Molecular Cell, Vol. 9, 241–251, February, 2002, Copyright 2002 by Cell Press
Direct Rescue of Stalled DNA Replication Forks via the Combined Action of PriA and RecG Helicase Activities Amanda V. Gregg, Peter McGlynn, Razieh P. Jaktaji, and Robert G. Lloyd1 Institute of Genetics Queen’s Medical Centre University of Nottingham Nottingham NG7 2UH United Kingdom
Summary The PriA protein of Escherichia coli plays a key role in the rescue of replication forks stalled on the template DNA. One attractive model of rescue relies on homologous recombination to establish a new fork via PriAmediated loading of the DnaB replicative helicase at D loop intermediates. We provide genetic and biochemical evidence that PriA helicase activity can also rescue a stalled fork by an alternative mechanism that requires manipulation of the fork before loading of DnaB on the lagging strand template. This direct rescue depends on RecG, which unwinds forks and Holliday junctions and interconverts these structures. The combined action of PriA and RecG helicase activities may thus avoid the potential dangers of rescue pathways involving fork breakage and recombination. Introduction The rescue of replication forks stalled on the DNA template is emerging as a feature of the cell cycle that is critically important for normal growth and survival and for maintaining the integrity of the genome. Although forks assembled at origins of replication have sufficient intrinsic processivity to complete chromosome duplication in a single round of origin firing, this may not always be achieved, at least in bacteria (Cox et al., 2000). Forks may stall at one of the many lesions generated in or on the DNA template by endogenous metabolism (Lindahl, 1996) and thus become inactivated with the result that other factors have to be recruited to complete the task. These are thought to include repair and recombination proteins to remove or bypass lesions, mechanisms for assembly of replication proteins at sites removed from normal origins, and checkpoint control systems to monitor progression. Defects in these processes may lead to increased sensitivity to DNA damage or allow inappropriate progression of the cell cycle, which may elevate the rates of mutation and chromosome rearrangements (Datta et al., 2000; Paulovich et al., 1997) and set human cells on the path to malignant transformation (Bell et al., 1999; Chakraverty and Hickson, 1999; Flores-Rozas and Kolodner, 2000). Links between DNA replication, recombination, and repair have been established over many years, but only recently has the extent of this interplay become widely appreciated (Cox, 2001). The close interplay of replica1
Correspondence:
[email protected]
tion and recombination in E. coli is thought to hinge around the formation of a Holliday junction from a stalled replication fork by regression of the fork and annealing of the nascent strands (McGlynn and Lloyd, 2000; Seigneur et al., 1998). Cleavage of this junction by the RuvABC Holliday junction resolvase would collapse the fork and generate a duplex DNA end that could be processed by RecBCD nuclease and RecA recombinase to initiate recombination with an intact sister duplex (Figure 1A, i–iii) or a homologous sequence (Kowalczykowski, 2000). Strand exchange creates a D loop that, based on studies in vitro, may be targeted by PriA to promote loading of the DnaB replicative helicase on the lagging strand template and the subsequent binding of DnaG primase. Both leading and lagging strand DNA synthesis by dimeric PolIII holoenzyme may then continue. Subsequent resolution of the (Holliday) junction at the D loop by RuvABC would restore a fully fledged replication fork that may replicate the remaining DNA, provided the original block to progression has been removed (Figure 1A, iii–iv). DnaB is sequestered in a complex with DnaC that prevents its inappropriate loading on ssDNA. Its transfer to the displaced strand of a D loop in vitro requires the combined activities of PriA, PriB, PriC, and DnaT (Liu and Marians, 1999). However, these factors can be bypassed by certain mutations in DnaC that allow direct loading from the mutant DnaB-DnaC complex (Liu et al., 1999; Xu and Marians, 2000) and which suppress the low viability and radiation sensitivity of a priA null strain (Sandler, 1996; Sandler et al., 1996, 1999). Thus, loading of DnaB on ssDNA is inferred to be the only function of PriA critical for replication restart. PriA has a 3⬘-5⬘ DNA helicase activity in vitro that can unwind both D loop and fork structures (Jones and Nakai, 1999; McGlynn et al., 1997; Nurse et al., 1999). A mutant protein (K230R) devoid of helicase activity retains the ability to bind D loops and load DnaB (Liu et al., 1999; McGlynn et al., 1997), and it suppresses the severe phenotype of a priA null strain (Kogoma et al., 1996; Sandler et al., 1996). However, overexpression of PriA, but not of the K230R mutant, reduces specifically the survival of UV-irradiated strains lacking RecG (AlDeib et al., 1996), a DNA branch migration protein implicated in recombination and DNA repair (Lloyd and Sharples, 1993). Conversely, certain mutant PriA proteins (SrgA) carrying substitutions in or close to the conserved helicase motifs function as suppressors of recG null strains (Al-Deib et al., 1996). These observations suggest that RecG counters some function associated with PriA helicase activity that can have a negative effect on cell survival. More recently, we proposed that RecG and PriA together act to rescue a stalled fork by facilitating loading of DnaB directly at the fork structure (McGlynn and Lloyd, 2000). The model was applied to a fork lacking a leading strand at the branch point and which therefore requires modification to position a leading strand 3⬘-OH at the branch point to prime leading strand synthesis before loading of DnaB can lead to productive fork pro-
Molecular Cell 242
Figure 1. Models of the Rescue of Replication Forks Stalled at Lesions in or on the DNA Template Mediated by Formation and Processing of Holliday Junctions (A) Rescue of a fork stalled at a replicative helicase block (red rectangle) mediated by junction cleavage, fork collapse, and recombination. (B) Rescue of a fork dislocated by a pyrimidine dimer (red triangle) blocking synthesis of the leading strand via manipulation of the fork structure and extension of the leading strand (a) or exonuclease trimming of the extended lagging strand (b). Possible protein functions mediating individual steps are indicated and discussed further in the text. Newly synthesized DNA is indicated in green.
gression. Two possible mechanisms of correction based on the formation of a Holliday junction by RecG (McGlynn and Lloyd, 2000, 2001; McGlynn et al., 2001) and subsequent processing of this structure are outlined schematically in Figure 1B, ii–iv. The first employs a polymerase to extend the leading strand at the junction and the second an exonuclease to trim the lagging strand. A complete fork is then established by branch migration of the junction in the reverse direction or by exonuclease digestion of the duplex spooled out during fork regression. The only major difference is in the timing of lesion repair. The role postulated for PriA is to unwind the lagging strand arm and then load DnaB on the exposed lagging strand template (Figure 1B, iv) (Jones and Nakai, 1999). Such PriA helicase activity is not required at a D loop since the displaced single strand is the one on which DnaB needs to be loaded (Figure 1A, iii). The models in Figure 1B suggest a precise function for PriA helicase in the loading of DnaB at fork structures and describe a possible scenario in which this might be essential for replication restart. Here, we present genetic and biochemical evidence supporting the idea that PriA promotes at least one pathway for the survival of UVirradiated cells, and thus by implication fork rescue, that does not rely on break-induced recombination and which requires PriA helicase function. The data suggest specific roles for both RecG and PriA helicase activities and explain how they may act sequentially to promote replication restart. Results Studies both in vivo and in vitro have suggested that the ability to rescue blocked replication forks is crucial for the survival of cells that have suffered damage to their DNA. The PriA, RuvABC, RecBCD, RecA, and RecG proteins are all involved in promoting survival and have
been implicated in fork rescue. To investigate how their functions might be related, we compared the effects of mutations that inactivate one or more of these proteins on cell viability and survival after irradiation with UV light. In doing so, we revisited some genotypes examined previously and tested their survival in parallel with new constructs. Evidence of at Least Two PriA-Dependent Pathways for Promoting Survival First, we found that a null allele of priA (priA2::kan) has a much more severe effect on survival and cell viability than mutations in recB, recG, or ruvA (Figures 2A and 2B; Table 1). Second, there is strong synergism between recB and recG and also between ruv and recG (Figures 2B and 2E). Indeed, survival of the double mutants is similar to that of a priA null strain. Furthermore, a ruv recG strain is as sensitive as a recA strain, and it is significant that it does not become more sensitive when RecA is eliminated (Figure 2E). These data indicate that all RecA-dependent mechanisms for promoting survival are blocked in strains lacking RecG and RuvABC. Third, mutations in recB and ruvA confer similar sensitivities to UV light (Figures 2B and 2C). The double mutant is no more sensitive (Figure 2C). We showed recently that a mutation in the  subunit of RNA polymerase called rpo*35 suppresses the sensitivity of a ruv strain (by a mechanism that is not yet fully understood) but has only a minor effect on the survival of a recB strain (McGlynn and Lloyd, 2000). The data in Figure 2C show that rpo*35 suppresses ruvA equally well, regardless of recB. This epistasis implies that RuvA acts upstream of RecB in the same pathway. These three sets of observations point to the involvement of PriA in two pathways for promoting the survival of UV-irradiated cells. One depends on RecG but not on RuvA or RecB. The other relies on RuvA and RecB but not RecG.
Rescue of Stalled Replication Forks 243
Figure 2. Epistasis Analysis of the Effect of priA, rec, and ruv Mutations on Survival of UV-Irradiated Cells The strains used (listed in Table 1 and Experimental Procedures) are identified by genotype and are: (A) AB1157, N3793, AG109, N2057; (B) N2101, N3789, N4703; (C) N2057, N4752, N4778, N5381; (D) RJ1073, RJ1075, RJ1086; (E) N4279, N5070, N5093, N5397.
To test this model directly, we investigated the effect of recG and ruvA mutations on a priA null strain. As reported previously (McGlynn and Lloyd, 2000), priA recG and priA ruv strains could be made by transducing the priA2::kan allele into recG or ruv derivatives of strains AB1157 or MG1665 and selecting for kanamycin-resistant colonies. Relative to control crosses with the rec⫹ruv⫹ AB1157 or MG1665 recipients, the yield of transductants was reduced ⵑ2-fold with the recG recipient and ⵑ6-fold with ruv recipients, which is to be expected from the mild recombination-deficient phenotype of these recipients (Lloyd, 1991). We also introduced priA2::kan into recG or ruv strains by cotransduction with argE. Using P1 from an arg⫹ priA2 donor strain and selecting Arg⫹ in crosses with argE3 recipients, cotransduction values of 42%, 30%, and 24% were obtained with rec⫹ruv⫹, recG265::cat, and ruvA60::Tn10 recipients, respectively. The values are based on the
number of Arg⫹ colonies tested (50 in each case) that proved resistant to kanamycin. In both types of crosses, the kanamycin-resistant transductants could be further subcultured to single colonies on LB agar, which in the case of the priA recG and priA ruv clones indicates that the constructs are viable, contrary to a recent report (McCool and Sandler, 2001). The number of transductants obtained also indicates that their recovery does not depend on the acquisition of unidentified suppressor mutations. However, the recG priA2 and ruvA priA2 constructs made do have a lower efficiency of plating on LB agar than a priA2 strain and form much smaller colonies (Table 1 and data not shown). Cultures grown in broth also tend to accumulate cells carrying suppressor mutations that improve plating efficiency and increase colony size, as happens with priA2 strains (see below) (Sandler, 1996; Sandler et al., 1996). Experiments based on cultures containing low numbers of such derivatives indi-
Table 1. Effect of priA, rec, and ruv Mutations on Cell Viability Straina
Relevant Genotype
Viabilityb
AB1157 N2057 N2101 N3793 AG109 AG156 AG257 RJ1075 RJ1073 AG216 AG181 AG182 AG197 AG198 AG199 AG200
rec⫹
1.0 0.65 ⫾ 0.076 0.41 ⫾ 0.0096 0.77 ⫾ 0.056 0.025 ⫾ 0.0056 0.008 ⫾ 0.0013c 0.0084 ⫾ 0.006c 0.10 ⫾ 0.014 0.033 ⫾ 0.013c 1.41 ⫾ 0.08 0.97 ⫾ 0.24 1.0 ⫾ 0.057 0.11 ⫾ 0.029 0.055 ⫾ 0.005 0.14 ⫾ 0.027 0.047 ⫾ 0.008
a
ruv⫹
priA⫹
dnaC⫹
ruvA60 recB268 ⌬recG263 priA2 priA2 ⌬recG265 priA2 ruvA60 priA4 priA4 ruvA60 dnaC212 priA2 dnaC212 priA2 dnaC212 ⌬recG264 priA2 dnaC212 recB268 priA2 dnaC212 recB268 ⌬recG264 priA2 dnaC212 ruvA60 priA2 dnaC212 ruvA60 ⌬recG264
All strains are derivatives of AB1157. Strains were grown in LB broth to an A650 of 0.4, and the number of colony-forming units (cfu) was determined by plating on LB agar. Values are relative to strain AB1157 (2.23 ⫻ 108 cfu/ml) and are based on measures of cfu in 3–12 independent cultures. c Cultures of these strains were examined by microscopy and the majority of cells were found to be highly filamentous. b
Molecular Cell 244
cate that recG263 priA2 is no more sensitive to UV than a priA2 single mutant and that ruvA60 priA2 is only marginally more sensitive (data not shown). These findings suggest that PriA is epistatic to both RecG and RuvA and are consistent with two PriA-dependent pathways, one relying on RecG and one on RuvA. However, the poor viability of the double mutants suggests that both RecG and RuvABC may have other roles that are independent of PriA. In an attempt to avoid the problem of suppressors arising in priA recG and priA ruvA strains, we turned to a partial function mutation of PriA encoded by priA4. This is a Tn10kan insertion generating a PriA fusion protein in which the final six amino acids at the C terminus are replaced with the sequence ADES PNDFGKNH (see Experimental Procedures). This allele confers reduced cell viability and increases sensitivity to UV light, but the effects are less severe than those seen with priA2 (compare Figures 2A and 2D; Table 1 and data not shown), which indicates that the PriA fusion protein retains some activity. The priA4 allele was used to test the effects of recG and ruvA null mutations in a strain with reduced PriA activity. The recG priA4 and ruvA priA4 strains made have a higher viability that the corresponding priA2 constructs (Table 1 and data not shown). Figure 2D shows they are also more sensitive to UV than a priA4 single mutant. However, the loss of RuvA clearly has a much more severe effect. Indeed, the survival of this strain was almost identical to that of a ruvA strain carrying the null priA2 allele (data not shown) or a ruvA recG strain (Figure 2B), which suggests that the mutant protein encoded by priA4 may be especially defective in an activity needed in a pathway that relies on RecG but not RuvABC. These data are therefore generally consistent with those obtained with the null priA2 allele. RuvABC and RecG Define Key Stages in Different PriA-Dependent Pathways Taken together, the data presented above indicate that one of two mechanisms can be employed to promote survival. The first depends on RuvABC and RecB, and the second depends on RecG. However, both rely on PriA. Since PriA is thought to rescue stalled forks by loading DnaB, it is possible that both pathways facilitate such loading. To investigate this possibility, we studied the effect of recB, recG, and ruvA mutations on a priA2 strain carrying a strong dnaC suppressor of priA2 of a type predicted to allow DnaB loading without the intervention of PriA, at least at D loops (Liu et al., 1999). This also allowed us to examine survival in the absence of PriA helicase activity. The suppressor chosen for this analysis has a L212V substitution in DnaC and is hereafter referred to as dnaC212 (see Experimental Procedures). On its own it has no detectable effect on survival of UV-irradiated cells, but it suppresses priA2 very well (Figure 3A, Table 1). However, the priA2 dnaC212 strain retains some sensitivity, which suggests that direct loading of DnaB from a mutant DnaB-DnaC complex cannot substitute fully for PriA-mediated loading. Significantly, elimination of RecG has very little additional effect, which indicates that RecG function is compromised already in priA2
Figure 3. Suppression of the priA Null Phenotype by dnaC212 The strains used (listed in Table 1 and Experimental Procedures) are identified by genotype and are: (A) AG109, AG181, AG182, AG216; (B) N2101, AG197, AG198; (C) N2057, N3789, AG199, AG200.
dnaC212. This result provides a second line of evidence that PriA has some specific activity that allows it to act with RecG to promote survival. In contrast, elimination of RuvA or RecB from the priA2 dnaC212 strain has an extreme effect (Figures 3B and 3C). Moreover, subsequent elimination of RecG has no additional effect. However, the ruvA recG derivative does seem a little more sensitive than the ruvA derivative (Figure 3C), and the recG mutation slightly reduces the viability of both the recB and ruvA constructs (Table 1). These results further support the hypothesis that there are at least two PriA-dependent pathways. Moreover,
Rescue of Stalled Replication Forks 245
Figure 5. Strand Specificity of Fork Unwinding by Wild-Type PriA
Figure 4. Effect of Plasmids Encoding Wild-Type PriA (PriA⫹) or the Helicase-Deficient K230R Derivative and pET3c Vector Controls on Survival of UV-Irradiated Strains
(A) Unwinding of complete and partial forks by 0, 0.05, 0.5, 5, and 50 nM protein. The fork structures shown are in lanes a–e, f–j, k–o, and p–t, and were at an initial concentration of 0.2 nM. The 3⬘ ends of the leading and/or lagging strands of the structures are indicated by arrows, and the position of the 5⬘ 32P labels are marked with a closed circle. Products of the unwinding reaction are identified on the right of the gel. (B) Rates of PriA-catalyzed accumulation of total DNA products from fork structures. The structures indicated were all at an initial concentration of 0.2 nM, and PriA was at 50 nM.
(A) Strains AG199 and N2057. (B) Strains AG182 and AG200.
since PriA helicase activity is missing in the priA2 dnaC212 strain and elimination of RecG has no additional effect, these findings suggest that PriA helicase acts with RecG to promote survival in a manner that is independent of RuvA and RecB. PriA Helicase Promotes Survival by a Mechanism that Requires RecG To see if the helicase function of PriA is the missing activity in a priA2 dnaC212 strain compromising the RecG pathway, we used plasmids expressing either wild-type PriA or the K230R derivative. The PriAK230R protein lacks helicase activity but is competent for loading DnaB at a fork. These plasmids and vector controls were introduced into priA2 dnaC212 strains lacking RuvA or both RuvA and RecG to see how they affected UV sensitivity. The PriA wild-type plasmid eliminates the extreme sensitivity of the ruvA construct (Figure 4A). Survival is identical to that of a ruvA single mutant (Figure 4A) and also of a ruvA dnaC212 strain (data not
shown) carrying the same plasmid, which implies that the presence of a mutant DnaB-DnaC complex does not interfere with PriA-mediated loading of DnaB. The K230R plasmid also improves survival of the priA2 dnaC212 ruvA strain. However, the effect is small by comparison with the wild-type plasmid, which suggests that direct loading from the DnaB-DnaC complex might be limiting in the absence of PriA helicase activity. In contrast, neither the K230R protein nor wild-type PriA is able to improve survival in the absence of RecG (Figure 4B). This finding establishes not only that PriA helicase activity has an important role in promoting survival but also that it acts specifically in a pathway that relies on RecG, but not RuvABC. PriA Helicase Activity at Fork Structures The results of our genetic studies indicate there are two PriA-dependent pathways promoting survival of UVirradiated cells. One depends on RuvABC and RecB but does not need PriA helicase activity. The other depends on RecG and does need the helicase function of PriA.
Molecular Cell 246
Figure 6. Fork Unwinding Activity of Hexa-Histidine-Tagged SrgA Mutants of PriA (A) Unwinding of a fork with a leading strand gap. Fork 5 DNA was at an initial concentration of 0.3 nM, and hexa-histidine tagged PriA or SrgA proteins were at concentrations of 0, 0.05, 0.5, 5, and 50 nM. (B) Rates of PriA- and SrgA-catalyzed accumulation of total DNA products from Fork 5. Proteins were at 2.5 nM and fork DNA at an initial concentration of 0.3 nM. (C) Unwinding of complete fork structures by hexa-histidine tagged PriA and SrgA proteins. The forks indicated were at initial concentrations of 0.3 nM, and tagged proteins were at 50 nM. (D) Rate of PriA- and SrgA-catalyzed accumulation of DNA products from Fork 1. Fork DNA was at an initial concentration of 0.3 nM, and proteins were at 5 nM.
To gain further insight into the interplay between the PriA and RecG helicases, we analyzed PriA helicase function in vitro. Wild-type PriA specifically unwinds the lagging strand from a fork structure, leaving the leading strand in place (Figure 5A, lanes a–j) (Jones and Nakai, 1999; Nurse et al., 1999). Unwinding is stimulated by a gap (4 nt) at the 5⬘ end of the lagging strand but is inhibited by a similar gap at the 3⬘ end of the leading strand (Figure 5B, compare Forks 6 and 7 with Forks 1 and 8). Similar findings were reported by Jones and Nakai (1999). PriA also removes the lagging strand of a fork missing a leading strand, albeit at a reduced rate, but not vice versa (Figure 5A, lanes k–t, and Figure 5B, substrate 4) (Jones and Nakai, 1999; McGlynn et al., 1997; Nurse et al., 1999). SrgA1 and SrgA2 Proteins Retain Helicase Activity at Fork Structures We next analyzed two mutant PriA proteins, each carrying a different amino acid substitution. They were chosen because they are the products of the srgA1 and srgA2 alleles of priA that were identified as strong suppressors of recG (Al-Deib et al., 1996). Their in vitro activities compared with wild-type PriA might thus shed
light on the interplay between PriA and RecG helicase activities in vivo. SrgA1 has a L557P substitution close to helicase motif V, and SrgA2 has a L425F substitution in motif IV. Both were designed to carry a hexa-histidine tag at the N terminus to facilitate their purification. Wildtype PriA with the same tag was purified using the same methods to provide an appropriate control. SrgA1 and SrgA2 retain the ability to unwind forks that have both leading and lagging strands and in which the 3⬘ end of the leading strand is at the branch point. With both proteins, activity is highest on a fork with a gap in the lagging strand, the substrate that in our hands best supports the helicase activity of wild-type PriA. SrgA1 is as active as the wild-type (tagged) PriA (Figure 6A, lanes a–j, and Figure 6B). SrgA2 is less active, but not substantially so (Figure 6A, lanes k–o, and Figure 6B). Both unwind the lagging strand only, as evident from the absence of product corresponding to the labeled leading strand (Figure 6A). The mutant proteins can also unwind complete forks with both leading and lagging strands abutting the branch point (Figure 6C, lanes c and d, and g and h). However, while SrgA1 accumulates product at nearly the same rate as the wild-type protein, the activity of SrgA2 is compromised (Figure 6D).
Rescue of Stalled Replication Forks 247
Figure 7. SrgA1 Protein Is Specifically Compromised in Its Ability to Unwind the Lagging Strand from a Fork Lacking a Leading Strand (A) Parallel analysis of unwinding of complete and partial fork structures by tagged wild-type PriA and SrgA proteins. The forks indicated were at an initial concentration of 0.3 nM, and PriA (lanes b, f, and j), SrgA1 (lanes c, g, and k), and SrgA2 (lanes d, h, and l) were at 50 nM. (B) Rate of PriA- and SrgA-catalyzed accumulation of DNA products from Fork 4. Fork DNA was at an initial concentration of 0.3 nM, and proteins were at 50 nM. (C) Band-shift assay showing binding of PriA and SrgA proteins to Fork 4 in the presence of EDTA. Fork DNA was at 0.2 nM, and the proteins indicated were at 0, 0.025, 0.1, 0.4, 1.6, 6.25, 25, and 100 nM. (D) Quantification of fork DNA binding. Values are based on repeats of band-shift assays as in (C).
SrgA1 Is Specifically Compromised in the Ability to Unwind a Partial Fork In contrast, both SrgA1 and SrgA2 are severely compromised in the ability to unwind the lagging strand from a fork lacking a leading strand. Figure 7A shows the results of a typical assay in which the unwinding activity of wild-type PriA, SrgA1, and SrgA2 was investigated in parallel on three fork structures. The lagging strand is clearly unwound from the complete fork by all three proteins (Figure 7A, lanes b–d), but none has any activity on a fork that has no lagging strand (Figure 7A, lanes f–g). Given enough protein, wild-type PriA can also dissociate the majority of a partial fork substrate lacking a leading strand (Figure 7A, lane j). As expected from the results in Figure 5A, the major labeled product is the unwound lagging strand, although a trace of partial duplex product is detectable. At the same high level of protein, very little dissociation is seen with either SrgA1 or SrgA2 (Figure 7A, lanes k and l). Measurement of the rates of unwinding confirmed that the two mutants have very low activity (Figure 7B). This failure to unwind the lagging strand fork is not due to reduced affinity for the structure. Both proteins bind the lagging strand fork with the same affinity as wild-type PriA, both in EDTA and in Mg2⫹ (Figures 7C and 7D; data not shown). Taken together, these findings indicate that SrgA2 has a general reduction in helicase activity on all forks, but
especially on those structures lacking a leading strand or without a gap on the lagging strand. SrgA1 has a specific defect related only to the absence of a leading strand at the branch point. These defects indicate that reduction of the ability of PriA to unwind a fork lacking a leading strand allows suppression of recG. Discussion We have presented evidence supporting the idea that the survival of UV-irradiated cells can be promoted by one of two radically different PriA-dependent pathways. One relies on both the RuvABC “resolvasome” complex and RecBCD nuclease, since it can be inactivated by mutation of either RuvA or RecB. The other relies on RecG helicase and is unaffected by RuvABC or RecBCD. They are further distinguished by their dependence on the helicase activity of PriA, which we found essential only for the RecG pathway. The existence of a second pathway promoted by the combined action of RecG and PriA helicase activities is supported by our finding that srgA suppressors of recG encode PriA proteins with altered helicase activities. The formation of a Holliday junction from a stalled replication fork and its subsequent processing by the RuvABC proteins is thought to play a key role in restarting replication (McGlynn and Lloyd, 2000; Seigneur et
Molecular Cell 248
al., 1998). Studies by Be´ne´dicte Michel and coworkers have shown that blocking advance of the DnaB replicative helicase results in chromosome breakage and that this breakage is mediated by RuvABC (Seigneur et al., 1998, 2000). These findings support models in which a Holliday junction forms at arrested forks. The data presented here are consistent with the idea that cleavage of such a junction by RuvABC and the subsequent recombination mediated by RecBCD and RecA provides an important mechanism for generating a substrate (a D loop) that PriA can employ to load DnaB and thereby promote cell survival after exposure to UV irradiation by enabling replication restart. The epistasis analysis presented here and in a previous study (McGlynn and Lloyd, 2000) using rpoB*35, ruvA, and recB mutations provides strong support for RuvABC-catalysed cleavage of a Holliday junction preceding RecBCD-mediated unwinding and degradation of a duplex DNA end. It also implies that RecBCD-catalysed degradation of the duplex DNA end spooled out during regression of a stalled fork is rarely employed to rescue forks blocked by UV damage. The second pathway we have identified also relies on PriA and presumably, therefore, on the formation of a substrate PriA can employ to load DnaB. However, this pathway acts independently of both RuvABC and RecBCD and thus, by implication, presumably does not involve fork breakage and recombination. What it does require is the helicase activity of both RecG and PriA. Our analysis of the SrgA mutants of PriA indicates that this helicase interaction may center on the targeting of damaged replication forks in which there is a lagging strand at the branch point, but no leading strand. These proteins also suppress the UV sensitivity of recG mutants (Al-Deib et al., 1996). Together, these findings imply that wild-type PriA helicase can target damaged forks lacking a leading strand and that this can be deleterious to survival of the cell. Why might such targeting by PriA reduce survival and, in particular, survival of cells lacking RecG? We assume the binding of any fork structure by PriA has the potential to subsequently catalyze loading of DnaB (Jones and Nakai, 1999). Targeting DnaB to the lagging strand of a fork lacking a leading strand may result in the formation of an abortive replication complex, because there would be no leading strand 3⬘-OH group at the branch point to prime continued leading strand synthesis. Furthermore, if DnaB helicase began to unwind the parental duplex, it would exacerbate replication problems by translocating the branch point still further from the 3⬘ end of the leading strand. This model implies RecG may normally prevent inappropriate targeting of such structures by PriA. This would explain how the specific defects in the helicase activities of the SrgA mutants could suppress the recG phenotype. The reduced abilities of these SrgA proteins to unwind the lagging strand specifically from a fork lacking a leading strand would reduce the chances of loading DnaB. In contrast, both SrgA proteins unwind the lagging strand from full fork structures. This implies they retain the ability to load DnaB at the preferred target of wild-type PriA. They may therefore be effective suppressors of recG because the ability to unwind the “wrong” fork structures is downregulated while the ca-
pacity to promote unwinding of forks with the potential to support productive replication is retained. They also probably retain the ability to load DnaB at D loops since srgA mutants support plasmids that require loading of DnaB at R loops to initiate replication (Al-Deib et al., 1996). The fact that RecG operates in a pathway that also requires PriA implies that RecG catalysis must ultimately facilitate PriA-mediated loading of DnaB at stalled forks. How could RecG cooperate with PriA? We have shown that RecG has a high affinity for forks lacking a leading strand and unwinds these structures efficiently. RecG also catalyses branch migration of both Holliday junctions and complete fork structures and allows their interconversion (Lloyd and Sharples, 1993; McGlynn and Lloyd, 2000, 2001; McGlynn et al., 2001). The requirement for RecG raises the possibility that interconversion of fork and junction structures at stalled replication intermediates may facilitate PriA helicase-mediated loading of DnaB. Figure 1B illustrates how the formation of a Holliday junction may be used to rectify a fork with a dislocated leading strand, either by extending the leading strand or by trimming the lagging strand. Subsequent processing of the junction establishes a complete fork. The only essential difference between the two examples presented is the timing of excision repair. This would need to occur before restoring a fork if restoration were to be achieved by exonuclease digestion of the DNA spooled out during fork regression (Figure 1B, iib–ivb). In contrast, extension of the leading strand via template switching would allow the lesion to be bypassed and then repaired some time after replication has restarted (Figure 1B, iia–iva). Thus, the essential function of RecG may be to allow the 3⬘ terminus of the leading strand to be brought into play at a stalled fork. The absence of RecG may therefore allow inappropriate loading of DnaB via PriA helicase activity at such forks. These activities may explain the recG mutant phenotype and its exacerbation by a plasmid encoding wild-type PriA, but not a helicase-deficient derivative (Al-Deib et al., 1996). We have suggested that stalled forks sometimes have a structure in which the lagging strand but not the leading strand is present at the branch point (Figure 1B). Such dislocation of the 3⬘ end of the leading strand has been observed, though not yet in E. coli (Cordeiro-Stone et al., 1999), and may result from uncoupled synthesis by the lagging strand polymerase when synthesis by the leading strand polymerase is blocked at a lesion in the template strand. None of the polymerase complexes associated with genome replication has a counterpart to the primase acting on the lagging strand template that could similarly prime leading strand synthesis. Thus, continued synthesis of the leading strand requires priming by the 3⬘ end of the leading strand and cannot be initiated downstream of a lesion. Leading strand blocks are therefore a serious threat to replication. Polymerase uncoupling has not been reported during replication in E. coli. However, the ability of both RecG and PriA helicases to target forks lacking a leading strand and the interplay we have described between RecG and PriA in UV-irradiated cells strongly suggest that such dislocated fork structures do arise in vivo.
Rescue of Stalled Replication Forks 249
We have suggested two possible mechanisms by which such forks might be rescued, both based on formation of a Holliday junction, but there are probably others. Indeed, the ability of SrgA proteins to alleviate the recG phenotype implies that other proteins can generate a Holliday junction from a stalled fork or may otherwise manipulate the structure of a dislocated fork to facilitate DnaB loading. For instance, regression of a stalled fork could be driven by RecA-mediated strand exchange (Robu et al., 2001) or by RecQ helicase, and any extension of the lagging strand could be removed by RecJ exonuclease. Functional overlaps may be necessary to cope with the potential threat to the genome of a lesion blocking synthesis of the leading strand. Translesion synthesis may also be employed, but at the risk of mutation. As a last resort, and particularly if advance of the replicative helicase is blocked, the fork may be collapsed, allowing its indirect rescue by recombination. However, direct rescue of a fork has one very important attribute. Replication resumes on the chromosome on which it stalls, and the rescued fork cannot be misplaced. There is no such guarantee with recombination (Barre et al., 2000; Michel et al., 2000). Thus, helicases such as RecG that act at damaged replication forks may be vital for maintaining the integrity of genomes and may be especially important in organisms whose genomes contain large numbers of repeats that could foster illegitimate exchanges (Chakraverty and Hickson, 1999; Constantinou et al., 2000; Karow et al., 2000; Myung et al., 2001). Indeed, recent studies suggest that members of the RecQ family of DNA helicases may act to remove Holliday junctions formed from blocked replication forks and thereby reduce the incidence of fork breakage and recombination (Boddy et al., 2001; Chen et al., 2001; Doe et al., 2000; Kaliraman et al., 2001). Experimental Procedures Strains and Plasmids The bacterial strains used are all derivatives of Escherichia coli K-12 strains AB1157 or MG1665 (Bachmann, 1996) or B strain BL21. The constructs described in Table 1 are all derivatives of AB1157. Additional constructs used in the experiments described in Figures 2–4 are as follows: N3789 (ruvA60 ⌬recG263), N4703 (recB268 ⌬recG263), N4752 (rpoB*35 ruvA60), N4778 (rpoB*35 ruvA60 recB270), N5381 (ruvA60 recB270), and RJ1086 (priA4 ⌬recG265) are derivatives of AB1157, whereas N4279 (recA269), N5070 (recA269 ⌬recG263 ⌬ruvABC), N5093 (⌬recG263 ⌬ruvABC), and N5397 (priA2) are derivatives of MG1665. Apart from the markers listed, all have the same genotype as the parent strains. Full derivations of all the constructs are available on request. The ruvA60::Tn10, ⌬ruvABC::cat, recA269::Tn10, recB268::Tn10, ⌬recG263::kan, and priA2::kan mutations have been described (Mahdi et al., 1996; Mandal et al., 1993; Sandler et al., 1996; Seigneur et al., 1998). recG264::cat is a null allele in which the 0.12 kb EspI restriction fragment has been deleted from the central coding region of recG and replaced with the 0.9 kb chloramphenicol transacetylase cassette from pKRP10. recG265::cat is a similar construct but has a 1.25 kb EspI-BglII deletion of the recG coding region (G.J. Sharples and R.G.L., unpublished data). The priA4 and recB270 alleles are mini-Tn10kan insertions in the coding sequence of priA and recB located after nucleotides 2585 and 3166, respectively. They were identified following infection of E. coli strain N4538 (McGlynn and Lloyd, 2000) with NK1327 (Kleckner et al., 1991) as Kmr derivatives that were sensitive to mitomycin C (R.J., unpublished data). The priA4 insertion generates a fusion protein in which the final six amino
acids at the C terminus of PriA (VDPIEG) are replaced with the sequence ADESPNDFGKNH (R.P.J., unpublished data). dnaC212 is a T634G transversion encoding a L212V mutant DnaC and was identified in a derivative of the priA2 strain AG109 (Table 1) selected for its resistance to UV light (A.V.G., unpublished data). rpoB*35 modifies the  subunit of RNA polymerase such that it alleviates the UV sensitivity of a ruv mutant (McGlynn and Lloyd, 2000) and was introduced by P1 transduction via linkage to argE. The priA⫹ plasmid construct pAD203 was made by PCR amplification of the priA coding sequence from strain AB1157 using primers incorporating flanking restriction sites and cloning the digested PCR product into the vector pET14b and the insert plus the upstream region encoding the hexa-histidine tag redirected into pET-3c cut with XbaI and BamHI (A.A. Mahdi, A. Al-Deib, and R.G.L., unpublished data). Plasmids encoding SrgA derivatives of PriA were made similarly by PCR amplification of the srgA1 and srgA2 alleles of priA from strains N3695 and N3696 (Al-Deib et al., 1996), respectively, and cloning the products in pET-14b via pET-3C (srgA1) or directly (srgA2), generating pAG108 and pAG118, respectively. DNA sequencing was used to verify the insert in each of the three constructs, and only the desired srgA1 and srgA2 mutations in pAG108 and pAG118, respectively, differed from the published sequence of priA. The three plasmids made encode PriA or SrgA proteins with a hexa-histidine tag at the N terminus. Full derivations are available on request. pET3 plasmids encoding wild-type PriA or the helicasedeficient K230R derivative were from Ken Marians (Zavitz and Marians, 1992). Media and General Methods LB broth and agar and 56/2 minimal salts media were used for bacterial culture. Strains carrying mutant alleles tagged with antibiotic resistance genes were made by P1vir transduction and selecting for the appropriate resistance. Strains carrying priA alleles together with recG or ruv null mutations were made by introducing priA into the appropriate recG or ruv strain. Media recipes and procedures for strain construction, determining sensitivity to mitomycin C, and measuring UV radiation survival have been cited (Al-Deib et al., 1996). UV survival values are means of at least two and usually three to eight independent experiments. Cultures of strains carrying priA2 and, to a lesser extent, priA4, alone or in conjunction with recG or ruv null alleles, can accumulate significant fractions of fast-growing derivatives carrying suppressor mutations in dnaC, which can be readily detected by their ability to form larger colonies. The presence of a dnaC suppressor in such derivatives was deduced from its tight linkage to the zji-202::Tn10 insertion located very close to dnaC and confirmed in several cases by sequencing the dnaC gene. Their accumulation in broth cultures used to measure UV sensitivity was limited by using large inocula from several small colonies to reduce the number of cell divisions needed to reach the required cell density. A priA2 derivative of strain AB1157 was also constructed and tested for UV sensitivity using minimal media throughout. The data obtained were essentially identical to those shown for strain AG109 in Figure 2A. Strains carrying plasmids were grown with appropriate antibiotic selection. Proteins Wild-type PriA was a gift from Ken Marians. The PriA, SrgA1, and SrgA2 proteins carrying a hexa-histidine tag at the N terminus were produced by IPTG induction from pAD203, pAG108, and pAG118, respectively, in strain BL21(DE3) carrying pLysS or, in the case of SrgA2, pLysE. Induced cells were lysed by sonication, and the PriA and SrgA were proteins recovered by chromatography on a Ni2⫹ affinity matrix and gel filtration using an S200HR Sephacryl column (Amersham Pharmacia) to remove trace contaminants before storing at ⫺80⬚C. Protein concentrations were estimated by a modified Bradford assay using a Bio-Rad assay kit and bovine serum albumin as a standard. Amounts are expressed as moles of the monomeric protein. Fork DNA Substrates Fork substrates were constructed using oligonucleotides, one of which in each structure was labeled with [␥-32P]ATP at the 5⬘ end, and each substrate was purified by gel electrophoresis (Parsons et
Molecular Cell 250
al., 1990). Sequences of the oligonucleotides, written 5⬘ to 3⬘, are: (a) GTCGGATCCTCTAGACAGCTCCATGATCACTGGCACTGGTAG AATTCGGC, (b) CAACGTCATAGACGATTACATTGCTACATGGAGC TGTCTAGAGGATCCGA, (c) TGCCGAATTCTACCAGTGCCAGTGAT, (d) TAGCAATG TAATCGTCTATGACGTT, (e) TGCCGAATTCTACCA GTGCCAG, (f) AATGTAATCGTCTATGACGTT, (g) GTCGGATCCTCT AGACAGCTCCATGTAGCCTGGCACTGGTAGAATTCGGC. Each fork was made by annealing combinations of these oligonucleotides, as follows: Forks 1 and 2 (a ⫹ b ⫹ c ⫹ d), Fork 3 (a ⫹ b ⫹ c), Fork 4 (a ⫹ b ⫹ d), Forks 5 and 6 (a ⫹ b ⫹ c ⫹ f), Fork 7 (a ⫹ b ⫹ e ⫹ d), and Fork 8 (g ⫹ b ⫹ e ⫹ f). Fork Dissociation and Binding Assays Dissociation of fork structures was performed as described (McGlynn and Lloyd, 1999), except that the buffer system was 20 mM Tris.HCl (pH 7.5), 20 mM potassium acetate, 2 mM dithiothreitol, and 0.1 mg/ml bovine serum albumin. ATP and magnesium acetate were both used at a concentration of 5 mM. Reactions in which increasing concentrations of PriA or SrgA were titrated against junction DNA were incubated for 30 min at 37⬚C prior to deproteinization. Band shift assays were used to measure DNA binding. Fork DNA (0.2 nM) and various concentrations of PriA or SrgA protein were mixed on ice in 50 mM Tris.HCl (pH 8.0), 1 mM DTT, 6% glycerol, and 100 mg/ml BSA in a total volume of 20 l and left for 15 min. Samples were then loaded onto a prechilled 4% native polyacrylamide gel in a low ionic strength buffer (6.7 mM Tris.HCl [pH 8.0], 3.3 mM sodium acetate) containing either 2 mM EDTA or 2 mM Mg2⫹ and electrophoresed at 160 V for 90 min at 4⬚C. For both DNA dissociation and binding assays, gels were dried onto filter paper and exposed to X-ray film and a storage phosphor screen (Molecular Dynamics). All quantitative data are the means of at least two independent experiments. Acknowledgments We are indebted to Carol Buckman and Lynda Harris for technical support, Ken Marians for wild-type PriA and priA plasmids, Akeel Mahdi for pAD203, and Gary Sharples for strains carrying recG264 and recG265 and for comments on the manuscript. This work was supported by a program grant from the MRC awarded to R.G.L and G.J. Sharples, the BBSRC (A.V.G.), and the Iranian Government (R.P.J.). P.M. is a Lister Institute-Jenner Research Fellow. Received September 17, 2001; revised December 12, 2001.
C.H. (2001). Human Mus81-associated endonuclease cleaves Holliday junctions. Mol. Cell 8, 1117–1127. Constantinou, A., Tarsounas, M., Karow, J.K., Brosh, R.M., Bohr, V.A., Hickson, I.D., and West, S.C. (2000). Werner’s syndrome protein (WRN) migrates Holliday junctions and co-localizes with RPA upon replication arrest. EMBO Rep. 1, 80–84. Cordeiro-Stone, M., Makhov, A.M., Zaritskaya, L.S., and Griffith, J.D. (1999). Analysis of DNA replication forks encountering a pyrimidine dimer in the template to the leading strand. J. Mol. Biol. 289, 1207– 1218. Cox, M.M. (2001). Historical overview: searching for replication help in all of the rec places. Proc. Natl. Acad. Sci. USA 85, 8173–8180. Cox, M.M., Goodman, M.F., Kreuzer, K.N., Sherratt, D.J., Sandler, S.J., and Marians, K.J. (2000). The importance of repairing stalled replication forks. Nature 404, 37–41. Datta, A., Schmeits, J.L., Amin, N.S., Lau, P.J., Myung, K., and Kolodner, R.D. (2000). Checkpoint-dependent activation of mutagenic repair in Saccharomyces cerevisiae pol3-01 mutants. Mol. Cell 6, 593–603. Doe, C.L., Dixon, J., Osman, F., and Whitby, M.C. (2000). Partial suppression of the fission yeast rqh1(⫺) phenotype by expression of a bacterial Holliday junction resolvase. EMBO J. 19, 2751–2762. Flores-Rozas, H., and Kolodner, R.D. (2000). Links between replication, recombination and genome instability in eukaryotes. Trends Biochem. Sci. 25, 196–200. Jones, J.M., and Nakai, H. (1999). Duplex opening by primosome protein PriA for replisome assembly on a recombination intermediate. J. Mol. Biol. 289, 503–516. Kaliraman, V., Mullen, J.R., Fricke, W.M., Bastin-Shanower, S.A., and Brill, S.J. (2001). Functional overlap between Sgs1-Top3 and the Mms4-Mus81 endonuclease. Genes Dev. 15, 2730–2740. Karow, J.K., Constantinou, A., Li, J.L., West, S.C., and Hickson, I.D. (2000). The Bloom’s syndrome gene product promotes branch migration of Holliday junctions. Proc. Natl. Acad. Sci. USA 97, 6504– 6508. Kleckner, N., Bender, J., and Gottesman, S. (1991). Uses of transposons with emphasis on Tn10. Methods Enzymol. 204, 139–180. Kogoma, T., Cadwell, G.W., Barnard, K.G., and Asai, T. (1996). The DNA replication priming protein, PriA, is required for homologous recombination and double-strand break repair. J. Bacteriol. 178, 1258–1264.
References
Kowalczykowski, S.C. (2000). Initiation of genetic recombination and recombination-dependent replication. Trends Biochem. Sci. 25, 156–165.
Al-Deib, A.A., Mahdi, A.A., and Lloyd, R.G. (1996). Modulation of recombination and DNA repair by the RecG and PriA helicases of Escherichia coli K-12. J. Bacteriol. 178, 6782–6789.
Lindahl, T. (1996). The Croonian Lecture, 1996: endogenous damage to DNA. Philos. Trans. R. Soc. Lond. B Biol. Sci. 351, 1529–1538.
Bachmann, B.J. (1996). Derivations and genotypes of some mutant derivatives of Escherichia coli K-12. In Escherichia coli and Salmonella Cellular and Molecular Biology, Second Edition, F.C. Neidhardt, R. Curtiss III, J.L. Ingraham, E.C.C. Lin, K.B. Low, B. Magasanik, W.S. Reznikoff, M. Riley, M. Schaechter, and H.E. Umbarger, eds. (Washington, DC: ASM Press), pp. 2460–2488.
Liu, J., Xu, L., Sandler, S.J., and Marians, K.J. (1999). Replication fork assembly at recombination intermediates is required for bacterial growth. Proc. Natl. Acad. Sci. USA 96, 3552–3555.
Barre, F.X., Aroyo, M., Colloms, S.D., Helfrich, A., Cornet, F., and Sherratt, D.J. (2000). FtsK functions in the processing of a Holliday junction intermediate during bacterial chromosome segregation. Genes Dev. 14, 2976–2988. Bell, D.W., Varley, J.M., Szydlo, T.E., Kang, D.H., Wahrer, D.C., Shannon, K.E., Lubratovich, M., Verselis, S.J., Isselbacher, K.J., Fraumeni, J.F., et al. (1999). Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286, 2528–2531.
Liu, J., and Marians, K.J. (1999). PriA-directed assembly of a primosome on D loop DNA. J. Biol. Chem. 274, 25033–25041.
Lloyd, R.G. (1991). Conjugational recombination in resolvase-deficient ruvC mutants of Escherichia coli K-12 depends on recG. J. Bacteriol. 173, 5414–5418. Lloyd, R.G., and Sharples, G.J. (1993). Dissociation of synthetic Holliday junctions by E. coli RecG protein. EMBO J. 12, 17–22. Mahdi, A.A., Sharples, G.J., Mandal, T.N., and Lloyd, R.G. (1996). Holliday junction resolvases encoded by homologous rusA genes in Escherichia coli K-12 and phage 82. J. Mol. Biol. 257, 561–573.
Boddy, M.N., Gaillard, P.L., McDonald, W.H., Shanahan, P., Yates, J.R., III, and Russell, P. (2001). Mus81-Eme1 are essential components of a Holliday junction resolvase. Cell 107, 537–548.
Mandal, T.N., Mahdi, A.A., Sharples, G.J., and Lloyd, R.G. (1993). Resolution of Holliday intermediates in recombination and DNA repair: indirect suppression of ruvA, ruvB and ruvC mutations. J. Bacteriol. 175, 4325–4334.
Chakraverty, R.K., and Hickson, I.D. (1999). Defending genome integrity during DNA replication: a proposed role for RecQ family helicases. Bioessays 21, 286–294.
McCool, J.D., and Sandler, S.J. (2001). Effects of mutations involving cell division, recombination, and chromosome dimer resolution on a priA2::kan mutant. Proc. Natl. Acad. Sci. USA 98, 8203–8210.
Chen, X., Melchionna, R., Denis, C., Gaillard, P.L., Blasina, A., Van de Weyer, I., Boddy, M.N., Russell, P., Vialard, J., and McGowan,
McGlynn, P., and Lloyd, R.G. (1999). RecG helicase activity at threeand four-strand DNA structures. Nucleic Acids Res. 27, 3049–3056.
Rescue of Stalled Replication Forks 251
McGlynn, P., and Lloyd, R.G. (2000). Modulation of RNA polymerase by (p)ppGpp reveals a RecG-dependent mechanism for replication fork progression. Cell 101, 35–45. McGlynn, P., and Lloyd, R.G. (2001). Rescue of stalled replication forks by RecG: simultaneous translocation on the leading and lagging strand templates supports an active DNA unwinding model of fork reversal and Holliday junction formation. Proc. Natl. Acad. Sci. USA 98, 8227–8234. McGlynn, P., Al-Deib, A.A., Liu, J., Marians, K.J., and Lloyd, R.G. (1997). The DNA replication protein PriA and the recombination protein RecG bind D-loops. J. Mol. Biol. 270, 212–221. McGlynn, P., Lloyd, R.G., and Marians, K.J. (2001). Formation of Holliday junctions by regression of stalled replication forks: RecG stimulates fork regression even when the DNA is negatively supercoiled. Proc. Natl. Acad. Sci. USA 98, 8235–8240. Michel, B., Recchia, G.D., Penel-Colin, M., Ehrlich, S.D., and Sherratt, D.J. (2000). Resolution of holliday junctions by RuvABC prevents dimer formation in rep mutants and UV-irradiated cells. Mol. Microbiol. 37, 180–191. Myung, K., Datta, A., Chen, C., and Kolodner, R.D. (2001). SGS1, the Saccharomyces cerevisiae homologue of BLM and WRN, suppresses genome instability and homeologous recombination. Nat. Genet. 27, 113–116. Nurse, P., Liu, J., and Marians, K.J. (1999). Two modes of PriA binding to DNA. J. Biol. Chem. 274, 25026–25032. Parsons, C.A., Kemper, B., and West, S.C. (1990). Interaction of a four-way junction in DNA with T4 endonuclease VII. J. Biol. Chem. 265, 9285–9289. Paulovich, A.G., Toczyski, D.P., and Hartwell, L.H. (1997). When checkpoints fail. Cell 88, 315–321. Robu, M.E., Inman, R.B., and Cox, M.M. (2001). RecA protein promotes the regression of stalled replication forks in vitro. Proc. Natl. Acad. Sci. USA 98, 8211–8218. Sandler, S.J. (1996). Overlapping functions for recF and priA in cell viability and UV-inducible SOS expression are distinguished by dnaC809 in Escherichia coli K-12. Mol. Microbiol. 19, 871–880. Sandler, S.J., Samra, H.S., and Clark, A.J. (1996). Differential suppression of priA2::kan phenotypes in Escherichia coli K-12 by mutations in priA, lexA, and dnaC. Genetics 143, 5–13. Sandler, S.J., Marians, K.J., Zavitz, K.H., Coutu, J., Parent, M.A., and Clark, A.J. (1999). dnaC mutations suppress defects in DNA replication- and recombination-associated functions in priB and priC double mutants in Escherichia coli K-12. Mol. Microbiol. 34, 91–101. Seigneur, M., Bidnenko, V., Ehrlich, S.D., and Michel, B. (1998). RuvAB acts at arrested replication forks. Cell 95, 419–430. Seigneur, M., Ehrlich, S.D., and Michel, B. (2000). RuvABC-dependent double-strand breaks in dnaBts mutants require RecA. Mol. Microbiol. 38, 565–574. Xu, L., and Marians, K.J. (2000). Purification and characterization of DnaC810, a primosomal protein capable of bypassing PriA function. J. Biol. Chem. 275, 8196–8205. Zavitz, K.H., and Marians, K.J. (1992). ATPase-deficient mutants of the Escherichia coli DNA replication protein PriA are capable of catalyzing the assembly of active primosomes. J. Biol. Chem. 267, 6933–6940.
