• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of embojLink to Publisher's site
EMBO J. Feb 8, 2006; 25(3): 596–604.
Published online Jan 19, 2006. doi:  10.1038/sj.emboj.7600941
PMCID: PMC1383526

RuvAB is essential for replication forks reversal in certain replication mutants


Inactivated replication forks may be reversed by the annealing of leading- and lagging-strand ends, resulting in the formation of a Holliday junction (HJ) adjacent to a DNA double-strand end. In Escherichia coli mutants deficient for double-strand end processing, resolution of the HJ by RuvABC leads to fork breakage, a reaction that we can directly quantify. Here we used the HJ-specific resolvase RusA to test a putative role of the RuvAB helicase in replication fork reversal (RFR). We show that the RuvAB complex is required for the formation of a RusA substrate in the polymerase III mutants dnaEts and holD, affected for the Pol III catalytic subunit and clamp loader, and in the helicase mutant rep. This finding reveals that the recombination enzyme RuvAB targets forks in vivo and we propose that it directly converts forks into HJs. In contrast, RFR occurs in the absence of RuvAB in the dnaNts mutant, affected for the processivity clamp of Pol III, and in the priA mutant, defective for replication restart. This suggests alternative pathways of RFR.

Keywords: DNA repair, genome stability, helicase, recombination, replication restart


Replication fork progression may be impeded by the encounter of obstacles such as DNA-bound proteins, or by a transient deprivation of an essential replication component. The consequences of replication fork arrest has been the subject of extensive studies in the past years and replication fork arrest is now recognized as an important source of DNA rearrangements in all organisms (Michel, 2000; Carr, 2002; Kolodner et al, 2002). A growing number of proteins known to participate in homologous recombination, DNA repair or replication restart have been proposed to act at blocked forks in vivo, and some of them directly bind Y-shaped DNA molecules that mimic fork structures in vitro (Jones and Nakai, 2000; Sandler and Marians, 2000; McGlynn and Lloyd, 2002; Hishida et al, 2004; Flores et al, 2005; Heller and Marians, 2005a). The rules that govern the action of different fork-binding proteins at inactivated replication forks in vivo presumably depend on the conditions of fork arrest and are still largely unknown.

Replication fork arrest leads in several Escherichia coli replication mutants to a reaction called replication fork reversal (RFR; Seigneur et al, 1998; reviewed in Michel et al, 2004). This reaction involves the annealing of leading- and lagging-strand ends, to form a Holliday junction (HJ) adjacent to a DNA double-strand end (Figure 1A). In E. coli, DNA double-strand ends are processed by the recombinase-exonuclease RecBCD and HJs by the resolvase RuvABC (Kuzminov, 1999). In the absence of RecBCD, resolution of the RFR-made HJ by RuvABC leads to fork breakage (Figure 1E). The RFR model was based on direct quantitative analysis of fork breakage and supported by genetic data. To date, the reaction was proposed to occur in: (i) two helicase mutants, the rep mutant defective for an accessory replicative helicase and dnaBts cells, in which the main replicative helicase is inactivated at high temperature (Seigneur et al, 1998); (ii) three Pol III mutants, dnaEts, dnaNts and holDQ10am, respectively, affected for the catalytic Pol III subunit, the processivity clamp and one of the subunits of the clamp loader (Flores et al, 2001; Grompone et al, 2002) and (iii) finally in the priA mutant defective for the main replication restart pathway (Grompone et al, 2004a).

Figure 1
The RFR model (adapted from Seigneur et al, 1998). In the first step (A), the replication fork is arrested by inactivation of dnaE or dnaN protein, causing fork reversal. The reversed fork forms a four-arm structure (HJ, two alternative representations ...

Although the processing of reversed forks by recombination proteins is well understood (Figure 1B–D), less is known about the mechanism of their formation. In vitro, both the homologous recombination protein RecA and the helicase RecG can reverse a particular fork structure (McGlynn and Lloyd, 2000; Robu et al, 2001, 2004). In vivo, RecA is required for RFR in the dnaBts mutant affected for the replicative helicase (Seigneur et al, 2000), but not in other replication mutants (Seigneur et al, 2000; Flores et al, 2001; Grompone et al, 2002, 2004a). RecG has been proposed to promote RFR at replication forks blocked by a UV lesion (McGlynn and Lloyd, 2002), a reaction which is now controversial (Donaldson et al, 2004; Wang, 2005). To date, in most conditions of replication inactivation, the mechanism by which a blocked replication fork is converted into an HJ remains unknown.

RuvA, RuvB and RuvC proteins form two complexes: a RuvAB complex with helicase and branch migration activities and a RuvABC complex that resolves HJs (reviewed in West, 1997). Forks are not broken when the RuvABC complex is inactive, because there is no other HJ resolvase in wild-type E. coli. This requirement for RuvABC for HJ resolution prevented us from testing whether the RuvAB complex could be able to bind blocked replication forks in vivo and to convert them into HJs. To address this question, we took advantage of the RusA protein (Mandal et al, 1993; Mahdi et al, 1996). RusA is a resolvase encoded by a cryptic E. coli prophage; it specifically resolves HJs in vitro (Sharples et al, 1994). In vivo, RusA is able to resolve HJs in E. coli as well as in heterologous organisms such as yeast and mammalian cells (Doe et al, 2000; Saintigny et al, 2002). The rusA gene is not expressed in wild-type E. coli due to the lack of a functional promoter, but mutations such as rus-1 activate the expression of a functional RusA protein and suppress the homologous recombination defect of ruvAB or ruvABC mutants (Mandal et al, 1993; Mahdi et al, 1996). In addition to HJs made by homologous recombination, RusA cleaves reversed forks in the helicase mutant dnaBts, in which RFR requires RecA (Seigneur et al, 2000).

In this work, we used the RusA HJ resolvase to test in different replication mutants whether RFR occurs in the absence of RuvAB. We first compared in detail two Pol III mutants, dnaNts affected for the processivity clamp and dnaEts affected for the polymerase subunit. RusA cleaves forks in dnaNts ruvABC mutant cells, indicating that reversal of dnaNts-blocked forks does not require RuvAB. In contrast, RusA is not able to cleave blocked forks in dnaEts ruvAB mutants, indicating that the RuvAB helicase is required for the conversion of dnaEts-blocked forks into HJs. The analysis of the other replication mutants in which RFR occurs confirms that RuvAB is required for fork reversal in some but not all of them.


RFR occurs in dnaNts ruvABC

Replication fork breakage is catalyzed in Pol IIIts mutants by the action of the HJ resolvase RuvABC on reversed forks (Grompone et al, 2002). Quantification of the amount of chromosomal DNA able to enter pulse-field gels is used to measure fork breakage (Seigneur et al, 1998; see Materials and methods). As only fully linear DNA is able to enter pulse-field gels, breakage of both replication forks in a circular chromosome is needed for a measurable linearization. Since the presence of RecB and RecC proteins prevents forks cleavage (Figure 1; Michel et al (2004) and references therein), RuvABC-dependent breakage can only be measured in cells in which either recB or recC is inactivated (see Supplementary Table S1 for strain constructions). In the dnaNts recB mutant, a high level of fork breakage occurs at 37°C (62.3%), which is lower when DnaN is fully inactivated at 42°C (36.7%) and is largely abolished by ruvABC inactivation (8.1% at 37°C; Figure 2A and Supplementary Table S2). The lower level of fork breakage in the dnaNts recB mutant at 42°C compared to 37°C has been previously described and is specific for this mutant (Grompone et al, 2002). Reasons for this observation are presently unknown (a partial DnaN activity may be required for RFR, or the DnaN mutant protein may block access to the replication fork at 42°C; Grompone et al, 2002).

Figure 2
RusA cleaves forks in the absence of RuvABC in a dnaNts mutant, but not in a dnaEts mutant. The histograms indicate the percentage of linear DNA in cultures propagated for 3 h at restrictive temperature (42°C) or semipermissive temperature (37°C). ...

In order to test whether the helicase action of the RuvAB complex is required for the conversion of inactivated replication forks into HJs, we used ruvABC rus-1 mutant cells, which lack RuvABC and express the RusA resolvase. dnaNts recB (RuvABC active), dnaNts recB ruvABC (no active resolvase) and dnaNts recB ruvABC rus-1 cells (RusA active) were compared. A high level of DNA breakage was also observed at 37°C in the ruvABC rus-1 mutant context, that is, when RusA is active (59.9%), indicating that HJs are formed at dnaNts-blocked forks in the absence of RuvAB (Figure 2A and Supplementary Table S2). Similar results were obtained when the RecBC complex is inactivated at high temperature by recBts and recCts mutations (called here recBCts) rather than a recB null mutation (Supplementary Table S2). It should be noted that 38% linear DNA forms in the recBCts ruvABC rus-1 mutant in which replication is not affected by a mutation (JJC2712; Figure 2B and Supplementary Table S2). The origin of spontaneous chromosome breakage is unknown. It may result from spontaneous replication arrest and resolution by RusA of Ruv-independent RFR occurring at these arrested replication forks. In addition, the level of linear DNA in replication-proficient cells is higher in ruvABC mutants when RusA is expressed (38.1% of linear DNA in recBCts ruvABC rus-1 cells at 42°C versus 20% in recB ruvABC mutant; Seigneur et al, 1998), presumably because RusA resolves recombination intermediates that would otherwise prevent some broken chromosomes from entering pulse-field gels. Nevertheless, the use of the rus-1 allele and the comparison of DnaN+ and dnaNts cells allow us to conclude that at dnaNts-blocked forks RFR does not require RuvABC. The occurrence of RFR in the absence of RuvAB suggests alternative pathways of RFR, but it does not exclude the involvement of RuvAB in RFR in the dnaNts mutant.

RFR does not occur in dnaEts ruvABC

The rus-1 allele was similarly used to test whether HJs are still formed at blocked forks in the dnaEts mutant when RuvABC proteins are absent. In contrast with the high level of fork breakage in the dnaEts recBCts mutant (64%; Figure 2B and Supplementary Table S3), the level of chromosome breakage in dnaEts recBCts ruvABC rus-1 was not significantly different from that in a DnaE+ recBCts ruvABC rus-1 mutant (37.2 versus 38%; Figure 2B and Supplementary Table S3). Similar results were observed when RecBCD was inactivated by a recB null mutation (recB268::Tn10, JJC2647; Supplementary Table S3). Consequently, the comparison of DnaE+ and dnaEts cells indicates that DnaE inactivation does not cause fork breakage when RusA is active and RuvABC absent. Similar results were obtained when only RuvA and RuvB were absent (ruvA60::Tn10 mutation, J1JC2725; Supplementary Table S3). The absence of RusA-dependent fork breakage in the dnaEts recBCts ruvAB(C) rus-1 mutant did not result from a peculiar inhibition of the RusA protein in this mutant, as the rus-1 allele fully suppressed the sensitivity to UV irradiation caused by ruvAB and ruvABC mutations in this background as in all mutants (data not shown).

In order to ascertain that the lack of dnaEts-induced fork breakage in a ruvABC rus-1 context results from the absence of RuvABC, a plasmid carrying functional ruvABC genes was introduced in the ruvABC mutant (the low copy plasmid routinely used for complementation experiments, pGB-ruvABC, has no deleterious effect in contrast with high copy plasmids carrying ruvABC genes; Seigneur et al, 1998; Grompone et al, 2002; Lopez et al, 2005). Expression of RuvABC from pGB-ruvABC in the dnaEts recBCts ruvABC rus-1 mutant restored more than 60% of linear DNA, whereas the plasmid vector had no effect (Figure 2B and Supplementary Table S3). This result confirms that the inactivation of ruvABC is responsible for the lower level of fork breakage dnaEts recBCts ruvABC rus-1 cells compared to dnaEts recBCts. The plasmid pGB-ruvAB, which expresses only RuvA and RuvB (Seigneur et al, 1998), was also tested. It was previously reported that cells that express RuvAB and RusA but not RuvC are unable to resolve HJ, presumably because, as shown in vitro, RuvAB prevents RusA action by masking the HJ (Chan et al, 1997). Accordingly, when pGB-ruvAB was introduced in dnaEts recBCts ruvABC rus-1 cells, it prevented homologous recombination at 30°C (rendering cells UV sensitive) and did not restore fork breakage (measured by pulse-field gel quantification, data not shown). In conclusion, the lack of dnaEts-induced fork breakage when RusA is active and RuvAB inactive indicates that RFR requires a functional RuvAB complex in the dnaEts mutant, in contrast with the dnaNts mutant.

RuvAB is required for RFR at dnaEts-blocked forks in recA or recFOR mutants

Previous results indicate that in both dnaEts and dnaNts mutants a futile reaction takes place at blocked forks prior to RFR: RecA binds to the fork with the help of the presynaptic proteins RecQ, RecJ and RecFOR, and the resulting recombination intermediate is then undone by the UvrD helicase (Flores et al, 2005). In order to determine whether the requirement for RuvAB for RFR is dependent on this ‘RecA-binding/RecA-removal' reaction, RusA-catalyzed fork breakage was tested in recF, recO and recA backgrounds (Figure 3, and Supplementary Tables S4 and S5). As in the RecAFOR+ context, the level of DNA breakage measured in a dnaEts recBCts ruvABC rus-1 context (RusA active and ruvABC inactive) was significantly lower (31–37%) than in the respective RuvABC+ isogenic strain (55–65%), and not significantly different from the DnaE+ recBCts ruvABC rus-1 recA (or recO) mutant (32.5–41%; Figure 3A, and Supplementary Tables S4 and S5). Expression of RuvABC from a plasmid in dnaEts recBCts ruvABC rus-1 recF (recO) cells restored a high level of fork breakage, indicating that the low level of breakage results from the absence of RuvABC (Figure 3B and Supplementary Table S5). These results show that RuvAB is required for the occurrence of RusA-catalyzed fork breakage, and hence for the formation of reversed forks in the dnaEts mutant, regardless of the presence of RecFOR or RecA. Similarly, breakage in a dnaEts recBCts recF uvrD mutant (60.5%) was decreased to 42% by the combination of ruvABC rus-1 mutations, a level similar to that of the DnaE+ recBCts recO uvrD ruvABC rus-1 mutant (JJC2722; Supplementary Table S5). This observation shows that, in a uvrD recFOR context also, RusA does not act at forks inactivated by the dnaEts mutation when the RuvAB complex is absent. We conclude that RuvAB is required for RFR in the dnaEts mutant regardless of the ‘RecA-binding/RecA-removal' reaction.

Figure 3
RusA does not cleave forks in dnaEts recBCts ruvABC rus-1 mutants regardless of the presence or absence of RecA or RecF. The histograms indicate the percentage of linear DNA in cultures propagated for 3 h at restrictive temperature (42°C). Bars ...

Preventing RFR is not lethal in dnaEts at 37°C

dnaEts cells can propagate at 37°C, but with a reduced ability to form colonies (Figure 4A; Grompone et al, 2002). dnaEts cells are chronically induced for the SOS response and their viability is strongly improved by mutations that abolish or decrease SOS induction, such as lexAind, recA, recFOR or recQ mutations (Flores et al, 2005; Figure 4). The dnaEts ruvABC mutant is not viable at 37°C (Figure 4A) and its lethality is suppressed by the inactivation of either recA or recFOR (Figure 4). However, the lethality of the dnaEts ruvABC mutant is not suppressed by the sole inactivation of the SOS induction, since dnaEts lexAind ruvABC and dnaEts recQ ruvABC mutants are killed upon shift to 37°C (data not shown). As RecFOR and RecA are required for recombinational gap repair as well as SOS induction, we propose that their inactivation alleviates the need for RuvABC by preventing recombinational gap repair (allowing, e.g., gap filling by a polymerase). Indeed, gaps may form during chromosome replication in the dnaEts mutant as a consequence of defects in lagging-strand synthesis, and recombine in a RecA- and RecFOR-dependent way, forming HJs that render RuvABC essential for viability. If RFR were required for the viability of the dnaEts mutant at 37°C, dnaEts ruvABC recF and dnaEts ruvABC recA strains would not be viable because of the lack of RFR. Since these mutants can propagate at 37°C, we conclude that the dnaEts mutant does not require RFR for growth at this temperature.

Figure 4
RFR is not required for growth of the dnaEts mutant at semipermissive temperature. Cells propagated at 30°C for 2 h were shifted to 37°C and appropriate dilutions were plated at the indicated times; plates were counted after 48-h incubation ...

We tested whether the inactivation of ruvABC, which suppresses fork breakage, suppresses the requirement for RecBC for growth of dnaEts cells at 37°C. To avoid the lethal effect of gap repair in dnaEts cells that lack RuvABC, the question was addressed in a recF background. The inactivation of ruvABC allowed a slow growth of the dnaEts recF recB mutant at 37°C in liquid medium (Figure 4B) and allowed colony formation (data not shown). This result supports the idea that RuvAB is responsible for the requirement for RecB upon replication impairment.

RecA, RecG and UvrD are not required for RFR in dnaNts

As RFR is RecA-dependent in the dnaBts mutant and RuvAB-dependent in dnaEts mutant, we tested a possible redundant role of RuvAB and RecA for the catalysis of RFR in the dnaNts mutant by measuring chromosome breakage in a dnaNts mutant that lacks both. As RecFOR can be required for RecA binding to block forks, the recO mutation was also tested. A high level of RusA-catalyzed fork breakage was observed in dnaNts recB ruvABC rus-1 recA (or recO) mutants (66–71%; Table I), indicating that dnaNts-blocked forks are reversed in the absence of both RuvAB and RecA (or both RuvABC and RecFOR).

Table 1
RecA, RecG and UvrD are not required for RFR in dnaNts

RuvAB and RecG share the property of being able to migrate HJs in vitro and in vivo (Lloyd and Sharples, 1993a). RecG has also been shown to reverse fork-like structures in vitro, and possibly in vivo (McGlynn and Lloyd, 2000; Robu et al, 2004). We tested whether RecG could be the helicase that reverses dnaNts-blocked forks in the ruvAB recA context, but we observed that RusA-mediated fork breakage occurs in dnaNts cells that lack RecA, RuvAB and RecG (58%; Table I). This indicates that forks are reversed by yet another activity. Finally, RusA-catalyzed fork cleavage occurs in the dnaNts mutant when RecFOR, UvrD and RuvABC are all inactive (JJC2741/JJC2729; Table I). Therefore, these experiments did not allow identification of the RFR pathway that operates at dnaNts-blocked forks.

RuvAB is required for RFR in holDQ10am and rep mutants, but not in the priA mutant

Based on the occurrence of RuvABC-dependent fork breakage in a recBC mutant context, RFR was proposed to occur in several replication mutants. The holDQ10am mutant carries an amber mutation in the gene encoding the HolD protein, one of the polypeptides of the Pol III clamp loader. This mutant was isolated in a background in which the mutation is poorly suppressed (Flores et al, 2001). In the AB1157 background used here, the amber mutation is only partially suppressed also since the holDQ10am mutant is lethal when RecBC is inactivated and the holDQ10am recBCts mutant suffers fork breakage at 42°C (Figure 5 and Supplementary Table S6). RFR also occurs in the priA mutant, in which spontaneously arrested replication forks persist because of the inactivation of the main replication restart pathway (Grompone et al, 2004a) and in the rep mutant (Seigneur et al, 1998). The role of the Rep helicase in vivo is not entirely elucidated; it is proposed to remove obstacles in front of replication forks and to participate in replication restart (Seigneur et al, 1998; Sandler, 2000; Heller and Marians, 2005b). In holDQ10am, priA and rep mutants, RFR was shown to be RecA-independent and the molecular mechanism of the reaction is still unknown (Seigneur et al, 2000; Flores et al, 2001; Grompone et al, 2004a). To test whether RuvAB is required for RFR in these three mutants, holDQ10am, rep and priA mutations were combined with the recBCts ruvABC rus-1 mutations. As these combinations of mutations affect growth, strain constructions were made in the presence of a complementing plasmid with conditional replication origin (see Materials and methods). Therefore, mutants were constructed in HolD+, PriA+ or Rep+ backgrounds, respectively, and holDQ10am-, priA- or rep-deficient cells were isolated after segregation of the plasmid (in addition, these mutants were propagated on minimal medium to further prevent the appearance of suppressor mutations).

Figure 5
RusA cleaves forks in holD recBCts ruvABC rus-1 mutants and rep recBCts ruvABC rus-1, but not in priA recB ruvABC rus-1 cells. The histograms indicate the percentage of linear DNA in cultures propagated for 3 h at 42°C. Bars indicate standard ...

Measures of fork breakage by pulse-field gels showed that RusA did not cause breakage in the holDQ10am recBCts ruvABC rus-1 mutant (28.9 versus 51.5% in RuvABC+ cells; Figure 5 and Supplementary Table S6). Therefore, similarly to dnaEts-blocked forks, holDQ10am-blocked forks are not converted to HJs in the absence of RuvAB. In contrast, the level of DNA breakage was similar in priA recB and priA recB ruvABC rus-1 cells, indicating that HJs form at priA-inactivated forks in the absence of RuvAB (Figure 5 and Supplementary Table S6). Finally, the level of fork breakage in rep recBCts ruvABC rus-1 cells (47.6%) was lower than in rep recBCts cells (67.7%), indicating that RuvAB is involved in the formation of reversed forks at rep-blocked forks. Nevertheless, this level was higher than in Rep+ recBCts ruvABC rus-1 cells (38.8%), suggesting that in the absence of ruvAB, a low level of RFR still occurs at rep-blocked forks (Figure 5 and Supplementary Table S6). The analysis of holDQ10am, priA and rep mutants confirms that the action of RuvAB at blocked forks depends on the cause of replication inactivation.


In this work, we show that, in addition to its central role in HJ branch migration, RuvAB participates in the formation of HJs at blocked replication forks. This finding reveals a new role for the RuvAB complex during chromosome dynamics, in the rescue of inactivated replication forks. Interestingly, RuvAB is not required in certain replication mutants for the conversion of blocked replication forks into HJs, which underlines that the processing of blocked replication is a complex reaction that depends on the cause of replication inactivation.

RFR has been proposed in E. coli, Saccharomyces cerevisiae and mammalian cells (Higgins et al, 1976; Seigneur et al, 1998; Sogo et al, 2002; Subramanian and Griffith, 2005), but little is known on the mechanism of the reaction. RuvAB is the paradigm of HJ-binding enzymes and its functional homologue in eukaryotes remains to be clearly identified (Kaliraman et al, 2001; Gaillard et al, 2003; Osman et al, 2003; Liu et al, 2004). We used here the HJ-specific RusA resolvase to investigate the role of RuvAB in the formation of reversed forks because of the high specificity of RusA for four-way junctions (Bolt and Lloyd, 2002; Rafferty et al, 2003). In two different Pol III mutants, the RusA resolvase does not act at blocked forks in the absence of RuvAB or RuvABC, which points to a need for RuvAB to convert a blocked fork into a HJ. Indeed, it is unlikely that forks could be reversed and protected from RusA in these mutants. RusA is able to resolve HJs made: (i) by homologous recombination, (ii) at dnaBts-blocked forks reversed by RecA and (iii) at dnaNts- and priA-blocked forks reversed by an unknown activity. Furthermore, it resolves HJs in heterologous environments such as yeast or mammalian cells, where apparently none of the endogenous protein prevents its action (Doe et al, 2000; Saintigny et al, 2002). Consequently, we favor a model according to which RusA does not cleave forks in the absence of RuvAB because RuvAB itself catalyzes the conversion of replication forks into HJs.

Extensive biochemical and structural studies have shown that RuvA proteins form tetramers that bind HJ in preference to any other substrate (reviewed in West, 1997). Two tetramers assemble as an octamer that sandwiches a HJ (reviewed in West, 1998). RuvB forms hexamers that target double-strand DNA on each side of the RuvA octamer and pull on DNA strands, causing branch migration of the RuvA-bound HJ (Stasiak et al, 1994). RuvA and RuvB also bind and unwind double-stranded Y-DNA molecules in vitro, although less efficiently than HJs (Lloyd and Sharples, 1993b; Hiom et al, 1996; McGlynn and Lloyd, 2001). In contrast, RuvC is specific for HJs as it does not act at RuvAB-bound Y-structures (Benson and West, 1994). In vitro, the nature of the RuvA–RuvB complex formed on Y-molecules seems to depend on the experimental conditions. It was originally observed that incubation of Y-DNA molecules with RuvA and RuvB leads to the formation of a complex composed of one RuvA tetramer and only one RuvB hexamer bound to one of the three branches (Hiom et al, 1996). Such RuvA–RuvB complexes could catalyze branch migration (George et al, 2000). Accordingly, a RuvA mutant protein that has lost the capacity to form octamers binds HJ poorly, but still binds Y-DNA as efficiently as wild-type RuvA protein, again forming a complex with one RuvA tetramer and one RuvB hexamer per Y-molecule (Privezentzev et al, 2005). In contrast, in a different study, two RuvB hexamers were proposed to bind to Y-molecules in the presence of RuvA (McGlynn and Lloyd, 2001). On a Y-molecule, the action of one RuvB hexamer bound to the template strand will promote RFR (Figure 6), whereas the action of two RuvB hexamers bound to the leading and lagging strand arms will unwind the template strands (McGlynn and Lloyd, 2001). Whatever the reasons for the reports of different complexes in vitro, the action of RuvAB at blocked forks in vivo is likely to be strongly influenced by the protein environment and our results indicate that this action leads to RFR. Consequently, our results are best accounted for by the model shown in Figure 6. We propose that conversion of replication forks to HJs can occur by the assembly of a tetramer of RuvA on a three-arm replication fork, followed by the binding of a single hexamer of RuvB to the template-strand arm. The helicase activity of the RuvB hexamer would convert the three-arm fork into a four-way structure to which a second RuvB hexamer and a second RuvA tetramer can bind. Next, in this bona fide branch migration complex, RuvC can resolve the junction. In vivo, the lagging-strand arm is likely to be single-stranded and covered by single-strand DNA-binding protein (SSB). In agreement, RuvA binds single-strand DNA more efficiently than double-strand DNA in vitro, and SSB can stimulate RuvAB activity (Shiba et al, 1991; Parsons et al, 1995). In support of the idea that SSB might be an important structural determinant of the fate of blocked forks, SSB strongly influences the helicase activity of several fork-binding proteins in vitro (Jones and Nakai, 1999; Cadman and McGlynn, 2004; Heller and Marians, 2005b).

Figure 6
Model of action of RuvAB at blocked forks. In the first step, a RuvA tetramer binds to the fork and drives the assembly of a RuvB hexamer on the template strands. The translocase action of this RuvB hexamer pulls the leading and lagging strands into the ...

Our results underline the existence of several pathways for RFR, which presumably depend on the structure of the fork and/or the nature of the replication proteins that remain bound after inactivation. RuvAB may reverse all types of blocked forks, and its action would not be detectable in the dnaNts and priA mutants because of the existence of another pathway of RFR. Alternatively, RuvAB may reverse forks only in dnaEts and holD mutants, either because it is targeted specifically to blocked forks in these mutants, or because it is prevented from binding to dnaNts- or priA-blocked forks. The action of RuvAB at dnaEts-blocked forks is independent of the presence of RecA or presynaptic recombination proteins (RecFOR, RecB), which are likely to be present in the vicinity of recombination intermediates during homologous recombination. RuvAB is therefore not targeted to forks by its homologous recombination partners. Replication in bacteria takes place at localized, highly ordered, multi-protein complexes called ‘replication factories' (Lemon and Grossman, 1998; Lau et al, 2003; Molina and Skarstad, 2004); it is conceivable that RuvAB binding to blocked forks is controlled by interactions with a protein that belongs to the replication factory. To date, the precise constituents of replication factory and, in turn, the control of the accessibility of replication forks to different repair proteins are still largely unknown (Sherratt, 2003).

Although RuvAB is nearly ubiquitous in prokaryotes (Rocha et al, 2005), its functional homologue(s) in unicellular or pluricellular eukaryotes has not been clearly identified so far. Several eukaryote proteins have been shown to recognize HJs in vitro, and the role of these enzymes in homologous recombination is the subject of intense research (Kaliraman et al, 2001; Gaillard et al, 2003; Osman et al, 2003; Liu et al, 2004). We show here that in E. coli the enzyme that catalyzes branch migration of HJs is also required for their formation at certain blocked forks, in a reaction that is dependent on the cellular context. The dual function of the prokaryotic RuvAB complex in homologous recombination and replication fork processing strengthens the links between these two processes and may provide clues for understanding the roles of its functional homologues in other organisms.

Materials and methods

Strains and plasmids

The strain background is JJC40, which is a hsdR Thr+ Pro+ derivative of AB1157 (leu-6 thi-1, his-4, lacY1, galK2, ara-14, xylS, mtl-1, tsx-33, rpsL31 and supE44). Most of the strains were constructed by P1 transduction. Antibiotics were used at the following concentrations: ampicillin 20 μg/ml for chromosome-carried genes and 50 μg/ml for plasmids, tetracyclin 10 μg/ml, chloramphenicol 10 μg/ml, spectinomycin 60 μg/l, kanamycin 50 μg/ml and rifampicin 50 μg/ml. Details of strain construction and strains genotypes are described in Supplementary Table S1. All mutants were checked for thermosensitivity at 42°C, for UV sensitivity (suppression of the sensitivity of ruvAB and ruvABC mutants to UV irradiation ascertains the functionality of RusA in all backgrounds) and for exonuclease V activity (the exonuclease V action of RecBC is inactivated at all temperatures in the recBCts mutant and exo V-deficient cells allow the growth of T4 gpII mutant phages). The plating efficiency at 37°C was systematically measured (recBCts and recB mutation prevented growth at 37°C for all mutants except priA (Grompone et al, 2004b). The mutator phenotype of uvrD mutants was tested by plating overnight cultures on rifampicin plates. In addition, the presence of insertional mutations was routinely checked by PCR amplification of the corresponding gene.

The low copy vector pGB2 carries the pSC101 replicon (Churchward et al, 1984). Its derivatives pGB-ruvAB (Seigneur et al, 1998) and pGB-ruvABC (Seigneur et al, 2000) have been described. rep recBCts ruvABC rus-1 cells were constructed in the presence of the pGBts-rep plasmid (Michel et al, 2000). Plasmidless cells were isolated by propagation at 42°C in liquid medium for 2 h and then plating at 30°C. The HolD+ and PriA+ derivatives of the IPTG-dependent plasmid pAM34 (Gil and Bouche, 1991) have been described (Flores et al, 2001; Grompone et al, 2004b; Veaute et al, 2005) and were used for construction of holDQ10am and priA mutants. Plasmidless cells were isolated by propagation in liquid medium devoid of IPTG for 2 h, followed by plating at 30°C on a medium devoid of IPTG. Plasmidless cells were always propagated on minimal medium. They formed smaller colonies than plasmid-containing cells and were checked for their sensitivity to UV irradiation and for their sensitivity to rich medium (priA mutants have a decreased plating efficiency on rich compared to minimum medium).

Growth curves

Overnight cultures grown at 30°C were diluted to OD 0.001 or 0.002 in Luria Broth medium containing additional thymine (LBT), grown at 30°C for 2 h and then shifted to 37°C. Aliquots were taken and dilutions were plated on LB every hour. Plates were incubated for 48 h at 30°C.

Measure of linear DNA by PFGE

Quantification of pulse-field gels was performed as described previously (Seigneur et al, 1998). Briefly, for chromosome labeling, cells were grown in minimal medium in the presence of tritiated thymidine and deoxyadenosine for 3 h at 30°C; then, part of the culture was shifted to 37 or 42°C for three more hours. Cells were collected, washed and embedded in agarose plugs. Gentle lysis was performed in plugs, which were then used for pulse-field gel electrophoresis (to avoid DNA damaging during PFGE, the apparatus was routinely washed with 0.1% SDS). The proportion of migrating DNA was determined by cutting each lane in slices and counting the tritium present in the wells and in the gel slices. Results were compared by a KHI2 test and were considered as (i) highly significantly different when P<0.001 (values that differ by more than 10% were always found highly significantly different, as, for example, the levels of chromosome breakage in dnaEts recBCts ruvABC rus-1 and dnaEts recBCts cells at 42°C), (ii) significantly different when 0.001<P<0.05 (rep recBCts and rep recBCts ruvABC rus-1 cells, which differ by 10%, were found significantly different) and (iii) not significantly different when P>0.05 (values that differ by less than 10% were found not significantly different, as, for example, the lack of effect of recG ruvABC rus-1 mutations in dnaNts recA (pBR-Gam)).

Supplementary Material

Supplementary Tables


We thank Dr David Leach, Dr Philippe Noirot and Dr Patrice Polard for very helpful reading of the manuscript. We also thank Nicolas Sanchez for excellent technical assistance. This work was supported, in part, by the ACI ‘Microbiologie 2003' and the ACI ‘Biologie Cellulaire, Moleculaire et Structurale 2004' of the Ministère de la Recherche Française. BM is on the CNRS staff.


  • Benson FE, West SC (1994) Substrate specificity of the Escherichia coli ruvc protein—resolution of three- and four-stranded recombination intermediates. J Biol Chem 269: 5195–5201 [PubMed]
  • Bolt EL, Lloyd RG (2002) Substrate specificity of RusA resolvase reveals the DNA structures targeted by RuvAB and RecG in vivo. Mol Cell 10: 187–198 [PubMed]
  • Cadman CJ, McGlynn P (2004) PriA helicase and SSB interact physically and functionally. Nucleic Acids Res 32: 6378–6387 [PMC free article] [PubMed]
  • Carr AM (2002) Checking that replication breakdown is not terminal. Science 297: 557–558 [PubMed]
  • Chan SN, Harris L, Bolt EL, Whitby MC, Lloyd RG (1997) Sequence specificity and biochemical characterization of the RusA Holliday junction resolvase of Escherichia coli. J Biol Chem 272: 14873–14882 [PubMed]
  • Churchward G, Belin D, Nagamine Y (1984) A pSC101-derived plasmid which shows no sequence homology to other commonly used cloning vectors. Gene 31: 165–171 [PubMed]
  • Doe CL, Dixon J, Osman F, Whitby MC (2000) Partial suppression of the fission yeast rqh1(−) phenotype by expression of a bacterial Holliday junction resolvase. EMBO J 19: 2751–2762 [PMC free article] [PubMed]
  • Donaldson JR, Courcelle CT, Courcelle J (2004) RuvAB and RecG are not essential for the recovery of DNA synthesis following UV-induced DNA damage in Escherichia coli. Genetics 166: 1631–1640 [PMC free article] [PubMed]
  • Flores MJ, Bidnenko V, Michel B (2004) The DNA repair helicase U or D is essential for replication fork reversal in replication mutants. EMBO Rep 5: 983–988 [PMC free article] [PubMed]
  • Flores MJ, Bierne H, Ehrlich SD, Michel B (2001) Impairment of lagging strand synthesis triggers the formation of a RuvABC substrate at replication forks. EMBO J 20: 619–629 [PMC free article] [PubMed]
  • Flores MJ, Sanchez N, Michel B (2005) A fork-clearing role for UvrD. Mol Microbiol 57: 1664–1675 [PubMed]
  • Gaillard PH, Noguchi E, Shanahan P, Russell P (2003) The endogenous Mus81–Eme1 complex resolves Holliday junctions by a nick and counternick mechanism. Mol Cell 12: 747–759 [PubMed]
  • George H, Kuraoka I, Nauman DA, Kobertz WR, Wood RD, West SC (2000) RuvAB-mediated branch migration does not involve extensive DNA opening within the RuvB hexamer. Curr Biol 10: 103–106 [PubMed]
  • Gil D, Bouche JP (1991) ColE1-type vectors with fully repressible replication. Gene 105: 17–22 [PubMed]
  • Grompone G, Ehrlich D, Michel B (2004a) Cells defective for replication restart undergo replication fork reversal. EMBO Rep 5: 607–612 [PMC free article] [PubMed]
  • Grompone G, Sanchez N, Dusko Ehrlich S, Michel B (2004b) Requirement for RecFOR-mediated recombination in priA mutant. Mol Microbiol 52: 551–562 [PubMed]
  • Grompone G, Seigneur M, Ehrlich SD, Michel B (2002) Replication fork reversal in DNA polymerase III mutants of Escherichia coli: a role for the beta clamp. Mol Microbiol 44: 1331–1339 [PubMed]
  • Heller RC, Marians KJ (2005a) The disposition of nascent strands at stalled replication forks dictates the pathway of replisome loading during restart. Mol Cell 17: 733–743 [PubMed]
  • Heller RC, Marians KJ (2005b) Unwinding of the nascent lagging strand by Rep and PriA enables direct restart of stalled replication forks. J Biol Chem 280: 34143–34151 [PubMed]
  • Higgins NP, Kato K, Strauss B (1976) A model for replication repair in mammalian cells. J Mol Biol 101: 417–425 [PubMed]
  • Hiom K, Tsaneva IR, West SC (1996) The directionality of RuvAB-mediated branch migration: in vitro studies with three-armed junctions. Genes Cells 1: 443–451 [PubMed]
  • Hishida T, Han YW, Shibata T, Kubota Y, Ishino Y, Iwasaki H, Shinagawa H (2004) Role of the Escherichia coli RecQ DNA helicase in SOS signaling and genome stabilization at stalled replication forks. Genes Dev 18: 1886–1897 [PMC free article] [PubMed]
  • Jones JM, Nakai H (1999) Duplex opening by primosome protein PriA for replisome assembly on a recombination intermediate. J Mol Biol 289: 503–516 [PubMed]
  • Jones JM, Nakai H (2000) PriA and phage T4 gp59: factors that promote DNA replication on forked substrates. Mol Microbiol 36: 519–527 [PubMed]
  • Kaliraman V, Mullen JR, Fricke WM, BastinShanower SA, Brill SJ (2001) Functional overlap between Sgs1-Top3 and the Mms4-Mus81 endonuclease. Gene Dev 15: 2730–2740 [PMC free article] [PubMed]
  • Kolodner RD, Putnam CD, Myung K (2002) Maintenance of genome stability in Saccharomyces cerevisiae. Science 297: 552–557 [PubMed]
  • Kuzminov A (1999) Recombinational repair of DNA damage in Escherichia coli and bacteriophage lambda. Microbiol Mol Biol Rev 63: 751–813 [PMC free article] [PubMed]
  • Lau IF, Filipe SR, Soballe B, Okstad OA, Barre FX, Sherratt DJ (2003) Spatial and temporal organization of replicating Escherichia coli chromosomes. Mol Microbiol 49: 731–743 [PubMed]
  • Lemon KP, Grossman AD (1998) Localization of bacterial DNA polymerase: evidence for a factory model of replication. Science 282: 1516–1519 [PubMed]
  • Liu Y, Masson JY, Shah R, O'Regan P, West SC (2004) RAD51C is required for Holliday junction processing in mammalian cells. Science 303: 243–246 [PubMed]
  • Lloyd RG, Sharples GJ (1993a) Dissociation of synthetic Holliday junctions by E-coli RecG protein. EMBO J 12: 17–22 [PMC free article] [PubMed]
  • Lloyd RG, Sharples GJ (1993b) Processing of recombination intermediates by the RecG and RuvAB proteins of Escherichia coli. Nucleic Acids Res 21: 1719–1725 [PMC free article] [PubMed]
  • Lopez CR, Yang S, Deibler RW, Ray SA, Pennington JM, Digate RJ, Hastings PJ, Rosenberg SM, Zechiedrich EL (2005) A role for topoisomerase III in a recombination pathway alternative to RuvABC. Mol Microbiol 58: 80–101 [PubMed]
  • Mahdi AA, Sharples GJ, Mandal TN, Lloyd RG (1996) Holliday junction resolvases encoded by homologous rusA genes in Escherichia coli K-12 and phage 82. J Mol Biol 257: 561–573 [PubMed]
  • Mandal TN, Mahdi AA, Sharples GJ, Lloyd RG (1993) Resolution of Holliday intermediates in recombination and DNA repair—indirect suppression of ruvA, ruvB, and ruvC mutations. J Bacteriol 175: 4325–4334 [PMC free article] [PubMed]
  • McGlynn P, Lloyd RG (2000) Modulation of RNA polymerase by (P)ppGpp reveals a RecG-dependent mechanism for replication fork progression. Cell 101: 35–45 [PubMed]
  • McGlynn P, Lloyd RG (2001) Action of RuvAB at replication fork structures. J Biol Chem 276: 41938–41944 [PubMed]
  • McGlynn P, Lloyd RG (2002) Recombinational repair and restart of damaged replication forks. Nat Rev Mol Cell Biol 3: 859–870 [PubMed]
  • Michel B (2000) Replication fork arrest and DNA recombination. Trends Biochem Sci 25: 173–178 [PubMed]
  • Michel B, Grompone G, Flores MJ, Bidnenko V (2004) Multiple pathways process stalled replication forks. Proc Natl Acad Sci USA 101: 12783–12788 [PMC free article] [PubMed]
  • Michel B, Recchia GD, PenelColin M, Ehrlich SD, Sherratt DJ (2000) Resolution of Holliday junctions by RuvABC prevents dimer formation in rep mutants and UV-irradiated cells. Mol Microbiol 37: 180–191 [PubMed]
  • Molina F, Skarstad K (2004) Replication fork and SeqA focus distributions in Escherichia coli suggest a replication hyperstructure dependent on nucleotide metabolism. Mol Microbiol 52: 1597–1612 [PubMed]
  • Osman F, Dixon J, Doe CL, Whitby MC (2003) Generating crossovers by resolution of nicked Holliday junctions: a role for Mus81–Eme1 in meiosis. Mol Cell 12: 761–774 [PubMed]
  • Parsons CA, Stasiak A, West SC (1995) The E. coli RuvAB proteins branch migrate Holliday junctions through heterologous DNA sequences in a reaction facilitated by SSB. EMBO J 14: 5736–5744 [PMC free article] [PubMed]
  • Privezentzev CV, Keeley A, Sigala B, Tsaneva IR (2005) The role of RuvA octamerization for RuvAB function in vitro and in vivo. J Biol Chem 280: 3365–3375 [PubMed]
  • Rafferty JB, Bolt EL, Muranova TA, Sedelnikova SE, Leonard P, Pasquo A, Baker PJ, Rice DW, Sharples GJ, Lloyd RG (2003) The structure of Escherichia coli RusA endonuclease reveals a new Holliday junction DNA binding fold. Structure (Cambridge) 11: 1557–1567 [PubMed]
  • Robu ME, Inman RB, Cox MM (2001) RecA protein promotes the regression of stalled replication forks in vitro. Proc Natl Acad Sci USA 98: 8211–8218 [PMC free article] [PubMed]
  • Robu ME, Inman RB, Cox MM (2004) Situational repair of replication forks: roles of RecG and RecA proteins. J Biol Chem 279: 10973–10981 [PubMed]
  • Rocha EP, Cornet E, Michel B (2005) Comparative and evolutionary analysis of the bacterial homologous recombination systems. PLoS Genet 1: e15. [PMC free article] [PubMed]
  • Saintigny Y, Makienko K, Swanson C, Emond MJ, Monnat RJ Jr (2002) Homologous recombination resolution defect in Werner syndrome. Mol Cell Biol 22: 6971–6978 [PMC free article] [PubMed]
  • Sandler SJ (2000) Multiple genetic pathways for restarting DNA replication forks in Escherichia coli K-12. Genetics 155: 487–497 [PMC free article] [PubMed]
  • Sandler SJ, Marians KJ (2000) Role of PriA in replication fork reactivation in Escherichia coli. J Bacteriol 182: 9–13 [PMC free article] [PubMed]
  • Seigneur M, Bidnenko V, Ehrlich SD, Michel B (1998) RuvAB acts at arrested replication forks. Cell 95: 419–430 [PubMed]
  • Seigneur M, Ehrlich SD, Michel B (2000) RuvABC-dependent double-strand breaks in dnaBts mutants require RecA. Mol Microbiol 38: 565–574 [PubMed]
  • Sharples GJ, Chan SN, Mahdi AA, Whitby MC, Lloyd RG (1994) Processing of intermediates in recombination and DNA repair: identification of a new endonuclease that specifically cleaves Holliday junctions. EMBO J 13: 6133–6142 [PMC free article] [PubMed]
  • Sherratt DJ (2003) Bacterial chromosome dynamics. Science 301: 780–785 [PubMed]
  • Shiba T, Iwasaki H, Nakata A, Shinagawa H (1991) SOS-inducible DNA repair proteins, RuvA and RuvB, of Escherichia-coli—functional interactions between RuvA and RuvB for ATP hydrolysis and renaturation of the cruciform structure in supercoiled DNA. Proc Natl Acad Sci USA 88: 8445–8449 [PMC free article] [PubMed]
  • Sogo JM, Lopes M, Foiani M (2002) Fork reversal and ssDNA accumulation at stalled replication forks owing to checkpoint defects. Science 297: 599–602 [PubMed]
  • Stasiak A, Tsaneva IR, West SC, Benson CJB, Yu X, Egelman EH (1994) The Escherichia coli RuvB branch migration protein forms double hexameric rings around DNA. Proc Natl Acad Sci USA 91: 7618–7622 [PMC free article] [PubMed]
  • Subramanian D, Griffith JD (2005) p53 monitors replication fork regression by binding to ‘Chickenfoot' intermediates. J Biol Chem 280: 45568–45572 [PubMed]
  • Veaute X, Delmas S, Selva M, Jeusset J, Le Cam E, Matic I, Fabre F, Petit MA (2005) UvrD helicase, unlike Rep helicase, dismantles RecA nucleoprotein filaments in Escherichia coli. EMBO J 24: 180–189 [PMC free article] [PubMed]
  • Wang TC (2005) Discontinuous or semi-discontinuous DNA replication in Escherichia coli? BioEssays 27: 633–636 [PubMed]
  • West SC (1997) Processing of recombination intermediates by the RuvABC proteins. Ann Rev Genet 31: 213–244 [PubMed]
  • West SC (1998) RuvA gets X-rayed on Holliday. Cell 94: 699–701 [PubMed]

Articles from The EMBO Journal are provided here courtesy of The European Molecular Biology Organization
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...