Sms and translesion polymerase, DinB

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/dnarepair

Replication arrest-stimulated recombination: Dependence on the RecA paralog, RadA/Sms and translesion polymerase, DinB Susan T. Lovett ∗ Department of Biology and Rosenstiel Basic Medical Sciences Research Center, Brandeis University, Waltham, MA 02454-9110, United States

a r t i c l e

i n f o

a b s t r a c t

Article history:

Difficulties in replication can lead to breakage of the fork. Recombinational reactions restore

Received 31 May 2006

the integrity of the fork through strand-invasion of the broken chromosome with its sister.

Received in revised form

If this occurs in the context of repeated DNA sequences, genetic rearrangements can result.

28 June 2006

We have proposed that this process accounts for stimulation of chromosomal rearrange-

Accepted 29 June 2006

ments by mutations in Escherichia coli’s replicative DNA helicase, DnaB. At its permissive

Published on line 9 August 2006

temperature for growth, a dnaB107 mutant is a 1000-fold more likely to experience a dele-

Keywords:

wild-type strain. We have previously shown that enhanced deletion in a dnaB107 strain is

DSB repair

reduced in recA, recB and recG102 (formerly known as radC102) derivatives. Here I show that

Replication fork repair

this enhanced recombination is dependent on other factors: the RuvA Holliday junction

Genetic rearrangements

helicase, the RecJ single-strand DNA exonuclease, the RadA/Sms RecA-paralog protein of

Translesion polymerase

unknown function and, surprisingly, the DinB translesion polymerase. The requirement for

tion of a 787 bp tandem repeated segment inserted in the E. coli chromosome than is a

these factors in DnaB-stimulated rearrangements is much greater than that observed for recombinational events such as P1 transduction. This may be because strand invasion into the repeats limits the extent of heteroduplex DNA that can be formed in the initial stage of recombination. I propose that RadA, RecG and RuvAB are critically required to stabilize the strand-invasion intermediate and that DinB polymerase extends the invading 3 strand to aid in re-initiation. The role of DinB in bacteria may be analogous to translesion DNA polymerase ␩ in eukaryotes, recently shown to aid recombination. © 2006 Elsevier B.V. All rights reserved.

1.

Introduction

Evidence suggests that replication forks frequently break and are restored by genetic recombination [1,2]. In Escherichia coli, broken forks are repaired by the action of a DNA exonuclease, RecBCD, and a strand exchange protein, RecA. RecBCD resects the breaks and loads RecA protein onto the revealed ssDNA [3]. Octameric DNA sequences that activate RecBCD’s recombina-

∗

Tel.: +1 781 736 2497; fax: +1 781 736 2405. E-mail address: [email protected]. 1568-7864/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.dnarep.2006.06.008

tional activities, known as Chi sites [4], are over-represented in the genome and oriented in the direction to facilitate repair of a chromosome broken during replication [5]. The ssDNA-RecA filament allows the broken end to invade its sister, producing an intermediate structure known as a D-loop. Processing of this intermediate is critical but not well understood. The Dloop is an intrinsically unstable structure: branch migration in one direction dissolves the joint, aborting recombination,

1422

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

Fig. 1 – Stabilization and destabilization of D-loop intermediates. (A) A broken fork. (B) Invasion of the broken end into the sister chromosome produces a D-loop structure, with a short region of heteroduplex DNA. Branch migration rightward can reverse this process and release the broken strand. (B ) Extension of the 3 invading strand by DNA polymerases can stabilize the joint. (C) Branch migration to extend the region of heteroduplex can occur to drive formation of a 4-strand Holliday junction. Branch migration in the opposite sense reduces the region of heteroduplex. (C ) Resolution of the Holliday junction (HJ) in the manner shown produces a fork structure.

whereas branch migration in the other direction stabilizes the structure by extending the region of heteroduplex DNA (Fig. 1), potentially producing a four-strand Holliday junction. In E. coli both the RecG and RuvAB DNA helicases can catalyze branch migration [6] and act on D-loop and Holliday junction structures. After repair of the broken end, replication must be re-established. E. coli possesses special systems to reload the replication machinery onto repaired forks involving the PriA or PriC primosomal assembly factors [7], which recognize slightly different structures in repaired forks. A double priA priC mutant is dead [8] and the viability of recA and recBC mutants is severely compromised [9], implying that spontaneous replication fork collapse and restart occur almost every replication cycle of E. coli. Difficulties in replication or replication of damaged DNA templates can further elevate the probability of fork breakage [10,11]. Although repair of collapsed replication forks is usually an error-free mode of DNA repair, it can lead to genetic rearrangements if strand invasion occurs within repeated sequences in the chromosome. Such a situation may be responsible for the elevation of the rate of repeat rearrangement observed for mutants of the DnaB fork helicase. The DnaB hexameric helicase promotes replication fork progression by unwinding DNA at the fork and recruitment of DNA primase to the lagging strand. DnaB also promotes replication processivity via interactions with DNA polymerase III through contacts with the ␶ subunit (reviewed in [12]). A dnaB107 temperaturesensitive strain is inviable at 37 ◦ C, but able to form colonies at 30 ◦ C. Although viable at 30 ◦ C, DnaB helicase function does not appear completely normal in dnaB107 strains, which grow more slowly and exhibit elevated genomic instability in assays measuring deletion or expansion of tandem repeated DNA sequences [13–15]. Our previous work impli-

cates homologous recombination in this instability because it was strongly dependent on the function of the RecA strand transfer protein of E. coli [13–15]. In addition, we provided evidence that rearrangements occur by recombination of a broken chromosome, because they require the double strand break-processing exonuclease, RecBCD. Although loss of the RecBCD was lethal in a dnaB107 strain, we demonstrated that deletion rearrangements were RecBCD-dependent in a strain heterozygous for dnaB107/dnaB+ , in which tandem repeat deletion rates were more modestly elevated [14]. Homologous strand invasion of the broken chromosome into the tandem repeat can lead to rearrangements involving loss or gain the number of repeats as shown in Fig. 2. This proposed mechanism invokes processing of the broken chromosome by RecBCD, followed by RecA mediated formation of a D-loop recombination intermediate in the repeat sequence. The invasion of the broken strand into the repeat followed by re-establishment of a replication fork can generate an unequal exchange, yielding deletion of one the repeats (Fig. 2A) or addition of a repeat (Fig. 2B). Because the DNA sequence of the invading strand is not homologous outside of the repeat, this constrains the D-loop intermediate to a rather short length, in the case of the rearrangement assay we commonly employ, maximally 787 bp. Because of this restricted homology in the recombination intermediate, factors that process D-loops or promote their conversion to replication forks should be particularly important in these reactions. In addition, recruitment of polymerases other than the replicative polymerase DNA polymerase III may be required to stabilize repair intermediates by 3 -end extension, prior to replication fork re-establishment. In this work, I complete a survey of other known recombination functions to determine if they are required for recovery of chromosomal deletions stimulated by the defective DnaB107 helicase. In addition, I examined the role of three

Fig. 2 – Repair of a collapsed replication fork can lead to repeat rearrangements. Shown are a broken chromosome and its intact sister with sequence ABCCD, C representing the tandem repeat sequence. (A) Invasion of the broken chromosome in the repeat region C to yield a tandem repeat deletion product. The healed end will produce the deleted product, with sequence ABCD. Note that branch migration leftward would be blocked by heterology flanking the repeat. (B) Invasion of the broken chromosome in repeat region C to yield a triplicated, expansion product. The healed chromosome will yield an ABCCCD sequence. The heteroduplex region is likewise constrained to the region of the repeat C.

1423

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

DNA polymerases, Pol II, Pol IV and Pol V, which are induced by blocked DNA replication as part of the cellular “SOS” response [16,17]. This analysis suggests that a number of functions, which have only weak effects on other double-strand break mediated recombination events such as P1 transduction, are nonetheless stringently required for recovery of repeat rearrangements associated with replication fork repair.

2.

Materials and methods

2.1.

Strains and plasmids

The strains used in this study were derived from STL695, a relative of AB1157 [18], and are listed in Table 1. Isogenic strains were produced by P1 transduction using LCG medium as described [19]. LB medium [19] was employed as the standard growth medium and included, when appropriate, antibiotics at the following concentrations: tetracycline (Tc), 10–15 ␮g/ml; ampicillin (Ap), 100 ␮g/ml; chloramphenicol (Cm), 15 ␮g/ml; spectinomycin (Sp) 100 ␮g/ml; kanamycin (Km) 30 ␮g/ml. For minimal medium selection, 56/2 medium [20] was supple-

mented with 0.4% glucose, 1 ␮g/ml Bl and 50 ␮g/ml required amino acids.

2.2.

Deletion assay

Rates for deletion of 787 bp tetA repeats integrated at lac were determined as described [13]. Growth was at 30 ◦ C. Briefly, whole colonies from overnight LB plates were resuspended in 1 ml LB and grown for 3 h at 30 ◦ C. At least eight independent cultures were assayed for each strain, plating for Tcr and total viable cell cfu/ml for each culture. The recombination rate was determined by the method of the median [21], with 95% confidence intervals.

2.3.

Transduction assay

Cell cultures were grown to OD600 of 0.4 at 30 ◦ C in LCG and titered by serial dilution in 56/2 buffer and plating on LB; this was used to derive a viability index (cfu at OD600 ). Cultures were concentrated 10-fold by centrifugation and were resuspended in LCG. A PI lysate prepared on strain MG1655 (prototrophic E. coli K12 [18]) was infected at moi of 0.1. The

Table 1 – Escherichia coli K-12 strains Strain STL695 STL753 STL1324 STL1353 STL2385 STL2923 STL2936 STL2962 STL3071 STL3074 STL3587 STL3646 STL3671 STL3717 STL3719 STL3768 STL3799 STL3946 STL3995 STL3999 STL4625 STL4664 STL4892 STL5029 STL5055 STL5068 STL5070 STL5072 STL5224 STL5282 STL5283 STL5284 STL5399 STL5401

Relevant genotype + (srlR-recA)304 dnaB107 malE::Tn10kan polB::aadA dnaBl07 (srlR-recA)304 malE::Tn10kan dnaB107 recN1502::Tn5 dnaB107 recN1502::Tn5 recO1504::Tn5 dnaB107 recO1504::Tn5 recG102 recF400::kan dnaBl07 recG102 dnaBl07 ruvC53 recR252::Tn10-9 dnaB107 recF400::kan recR252::Tn10-9 dnaB107 recB2053::Tn10kan ruvC53 dnaBl07 malE::Tn10kan recB2053::Tnl0kan radAl::kan radAl::kan dnaB107 polB::aadA dnaB107 ruvA63 recJ2003::Tnl0-9kan recJ2003::Tn10-9kan dnaB107 recQkan dnaB107 ruvA63 dnaB107 malE::TnlOkan yicR::kan dnaB107 yicR::kan umuC::cat dnaBl07 dinB::kan dnaB107 dinB::kan umuC::cat

Origin or reference [22] [22] Apr transductant P1 STL695 × STL1336 [37] [13] [13] [14] [14] [14] [14] [14] [14] [14] [14] [14] [14] [14] [14] [14] Kmr transductant P1 STL3839 [14] × STL695 Kmr transductant P1 STL1815 [29] × STL695 Kmr transductant P1 STL1815 [29] × STL2923 Spr transductant P1 STL1336 [37] × STL2923 Apr transductant P1 STL695 × N2096 (R. Lloyd) Kmr transductant P1 STL114 [38] × STL695 Kmr transductant P1 STL114 [38] × STL2923 Kmr transductant P1 STL4911 × STL2923 Kmr ts transductant P1 STL1324 × STL5029 Kmr transductant P1 GS1846 (G. Sharpies) × STL2923 Kmr transductant P1 GS1846 (G. Sharpies) × STL695 Cmr transductant P1 YG2224 (H. Ohmori) × STL2923 Kmr transductant P1 YG7207 (H. Ohmori) × STL2923 Kmr transductant P1 YG7207 (H. Ohmori) × STL695 Cmr transductant P1 YG2224 (H. Ohmori) × STL695

Genotype of strains derived STL695 includes: F− ␭− lacZ::[bla+ tetAdup787 ] hisG4 argE3 leuB6 (gpt-proA)62 thr-1 thi-1 rpsL31 galK2 lacYl ara-14 xyl-5 mtl-1 kdgK51 supE44 tsx-33 rfbDl mgl-51 rac− qsr− .

1424

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

mixture was incubated at 30 ◦ C for 30 min, then washed in 56/2 buffer and plated on selective plates, minimal complete 56/2 plates lacking arginine in the presence of 3 mM sodium citrate. Colonies were counted after 3 days. Transduction efficiency as calculated as the number of Arg+ colonies/viable cfu normalized to wild-type strain, STL695, for duplicate determinations performed in parallel.

3.

Results

3.1.

The deletion assay

In previous work, we engineered a 787 bp tandem duplication in the tetA gene, encoding tetracycline-resistance [22], inserted on the E. coli chromosome in the lac operon. The duplication inactivates the gene and deletion of one of the repeats restores function. Deletion rates as measured in this assay are normally low, at approximately 10−6 per cell generation. However, dnaB107 ts strains, at their “permissive” temperature of 30 ◦ C, experience deletion at rates of over 10−3 [13,14]. Mutations in functions known to be required for genetic recombination were moved into the dnaB107 and dnaB+ backgrounds and the strains were assayed in fluctuation tests for their rate of deletion of the 787 bp tandem tetA repeats.

3.2.

Genotype

Rate (×10−3 ) (95% CI)

Relative rate

dnaB107 dnaB+ dnaB107 recA304 dnaB107 recB2053 dnaB107 recF400 dnaB107 recO1504 dnaB107 recR252 dnaB107 recJ2003 dnaB107 recQkan dnaB107 recN1502 dnaB107 recG102 dnaB107 yicR::kan dnaB107 ruvA63 dnaB107 ruvC53 dnaB107 radA::kan dnaB107 polB dnaB107 dinB dnaB107 umuC

1.7 (0.82–2.5) 0.0039b (0.0022–0.0068) 0.0074b (0.0066–0.033) Inviableb 2.0b (1.2–4.0) 3.8b (2.9–4.4) 3.2b (2.6–3.9) 0.11 (0.033–0.37) 4.6 (0.85–8.8) 0.63b (0.58–1.0) 0.015b (0.0066–0.033) 2.1 (1.2–2.3) 0.0061 (0.0012–0.0079) Inviableb 0.025 (0.020–0.042) 4.1 (1.5–9.5) 0.30 (0.22–0.41) 4.7 (2.4–10)

1 0.0023 0.0044 NA 1.2 2.2 1.9 0.065 2.7 0.37 0.0088 1.2 0.0036 NA 0.015 2.4 0.18 2.8

CI = 95% confidence interval [14]; NA = not applicable. a

b

Presumed early functions in recombination

As shown previously [13,14], loss of the RecA strand transfer protein reduces rearrangements by over 300-fold in dnaB107 strains. E. coli has two systems that load RecA onto ssDNA to initiate recombination [3]: one dependent on RecBCD that promotes recombination at double strand ends and the other dependent on RecFOR that appears to promote recombination at single-strand DNA gaps. Mutations in the RecFOR pathway do not affect deletion rates in the dnaB107 strains [14]; if anything, rates are slightly elevated. However, mutation in the single-strand exonuclease RecJ reduced deletion rates 15fold (Table 2). Loss of the RecQ helicase, with which RecJ has been proposed to act presynaptically (see [3]), had no effect (Table 2). As previously reported [14], mutation of the Smclike RecN protein, implicated in repair of double-strand breaks [23–25] had a slight effect, reducing deletion rates three-fold in the dnaB107 background. None of these functions were required for deletion in a dnaB+ strain (Table 3), although deletion rates may be slightly elevated in recB and recJ mutants. Rearrangements in the dnaB+ background are independent of homologous recombination and occur via a distinct pathway dependent on dnaK [26].

3.3.

Table 2 – Recombination gene dependence of tandem repeat deletiona in dnaBts strains

Putative late functions in recombination

Mutants in functions implicated in recombination intermediate processing were also assayed. A mutation in the RecG branch migration helicase produced a strong reduction in deletion rates, over 100-fold. We reported this observation previously [14], although at the time the recG102 allele was reported to be in the radC/yicR gene [27]; subsequent work identified this mutation as an allele of recG [28], which we have confirmed. Deletion of the yicR gene (the opening read-

Deletion of 787 bp tetA repeats integrated into chromosomal lacZ [22]. Data from Saveson and Lovett [14]. In that experimental set, the deletion rate for wild-type control was 3.7 × 10−6 , with 95% confidence intervals of (2.8–5.3) × 10−6 and that for dnaB107 was 1.3 × 10−3 , with confidence interval of (0.98–1.9) × 10−3 , both in good agreement with this determination.

ing frame formerly denoted as “radC”) had no effect on deletion rate in the dnaB107 background. A mutation in ruvA, a component of the RuvAB branch migration helicase caused a strong reduction in dnaB107-stimulated rearrangements, to a level comparable to that seen for recA mutants. A muta-

Table 3 – Recombination gene dependence of tandem repeat deletion in dnaB+ strains Genotype a

wt recA304 recB2053 recF400 recO1504 recR252 recJ2003 recQkan recN1502 recG102 yicR::kan ruvA63 ruvC53 radA::kan polB::aadA dinB::kan umuC::cat

Rate (×10−6 ) (95% CI) 3.9 (2.2–6.8) 3.7 (1.7–9.0) 9.4 (4.4–18) 6.9a (5.9–8.0) 5.7a (3.1–8.4) 4.9a (2.7–6.2) 11 (4.0–16) 3.2 (1.9–4.8) 3.6a (2.0–4.9) 17a (13–19) 3.8 (1.5–8.5) 2.8 (1.4–14) 21 (7.5–41) 2.1 (1.4–8.4) 3.1 (1.9–6.7) 3.9 (1.4–6.4) 3.1 (1.7–5.8)

Relative rate 1 0.95 2.4 1.8 1.5 1.3 2.8 0.82 0.92 4.4 0.97 0.72 5.4 0.54 0.95 1.0 0.95

STL695, assay strain for deletion of 787 bp tetA repeats integrated into chromosomal lacZ [22]. CI = 95% confidence interval. a

Data from [14]; deletion rate for wild-type in that study was 3.7 × 10−6 , with 95% confidence intervals of (2.8–5.3) × 10−6 , in good agreement with this set.

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

tion in the RadA gene, previous shown to exhibit synergistic phenotypes with recG or ruvABC mutations [29], also strongly reduced rearrangement rates, about 70-fold. This is the first reported instance of recombination dependent on radA in an otherwise wild-type strain background. Mutations in radA and ruvA had little effect on deletion in the dnaB+ strain and the recG mutation promoted four-fold higher levels of deletion in a dnaB+ background (Table 3). Interesting, a mutation in the RuvC Holliday junction endonuclease, also elevated deletion events substantially in dnaB+ strains, by five-fold (mutants in RuvC could not be tested in the dnaBl07 background because of synthetic lethality [14]). We have previously observed a hyperrecombinational phenotype associated with ruvC mutations [30].

3.4.

SOS polymerases

DNA damage induces three DNA polymerases that can act to bypass lesions in DNA templates: polymerase II (the product of the polB gene), polymerase IV (encoded by dinB) and polymerase V (encoded by umuCD) [16,17]. We introduced disruption or deletion mutations in genes encoding these polymerases into dnaB107 and dnaB+ in which tandem repeat deletion rates could be measured. Mutations in umuC and polB did not affect deletion rates in either wild-type or dnaB mutant backgrounds. However, the dinB mutation significantly reduced recovery of deletion in the dnaB107 background, by over five-fold.

3.5. Effects of recombination and replication functions on RecBCD-dependent P1 transduction Phage P1 transduction involves the transfer of ca. 80 kb chromosomal dsDNA fragments that can be inherited by recombination with the recipient chromosome [19]. For those functions that were required for optimal levels of dnaB107-stimulated deletion rearrangements, I determined their effects on P1 transduction (Table 4). Levels of Arg+ transductants were significantly reduced, 30-fold, in recB mutants. The recG102 allele also reduced transduction inheritance about 10-fold in these assays, a slightly stronger effect than has previously been reported for recG [31], which may be a consequence of this particular allele. Mutants in ruvA and radA showed modestly reduced transduction levels, three and twofold, respectively. Other mutants, including recJ, dinB, and recN

Table 4 – Recombination efficiency, P1 transduction Genotype

Relative Arg+ transduction efficiency

S.D.

Wild-type recB2053 recJ2053 recN1502 recG102 ruvA63 radAl::kan dinB::kan

1 0.043 1.2 0.78 0.093 0.37 0.50 0.68

0.025 0.03 0.38 0.04 0.12 0.30 0.16

S.D. = standard deviation.

1425

showed little or no reduction in Arg+ inheritance by P1 transduction.

4.

Discussion

A mutation in the DnaB fork helicase, dnaBl07, causes greatly elevated rates of deletion and expansion of tandem repeat sequences at its permissive temperature of growth [13–15]. Presumably difficulties in replication trigger breakage of the fork, with its subsequent repair giving rise to the observed repeat rearrangements. Other alleles of DnaB, including dnaB8, have been deduced to increase the occurrence of regressed and broken replication forks [11,32]. Investigation of the genetic requirements for the rearrangements stimulated by dnaB107 gives us some insight in the process of replication fork collapse and repair. The genetic dependence of deletion rearrangements that are stimulated by the DnaB107 allele of the replication fork helicase are consistent with a mechanism involving replication fork repair by RecA RecBCD-dependent homologous recombination [14]. Because RecBCD enzymatic processing requires a free double-strand end [3,4], this suggests that broken chromosomes accumulate in dnaBl07 mutants and that their repair can lead to repeat rearrangements in one cell out of a thousand. Recovery of these repeat rearrangements also requires several other branched DNA processing enzymes including RuvA and RecG as well as the RecA paralog protein, RadA. Mutants in these functions have particularly strong effects, exhibiting reductions of about 100-fold, to levels approximating that of RecA mutants. The requirement for RadA is particularly striking, since in other assays for recombination, mutations in RadA have effects only when combined with those in RecG or RuvABC [29]. I could not test the requirement for the Holliday junction resolving endonuclease, since ruvC mutants are lethal in combination with dnaB107 [14]. Mutants in the ssDNA exonuclease RecJ also yielded a strong reduction in repeat rearrangements. Defects in Smc-like protein RecN showed a slight reduction, consistent with its role in promoting double strand break repair. Other recombination factors had no effect on the recovery of repeat rearrangements, including the RecFOR proteins that mediate ssDNA gap recombination and the RecQ helicase. YicR, formerly known as RadC, the ORF misidentified as mutated in the radC[recG]102 mutant, also had no effect. Misinvasion of a broken chromosome at the site of repeated sequence homology can account for the occurrence of repeat rearrangements concomitant with repair of collapsed replication forks (Fig. 2). As such, this invasion necessarily limits the amount of heteroduplex DNA that can be formed to 787 bp or less, the length of homology between the repeats. Factors that stabilize D-loop intermediates and promote their conversion to replication forks are expected to be especially important for recovery of these rearrangements. It is potentially this limited amount of homology in a critical intermediate that explains the fact that many recombination factors, including those known to process D-loops, are more stringently required for dnaB107-stimulated rearrangements than they are for other recombination events. Although RecG, RuvA and RadA have very modest effects and appear genetically redun-

1426

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

dant for recombination in standard genetic crosses [29,31], they show here very strong effects on DnaB107-stimulated rearrangements. These factors may affect recombination in this assay by mediating branch migration of D-loop or Holliday junction intermediates to extend heteroduplex DNA (Fig. 1C), stabilize the joint through DNA binding, promote cleavage of D-loops or Holliday junctions (Fig. 1C ) or aid recruitment of DNA polymerases, such as DinB (Fig. 1B ), all of which stabilize the strand invasion product. A similar strong effect on DnaB107-stimulated rearrangements is seen for RecJ exonuclease. Because pre-synaptic DNA digestion of broken forks is likely to occur via RecBCD, RecJ must perform some aspect of post-synaptic DNA processing to promote recovery of these rearrangements. In RecBCDmediated recombination events where homology is limited, RecJ exonuclease activity contributes to recovery of products [33]. In vitro, RecJ promotes RecA-mediated branch migration by post-synaptic degradation of the DNA strand that competes for pairing with the invading DNA strand [34], which may explain its requirement here. Perhaps the most surprising result of this study is that recovery of dnaB107-stimulated deletions requires Polymerase IV, the product of the SOS-induced dinB gene. The other translesion polymerases, Polymerase II (polB) and Polymerase V (umuCD) were dispensable. Extension of the invading 3 strand by PolIV may be required to stabilize the D-loop, prior to re-establishment of a DNA polymerase III-dependent replication fork. This is highly reminiscent of recent work showing a similar property for eukaryotic translesion polymerase ␩. Pol ␩ can extend D-loops formed in vitro [35] and is required for optimal recovery of gene conversion and other double-strand break induced recombination events in DT40 chicken cells [36]. Therefore, in both prokaryotic and eukaryotic cells, translesion polymerases may aid double-strand break repair by stabilization of strand invasion intermediates. Recombination proteins and DNA recombination intermediates may actively recruit specialized repair polymerases to facilitate repair of broken forks by 3 -end extension, prior to the more elaborate reassembly of bona-fide replication forks.

Acknowledgments This work was supported by a grant from the General Medical Sciences Institute GM51753. I thank G. Sharpies, H. Ohmori and R. Lloyd for providing E. coli strains and C.J. Saveson for initially constructing and testing the radA mutant.

references

[1] M.M. Cox, M.F. Goodman, K.N. Kreuzer, D.J. Sherratt, S.J. Sandler, K.J. Marians, The importance of repairing stalled replication forks, Nature 404 (2000) 37–41. [2] A. Kuzminov, Collapse and repair of replication forks in Escherichia coli, Mol. Microbiol. 16 (1995) 373–384. [3] S.C. Kowalczykowski, D.A. Dixon, A.K. Eggleston, S.D. Lauder, W.M. Rehrauer, Biochemistry of homologous recombination in Escherichia coli, Microbiol. Rev. 58 (1994) 401–465. [4] R.S. Myers, F.W. Stahl, Chi and the RecBCD enzyme of Escherichia coli, Annu. Rev. Genet. 28 (1994) 49–70.

[5] V. Burland, G. Plunkett, D.L. Daniels, F.R. Blattner, DNA sequence and analysis of 136 kilobases of the Escherichia coli genome: organizational symmetry around the origin of replication, Genomics 16 (1993) 551–561. [6] G.J. Sharples, S.M. Ingleston, R.G. Lloyd, Holliday junction processing in bacteria: insights from the evolutionary conservation of RuvABC, RecG, and RusA, J. Bacteriol. 181 (1999) 5543–5550. [7] R.C. Heller, K.J. Marians, The disposition of nascent strands at stalled replication forks dictates the pathway of replisome loading during restart, Mol. Cell 17 (2005) 733–743. [8] S.J. Sandler, Multiple genetic pathways for restarting DNA replication forks in Escherichia coli K-12, Genetics 155 (2000) 487–497. [9] F.N. Capaldo, G. Ramsey, S.D. Barbour, Analysis of the growth of recombination-deficient strains of Escherichia coli K-12, J. Bacteriol. 118 (1974) 242–249. [10] A. Kuzminov, Instability of inhibited replication forks in E. coli, Bioessays 17 (1995) 733–741. [11] B. Michel, S.D. Ehrlich, M. Uzest, DNA double-strand breaks caused by replication arrest, EMBO J. 16 (1997) 430–438. [12] K.J. Marians, Prokaryotic DNA replication, Annu. Rev. Biochem. 61 (1992) 673–719. [13] C.J. Saveson, S.T. Lovett, Enhanced deletion formation by aberrant DNA replication in Escherichia coli, Genetics 146 (1997) 457–470. [14] C.J. Saveson, S.T. Lovett, Tandem repeat recombination induced by replication fork defects in Escherichia coli requires a novel factor, RadC, Genetics 152 (1999) 5–13. [15] A.S. Morag, C.J. Saveson, S.T. Lovett, Expansion of DNA repeats in Escherichia coli: effects of recombination and replication functions, J. Mol. Biol. 289 (1999) 21–27. [16] M.F. Goodman, B. Tippin, Sloppier copier DNA polymerases involved in genome repair, Curr. Opin. Genet. Dev. 10 (2000) 162–168. [17] P. Pham, S. Rangarajan, R. Woodgate, M.F. Goodman, Roles of DNA polymerases V and II in SOS-induced error-prone and error-free repair in Escherichia coli, Proc. Natl. Acad. Sci. U.S.A. 98 (2001) 8350–8354. [18] B.J. Bachmann, Derivations and genotypes of some mutant derivatives of Escherichia coli K-12, in: F.C. Neidhardt (Ed.), Escherichia coli and Salmonella: Cellular and Molecular Biology, ASM Press, Washington DC, 1996, pp. 2460–2488. [19] J.H. Miller, A Short Course in Bacterial Genetics, Cold Spring Harbor Press, New York, 1992. [20] N.S. Willetts, A.J. Clark, B. Low, Genetic location of certain mutations conferring recombination deficiency in Escherichia coli, J. Bacteriol. 97 (1969) 244–249. [21] D.E. Lea, C.A. Coulson, The distribution of the numbers of mutants in bacterial populations, J. Genet. 49 (1949) 264–285. [22] S.T. Lovett, P.T. Drapkin, V.A. Sutera Jr., T.J. Gluckman-Peskind, A sister-strand exchange mechanism for recA-independent deletion of repeated DNA sequences in Escherichia coli, Genetics 135 (1993) 631–642. [23] J.L. Kosa, Z.Z. Zdraveski, S. Currier, M.G. Marinus, J.M. Essigmann, RecN and RecG are required for Escherichia coli survival of Bleomycin-induced damage, Mutat. Res. 554 (2004) 149–157. [24] T.R. Meddows, A.P. Savory, J.I. Grove, T. Moore, R.G. Lloyd, RecN protein and transcription factor DksA combine to promote faithful recombinational repair of DNA double-strand breaks, Mol. Microbiol. 57 (2005) 7–110. [25] S.M. Picksley, P.V. Attfield, R.G. Lloyd, Repair of DNA double-strand breaks in Escherichia coli K12 requires a functional recN product, Mol. Gen. Genet. 195 (1984) 267–274. [26] S.J. Goldfless, A.S. Morag, K.A. Belisle, V.A. Sutera Jr., S.T. Lovett, DNA repeat rearrangements mediated by DnaK-dependent replication fork repair, Mol. Cell 21 (2006) 595–604.

d n a r e p a i r 5 ( 2 0 0 6 ) 1421–1427

[27] I. Felzenszwalb, S. Boiteux, J. Laval, Cloning of the Escherichia coli radC gene: identification of the RadC protein, Braz. J. Med. Biol. Res. 26 (1993) 1261–1268. [28] M.J. Lombardo, S.M. Rosenberg, radC102 of Escherichia coli is an allele of recG, J. Bacteriol. 182 (2000) 6287–6291. [29] C.E. Beam, C.J. Saveson, S.T. Lovett, Role for radA/sms in recombination intermediate processing in Escherichia coli, J. Bacteriol. 184 (2002) 6836–6844. [30] S.T. Lovett, R.L. Hurley, V.A. Sutera Jr., R.H. Aubuchon, M.A. Lebedeva, Crossing over between regions of limited homology in Escherichia coli, RecA-dependent and RecA-independent pathways, Genetics 160 (2002) 851–859. [31] R.G. Lloyd, Conjugational recombination in resolvase-deficient ruvC mutants of Escherichia coli K-12 depends on recG, J. Bacteriol. 173 (1991) 5414–5418. [32] M. Seigneur, S.D. Ehrlich, B. Michel, RuvABC-dependent double-strand breaks in dnaBts mutants require recA, Mol. Microbiol. 38 (2000) 565–574. [33] L. Miesel, J.R. Roth, Evidence that SbcB and RecF pathway functions contribute to RecB CD-dependent transductional recombination, J. Bacteriol. 178 (1996) 3146–3155.

1427

[34] S.E. Corrette-Bennett, S.T. Lovett, Enhancement of RecA strand-transfer activity by the RecJ exonuclease of Escherichia coli, J. Biol. Chem. 270 (1995) 6881–6885. [35] M.J. Mcllwraith, A. Vaisman, Y. Liu, E. Fanning, R. Woodgate, S.C. West, Human DNA polymerase eta promotes DNA synthesis from strand invasion intermediates of homologous recombination, Mol. Cell 20 (2005) 783–792. [36] T. Kawamoto, K. Araki, E. Sonoda, Y.M. Yamashita, K. Harada, K. Kikuchi, C. Masutani, F. Hanaoka, K. Nozaki, N. Hashimoto, S. Takeda, Dual roles for DNA polymerase eta in homologous DNA recombination and translesion DNA synthesis, Mol. Cell 20 (2005) 793–799. [37] M. Escarceller, J. Hicks, G. Gudmundsson, G. Trump, D. Touati, S. Lovett, P.L. Foster, K. McEntee, M.F. Goodman, Involvement of Escherichia coli DNA polymerase II in response to oxidative damage and adaptive mutation, J. Bacteriol. 176 (1994) 6221–6228. [38] S.T. Lovett, V.A. Sutera Jr., Suppression of recJ exonuclease mutants of Escherichia coli by alterations in DNA helicases II (uvrD) and IV (helD), Genetics 140 (1995) 27–45.