J. MOE. Biol.

(1991)

219,

605-613

The Stringent Response Blocks DNA Replication Outside the ori Region in Bacillus subtilis and at the Origin in Escherichia coli Alain Levine, Fraqoise Vannier, Mohammed Dehbi Gilles Henckes and Simone J. S4xor-f Institut de Ge’ne’tique et de Microbiologic Unite’ de Recherche Associe’e au Centre National de la Recherche Scientijque 1354 Bdtiment 409, Universite’ Paris XI 91405 Orsay Cedex 05, France (Received

4 September

1990; accepted 18 February

1991)

When the Bacillus subtilis dnaB37 mutant, defective in initiation, is returned to permissive temperature after growth at 45°C DNA replication is synchronized. Under these conditions, we have shown previously that DNA replication is inhibited when the Stringent Response is induced by the amino acid analogue, arginine hydroxamate. We have now shown, using that substantial replication of the oriC region DNA-DNA hybridization analysis, nevertheless occurs during the Stringent Response, and that replication inhibition is therefore implemented downstream from the origin. On the left arm, replication continues for at least 190 x lo3 base-pairs to the gnt gene and for a similar distance on the right arm to the gerD gene. When the Stringent Response is lifted, DNA replication resumed downstream from or&’ on both arms, confirming that DNA replication is regulated at a post-initiation level during the Stringent Response in B. subtilis. Resumption of DNA synthesis following the lifting of the Stringent Response did not require protein or RNA synthesis or the initiation protein DnaB. We suggest, therefore, that a specific control region, involving Stringent Control sites, facilitate reversible inhibition of fork movement downstream from the origin via modifications of a replisome component during the Stringent Response. In contrast, in Escherichia coli, induction of the Stringent Response appears to block initiation of DNA replication at oriC itself. No DNA synthesis was detected in the oriC region and, upon lifting the Stringent Response, replication occurred from oriC. Post-initiation control in B. subtilis therefore results in duplication of many key genes involved in growth and sporulation. We discuss the possibility that such a control might be linked to differentiation in this organism. Keywords:

Bacillus subtilis;

Escherichia coli; initiation Stringent Response

The co-ordination of overall bacterial gene expression involves the interplay of complex regulatory systems named regulons (Gottesman, 1984). Individual regulons are defined by the existence of a regulatory gene that responds to stress by regulating the expression of various unlinked genes. The stringent system is one of the most important regushould

be

Control; kb, h.p.l.c., high-

lo3

605 0022-2836/91/120~5&09

$03.00/O

replication

control;

lons in bacteria. When the level of any aminoacyl transfer RNA becomes limiting, the Stringent Response is triggered. This is mediated through the nucleotide guanosine-5’-diphosphate-3’-diphosphate (ppGpp) and leads to the inhibition of synthesis of several major macromolecules, including rRNAs (Gallant, 1979; Cashel & Rudd, 1987). In order to test the possibility that DNA replication in bacteria is subject to Stringent Control (SCf), we have employed a mutant of B. subtilis (dnaB37), in which the requirement for protein synthesis and RNA transcription can be completely separated. Thus, after return to permissive temperature, the first round of DNA replication is resistant to chloramphenicol but completely sensitive to rifampicin

1. Introduction

t Author to whom all correspondence addressed. $ Abbreviations used: SC, Stringent base-pairs; TCA, trichloroacetic acid; pressure liquid chromatography.

mutant;

0

1991 Academic

Press Limited

A. Levine et al.

606

(Laurent, 1973). In these properties, the dnaB37 and the dnaA46 (Tippe-Schindler et al., 1979) behave similarly. We have previously shown for dnaB37 and dnaA46, in B. subtilis and E. coli, respectively, that DNA replication is blocked in a rel+ strain but not in a rel- strain. This inhibitory effect in B. subtilis was not due to inhibition of uptake of the radiolabel or of inhibition of DNA elongation. Consequently, we concluded that the initiation of DNA replication in both B. subtilis and E. coli is subject to Stringent Control (S&or et al., 1986). In agreement with our results for E. coli, Rokeach & Zyskind (1986) have shown that, in this organism, the mioC promoter, the expression of which is apparently required for efficient oriC activity, is stringently controlled. More recently, the dnuA promoters have been found to be stringently regulated (Chiaramello & Zyskind, 1990). Previously we have shown that over-initiation occurs in dnaB37 synchronized for DNA replication during the first replication round (Henckes et al., 1989). This over-replication is limited to a small region of about 400 kb around the origin. We have proposed that, in B. subtilis, DNA replication is subject to control at two levels on the chromosome, oriC itself and a second region outside the origin, which presumably acts to limit over-replication of the chromosome after a premature initiation. This has been termed post-initiation control (Henckes et al., 1989). In E. coli, over-replication of oriC can be achieved by raising the level of DnaA. However, under these conditions, the over-replicated DNA was degraded and there was no evidence for a specific post-initiation control downstream from the origin (Atlung et al., 1987). In view of these results, it was important to investigate at which level the Stringent Response operates in both B. subtilis and E. eoli. We show here that, in B. subtilis, after induction of the Stringent Response, DNA replication occurs in the origin region. However, this replication is limited to a relatively small region of about 400 kb around the origin, as shown by DNA-DNA hybridization. Moreover, after reversal of the Stringent Response, elongation of DNA replication resumes from regions close to the sites of blocked forks and not from the origin. In contrast, in E. coli, induction of the Stringent Response prevents any detectable replication of the or& region. Possible mechanisms of Stringent Control of DNA replication in these organisms are discussed.

2. Materials and Methods (a)

Bacterial

Bacterial

strains,

growth

conditions

and plasmids

strain used were: (1) Bacillus subtilis OMM242 dnaB37 (thyA thyB trp) derived from W168 (Laurent & Vannier, 1973). Bacteria were grown at 30°C in a shaking water-bath in minimal Spizizen medium (Spizizen, 1958) supplemented with @5’$” (w/v) glucose, 002% (w/v) casein hydrolysate, 5 pg of thymine/ml and 40 pg of tryp-

tophan/ml. The mass doubling time in this medium was approximately 60 min at 30°C and 25 to 30 min at 45°C’. (2) Escherichia coli dnaA46 (thrleuthisupE ~1’ ) provided by Dr E. Orr, Leicester University, and dnaAd(i (thrleu- thi- supE reEA1) our work. were used. Bacteria were grown at, 30°C in a shaking water-bath in minimal M63 medium (Monod et al., 1951) supplemented wit,h @S% (w/v) glucose, 1 miw-MgSO,, @I mM-Call,. 50 p’p of leucine/ml, 50 pg of threonine/ml and 1.5 fig of thiaminej ml. The mass doubling time in this medium was approximately 90 min at 30°C. All plasmids used are lisbed in Table 1. E. coli clones from the E. coli chromosome genomica library (Kohara et al., 1987) were obtained from Professor B. Holland, Institut de GCn&ique et MicrobiologitJ. Universitb Paris XI. (b) Chemicals [2-3H]adenine

(specific activity

20 Ci/mmol,

equivalent,

to 0.74TBq/mmol) and [methyL3H]thymidine (specific activity

45 Ci/mmol,

equivalent

to 1.66 TBq/mmol)

were

from C.E.A. Saclay, France. Rifamycin, chloramphenico]. tetrabutylammonium phosphate and arginine hydroxamate were from Sigma. Dodecyl sulphate sodium salt, (SDS) was from Merck. Proteinase K, lysozyme and restriction

endonucleases

were

purchased

from

Boehringer. Acetonitrile wasfrom Prolabo. (c)

Label&q

of DNA

after

synchronization

DKA replication of B. subtilis was synchronized as described (Levine et al., 1987). In brief, cells were grown at 30°C up to a density of 3 x lo7 cells/ml and shifted to 45°C for 30 min. The culture was then returned to 30°C’ for subsequent analysis. Under these conditions exrellent synchrony was observed, as shown by measuring the rate of DNA synthesis. Cells were labelled for times indicated in the text with [methyl-3H]thymidine at a concentration of 20 @J/ml (@74 MBq/ml). DNA replication of E. coli was synchronized as indicated. Cells were grown at 30°C to a density of 5 x 107 cells/ml and shifted to 42°C for 1 h. The culture was then returned to 30°C for analysis. Cells were labelled for times indicated in the text with [methyl-3H]thymidincx at a concentration of 20 &i/ml (074 MBq/ml). (d) Measurement

of the rate oj’ IINA

synthesis

Tn lJ. ,subtilis samples (2 ml) were withdrawn from thr cultures at appropriate intervals and added to 50 ~1 of [2-3H]adenine (80 &i/ml. equivalent’ t)o 2.96 MBq/ml). After 10 s. 2 min and 3.5 min. samples (0.5 ml) were witlrdrawn: incorporation was st,opped by adding (PI volume of I mM-NaX;,. Samples were mixed with VEi ml of ti4 M-EaOH. incubated at 80°C’ for 30 min. neutralized b) addition of 0.5 ml of e.5 M-HCI and precipitated with IO?,, (w/v) trichloroacetic acid (TCA). The rate of DNA sgnthrsis was expressed as [2-3H]adenine cts/min per ml of culture per min of incubation at 30°C. In E. coli samples (2 ml) were withdrawn from thr caultures at appropriate intervals and added to 50 ~1 ot [methyZ-3H]thymidine (80 &i/ml, cbyuivalrnt to 2.96 MBq/ml). After 10 s, 2 min and 3.5 min. samples (0.5 ml) were withdrawn: incorporation was stopped b? adding 4 ml of 10% (w/v) TCA. The rate of DPU’A synthrsis was expressed as [methyl-‘Hlthymidine cts/min per ml of culture per min of incubation at 30”(‘,

Stringent

Control of DNA

Replication

607

in Bacteria

Table 1 Sources of DNA probes

Gene sources

Name

A. For pDG120

DNA probes (size in kb)

B. subtilis

experiments spoIID sacP

Map positiont (deg. or min)

Distance from the origin1 W)

0.9 HindIII-Hind111

316

520

2 EcoRI-EcoRI 6 SmuI-Sal1

330 333

355 320

344 351 0

190 106 0

3

35

7

pBSG8 pMF2

sacA, sad3

pCG8 IC4BsF14 pJHlOl-E6’

Snt dnaC w&wA

36 HindIII-Hind111 135 EcoRI-EcoRI 2.2 BarnHI-EcoRI

p20KRl

alWB

(E6’ fragment) 58 EcoRI-Sal1

P63 pSD15

spo VC-tms-spo spoIIE

pIS139 pBMD16 pDH5 PX4 B. For E. pFHC54Y pBJC917 pFHC270 pFHC247 pTAC1674

3 HindIII-Hind111 3.2 AvaI-AvaI

10

85 120

.ypoOH

2.5 Sma-EcoRI

11

130

gerD amyR da1

1% PstI-P&I 1.2 HhaI-HhaI 2 EcoRI-KpnI

16 25 38

190 295

3 HindIII-Hind111 93 HindIII-BgZII 2 HindIII-Hind111 34 HindIII-EcoRI 45 HindIII-Hind111

83 84 84 84 85

VG

450

References sources

01

Stragier et al (1988) Fouet et al. (1982) DBbarbouillB et al. (1987) Fujita et al. (1986) Perego et al. (1987) Lampe & Bott (1984) Zuber & Losick (1987) Moran et al. ( 1980) Stragier etal (1988) Dubnau et al. (1988) Moir et al. (1978) D. Henne4 Martin et al. (1987)

coli experiments rnpA uncA-glmS gid asnA ilVG

47 8 2 3 47

Atlung et T. Atlung Atlung et Atlung et T. Atlung

al. (1987) [( al. (1987) al. (1987) 11

t Position of the genes on the map of the B. subtilis chromosome (Piggot et al., 1990) or the map of the E. coli chromosome (Bachman, 1983). For B. subtilis, the origin of the chromosome replication is located at 0”. For E. coli, the origin of the chromosome replication is located at 84 min. $ Assuming a size of 4250 kb for the B. subtilis chromosome and 4700 kb for the E. coli chromosome. (i Genetech, San Francisco. 1) University of Copenhagen.

(e) PuriJication Nucleic acids were 1987) and sonicated

reduce the DNA 400 nucleotides. (f) Hybridization DNA

DNA

of chromosomal

extracted

as described

(Levine

et al..

using an MSE Ultrasonicator fragments to a single-strand length

of radioactive chromosomal probes immobilized on jilters

DNA

to of

to

Probe DNA (linearized plasmids) was loaded onto @45 mm BA85 nitrocellulose filters (Schleicher & Schuell) as described (Henckes et al., 1989). Each filter contained about 200 ng of DNA per lo3 bases of probe. Blank filters contained 100 pg of denatured calf-thymus DNA. Hybridizations were carried out as described (de Massy et al., 1984) and were for 4 days at 42°C. Background values given by calf-thymus DPjA filters were 2 x 10m5 to 5 x 10e5 of the inputs. The cts/min hybridizing to all probes were proportional to the amount of 3H-1abe11ed chromosomal DNA added. (g) Quantification

of

guanosine-%diphosphateJ-diphosphate

A combination Bremer (1982)

of the methods and by Payne

described by Little & & Ames (1982) for

measuring ppGpp in E. coZi extract was used with modifications. Samples (30 ml) were withdrawn during the synchronous replication of dnaB37 mutant and treated with 3 ml of 1.9% (v/v) formaldehyde for 25 min at 0°C and cells were harvested by centrifugation (15,000 g for 30 min). Cells were suspended in 1 ml of @I M-KOH for 30 min at 0°C. KOH was neutralized by the addition of 5~1 of 88% (v/v) phosphoric acid and 1 ml of h.p.1.c. buffer A. Cellular debris were removed by centrifugat,ion (30,OOOg for 45 min), and the supernatant was retained for h.p.1.c. analysis. Samples of the extract (200~1) were fractionated by ion-pair reverse-phase h.p.1.c. The column used was an Ultrasphere-IP-Cl8 column (46 mm x 250 mm) with a (3.2 mm x 4.6 mm) Cl8 precolumn, both from Beckman. Buffer A was 30 mM-KH,PO,, 5 mM-t,etrabutylammonium phosphate, 4% (v/v) acetonitrile (pH 6). Buffer B was 100% acetonitrile. The elution gradient was a 60 min convex gradient curve no. 3, 4 to 80% B. The flow-rate of the h.p.1.c. buffer was 1 ml/min at about of the eluant was continuous13 13,800 kPa. The A,,, monitored. For quantification, the area of the peak in the elution profile that corresponded to ppGpp was determined and converted to the total A,,, units in the samples analysed. The amount of ppGpp was calculated using an extinction coefficient corresponding to 74.6 nmol/A,,, unit (Little & Bremer, 1982). The intra-

608

A. Levine

et al.

Table 2 Effect

Time Conditions A. Exponentially growing Control + Arginine hydroxamate B. Synchronized Control + Arginine

hydroxamate

of arginine

after return 30°C (mm)

to

on the synthesis

of ppGpp

Time after addition of the drug (min)

Amount of ppGpp (pm(W570)t

cells 30

21 570

15 30

48 35 603 720

r&f 0 30

hydroxamate

t Average amounts of ppGppj200 ~1 of supernatants from triplicate culture samples, calculated from the area corresponding to the ppGpp peak in h.p.1.c. elution profiles as described in Materials and Methods, expressed in pmol/unit of absorbance at 570 nm, with a standard deviation of about 10°C. $ An exponential culture of OMM242 grown at 30°C was transferred to 45°C for 30 min and then returned to 30°C (zero time) in the presence or absence of the drug. Samples were analysed 0, 15 or 30 min after the shift to 30°C.

cellular pmol/A570 analysed.

concentration unit of

of culture

ppGpp was expressed in represented in the samples

3. Results (a) Localization

of the Stringent

B. subtilis

Control

block on the

chromosome

sequences located further from the origin such as sad? (-320 kb), sacA (-355 kb) or spoIID (- 520 kb) on the left side and amyR ( + 295 kb) or da1 (+450 kb) on the right side were not replicated. In a control experiment without drug, all these chromosomal sequenceswere fully replicated during this

period.

In

order

to

confirm

that

replication

forks were completely inhibited downstream from

previously shown that in the Bacillus initiation mutant dnaB37, initiation of chromosomal replication is subject to Stringent We

have

subtilis Control

(S&or

et al., 1986).

In

order

to

induce

the

Stringent Response, we used arginine hydroxamate, which inhibits synthesis of stable RNA, presumably through accumulation of ppGpp (Price & Gallant, 1982). To confirm this, we have measured ppGpp under various conditions. When OMM242 is shifted from 30°C to 45°C for 30 minutes, which allows completion of the replication round, there is an increase in ppGpp of about twofold (Table 2). After return to 3O”C, the ppGpp level gradually declined. Nevertheless, following the addition of the drug (250 pg/ml), the level of ppGpp increased at least 20-fold in exponentially growing cells or in cells synchronized for DNA replication (Table 2). In order to determine whether there was any replication under these conditions, we hybridized DNA labelled during the Stringent Response with unlabelled DNA probes representing different regions of the chromosome, in particular, those close to the origin. A culture of OMM242 grown at 30°C was transferred to 45°C for 30 minutes and then returned to 30°C in the presenceor in the absenceof arginine hydroxamate and labelled with [methyZ-3H]thymidine for appropriate times. Figure 1 shows the results of such an experiment. During the first 34 minutes after the return to 3O”C, in the presence of arginine hydroxamate, the chromosomal region around the origin continued to replicate. The extent of this replication was roughly between - 190 kb (gnt) on the left side of the chromosome and + 190 kb (gerD) on the right side. On the contrary,

g Fi a 2i 2 0 ae 600

400

200

0

200

400

600

kb

Chromosbmoloriqrn

Figure 1. Analysis of DPU’A replication during the Stringent Response in B. subtilis. An exponential culture of OMM242 grown at 30°C was transferred to 45°C for 30 min and then returned to 30°C in the presence or absence of arginine hydroxamate (250 pg/ml). Samples were pulse-labelled with [methyl-, Jthymidine (45 Ci/mmol, 20 &i/ml) for 34 min (open bars). A sample of the arginine hydroxamate-treated culture was also pulse-labelled with [methyL3H]thymidine between 34 and 54 min after return to the permissive temperature (filled bars). For each sample, the total [nzethyL3H]thymidine incorporated into DNA was in agreement with the rate of DKA synthesis. An identical portion of each sample (05 x lo6 to 2 x lo6 cts/min) was allowed to reassociate in the presence of an excess of probe DKA fixed on nitrocellulose filters. Results were corrected for background (amount of DF;A binding to calf-thymus DKA) and usually the quantification of a given sequence was the mean value of 3 to 5 determinations. The data are presented as To of hybridization relative to the control experiment without drug versus probe positions of the chromosomal origin region.

Stringent

Control

of DNA

Replication

in Bacteria

609

the origin, the arginine hydroxamate-treated culture was also pulse-labelled with [methyZ-3H]thymidine between 34 and 54 minutes after return to the permissive temperature. The results presented in Figure 1 demonstrate that, indeed, DNA replication was completely blocked under these conditions. In an additional control experiment, the drug was added to the culture at 45”C, 20 minutes before the return to 30°C. Under these conditions, replication was still blocked outside the origin, clearly indicating that the effect was not due to delayed uptake of the drug (data not shown).

0 kb

Chromoso&

(b) Analysis of DNA li$%tg the Stringent

sequences replicated after Response in B. subtilis

The Stringent Response is a reversible phenomenon. After elimination of arginine hydroxamate from treated cultures by filtration and washing, DNA replication resumes with a few minutes (data not shown). Therefore, this allowed us to identify the first fragments replicated after lifting the Stringent Response. In the experiment described in Figure 2, bacteria were synchronized for DNA replication as before and treated at 30°C with arginine hydroxamate for 34 minutes. Bacteria were then filtered, washed to remove the inhibitor and labelled with [methyl-3H]thymidine for 17 minutes at 30°C. The pattern of replicated DNA sequences was then analysed. After resumption of replication, the chromosomal sequences closest to the origin were not replicated (E/J, gnt, abrB, spoVC, spoIIE, spoOH). In contrast, sequences further from the origin, such as sacs, sacA and spoZZD on the left side and gerD, amyR and da1 on the right side were fully replicated. These results indicated therefore that, under these conditions, DNA replication resumed from regions close to the sites of original inhibition and did not resume from the origin itself.

600

400

200

0

Chromos&ol

200

400

600

origin

Figure 3.

kb

origm

Figure 2. Analysis of DNA sequences replicated after reversing the Stringent Response in B. subtilis. Bacteria were synchronized for DPU’A replication as before and treated at 30°C with arginine hydroxamate (250 pg/ml) for 34 min. Bacteria were then filtered, washed to remove the inhibitor and labelled with [methyl-‘Hlthymidine for 17 min at 30°C. The data are presented as y0 of hybridization relative to the control experiment without drug versus probe positions of the chromosomal origin region.

Requirements for the resumption of DNA synthesis after lifting the Stringent Response in B. subtilis. Bacteria were synchronized for DPU’A replication as before and treated at 30°C with arginine hydroxamate (250 mg/ml) for 34 min. Bacteria were then filtered. washed to remove the inhibitor and labelled with [methyL3H]thymidine without any drug ([I), or in the

presence of chloramphenicol

(100 pg/ml: a), or rifampicin

(250pg/ml: H), or at 45°C (m), for 17 min at 30°C (or 45°C). The data are presented as y0 of hybridization relative to the control experiment without arginine hydroxamate versus probe positions of the chromosomal origin region.

Indeed, on the left of the origin (compare gnt in Fig. 1 and in Fig. 2), replication appeared to resume from the blocked forks. On the contrary, on the right of the origin, the data shown in Figure 2 indicate that replication apparently resumed from a region close to gerD, upstream from the blocked forks. This result is not due to the presence of duplicated gerD sequences in the B. subtilis chromosome (data not shown). This result may therefore indicate resumption of DNA replication from a secondary origin upstream from gerD, perhaps in the cluster of the three rrn genes located close to gerD (Widom et al., 1988). Nevertheless, the data of Figure 2 clearly demonstrate that replication does not resume from the origin after lifting the Stringent Response. These data therefore provide further confirmation that the inhibition of DNA replication in B. subtilis during the Stringent Response occurs well downstream from the origin. We have also analysed the requirements for the resumption of DNA synthesis after lifting the Stringent Response. The results shown in Figure 3 demonstrate that the resumption of DNA synthesis after removal of arginine hydroxamate was not dependent upon RNA synthesis, protein synthesis or apparently DnaB. This indicated a simple model of stalled replisomes with a reversibly inhibited subunit, which then resume polymerization upon removal of the drug. (c) Analysia To induce used valine

of the effect of the Stringent on DNA synthesis in E. coli

Response

the Stringent Response in E. coli, we as an inhibitor of isoleucine tRNA

A. Levine et al.

610

50

a 5 &

60

E (L 40

Time after

shift

bock

to 30°C

(min)

(b)

Figure 4. Effect of valine on initiation of DNA synthesis in cultures of E. coli dnaA46 and dnaA46relA 1. (a) An exponential culture of strain dnaA46 and (b) its derivative dnaA46reEAl at an A 570 = 0.18, were shifted to 42°C for 60 min, diluted 3-fold and shifted back to 30°C. Under these conditions, DNA replication is synchronized. Pulses of [methyL3H]thymidine were performed as indicated in Materials and Methods. The rate of DNA synthesis is expressed as [met/$‘Hlthymidine counts incorporated per ml of culture per min of incubation. Open symbols. control culture. Filled symbols, plus valine (1 mg/ml). Valine was added at 42°C 3 min before the return to 30°C.

synthetase (Lamond & Travers, 1985). In order to test the effects of the Stringent Response in E. co&, it was first necessary to synchronize DNA replication under conditions where protein synthesis was no longer required for initiation. This was achieved by using a dnaAts mutant (dnaA46) after return to permissive temperature as shown in Figure 4. The results showed that the Stringent Response blocks any detectable DNA synthesis in a Tel+ strain (Fig. 4(a)), whilst in the dnaA46 reZA1 strain, DNA synthesis continued in the presence of valine (Fig. 4(b)) at a slightly reduced rate. It is also noticeable in Figure 4 that the rate of DNA synthesis is somewhat reduced in the relA1 mutant even in the absence of valine. In order to localize the Stringent Control block on the chromosome, similar DNA-DNA hybridization experiments were carried out as described for B. subtilis in Figure 1. Results

25

0

25

50

kb

Figure 5. Analysis of DNA replication during the Stringent Response and after reversing the Stringent, Response in E. coli. An exponential culture of dnaA46 grown at 30°C was transferred to 42°C for 60 min and then returned to 30°C in the presence of valine (1 mg/ml). Samples were pulse-labelled with [methyL3H]-thymidine (45 Ci/mmol, 20 @/ml) for 30 min (filled bars). A sample of the valine-treated culture for 30 min was also filtered. washed to remove the inhibitor and labelled with [methyL3H]thymidine for 20 min at 30°C (open bars). An identical portion of each sample (03 x IO6 to 1.5 x lo6 cts/min) was allowed to reassociate in the presence of an excess of probe DNA fixed on nitrocellulose filters. Results were corrected for background (amount of DNA binding to calf-thymus DNA) and usually the quantification of a given sequence was the mean value of 3 determinations. The data are presented as “/b of hybridization relative to the control experiment without valine versus probe positions on the physical map of the or& region, using Kohara et al. (1987) ordinates corrected according to MBdigue et al. (1990). Probes are designated by the name of the gene on the E. coli chromosome (Bachman, 1983) or by the name of the phage clone in the Kohara bank.

obtained are shown in Figure 5. In these experiments, in order to maximize detection of any replication within the origin region, cells were labelled for 30 minutes after return to t’he permissive temperature in the presence of valinr or for 20 minutes following the removal of valine exactly as in Figure 1 with B. subtilis. Tn contrast t.o R. subtilis, we were unable to detect any significant, replication of DNA within the origin region after the addition of valine. This inhibition of DNA synthesis can be reversed by removing valine. The results shown in Figure 5 demonstrated that all the markers in the origin region (asnA. 4F4 and yid, uncA and 2Al) were replicated under the labellinp conditions; as expected, the outside markers 2E6, iEvG (+50 kb) and 3D1, rpnA (-50 kb), were also extensively replicated.

4. Discussion We have previously shown that the onset of neu rounds of DNA replication in both R. subtilis and E. coli is inhibited by the Stringent Response in a

Stringent Control of DNA reZA-dependent manner (S&or et al., 1986). In this study we were concerned to analyse further the nature of such controls and, in particular, whether the inhibition of DNA synthesis observed during the Stringent Response operated at the first or at a second (post-initiation) level of control. In this paper, we have clearly shown that the addition of arginine hydroxamate to B. subtilis OMM242 (dnaB37) synchronized for DNA replication, leads to rapid inhibition of DNA replication not at oriC but at, a minimum of 190 kb on the left arm and a minimum of 190 kb on the right arm of the chromosome. We propose that the Stringent Response is mediated via Stringent Control (SC) sites, SC1 on the left and SC2 on the right, which act to halt movement of the replication forks. Replication terminator sequences have been identified near the terminus of replication in E. coli, B. subtilis and some plasmids (Kuempel et al., 1989). However, it is not clear whether SC sites are in any way related to such sequences. We have described a post-initiation control mechanism in B. subtilis that limits replication after premature initiation (Henckes et aE., 1989). In these experiments, we mapped the extension of overreplication at a minimum of 190 kb on the left (gnt) and 130 kb on the right (cysA). We have now found that the limit of over-replication on the right arm is apparently just upstream from the gerD marker, 190 kb on the right arm (data not shown). These results increase the possibility that Stringent Control and Post-initiation Control may operate at the same sites, through a common mechanism involving ppGpp. SC sites may be designed to prevent over-replication and therefore play a role in second level of control of DNA replication. Inhibition of replication fork movement outside the origin in B. subtilis during the Stringent Response is clearly reversible and does not require protein synthesis, RNA synthesis, or the initiation protein DnaB. This would be more consistent with the reversible inhibition of a replisomal protein at stalled forks. The apparent resumption of replication from a stalled fork at an SC site, at least on the left of the origin (see Fig. 2), supports this view. One possible candidate for reversible inhibition of the replisome is the DnaE protein (equivalent in B. subtilis to DnaG protein in E. coli), since some studies have indicated that the expression of the dnaG gene is subject to Stringent Control in E. coli (Grossman et al., 1985). Results of the analysis of DNA replicated on the right arm of the chromosome were more difficult to interpret. In this case, replication appeared to resume upstream from the blocked fork upon removal of arginine hydroxamate with the gerD marker being replicated. This may represent re-initiation of DNA replication from a secondary origin, under these conditions, upstream gerD and in the vicinity of a cluster of three ribosomal operons: rrnJ, rrnH and rrnG (Widom et al., 1988). However, we did not detect an excess of replication and, therefore, whatever the mechanism, a single replisome initiating replication in the gerD

Replication

in Bacteria

611

region appears to be responsible for the resumption of DNA synthesis on the right arm. In contrast to the results obtained with B. subtilis, induction of the Stringent Response in E. coli (dnaA46) through the addition of valine completely inhibited DNA replication in the origin region as determined by DNA-DNA hybridization analysis. This indicates that Stringent Control of chromosomal replication operates at quite different sites in the two organisms. This may reflect the absence of a major post-initiation control mechanism, downstream from the origin in E. coli. Inhibition of replication from oriC during the Stringent Response might be explained, at least in part, by a direct effect upon transcription from mioC. Transcription from the mioC promoter has been implicated in regulating initiation from oriC in several studies (Messer, 1987; Shauzu et al., 1987). Moreover, transcription from mioC has been shown to be stringently controlled (Rokeach & Zyskind, 1986). In order to test this possibility, we have attempted to analyse the effect of Stringent Control on initiation in a dnaA46 mutant in which mioC is deleted or its transcription is blocked (Stuije et al., 1986). However, such a double mutant is so perturbed that DNA replication is not synchronized after a temperature shift and the experiment could not be carried out. The different levels or sites at which Stringent Control operates in the two organisms is intriguing and raises questions about the possible physiological implications. It is noteworthy that, in addition to genes encoding DNA gyrase, DnaA, and seven out of ten ribosomal operons, the region of the B. subtilis chromosome between oriC and the postulated SC1 (left) and SC2 (right) contains several key sporulation genes including spoOH, abrB and spoVG, three genes involved in the early stages of sporulation. Under conditions of nutrient starvation or other stresses, which normally induce sporulation and may also trigger the Stringent Response, it may be critical to allow limited replication from oriC. This could lead to higher levels of expression of these gene products through gene-dosage effects or possibly through local changes in DNA supercoiling resulting from limited replication of this strategic portion of the chromosome. The differential expression of this group of duplicated genes under starvation conditions could be a crucial mechanism in determining the final decision between continued vegetative growth or commitment to sporulation. We are very grateful to T. Atlung, K. Bott, E. Dubnau, Y. Fujita, A. Galizzi, D. Henner, A. Moir, G. Rapoport, A. Sonenshein, P. Stragier and P. Zuber for providing us with plasmids used as sources of DNA probes. We thank Y. Kohara for providing us with E. coli clones from the E. coli chromosome genomic library. We thank Dr De Wind for supplying MR 489 and MR 490 strains of E. coli. We are very grateful to D. Gadelle for her help in the h.p.1.c. experiments. We thank I. B. Holland for stimulating discussions during this work. These studies were supported in part by the Institut National de la Sante et

612

A. Levine et al

de la Recherche MBdicale (grant no. 861024), the Ligue Nationale Franpaise Contre le Cancer, the Association pour la Recherche sur le Cancer, and the Fondation pour la Recherche MBdicale Franqaise.

References Atlung, T.. Lobner-Olsen, A. & Hansen, F. G. (1987). Over-production of DnaA protein stimulates initiation of chromosome and minichromosome replication in E. coli. Mol. Gen. Genet. 206, 51-59. Bachmann, B. (1983). Linkage map of Escherichia coli K-12. Microbial. Rev. 41, 180-230. Cashel, M. & Rudd, K. (1987). The Stringent Response. In Escherichia coli and Salmonella typhimurium. Cellular and Molecular Biology (Xeidhardt, F. C., Ingraham, J. L., Low, K. B., Magasanik. B., Schaechter. M. & Umbarger, H. E., eds). pp. 1410-1438, American Society for Microbiology:y. Washington, DC. Chiaramello, A. E. & Zyskind, J. (1990). Coupling of DNt\ replication to growth rate in Escherichia coli: a possible role for guanosine tetraphosphate. J. Bacterial. 171, 4272-4280. DkbarbouillB, M., Kunst, F.. Klier, A. & Rapoport, (:. (1987). Cloning the sacS gene encoding a positive regulator of the sucrose regulon in Bacillus subtilis. FEMS Microbial. Letters, 41; 137-140. De Massy, B., Patte, J. C.. Louarn, ,J. M. & Bouchk. ,J. I’. (1984). oriX: a new replication origin in Escherichin coli.

Cell,

36, 221-227.

Dubnau, E., Weir, J., Nair, G., Carter, I,.. Moran, (1. CYL Smith, I. (1988). Bacillus sporulation gene spoOH codes for a3’ (c?). J. Bacterial. 170, 1054-1062. Fouet, A., Klier, A. & Rapoport, G. (1982). Cloning and expression in Escherichia coli of the sucrase gene from Bacillus subtilis. Mol. Gm. &net. 186. 399-404. Fujita. Y.. Nihashi, J. 8: Fujita. T. (1986). The characterization and cloning of a gluconate (gnt) operon of Bucillus au&&s. 1. Gen. Microbial. 132. 161- 169. Gallant. cJ. A. (1979). Stringent control in E. coli. Ke,,. Genet. 13, 393-415. Gottesman. S. (1984). Bacterial regulation: global regulatory networks. Annu. Rev. Genet. 18. 415-441. Grossman, A. D., Taylor, W. E., Burton. Z. F.. Burgess. R. R. & Gross. C. A. (1985). Stringent response in &scherichia, coli induces expression of heat, shock proteins. J. Mol. Biol. 186, 357-365. Henckes, G.. Harper. F.. Levine, A.. Vannier. F. &, SF;ror. S. (1989). Over-replicat,ion of the origin region in the dnaB37 mutant of BucilZus xubtilis: post,-init,iation cont,rol of chromosomal replication. Proc. ,Vat. A cold. Ski.,

Kohara,

U.S.A.

86. 8660-8663.

Y.. Akiyama. K. & Isono, K. (1987). Thr physical map of the E. coli cshromosome: application of a new strategy for rapid analysis and sorting of a large genomic library. Cell, 50. 495-508. Kuempel, P. L., Pelletier, A. ,J. & Hill. T. M. (1989). 7’~~s and the terminators: the arrest of replication in prokaryotes. Cell, 59, 581-583. Lamond, A. *J. & Travers. A. A. (1985). Genetically srparable functional elements mediate the opt,imal expression and stringent regulation of a bacterial tRNA gene. Cell, 40, 319-326. Lampe, M. F. & Bott, K. F. (1984). Cloning the yyrd gene of Bacillus suhtilis. Nucl. Acids Res. 12. 6307-6323.

Laurent. 8. J. (1973). Initiation of deoxyribonucleic avid replication in a temperature mutant of R. .szr.bfilis: evidence for a transcriptional step. .J. Bactrriol. 116. 141-145. Laurent, S. J. d Vannier. F. (1973). Temperaturtl~ sensitive init.iation of rhromosome replic*ation in a mutant of Bncillws subtilis. ,J. Bactwiol. 114. 474-484. I .evine. A.. Henckes, (i.. Vannier. F. 8r S&or. S. (19X7). Chromosomal initiation in Bacillus .subtilis ma)involve two closely linked origins. ;Mol. Gm. Grnrt. 208, 37-44. Little, R. & Bremer. H. (1982). Quantitation of guanosint 5’,3’-bisdiphosphatr in extracts from bacterial crlls by ion-pa,ir reverse-phase high-performance liquid chromatography. And. Biochem. 126. 381-388. Martin, I.. DCbarbouillB. M.. Klier, A. & Rapoport. (:. (1987). Tdentifiration of a new locus. .SQ~v. involved in the regulation of levansucrase synthesis in I~aarillus .rubtilis. FEMS Microbial. Lettws, 44. 39-43. Mkdigue, c’.. BouchB. J. P.. HBnaut, A. 8; Danchin. ,\. (1990). Mapping of seyuenced genes (700 kbp) in the restriction map of the Esch.erichia coli (~hromosomt~. Mol. Microhiol. 4. 169-187. Messer. W. (1987). Initiation of l),l;A rt*plicaation ir) Kschrrichicc coli. J. Bacterial. 169. 339.5 -3399. Moir. A.. Laffert,v. E. & Smith, I). :\. (1978). (itlnetic. analysis of spore germination niutants of Hncill/c.s subtilis 168: the correlation of phenotype wit.h malt location. Pror. Sat. Acad. Sci.. I ‘.S.,4. 84. 6X-657. Monod, .I.. (lohen-Bazire. G. & Cohn. M. (1951). Sur Ia biosynthPse dr la /&galactosidasr (lactose) clhrz h’scherichia coli. La sp&ific+b (ita I‘induetioll. Kiochinr. Bioyhys. Acta. 7. 585-589. Moran. (‘ I’.. J,osick. IZ. & Sonenshein. A. I,. (I!#(JJ Identifkation Bacillus subtilis

of

a

sporulation

deoxyribonucleic

10~~s

in

c4r~nrd

acid. -1. Barkrio/.

142, 33lL334. Pa.vne, S. M. 81.Ames. B. N. (1982). A procedure for rapid t,xtraction and high-pressure liquid chromatographic separation of t,he nucleotides and other small molecules from bacterial cells. Anal. &n&m. 123, 151.-161. Perego. M., Ferrari, E.. Bassi. M. T.. Galizzi. A. & Mazza, P. (1987). Molecular cloning of Bucillu~s su6tilis genes involved in DNA metabolism. Mol. Gen. Gmet. 209, 8---14. Piggot, P. ,J., Amjad, M.. Wu, ?J. ,J.. Sandoval, H. & Castro. ,J. (1990). Genetic and physical map of Bacillus subtilis. lb??. In Molecular Biological Methods for Bacillus (Harwood. C. R. & Cutting. S. M., eds). pp. 493$540, ,John Wiley and Sons, Xew York. Price, V. L. & Gallant, -1. A. (1982). A new relaxed mut’ant of Bacillus subtilis. J Bacterial. 149. 635-641. Rokeach. I,. A. & Zyskind, *I. W. (1986). RKA trrmnating within the E. coli origin of replication: stringent regulation and control by DnaA protein. Crll, 46. 763-77 1. Schauzu, M. A., Kiicherer, C., Kiilling, R.. Messer, W. 1(: Lother, H. (1987). Transcripts within the replication origin, oriC, of Escherichia coli. AVucl. Acids Res. 15, 2479-2497. S&or, S.. Vannier, F., Levine. A. & Henekes. (:. (1986). Stringent control of initiation of chromosomal replication in B. suhtilis. Nature (London), 321, 709-710. Spizizen, .J. (1958). Transformation of biochemically delirient strains of Bacillus subtilis by deoxyribonucleate. Proc. Nat. Acad. Sci., U.S.A. 48. 1072-~1078. Stragier, P.. Ronamy. (1. Sr Karmazyn-Campelli. (‘.

Stringent Control of DNA (1988). Processing

of a sporulation sigma factor in how morphological structure could control gene expression. Cell, 52, 697-704. Stuije, A. R., De Wind, K., Van Des Spek, J. C., Pors, T. H. & Meijer, M. (1986). Dissection of promoter sequences involved in transcriptional activation of the Escherichia coli replication origin. Nucl. Acids Res. 14, 2333-2344. Tanaka, M. & Hiraga, S. (1985). Negative control of oriC plasmid replication by transcription of the oriC region. Mol. Gen. Genet. 200, 21-26. Bacillus

subtilis:

Replication

in Bacteria

613

Tippe-Schindler, R., Zahn, G. & Messer, W. (1979). Control of the initiation of DNA replication in Escherichia coli. Mol. Gen. Genet. 168, 185-195. Widom, R. L., Jarvis, E. D., Lafauci, G. $ Rudner, R. (1988). Instability of rRNA operons in Boxillus subtilis. J. Bacterial. 170, 605-610. Zuber, P. & Losick, R. (1987). Role of ArbB in &ooOAand SpoOB-dependent utilization of a sporulation promoter in Bacillus subtilis. J. Bacterial. 169, 2223-2230.

Edited by J. H. Miller