J. Mol. Biol.

(1979) 127, 266-283

Gyrase-dependent Initiation of Bacteriophage T4 DNA Replication: Interactions of Escherichia coli Gyrase with Novobiocin, Coumermycin and Phage DNA-delay Gene Products DAVID

MCCAF~THY t

Molecular Biology Program Department of Microbiology College of Medicine University of Arizona Tucson, Ariz. 85724, U.S.A. (Received 16’ January

1978, and in revised form

19 September 1978)

Wild-type bacteriophage T4 and DNA-delay am mutants defective in genes 39, 52, 60 and 58-61 were tested for intracellular sensitivity to the antibiotics coumermycin and novobiocin, drugs which inhibit the DNA gyrase of Ewherichia coli. Treatment with these antibiotics drastically reduced the characteristic growth of gene 39,52 and 60 DNA-delay awa mutants in E. COG laoking an amber suppressor (ea-). Wild-type phage-infected cells were unaffected by the drugs while the burst size of a gene 58-61 mutant w&s affected to an intermediate extent. A su- E. coli strain which is resistant to coumermycin due to an altered gyrase permitted growth of the DNA-delay am mutants in the presence of the drug. Thus, the characteristic growth of the DNA-delay am mutants in an SUhost apparently depends on the host gyrase. An E. coli himB mutant is defective in the coumermycin-sensitive subunit of gyrase (H. I. Miller, personal communication). Growth of the gene 39, 52 and GO am mutants was inhibited in the himB mutant whilo the gene 58-61 mutant and wild-type T4 showed small reductions in burst size in this host. Experiments with nalidixic acid-sensitive and resistant strains of E. coli show that wild-type phage T4 requires a functional r&A protein for growth. Novobiocin and coumermycin inhibit phage DNA synthesis in DNA-delay mutant-infected su-E. co% if added during the early logarithmic phase of phage DNA synthesis. The gene 58-61 mutant showed the smallest inhibition of DNA synthesis in the presence of the drugs. Addition of the drugs during the late linear phase of phage DNA synthesis had no effect on further synthesis in DNA-delay mutant-infected cells. Coumermycin and novobiocin had no effect on DNA synthesis in wild-type-infected cells regardless of the time of addition of the antibiotics. Models are considered in which the DNA-delay gene products either form an autonomous phage gyrase or interact with the host gyrase and adapt it for proper initiation of phage DNA replication.

1. Introduction Amber mutants defective in bacteriophage T4 genes 39, 52, 60 and 58-61, the so-called DNA-delay genes, share a common phenotype characterized by slow rates of DNA synthesis in Escherichiic coli lacking an amber suppressor (au-) (Epstein t Present address : Biology 11973, U.S.A.

Department,

Brookhaven

National

Laboratory,

Upton,

N.Y.

266 0022-2836/79/0302a6-19

$02.00/0

6 1979 Aoademio

Press Ino. (London)

Ltd.

266

D. MCCARTHY

et al., 1963; Warner 8t Hobbs, 1967; Yegian et al., 1971). This is a reflection of a specific defect in DNA synthesis rather than the indirect result of defects in early protein or RNA synthesis (Yegian et al., 1971). McCarthy et al. (1976) found that a gene 52 am mutant showed normal rates of DNA elongation even while it had a slow rate of DNA accumulation in the cell. This observation led to the interpretation that the slow rate of DNA synthesis of the mutant is due to a low incidence of replicat,ive growing points per DNA template and that the DNA-delay mutants may be defective in the initiation of growing points. DNA-delay am mutants infecting su- E. coli characteristically produce a substantial burst of progeny phage (Mufti & Bernstein, 1974; Yegian et al., 1971) whereas am mutants defective in most other essential genes, under similar conditions, produce virtually no progeny phage. Translational ambiguity, which weakly suppresses most am mutations, cannot account for the levels of growth observed for mutants in the DNA-delay genes (Karam & O’Donnel, 1973). Mufti & Bernstein (1974) have shown that growth of the DNA-delay mutants on su- E. coli probably depends on a function provided by the host cell which is only partially effective in the mutant infections. Our results imply that the compensating host component is DNA gyrase. In vitro, E. c&i DNA gyrase introduces negative superhelioal turns into closed circular DNA of phage h, simian virus 40 virus, and the ColEl plasmid (Gellert et al.. 1976a). Gyrase has also been shown to play a role in the cell-free replication system of $X174 replicative form DNA (Marians et al., 1976) as well as the replication of phage T7 in vivo (Itoh t Tomizawa, 1977). Evidence is presented here indicating that the DNA-delay mutants of phage T4 require a functional host gyrase for growth and efficient initiation of replicative growing points in DNA.

2. Materials and Methods (a) Phuge &-a&w T$D(wild-type), r7l(gene rIIA), tsAl4(gene 41) and DNA-delay mutants amN116(gene 39), omE1240(gene 52), amNG576(gene 52), amE300(gene 60) and amHL627(gene 58-61) were obtained from the California Institute of Technology collection. Rev&ants of mutant alleles amNllG(gene 39), amNG576(gene 52), and amE300(gene 60) were obtained by plating 2 x 10s to 5 x 10s phage on E. coli S/6/5 at 25°C and selecting isolates able to form plaques. The inability of DNA-delay am mutants to plate under these conditions has been shown to map in the respective DNA-delay genes (Mufti & Bernstein, 1974). These revertants are designated by the original mutant allele number followed by R (e.g. amNl16R is a revertant of amNll6). (b) E. coli strains The E. wli strains used are listed in Table 1. E. coli AZ100 and AZ101 jugates from crosses of Hfr6 to strains N1747 and N1748, respectively. (c) Chemicals

are SU- excon-

and media

The antibiotic coumermycin Ai, referred to as coumermycin, was a gift from W. F. Minor (Bristol Laboratories, Syracuse, N.Y.). Coumermycin was stored at 25 mg/ml in dimethylsulfoxide. Novobiocin (Sigma Chemical Co., St. Louis, MO.) was stored in solution at 10 mg/ml in dilute NaOH (pH 7.5). Nalidixic acid (a gift from D. Mount) was stored at 20 mg/ml in 0.12 rvr-NaOH. [MethyZ-3H]thymidine ([3H]thymidine), 6.7 Ci/mmol was purchased from New England Nuclear, Boston, Mass. Hershey broth (H-broth) (Steinberg & Edgar, 1962) was used as the growth medium in determining phage burst size and also for the growth and suspension of indicator bacteria. H-broth was adjusted to pH 7.5 when coumermyoin, novobiooin or nalidixio acid were used in burst size measurements.

Standard permissive mutants Derivative Nalidixic su- himB+ K37himBCoumermycin-sensitive Coumermycin-resistant

thy-0 dra-1 COG

Hfr metBl

nalA

galK2 StP

galK2 St@ himBll4

supE44 lacY1

supD60

thy-6 &a-l

thy-6 dra-1 supE44 lacy1 COUR

CR63

Hfr6

HG3

K37

KS07

N1747

N1748

Comments

SU-

Nalidixic

W3360

acid-sensitive;

B strain of E. coli; standard for phage T4 am mutants

gyrase host

from cross

from cross

au- restrictive

altered

gyrese

W3350 nnlA

W/5

; defective

acid-resistant

of W3807

segregant

segregant

host for phage T4 am

8~~, coumermycin-resistant N1748 x Hfr6

ooumermycin-sensitive N1747 x Hfr6

AZ101

sue,

thy-6 dra-1 Cou”

AZ100

genotype

Relevant

Strain

1 Bacterial strains TABLE

in this laboratory

Constructed

H. Ginsburg

M. Gellert

M. Gellert

H. I. Miller

H. I. Miller

H. Ginsburg

K. B. Low via D. Mount

in this laboratory

Constructed

source

et al. (1963)

et al. (19766) et al. (1963)

Epstein

et al. (19766) Gellert

Gellert

Low (1973)

Epstein

This paper

This paper

Reference

268

D. MCCARTHY

EHA top and bottom agar were used for plaque assays and plating of bacteria (Steinberg & Edgar, 1962). M9 medium (Adams, 1959) supplemented with FeCI,~6H,O (2*7pg/ml) was used in DNA labeling experiments. M9 medium supplemented with 2.7 pg FeCl, * 6H,O/ ml, 0.25% Casamino acids (vitamin-free) and 0.1 pg thiamine/ml (M9C) was used as the growth medium when E. coli AZ100 and AZ101 were employed as hosts in DNA labeling experiments. Both M9 and M9C media were at pH 7.5. Additional supplements were as described in the following section. (d) DNA synthek Cells from an overnight culture were grown at 37°C to a titer of between 2 x lOa and 4~ lo* cells/ml in either M9 medium for E. co& S/6/5 or M9C supplemented with 10 pg thymidine/ml in the case of E. coli AZ strains. Phage were then added to the cells at a multiplicity of infection of about 10 phage/cell. Since infections were carried out without prior addition of a metabolic inhibitor (e.g. KCN), delayed lysis and superinfection inhibition would not be prevented. At 1 min after addition of the phage, the infected cell culture was diluted 3-fold into labeled growth medium at 37°C. When E. COG S/6/5 was used as the host the growth medium was M9 with final concentrations of 250 rg/ml 2’deoxyadenosine and 8.1 pg/ml [3H]thymidine (0.27 Ci/ mmol). When E. coli AZ strains were used as hosts the growth medium was M9C containing a final concentration of 8.7 pg [3H]thymidine/ml (O-27 Ci/mmol). At various times, portions of a solution of either novobiocin (1 mg/ml) or coumermycin (375 pg/ml) in M9 medium were diluted lo-fold into portions of the infected cell culture. The untreated control culture received a proportionate volume of M9 medium without the drug. At 375 pg/ml, coumermycin was only partially soluble in M9 so that the final concentrations of the drug in the growth tubes might have been less than 37.5 pg/ml. The potency of each coumermycin preparation used to inhibit DNA synthesis w&s measured in a parallel burst size control. At various times O-l-ml samples from the drug-treated and untreated cultures were diluted into equal volumes of sampling medium (1.0 M-KCN, 1.0 mg thymidine/ml) to inhibit further metabolism and uptake of labeled thymidine. Portions (0.1 ml) of the diluted samples were spotted onto Whatman 3MM filter papers (2.3 cm) and dried at 85°C for 15 min. After drying, the filters were washed 3 times in 5% (w/v) trichloroacetic acid for 10 min, once in 95% ethanol for 10 min and once in acetone for 10 min. All washes were done at 5°C. The washed filters were dried and placed in glass scintillation vials containing 6 ml Spectrafluor/PPO/POPOP (Amersham Searle, Arlington Heights, Ill.). Samples were counted in a Beckman LS-250 Scintillation Counting System. (e) Burst

size measurements

Cells from an overnight culture were grown in H-broth to a cell titer of 2 x lOa t,o 3 x lO*/ml. Prior to phage adsorption 0.02 ml of 0.1 M-KCN w&s added per ml of cells to inhibit further metabolism during the adsorption period. At 2.5 min after the addition of KCN, phage were added to cells at a multiplicity of infection of 7 to 10 phage/cell. Adsorption was allowed to proceed for 10 min at 37°C after which growth was initiated by diluting the infected culture 4 x IO*-fold into parallel growth tubes containing H-broth with and without the drugs to be tested. When infections of tsA14 were carried out at a semipermissive temperature, adsorption was also at 37°C for 10 min followed by dilution into growth tubes incubated at the appropriate temperature. Drug-treated cells were exposed to the drugs throughout infection. When coumermycin was used, it was diluted from a stock solution (25 mg/ml in dimethylsulfoxide) to the appropriate concentration in H-broth at pH 8.5. After dissolution of the coumermycin, the pH of the H-broth was readjusted to 7.5. The final concentration of dimethylsulfoxide in control and drugtreated growth tubes was never more than 0.06% (v/v). In a control experiment a dimethylsulfoxide concentration of 0.2% (v/v) had no effect on phage burst size. Novobiocin and nalidixic acid were diluted directly from stock solutions into H-broth at pH 7.5. Immediately after addition of the phage-infected cell mixture to the growth tube, the titers of unadsorbed phage and infective centers were measured. A portion from the growth tube was transferred to 0.1 ml of chloroform and plated on permissive indicator bacteria. to determine the titer of unadsorbed phage. The titer of infective centers in the growth tube was measured by plating a portion from the untreated control growth tube on

GYRASE

AND

PHAGE

T4 DNA

REPLICATION

269

permissive bacteria. The subsequently obtained titer of infective centers was corrected for urn&sorbed phage. In some cases infective centers were also assayed from the drugtreated growth tube immediately after dilution from the adsorption tube. These measurements agreed with measurements from the untreated growth tubes. At 120 min after the initiation of phage growth, chloroform was added to the growth tubes and the titer of progeny phage was determined by plating. Phage burst size was calculated as the ratio of progeny phage titer to the titer of infective centers.

3. Results (a) Sensitivity of wild-type

phage T4 and DNA-delay to novobiocin

am mutants

Novobiocin and the related drug coumermycin have been shown to inhibit the DNA gyrase of E. wli (Gellert et al., 19766). If the E. coli gyrase provides the function missing in infection by DNA-delay am mutants, then growth of these mutants should be inhibited by novobiocin. The data in Table 2 show that in the absence of the drug, gene 39,52,60 and 58-61 mutants gave average burst sizes of 13, 24,23 and 29 phage/cell, respectively, while wild-type T4 gave an average burst of 117 phage/cell. The burst sizes of the DNA-delay am mutants are somewhat lower than previously reported by Mufti & Bernstein (1974). These differences may reflect differences in growth conditions which may affect the ability of S/6/5 to support the growth of the DNA-delay mutants (Mufti & Bernstein, 1974). Infected cells were exposed throughout infection to 109 pg novobiocin/ml. This concentration of the drug is sufficient to block the growth of S/6/5 in liquid culture. In the presence of novobiocin growth of the gene 39, 52 and 60 am mutants was reduced to 1 o/oor less of the untreated cultures. However, the burst size of the gene 58-61 mutant was only reduced to 40% of the untreated culture after drug treatment. Wild-type phage T4-infected cells showed no reduction in burst size when exposed to novobiocin. The large reduction in burst size of the gene 39, 52 and 60 mutants by novobiocin suggests that the host gyrase may be required for the growth TABLE

Effect of novobiocin

Phage strain

T4D wild-type Gene 39 aWaN Gene 52 amE 1240 Gene 60 nmE300 Gene 58-61 amHL627

2

on burst size of wild-type phuge T4 and DNA-delay in E. coli S/6/5 (su- and no&h&a-sensitive)

Untreated

Burst size Novobiooin

(100 pggiml)

am mutants

Relative burst size: novobiooin/untreated

11wt

109(3)

0.93*0-10

13(3)

0*1(3)

0~010~0~002

24(2)

0*2(2)

0.008 kO.007

23(l)

O.l( 1)

0.0043

2W)

11ca

0.40 f 0.005

t The values shown are the average of the number of experiments indioated in parentheses. $ Where relative burst size determinations were repeated, the mean value is given with estimated standard error of the mean.

the

270

D. MCCARTHY

of these DNA-delay am mutants. The observation that growth of wild-type T4 is unaffected by the drug, even though host gyrase is normally inhibited by it, suggests that the wild-type products of the DNA-delay genes either protect the host gyrase or can compensate for its loss of function. (b) Effect of wumermycin-resistant host gyrate on sensitivity of DNA-deluy am mutants to coumermycin Coumermycin, like the related drug novobiocin, inhibits the gyrase of E. coli (Gellert et al., 19763). E. coli K12 strains have a permeability barrier to novobiocin but not coumermycin (Gellert et al., 1976b) so that genetic experiments involving E. coli K12 derivatives can be more easily done with coumermycin than novobiocin. The E. co.5 K12 strains N1747 (coumermycin-sensitive) and N1748 (coumermycinresistant) were gifts from Dr M. Gellert. Growth of N1747 is inhibited at concentrations of coumermycin as low as 15 PgLglml. N1747 yields gyrase extracts that are both E. coli N1748 carries a spontaneous mutation coumermycinand novobiocin-sensitive. conferring resistance to concentrations of coumermycin as high as 60 pg/ml in liquid culture. N1748 also yields gyrase extracts which are resistant to both coumermycin and novobiocin (Gellert et al., 1976b). Since these E. coli strains contain an amber suppressor, su- derivatives were constructed by crossing N1747 and N1748 to Hfr6 (Table 1). E. coli AZ100 and AZ101 are, respectively, coumermycin-sensitivesuand coumermycin-resistant su- segregants from these crosses. Table 3 shows the effects of different concentrations of coumermycin on DNA-delay am mutants infecting the coumermycin-sensitive strain AZ100 and the gyrasealtered, coumermycin-resistant strain AZlOl. In infections of AZlOO, the DNA-delay mutants give burst sizes similar to those seen on E. coli S/6/5 (Table 2). If infected AZ100 cells were exposed to coumermycin (5 pg/ml) throughout infection, the growth of gene 39, 52 and 60 am mutants was reduced to 7% or less of the untreated controls. The gene 58-61 mutant showed a small reduction in burst size, to 73% of the control value. At a coumermycin concentration of 15 pg/ml, the burst sizes of the gene 39, 52 and 60 mutants were reduced to less than one phage per cell while burst size of the gene 58-61 mutant was reduced to 51 y0 (19 phage/cell) of the untreated control. The wild-type T4 infection was unaffected by growth in the presence of coumermycin. Thus, the effects of coumermycin on the growth of the DNA-delay mutants infecting AZ100 parallel the effects of the related antibiotic novobiocin on mutant-infected E. coli S/6/5. The data on Table 3 also indicate that sensitivity to coumermycin appears to he specifically associated with the DNA-delay am mutations. Reversion of the gene 39, 52 and 60 am mutations restores the coumermycin resistance of the formerly sensitive strains to wild-type levels of coumermycin resistance. In addition, a temperaturesensitive gene 41 mutant (DNA-negative) at a semipermissive temperature and an rIIA mutant are fully resistant to coumermycin. When coumermycin-resistant E. coli AZ101 was infected by mutants representing the four DNA-delay genes, the characteristic growth of the DNA-delay mutants was again evident with burst sizes ranging from 9*9 to 36 phage/cell (Table 3, column 7). The mutant infections of AZ101 in the presence of 5 pg of coumermycin gave burst sizes that were 78 to 111 o/o of the untreated control values. At 15 pg of coumermycin per ml the DNA-delay mutants show a reduction of burst size from 36 to 86% of the untreated controls. These reductions appear to reflect the sensitivity of E. coli AZ101

272

D. MCCARTHY

to high concentrations of coumermyein. At 15 pg of coumermyoin per ml the generation time of AZ101 is twofold longer than the generation time in the absence of the drug. Thus, these DNA-delay mutants were relatively insensitive to coumermycin in the coumermycin-resistant host. Since AZ101 should differ from AZ100 only in having a coumermycin-resistant gyrase, the sensitivity of the DNA-delay mutants to coumermycin should reflect the sensitivity of the host gyrase to the drug. This again implies that the growth of the DNA-delay am mutants is dependent on a functional host DNA gyrase. (c) Effect of host himB mutation on burst size of wild-type phage T4 and the DNA-delay am mutants E. c&i mutations, called himB-, have been isolated (H. I. Miller, munication) which are defective in gyrase-catalyzed supercoiling of DNA. The himB mutations cotransduce greater than 99.5% with resistance (H. I. Miller, personal communication). Burst size experiments were carried out on wild-type (himB+) E. himB- E. coli K807. These strains are thought to be isogenic except locus (H. I. Miller, personal communication). TABLE

personal comclosed circular coumermycin coli K37 and for the himB

4

Effect of host himB mutation on burst size of wild-type phage T4 and DNA-delay am mutants Phage strain

T4D Wild-type Gene 39 anaN Gene 52 amNG676 Gene 60 amE Gene 58-61 anzHL627

Burst size K37 (himB+) KS07 (himB-)

11v‘4t

47w

Relative burst size himB-lhimB+

0.46&0*20

21

o-30

o-01

40

0.16

0.006

18

0.44

0.024

31

9.7

0.31

t The value shown is the average of 2 experiments.

Column 2 of Table 4 shows the burst sizes of wild-type phage T4 and gene 39, 52, 60 and 58-61 am mutants in himB+. Wild-type gave a burst of 117 phage per cell while the DNA-delay mutants gave burst sizes varying bet,ween 18 and 40 phage per cell on himB+ E. wli. When infections were carried out in the himB- host, wild-type phage and the gene 58-61 am mutant showed only a small reduction in burst size (to 46% and 31%, respectively, of the burst size in himB+). For the gene 39, 52 and 60 am mutants, burst size was reduced to less than one phage per cell in the himBhost. The reductions in growth of the DNA-delay mutants in the himB- host parallel the effects of coumermycin and novobiocin on these mutants. This again suggests that growth of the DNA-delay mutants is dependent on a fully functional host gyrase.

GYRASE

(d) sensitivity

AND

PHAGE

T4 DNA

REPLICATION

of wild-type T4 and DNA-delay nalidixic acid

273

am mutants to

Nalidixic acid is a specific inhibitor of DNA synthesis in E. coli (Gross et al., 1965). The E. coli n&A gene codes for the target protein of nalidixic acid (Pnal). Mutations allowing growth in high concentrations of nalidixic acid map in the n&A gene. Pnal appears to interact with the target protein of coumermycin to form the gyrase of E. coli. Nalidixic acid inhibits the gyrase-catalyzed supercoiling of closed circular DNA (Sugino et aZ., 1977; Gellert et al., 1977). In the experiments described in previous sections, a reduction in burst size of the DNA-delay am mutants resulted when the coumermycin-sensitive component of gyrase was inhibited. To test the effect of inhibition of the product of the E. coli n&A gene on the burst size of wild-type T4 and the DNA-delay mutants, burst size measurements were made in the presence of 100 pg nalidixic acid per ml and in the absence of the drug in a nalidixic acid-sensitive su- host (E. coli W3350). When infecting su- E. coli W3350 in the absence of nalidixic acid, wild-type T4 gave a normal burst of phage (106 phage/cell) while the DNA-delay urn mutants tested showed low burst sizes ranging from three to nine phage per cell (Table 5). Because of the low untreated burst sizes of the DNA-delay mutants tested, measurements of the inhibition of phage production for these mutants are approximate. When infected cells were exposed to nalidixic acid throughout infection, the burst sizes of wild-type and the DNA-delay mutants were apparently reduced to a comparable extent, with relative burst sizes (nalidixic acid treated/untreated) varying from O-01 to 0.13. The relative burst size of wild-type phage T4 treated with 100 pg nalidixic acid/ml was 10% of the untreated control. This is consistent with the relative burst sizes of wild-type phage T4 observed by Baird et a2. (1972) at 10 pg/ml (65% of untreated control) and 50 pg/ml (35% of untreated control). To determine if the sensitivity of phage T4 to nalidixic acid was a reflection of the drug sensitivity of the r&A gene product of the host, a nalidixic acid-resistant derivative (E. coli HG3) of E. coli W3350 was used as a host in burst size experiments. When infecting E. coli HG3, the nalA host, in the absence of nalidixic acid, gene 39, 60 and 58-61 am mutants gave burst sizes ranging from 13 to 36 phage per cell while wild-type T4 gave a burst of 150 phage per cell. Wild-type T4 and the DNA-delay mutants infecting the nalA host in the presence of nalidixic acid were relatively resistant to the effects of the drug compared to infections of the nalidixic acidsensitive host under similar condit’ions. The burst sizes of wild-type phage and the DNA-delay mutants in the presence of the drug were O-34 to 0.65 of the untreated control values. Relative burst sizes were on the average, sevenfold higher in the n&A host than in the wild-type E. coli strain (Table 5). It appears that inhibition of phage growth by nalidixic acid, in the nalidixic acid-sensitive host, is primarily a result of inhibition of the n&A gene product. However, as we will discuss below, at this concentration of nalidixic acid, the inhibition of phage production was probably not primarily a result of inhibition of DNA gyrase activity. (e) Effect of novobiocin on DNA and DNA-delay

synthesis by wild-type

pluqe

am mutants

Novobiocin is a specific inhibitor of DNA synthesis in E. coli (Smith & Davis, 1967 ; Staudenbauer, 1975) and the cell-free replication systems of ColEl (Staudenbauer,

3

7

9

Gene 60 amE

Gene 58-61 amHL627

106

0.50

0.08

0.40

11

0.06

o-01

0.13

0.10

acid/untreated

Nalidixic

32

36

13

160

11

20

8

98

0.34

0.66

0.62

0.66

HG3 (nalidixic acid-resistant) Burst size Relative burst size Nalidixic acid Untreated Nalidixic acid/untreated (100 &ml)

burst size

acid-sensitive) Relative

W3350 (nalidixic Burst size Nalidiiic acid Untreated (100 w/ml)

mutants in E. coli W3350 and E. coli HG3 (nalA)

am

5

acid on burst size of wild-type p?qe T4 and DNA-delay

Gene 39 umNl16

T4D Wild type

Phage strain

Efect of ndidixic

TABLE

GYRASE

AND

PHAGE

T4 DNA

REPLICATION

275

1976) and +X174 replicative form DNA (Sumida-Yasumoto et al., 1976). To determine the effect of novobiocin on phage T4 DNA synthesis, DNA synthesis of wild-type T4 and DNA-delay mutant-infected E. coli S/6/5 was measured as incorporation of C3H]thymidine into acid-insoluble material at 37°C in the presence and absence of novobiocin. Figure l(a) shows the DNA synthesis of wild-type phage T4 in E. coli S/6/5. In the absence of novobiocin, incorporation began at about 5 minutes after infection and the amount incorporated increased exponentially until 25 minutes after infection, as shown by the linear shape of the DNA incorporation curve on a semi-logarithmic plot (insert Fig. l(a)). At 25 minutes after infection there was a transition from exponential to linear DNA synthesis as shown by the linear shape of the incorporation curve on the linear plot of Figure l(a). The point of transition from exponential to linear synthesis is indicated as the point at which the broken line diverges from the DNA synthesis curve in the insert of Figure l(a). As discussed previously (McCarthy et al., 1976) exponential DNA synthesis occurs when the number of DNA growing points in the cell increases to remain in a constant proportion to the enlarging content of phage DNA in the cell. For wild-type T4 there is about one growing point per chromosome equivalent of template DNA during the exponential phase of DNA synthesis (Werner, 1968; McCarthy et al., 1976). Thus, the exponential phase of phage DNA synthesis represents a period of frequent initiations of DNA growing points as each generation of newly made DNA enters the replicating pool. Previous results suggest that the transition to linear incorporation represents the point at which the number of growing points in the cell reaches an upper limit (Werner, 1968; McCarthy et al., 1976). From this point DNA increases in a linear fashion from a constant number of growing points (about one growing point/chromosome equivalent of template DNA, McCarthy et al., 1976). When novobiocin was added at a final concentration of 300 pg/ml to wild-type T4-infected cells at various times after infection there was no inhibition of DNA synthesis, as shown in Figure l(a). The curves representing wild-type phage T4 DNA synthesis in the presence of novobiocin are superimposable on the untreated control curve of DNA synthesis. This result is in agreement with the previous finding that the burst size of wild-type T4 is unaffected by novobiocin (Table 2). DNA synthesis was also measured in phage strains carrying mutations in genes 39(amN116), 52(umNG576), 6O(amE300) and gene 5%61(amHL627) as shown in Figure l(b), (c), (d) and (e). In the absence of novobiocin DNA synthesis by each mutant began at about the same time as in a wild-type infection, 5 to 7 minutes after infection, and proceeded exponentially (linear curves on semilogarithmic coordinates in inserts of Fig. l(b), (c), (d) and (e)). Th e rates of exponential synthesis were lower than the rates observed in wild-type T4-infected cells and in every case the transition to linear incorporation was delayed relative to wild-type infection. In the case of amE300, DNA increased exponentially throughout the experiment showing no transition to linear synthesis for at least 100 minutes after infection (insert, Fig. l(d)). Novobiocin was added at various times after infection to gene 39, 52, 60 and 58-61 mutant-infected cells. In the case of the gene 39 mutant the final concentration of novobiocin added at various times after infection was 100 pg/ml. While this concentration of novobiocin fully inhibited growth of the DNA-delay am mutants (Table 2), it caused partial inhibition of DNA synthesis (Fig. l(b)). Inhibition of 11

276

D. MCCARTHY

6 t

x

I

PtE z T I5

,,I,,,,,,

0

20

0

20

1

40

60

11 40

I 60

0

I

00

lo(

11 00

II loo

20

40

0

20

60

00

I 40

I

I 60

I

I 00

I

II IO0

100

Time (mm)

FIQ. 1. Phage DN-4 synthesis in the presence of novobiocin. DNA synthesis was measured as incorporation of [3H]thymidine into acid-insoluble material as described in Materials and Methods. Determinations of DNA synthesis were made after

GYRASE

AND

PHAGE

T4 DNA

REPLICATION

277

burst size appears to be a more sensitive measure of novobiocin activity than is inhibition of DNA synthesis, presumably because burst size reflects not only inhibition of phage DNA synthesis, but also the probable effects of this inhibition on late protein synthesis and DNA degradation (Bolle et al., 1968; Yegian et al., 1971; Naot & Shalitin, 1973). In subsequent DNA incorporation experiments gene 52, 60 and 58-61 mutant infections were exposed to 300 pg of the drug per ml to achieve a more complete inhibition of phage DNA synthesis. It can be seen that the expected increases in the rates of DNA synthesis characteristic of the exponential phase were inhibited in the mutants if the drug was added early in infection. The gene 58-61 mutant amHL627 (Fig. l(e)) showed the smallest inhibition of DNA synthesis after addition of the drug consistent with the low inhibition of burst size of amHL627 by novobiocin (Table 2). In each mutant infection, with the exception of the gene 60 mutant amE300, novobiocin was added during both the exponential and linear phases of incorporation. In the case of amE300, because of the long exponential phase, novobiocin was only added during the exponential phase. The transitions from exponential to linear DNA synthesis of the untreated controls are indicated in the inserts to Figure l(b), (c), (d) and (e) as the points at which the broken lines diverge from the curves of DNA synthesis. In the case of each of the DNA-delay mutants, DNA synthesis was inhibited if novobiocin was added during the exponential phase of DNA synthesis. Addition of the drug at progressively later times during the exponential phase resulted in progressively less inhibition of further DNA synthesis as shown by the increasing slopes of the drug-treated DNA synthesis curves representing progressively later addition of the drug. Addition of novobiocin after the transitions to linear DNA synthesis in the mutant-infected cells resulted in no inhibition of further DNA synthesis (Fig. 1(b), (c) and (e)). Thus, novobiocin only inhibits DNA synthesis in the DNA-delay mutant-infected cells if added during the exponential period when initiation of growing points is expected to be occurring. In the case of the gene 60 defective mutant which exhibited exponential synthesis throughout infection, novobiocin inhibited further DNA synthesis following each addition of the drug during the course of infection (Fig. l(d)). In the gene 52 and 60 mutants, DNA synthesis underwent an immediate transition from exponential to linear increase after addition of novobiocin during the exponential phase of DNA synthesis (Fig. l(c) and (d)). Th is immediate shift from exponential to linear DNA synthesis suggests that novobiocin prevents further increase in the number of growing points. In the gene 39 mutant-infected cells (Fig. l(b)) the inhibition of DNA synthesis by novobiocin was neither as rapid nor as severe as with the gene 52 and 60 mutants. This is probably a reflection of the lower concentration of the drug used in the case of the gene 39 mutant since this infection was exposed to 100 pg of novobiocin per ml while the gene 52 and 60 mutant infections were exposed to 300 tug infection of E. co& S/6/5 in the absence of novobiocin (0) end subsequent to addition of the drug ot 1.6 mm (A), 6 min (A), 12 min ( q ), 26 min (0) and 60 min ( n ). The data are plotted using linear co-ordinates. In the inserts DNA synthesis in the absence of novobiocin is replotted on s semilogarithmic scale. The semilogarithmic plots show that the initial incretlses in DNA are exponential. This is followed in each ease by a transition to the linear increase 8s shown in the linear plot. The phage strains used, in the final concentrations of novobiocin in the drug-treated cultures, and the times of trrtnsition from exponential to linear DNA increase are, respectively, (a) T4D(wild-type), 300 pg/ml, 25 min; (b) amN116(gene 39), 100 pg/ml, 30 mm; (c) amNG676(gene 52), 300 pg/ml, 60 min; (d) amE300(gene 69), 300 yg/ml, DNA increase was exponential throughout the experiment; (e) amHL627(gene 58-61), 300 pg/ml, 60 min.

D. MCCARTHY

278

of the drug per ml. It is likely that novobiocin does not inhibit DNA elongation in the DNA-delay mutant-infected cells since addition of the drug during the linear phase of DNA synthesis has no effect on subsequent DNA synthesis. (f) Effect of wumermycin

on DNA synthesis of wild-type and a gene 52 am mutant

phage T4

To test the sensitivity of phage DNA synthesis to coumermycin, the incorporation of [3H]thymidine into acid-insoluble material in the presence and absence of coumermycin was measured in wild-type and a gene 5.2mutant (amNG576) infection (Fig. 2). Figure 2(a) represents the DNA synthesis of wild-type phage T4 in E. coli AZIOO. AZ100 contains a coumermyoin-sensitive gyrase. DNA synthesis in the wild-type T4-infected cells began at about 7 minutes after infection and proceeded exponentially until the transition to linear DNA synthesis at 13 minutes after infection (linear curve on semilogarithmic insert in Fig. 2(a)). Addition of coumermycin at a fiual conoentration of 37.5 pg/ml at any time during either the exponential or linear phase of DNA synthesis had no effect on further DNA synthesis as expected from the lack of coumermycin inhibition of wild-type burst size (Table 3). In the case of amNG576-infected E. coli AZIOO, incorporation of label appeared to begin at about 12 minutes after infection and proceeded exponentially until about 25 minutes after infection as shown by the semilogarithmic plot of DNA synthesis in the insert of Figure 2(b). In the mutant-infected cells, the exponential and the subsequent linear DNA synthesis occurred at slower rates than in the wild-typeinfected AZlOO. The transition from exponential to linear DNA synthesis was also delayed relative to the wild-type infection. When coumermycin was added at 37.5 pg/ml to amNG576-infected E. coli A!ZlOO, further DNA increase was inhibited if the drug was added during the exponential phase. Addition of the drug during the linear phase of DNA synthesis had no effect on further DNA synthesis. The curves of DNA synthesis subsequent to addition of the drug at 25 and 60 minutes (during the linear phase) are superimposable on the linear untreated control curve of DNA synthesis (Fig. 2(b)). Addition of the drug during the exponential phase did not result in an immediate shift to linear incorporation in the drug-treated culture. Since the concentration of coumermycin used in the incorporation experiment gave only a partial reduction in burst size (data not shown), this may reflect an incomplete effect of the drug on DNA synthesis (see Materials and Methods, section (d)). When coumermycin was added at progressively later times during the exponential phase of DNA synthesis the resulting drug-treated DNA synthesis curves showed progressively greater slopes indicating that, like novobiocin, coumermycin inhibited DNA synthesis to a progressively smaller extent as the exponential phase progressed. To eliminate the possibility that the inhibition of phage DNA synthesis by coumermycin was the result of the inhibition of some phage or host factor other than gyrase, the effect of coumermycin on DNA synthesis of amNG576 was measured in E. coli AZlOl, the gyrase-altered, coumermycin-resistant strain (Fig. 2(c)). In the absence of coumermycin, incorporation of label began at about 7 minutes after infection and proceeded exponentially until about 30 minutes after infection, when incorporation began to level off (semilogarithmic insert in Fig. 2(c)). Addition of coumermycin at any time during the course of infection had no effect on subsequent DNA synthesis. The curves of DNA synthesis in the drug-treated cells are superimposable on the

GYRASE

AND

PHAGE

T4 DNA

REPLICATION

279

24 21 18 15 12 9 6 3

FIG. 2. Phsge DNA synthesis in the presence of ooumermycin. DNA synthesis was followed as incorporation of [3H]thymidine into acid-insoluble material &s described in Mrtteriels and Methods. Determinations of DNA synthesis were made in the absence of coumermyoin (0) and after addition of the drug to 37.6 &ml at 1.6 min (A), 6 min (A), 12 min ( q ), 26 min (0) and 60 min ( n ) after infection. In the inserts, DNA synthesis in the absence of coumermycin was plotted on e semilogarithmic scale. The semilogarithmic plots show that the initial increases in DNA are exponential end are followed by trsnsitions to the linear increases shown in the linear plots. The phage &sins and E. coli strains used and the times of (8) T4D(wild-type), E. coli transition from exponential to linear synthesis were, respectively, AZ100 (Cous), 13 min; (b) amNGK’IB(gene 52), E. coli AZ100 (Cous), 26 min; (c) umNGS’lB(gene 52), E. coli AZ101 (CouR), incorporation leveled off before transition to linear synthesis.

curve of the untreated infection, although the drug preparation inhibited phage growth in a parallel burst size control using a Cous host (data not shown). The coumermycin resistance of DNA synthesis by the gene 52 mutant in the coumermycin-resistant, gyrase-altered host indicates that the coumermycin inhibition of mutant phage DNA synthesis in the coumermycin-sensitive host resulted from inhibition of the sensitive host gyrase. Thus, efficient DNA synthesis during the exponential phase of DNA-delay mutant DNA synthesis depends on a fully functional host gyrase.

280

D. MCCARTHY

4. Discussion (a) DNA-delay

gene 39, 52 and 60 am mutants require a functiolzal DNA gyrate for growth in su- E. coli

host

Unlike

most am mutants defective in essential genes, DNA-delay am mutants can E. coli (Mufti 8t Bernstein, 1974; Yegian et al,, 1971). This characteristic growth of gene 39,52 and 60 DNA-delay am mutants is greatly reduced by novobiocin and coumermycin (Tables 2 and 3). These drugs have been shown to inhibit the gyrase of E. coli (Gellert et al., 1976b). DNA-delay am mutants infecting a host that is coumermycin-resistant due to an altered gyrase are able t’o grow in the presence of the drug (Table 3). Thus, it appears that the coumermycin sensitivity of the DNAdelay am mutants reflects the drug sensitivity of the host gyrase. That growt,h of the DNA-delay am mutants depends on a functional host gyrase is also indicated by the large reduction in burst size observed when the phage mutants infect an E. coli himB- mutant which has a defective gyrase (Table 4). The requirement of phage production for a functional host gyrase appears to be specifically associated with the DNA-delay mutations. The rIIA gene like DNA-delay genes 39, 52 and 60 maps in a cluster of genes with membrane-related functions but’ growth of an rIIA mutant is fully resistant to coumermycin (Table 3). In addition, a temperature-sensitive gene 41 (DNA-negative) mutant at a semipermissive temperature shows no sensitivity to coumermycin (Table 3). The relative insensitivit,y of the growth of the gene 58-61 mutant to inhibition of the host gyrase suggests that t,here may be two classes of DNA-delay mutants. One class, represented by gene 39, 52 and 60 am mutants, is dependent on the host gyrase for growth while the other class, represented by the gene 58-61 am mutant, exhibits the DNA-delay phenotype but grows relatively well under conditions that inhibit the host gyrase. Yegian et al. (1971) have also obtained evidence implying that the gene 39, 52 and 60 mutants affect the same pathway while the gene 58-61 mutants affect an independent function. Broker (1973) has observed that the gene 58-61 product may limit the activity of the putative exonuclease controlled by genes 46 and 47. Thus, the DNA-delay phenotype of the gene 58.61 mutant may be caused by the uncontrolled activity of a T4 exonuclease, whereas the phenotype of the gene 39, 52 and 60 mutants apparently results from an inefficient gyrase activity. grow

011 su -

(b) Models of DNA-delay

gene function

Our results imply that the host gyrase can partially substitute for the lack of the products of phage genes 39, 52 and 60. In addition, the products of phage genes 39 and 52 have been shown to bind specifically to DNA (Huang & Buchanan, 1974) as well as to the cell membrane (Huang, 1975; Takacs & Rosenbusch, 1975). Mufti & Bernstein (1974) have obtained evidence consistent with the association of the gene 39 product with the products of genes 52,60 and a compensating host component. Taken together this evidence can be used to formulate models of DNA-delay gene function. In an adaptation of a proposal by Mufti & Bernstein (1974) the wild-type products of genes 39, 52 and 60 could modify the host gyrase at the membrane for interaction with phage DNA and in the process render it resistant to coumermycin and novobiocin. The characteristic low levels of growth of the DNA-delay am mutant would stem from the inefficient activity of the unaltered host gyrase when one of the modifying phage DNA-delay gene products is incomplete.

GYRASE

AND

PHAGE

T4 DNA

281

REPLICATION

Alternatively, the product of gene 39 could interact with the products of genes 52 and 60 at the membrane to form a ooumermyoin-resistant autonomous phage gyrase. According to this model, the characteristic growth of the DNA-delay am mutants would result from the substitution of the host enzyme for the inactive phage gyrase. The reduction of the burst size of wild-type T4 in the gyrase-altered himB- E. coli mutant to 46% of the control (Table 4) could be taken to favor the model in which the DNA-delay gene products interact with the host gyrase since it implies that wild-type phage requires a fully functional host gyrase for maximum growth. The fact that wild-type T4 requires a functional E. wli n&A gene product for efficient growth might also be taken to imply a role for DNA gyrase when the DNA-delay gene products are present. However, this latter evidence should be interpreted with caution. From the concentration of oxolinio acid required to inhibit in vivo supercoiling of phage h DNA (Gellert et al., 1977), one can infer that although the oonoentration of nalidixic acid used in the burst size experiment (Table 5) reduced phage production, it might be insufficient to inhibit the host gyrase in phage TCinfeoted cells. Based on the reversible association of gyrase subunits, Higgins et al. (1978) hypothesized that Pnal might be involved in reactions other than supercoiling. The sensitivity of phage T4 and E. wli (Gross et al., 1964) to a low concentration of nalidixic acid (196 pg/ml) might reflect inhibition of Pnal in a form other than as a gyrase component. In addition to the above models postulating that the DNA-delay gene products provide a gyrase activity for phage DNA, it is possible that the relationship between DNA-delay functions and gyrase activity is more indirect. For example, if one views the DNA-delay gene products as membrane-bound DNA-binding proteins, the requirement for gyrase might reflect the extent to which phage DNA is physically constrained at the membrane. In the presence of coumermyoin residual gyrase activity might be sufficient to support the growth of wild-type T4 but not the DNAdelay mutants. While we cannot offer a detailed description of the function of the DNA-delay gene products, our data suggest that the products of genes 39, 52 and 60 either provide a gyrase activity adapted specifically for phage T4 DNA or in some way affect the normal contribution of the host gyrase to phage production. of initiation of DNA replication coumermycin and novobiocin

(c) Probable inhibition

by

McCarthy et al. (1976) have shown that under conditions of partial compensation by the host gyrase a gene 52am mutant (amNG576) exhibited normal ratesof DNA elongation. This finding implies that the low rate of DNA synthesis characteristic of the DNA-delay gene 52 mutant is due to a low incidence of growing points caused by a defect in the initiation of growing points. Other DNA-delay mutants were not studied. Based on the apparent defect in initiation of replication by a gene 52 mutant under semipermissive conditions, the results obtained under fully restrictive conditions, caused by coumermycin and novobiooin inhibition of the remaining host gyrase function, indicate that the host gyrase may be required for the remaining initiations of phage DNA replication. If ooumermycin and novobiooin are added to mutant-infected cells during the exponential phase of DNA synthesis they inhibit subsequent increases in the rates of DNA synthesis. The exponential phase is the period when initiation of growing points probably occurs frequently as each generation

282

D. MCCARTHY

of newly made DNA enters the replicating pool (Werner, 1968). In the gene 52 and 60 mutant infections, addition of novobiocin during the exponential phase caused an immediate transition to linear DNA synthesis, suggesting that the drug acted to prevent further increase in the number of growing points. The progressively higher linear rates of DNA synthesis obtained following successively later additions of the drug probably reflects the progressive establishment of growing points prior to each subsequent addition of the drug. Coumermycin and novobiocin do not appear to inhibit DNA elongation since addition of the drugs during linear phase of DNA synthesis had no effect on subsequent synthesis even though synthesis was measured in some cases for up to 40 minutes after addition of the drug (Fig. I(b), (c) and (e) and Fig. 2(b)). The lack of inhibition of DNA synthesis following addition of the antibiotics during the linear phase implies that once the cellular content of growing points reaches a maximum, addition of the drugs no longer inhibits DNA synthesis. Novobiocin and coumermycin do not appear t,o affect phage DNA synthesis indirectly by inhibiting protein synthesis, since blockage of lete protein synthesis with chloramphenicol (Bolle et aE., 1968; Naot & Shalitin, 1973) or by additional maturation defective mutations (genes 33 and 55) (Bolle et al., 1968; Yegian et al., 1971) tends to suppress the DNA synthesis phenotype of the DNA-delay mutants, enhancing this synthesis to a level closer to wild-type. I thank B. Alberts for suggesting the use of novobiocin to study the DNA-delay mutants. I am also grateful to M. Gellert, H. Ginsburg, H. I. Miller and D. W. Mount for their generous gifts of E. coli strains; to N. Cozzarelli and H. I. Miller for communicating information; to C. Bernstein, H. Bernstein, V. Johns, S. Mufti and D. Yarosh for critically reading the manuscript. Special thanks are extended to H. Bernstein and C. Bernstein for helpful advice and discussion. This work was supported by National Science Foundation grant PCM-7714971 to Harris and Carol Bernstein. REFERENCES Adams, M. H. (1959). Bacterioph4~gea, Interscience, London. Baird, J. P., Bourguignon, G. J. $ Sternglanz, R. (1972). J. V’irol. 9, 17-21. Bolle, A., Epstein, R. H., Salser, W. & Geiduschek, E. P. (1968). J. Mol. Biol. 33, 339-362. Broker, T. R. (1973). J. Mol. BioZ. 81, I-16. Epstein, R. H., Bolle, A., Steinberg, C. M., Kellenberger, E., BoydelaTour, E., Chevalley, R., Edgar, R. S., Susman, M., Denhardt, D. t Leilausis, A. (1903). CoZd Spring Harbor Symp. Quad. BioZ. 28, 375-394. Gellert, M., Mizuuchi, K., O’Dea, M. H. C Nash, H. A. (1976a). Proc. Nat. Acad. Sci., U.S.A. 73, 3872-3876. Gellert, M., O’Dea, M. H., Itoh, T. & Tomizawa, J. (197ti). Proc. Nat. Acad. Sci., U.S.A. 73, 4474-4478. Gellert, M., Mizuuchi, K., O’Dea, M. H., Itoh, T. t Tomizawa, J. I. (1977). Proc. Nat. Acad. Sk., U.S.A. 74, 4772-4776. Gross, W. A., Deitz, W. H. L Cook, T. M. (1964). J. Baoteriol. 88, 1112-1118. Gross, W. A., Deitz, W. H. & Cook, T. M. (1965). J. Bucteriol. 89, 1068-1074. Higgins, N. P., Peebles, C. L., Sugino, A. & Cozzarelli, N. R. (1978). Proc. Nat. Acad. Sci., U.S.A. 75, 1773-1777. Huang, W. M. (1975). Virology, 66, 508-521. Huang, W. M. & Buchanan, J. M. (1974). Proc. Nat. Acd Sci., U.S.A. 71, 2226-2230. Itoh, T. & Tomizawa, J. (1977). Nature (Lo&on), 270, 78-80. Karam, J. D. & O’Donnel, P. V. (1973). J. ViroZ. 11, 933-945. Low, B. (1973). J. Bactwiol. 113, 798-812. Marians, K. J., Ikeda, J., Schlagman, S. t Hurwitz, J . (1976). Proc. Nat. Acad. Sk., U.S.A. 74, 1965-1968.

GYRASE

AND

PHAGE

T4 DNA

REPLICATION

283

McCarthy, D., Minner, C., Bernstein, H. t Bernstein, C. (1976). J. Mol. BioE. 106, 963-981. Mufti, S. & Bernstein, H. (1974). J. Fi’irol. 14, 860-871. Naot, Y. & Shalitin, C. (1973). J. l’irol. 11, 8622871. Smith, D. H. BEDavis, B. D. (1967). J. BacterioZ. 93, 71-79. Staudenbauer, W. (1975). J. Mol. Biol. 96, 201-205. Staudenbauer, W. (1976). Mol. Gen. Genet. 145, 273-280. Steinberg, C. & Edgar, R. S. (1962). Genetics, 47, 187-208. Sugino, A., Pebbles, C. L., Kreuzer, K. N. & Cozzarelli, N. R. (1977). Proc. Nat. Acad. Sci., U.S.A. 74, 4767-4771. Sumida-Yasumoto, C., Yudelevich, A. & Hurwitz, J. (1976). Proc. Nat. Acad. Sci., U.S.A. 73, 1887-1891. Takacs, B. J. & Rosenbusch, J. P. (1975). J. Biol. Chem. 250, 2339-2350. Warner, H. R. & Hobbs, M. D. (1967). Fiirology, 33, 376384. Werner, R. (1968). Cold Sp&g Hwbor Symp. Quant. Biol. 33, 501-507. Yegian, C. D., Mueller, M., Selzer, G., Russo, V. t Stahl, F. W. (1971). Birology. 46. 900-919.