucleic acid replication. J. Bact. 120, 1227-1237 (1974)

Communicated by B.A. Bridges Received August 12, 1975