Stabilization of a Stalled Replication Fork by Concerted Actions of Two Helicases*

PriA helicase plays crucial roles in restoration of arrested replication forks. It carries a “3′ terminus binding pocket” in its N-terminal DNA binding domain, which is required for high affinity binding of PriA to a fork carrying a 3′-end of a nascent leading strand at the branch. We show that the abrogation of the 3′ terminus recognition either by a mutation in the 3′ terminus binding pocket or by the bulky modification of the 3′-end leads to unwinding of the unreplicated duplex arm on this fork, causing potential fork destabilization. This indicates a critical role of the 3′ terminus binding pocket of PriA in its “stable” binding at the fork for primosome assembly. In contrast, PriA unwinds the unreplicated duplex region on a fork without a 3′-end, potentially destabilizing the fork. However, this process is inhibited by RecG helicase, capable of regressing the fork until the 3′-end of the nascent leading strand reaches the branch. PriA now stably binds to this regressed fork, stabilizing it. Using a model arrest-fork-substrate, we reconstitute the above process in vitro with RecG and PriA proteins. Our results present a novel mechanism by which two helicases function in a highly coordinated manner to generate a structure in which an arrested fork is stabilized for further repair and/or replication restart.

It has become widely recognized that progression of replication forks is stalled or attenuated not only by various lesions on the template DNA but also during the normal course of DNA replication (1) and that cellular responses to the arrested replication forks are crucial for maintenance of genetic integrity and sometimes for the cellular viability (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13). The first step for responses to stalled replication forks is the stabilization of arrested forks (14). Several checkpoint proteins have been implicated in this process in eukaryotes (15,16). In the absence of these proteins, the replication fork may not be stably arrested, and the fork may progress in the absence of DNA synthesis, generating the single-stranded DNA regions (17). This results in "unstable" replication forks, which may be degraded or fail to activate subsequent cellular responses required for restoration of replication forks. Helicases play central roles at a replication fork not only in its advancement but also in its proper processing in the event of fork arrest or pausing. In addition to the MCM helicase, a likely replicative helicase at eukaryotic replication forks, several helicases have been implicated in processing of replication forks, including RecQ-type helicases. However, precise roles of these helicases in recognition and stabilization of arrested replication forks are still elusive.
In bacteria, DEXH/DEAD-type helicases have been implicated in cellular responses to replication fork arrest (18 -20). Among them, RecG is known to generate a chicken foot structure by regressing the replication forks (18,21). PriA, originally identified as a component essential for a "primosome" responsible for DNA unwinding and primer RNA synthesis (22)(23)(24), can also recognize arrested fork or recombination intermediate structures derived from double-stranded DNA breaks and is believed to facilitate the reassembly of the replication fork (24 -32). PriA, widely conserved in eubacteria, is composed of two domains, namely the N-terminal DNA recognition domain and the C-terminal DEXH-type helicase domain (33,34). The former contains a 3Ј terminus binding pocket, which specifically recognizes and binds to a 3Ј-end of DNA. The presence of this 3Ј-end binding structure facilitates recognition of D-loop or arrested fork structures carrying a 3Ј-end at the branch point (35).
We discovered that the presence of a 3Ј-end at the branch point of an arrested fork can stabilize the bound PriA and prevents the unwinding of duplex DNA by PriA, thus permitting the assembly of a primosome for fork restoration. On the other hand, it unwinds the unreplicated duplex arm on a fork without a 3Ј terminus at the branch. This would potentially cause the fork destabilization because of abortive unwinding. We then discovered that this potentially fork-destabilizing effect of PriA is inhibited by RecG, which reverses the fork and generates a structure that would now be bound and protected by PriA protein. On the basis of the in vitro studies using a novel arrest-fork-substrate, we propose a novel mechanism of stabilization and restart of arrested replication forks by the concerted actions of two helicases.

EXPERIMENTAL PROCEDURES
Proteins-Wild-type and mutant PriA proteins and RecG protein were expressed and purified as previously described (34 -37). Protein concentrations were measured by Bradford assay using bovine serum albumin as standard and the purity was confirmed by running the preparations on SDS-PAGE followed by staining with Coomassie Brilliant Blue.
DNA Substrates-Forked, tailed, or other substrates were constructed by annealing the oligonucleotides as indicated in Table 1. Substrates termed [3Ј-P] carry the leading strand whose 3Ј-end is phosphorylated. The 5Ј-end of the oligonucleotide indicated was phosphorylated by T4 polynucleotide kinase (New England Biolabs) and [␥-32 P]ATP prior to annealing. Each substrate was purified from polyacrylamide gel as described previously (34). To prepare the partially complementary arrested fork (pcA-fork 2 [3Ј, 5Ј] 49 L10), leading and lagging arms were independently annealed as described above. After isolation from polyacrylamide gel, both partial duplex constructs were mixed at room temperature for 4 h, and annealed arms were reisolated from polyacrylamide gel.
Gel Shift Assay-Reactions were conducted as described previously (34). In the assays containing both proteins, two proteins were mixed prior to addition to the reaction mixtures.
Nuclease Protection Assay-Reaction mixtures (25 l) contained 40 mM Hepes/KOH (pH 7.6), 8 mM magnesium acetate, 0.5 mM CaCl 2 , 40 mM potassium glutamate, 20 g/ml bovine serum albumin, 3 nM substrate, and 16 and 32 nM PriA or RecG protein as indicated in the figures. After incubation for 20 min at 30°C, 0.002 units of DNase I (TAKARA) was added to the reaction, and incubation was continued for 1 min, followed by the addition of an equal volume of 2 ϫ stop solution containing 20 mM EDTA, 1% SDS, 200 mM NaCl, and 125 g/ml yeast tRNA. In the assays containing both proteins, two proteins were mixed prior to addition to the reaction mixtures in the presence or absence of 0.05 mM ATP. Digested fragments were extracted by phenol, recovered by ethanol precipitation, and were resolved in 80% formamide. After heating at 96°C for 5 min, each sample was applied onto 12% denaturing polyacrylamide sequencing gel (40 cm) containing 8 M urea in 0.5 ϫ Tris-borate EDTA. High resolution electrophoresis was conducted at 40 W of constant electric power for 1 h. Digestion patterns were visualized by autoradiography. Each lane was quantified by Multi Gauge version 3.0 (Fuji Photo Film) software.
Helicase Assay-Reactions were conducted in 20-l mixtures containing 20 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 5 mM ATP, 2 mM dithiothreitol, 100 g/ml bovine serum albumin, and 0.4 nM substrate. After 37°C incubation for 30 min, 5 l of stop solution containing 100 mM Tris-HCl (pH 7.5), 200 mM EDTA, 2.5% SDS, and 10 mg/ml proteinase K was added, and incubation was continued for 10 min. Reactions were directly applied onto 10% native polyacrylamide gel. Electrophoresis was conducted at 200 V of constant voltage for 1 h at 25°C. Gels were dried and autoradiographed. Each lane was quantified by Image Gauge version 3.2 (Fuji Photo Film) software, and the fraction of the displaced oligonucleotide relative to the total substrate was calculated as the helicase activity. The values were plotted against protein concentration.

RESULTS
Helicase Action of PriA Protein on Various Substrates-PriA is a DNA helicase with a 3Ј-5Ј polarity (38,39). We have examined the helicase actions of PriA on various substrates. On A-fork [3Ј], very little helicase activity was observed, as previously reported (Fig. 1A) (25,31). Only a very low level of unwinding of the unreplicated duplex segment was observed. In contrast, efficient displacement of the unreplicated duplex and the annealing lagging strand was detected on A-fork [5Ј]. Displacement of the annealing lagging strand was previously reported on similar fork structures (19,25,26). On A-fork [3Ј, 5Ј], on the other hand, displacement of only the annealing lagging strand was observed (Fig. 1B) (19). The displacement of the lagging strand is likely to be mediated by translocation of PriA on the lagging strand template DNA (lower strand) (29,31), because PriA is not capable of unwinding DNA from a duplex end (25). This implies that PriA is loaded onto these substrates near the fork, as was predicted from the previous binding studies (27,30,31).
We also examined the helicase actions on the 3Ј-extension and 5Ј-extension substrates and again confirmed previous reports. Although both substrates are bound with PriA with similar affinity under our experimental conditions (data not shown), only 3Ј-extension was displaced, as was reported (Fig. 1C) (25,31). This result may be related to the ability of PriA to displace the lagging strand on A-fork [5Ј] but not the leading strand DNA on A-fork [3Ј] (31). PriA, loaded at the single-to-double-  strand DNA junction on the 3Ј-extension (data not shown), can translocate on the lower strand to displace the annealing DNA (25, 31, 32) but cannot do so on the 5Ј-extension (25). In the latter, PriA is probably bound to the junction through the 3Ј terminus recognition, and this binding mode may not permit the translocation of PriA on the lower strand.

Inhibition of PriA Helicase Action on A-fork [3Ј] by Its 3Ј Terminus Binding Ability-The unwinding of the unreplicated duplex in A-fork
[5Ј] indicates that PriA translocates on the leading strand template (upper strand), as was shown before (31). The absence of this activity in A-fork [3Ј, 5Ј] suggests that the presence of the nascent leading strand (providing a 3Ј-end at the branch) prevents loading and/or helicase action of PriA on this substrate. Because gel shift assays and footprinting analyses indicate that PriA binds to A-fork [3Ј, 5Ј] with affinity higher than that to A-fork [5Ј] (35) and that it interacts with leading strand template on A-fork [3Ј, 5Ј] (data not shown), it is not the defect in the loading of PriA that causes the inhibition of helicase action. Therefore, we speculated that the presence of a 3Ј-end at the branch point may prevent the displacement of the unreplicated duplex.
We tested this idea by comparing a simple Y-fork and A-fork [3Ј]. The duplex segment of the simple Y-fork was efficiently displaced by PriA, whereas that of A-fork [3Ј] was resistant to helicase action ( Fig. 2A, lanes  2-6 and 15-19). Similar results were previously reported (31), and it is predicted that PriA translocates on the leading strand template of Y-fork to displace it. The inability of PriA to unwind the unreplicated duplex of the A-fork [3Ј] in the presence of the hybridizing leading strand is consistent with the idea that the presence of a 3Ј terminus at the fork is inhibitory for the helicase action of PriA.
To examine whether the interaction between the 3Ј terminus binding pocket and the 3Ј-end exerts any negative effect on the displacement of the duplex portion of this fork, we next constructed A-fork [3Ј-P] in which the 3Ј terminus of the leading strand is blocked by phosphorylation. This prevents the above interaction and lowers the binding affinity of PriA to A-fork [3Ј] by more than one order of magnitude (35). On A-fork [3Ј-P], the unreplicated duplex was displaced by PriA as efficiently as that on a simple Y-fork (Fig. 2, A, lanes 28 -32, and B). This supports the idea that 3Ј terminus binding of PriA is inhibitory for the helicase action on A-fork [3Ј].
To further confirm this notion, we next examined helicase activities of a mutant PriA protein defective in the 3Ј terminus binding. The mutant carries three amino acid substitutions in the 3Ј terminus binding pocket, which result in greatly reduced 3Ј terminus binding and one order of magnitude lower affinity to A-fork [3Ј], and is compromised in some of the biological functions including decreased growth rate (35). The mutant protein is now able to unwind the unreplicated duplex of the A-fork [3Ј] (Fig. 2, A, lanes 21-25, and B). It was more active than the wild-type protein also on A-fork [3Ј-P] (Fig. 2, A, lanes 34 -38, and B). These results further strengthen the conclusion that the interaction of the 3Ј terminus binding pocket with a 3Ј terminus present at the fork inhibits the fork (unreplicated duplex) unwinding activity of PriA helicase. This would lead to stable binding of PriA at the fork with a 3Ј terminus, permitting assembly of a primosome for replication restart.
We previously showed that D-loop substrates can stimulate ATPase activity of PriA, whereas a bubble substrate, without a 3Ј terminus at the fork junction, can only weakly activate the ATPase (35,40). We also showed that the 3Ј phosphorylation in D-loop decreased the ATPase activity and that the 3Ј terminus binding pocket mutant exhibited decreased ATPase activity with a normal D-loop (35). Thus, there is correlation between the 3Ј terminus binding and ATPase activation. The reason why PriA does not show the helicase activity on A-fork [3Ј] is therefore not because of its inability to activate ATPase but because of the configuration of PriA on the substrate in which it is not poised to unwind the unreplicated arm or simply because of the lack of the strand to be displaced (on the lagging strand template).

RecG Protein Inhibits the Fork Unwinding Activity of PriA Protein on A-fork [5Ј]-
The ability of PriA to unwind the unreplicated duplex DNA on A-fork [5Ј] indicates that an arrested fork could be destabilized by PriA. This helicase action of PriA needs to be somehow inhibited to have a stabilized fork and reassembly of a replisome. Therefore, we examined the effect of RecG protein, another conserved helicase implicated in the processing of arrested replication forks in bacteria, on the helicase activity of PriA (12,18,19,41). First, we examined the effect of RecG on PriA binding to A-fork [5Ј]. RecG binds to A-fork [5Ј] with affinity nearly one order of magnitude higher than PriA does (k D ϭ 4 nM for RecG and 22 nM for PriA; Fig. 3A, lanes 1-11). When increasing amount of RecG is added to A-fork [5Ј] together with a fixed amount of PriA, the PriA⅐A-fork [5Ј] complex is lost and taken over by the RecG⅐Afork [5Ј] complex (Fig. 3A, lanes 12-18). We then examined whether the fork unwinding activity of PriA is affected by RecG protein. Although RecG displaces the lagging strand in A-fork [5Ј] with efficiency similar to that of the PriA (25), it unwinds the unreplicated duplex much less efficiently than PriA does (Fig. 3, B, lanes 1-9, and C). The unreplicated duplex unwinding activity of PriA is significantly inhibited by the presence of equimolecular amount of RecG protein (Fig. 3, B, lanes 11 and  12, and D). The restimulation of unwinding at higher concentrations of RecG may be because of displacement of the hybridizing nascent lagging strand by RecG, which leads to generation of a Y-fork structure that is not bound by RecG but is unwound by PriA. In the cells, the numbers of RecG and PriA molecules are ϳ10 (42) and ϳ50 (40, 43) per cell, respectively. RecG may unwind the lagging strand, but this will generate chicken foot structures that may be rapidly and stably bound by PriA without unwinding. Therefore, we may expect more efficient protection of the unreplicated duplex from unwinding in vivo. Alternatively, excess RecG may form an aggregate in vitro that is not able to counteract PriA helicase. The above result indicates that potentially fork-destabilizing effect of PriA on arrested forks can be counteracted by RecG protein.
RecG Protein Regresses an Arrested Fork to Permit Fork Stabilization by PriA Protein-We next examined whether PriA protein can ultimately stabilize arrested replication forks in collaboration with RecG protein. For this purpose, we designed a novel arrested fork substrate (pcA-fork [3Ј, 5Ј] 49 L10) carrying a 5Ј-end of the lagging strand at the branch and a leading strand with a single-stranded gap of 10 nucleotides. The two strands are complementary on the gap so that reannealing of the template strands is permitted during the displacement of the lagging strand by RecG protein. RecG protein exhibited higher affinity to this substrate than PriA did (k D ϭ 4 nM for RecG and 12 nM for PriA). In contrast, in the presence of a fixed amount of RecG protein and ATP, PriA bound to the pcA-fork [3Ј, 5Ј] 49 L10 substrate with increased efficiency (Fig. 4, A, lanes 18 -25, and B in the presence of 2 nM RecG), indicating that ATP hydrolysis mobilizes RecG protein and facilitates the binding of PriA.
DNase I footprinting analyses show that RecG protein mainly binds to the unreplicated duplex arm, whereas PriA bound near the duplexto-single strand junction on the leading strand, as was shown for 5Ј-extension substrate (Fig. 5A and data not shown). The presence of both footprints in the presence of RecG and PriA suggested simultaneous binding of both proteins on the substrate. Indeed, a slow migrating band appeared in the gel shift assays on the pcA-fork [3Ј, 5Ј] 49 L10 substrate in the presence of both proteins (Fig. 5B). Upon addition of ATP, the protection by RecG disappeared, consistent with the expected translocation of RecG protein on the substrate, whereas the protection by PriA largely remained (Fig. 5A, lanes 4 -7 versus 9 -12). Similarly, in the presence of both PriA and RecG, only the PriA protection remained intact, but the RecG protection disappeared in the presence of ATP (Fig. 5A,  lanes 14 and 15 versus 17 and 18). In gel shift assays, the trimolecular complex disappeared in the presence of ATP, and converted to the PriA-complex (Fig. 5B, lanes 6 and 12), consistent with the loss of RecG protein after translocation. PriA protein strongly protected near the 3Ј-end at the branch point, in a manner dependent on the 3Ј terminus binding pocket and in a stable configuration. These observations are consistent with displacement of the lagging strand by RecG protein on this model arrest-fork-substrate, and the stable binding of PriA at the fork junction. PriA protection over the fork junction in the presence of ATP was more prominent in the presence of RecG than in its absence (Fig. 5A, compare lanes 11 and 12 with lanes 17 and 18), suggesting that unwinding by RecG generates a favorable substrate for PriA. These results support our model in which coupled actions of two helicases stabilize the arrested replication forks.

DISCUSSION
PriA is a helicase with unique DNA binding specificity. A number of studies have been conducted on binding and helicase actions of PriA on various DNAs. It binds to structures mimicking D-loop or arrested replication forks in vitro (24, 25, 29 -32). It was concluded that PriA binds to fork structures with a nick at the junction and that it efficiently displaces the lagging strand arm fragment. It was also suggested that PriA recognizes bent DNA at the fork (30). We reported the presence of a structure-specific DNA binding domain within the N-terminal 181 amino acid segment of PriA as well as involvement of the C-terminal helicase domain in specific and stable interaction of PriA with D-loop structures (34). We also reported the presence of a 3Ј terminus binding pocket within the N-terminal DNA binding domain and showed that it plays a critical role in specific binding of PriA to the arrested fork structures carrying a hybridizing nascent leading strand as well as in the  FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6  (35). In this report, we attempted to understand the roles of the 3Ј terminus binding pocket in processing of arrested replication forks.

Stabilization of a Stalled Replication Fork
Critical Role of the 3Ј Terminus Binding Pocket in Stabilization of an Arrested Replication Fork-PriA can displace a simple Y-fork, but the presence of the nascent leading strand with a 3Ј-end inhibits this helicase action. On A-fork [3Ј], the DNA helicase domain presumably is localized on the unreplicated strand, as was indicated from the structure analyses of RecG-fork DNA complex (44), and this mode of binding is inert in displacing a fork, because the helicase domain is placed in a "unwinding-deficient" orientation. The modification of the 3Ј-end of A-fork [3Ј] or mutagenesis of the 3Ј terminus binding pocket of PriA leads to binding to A-fork [3Ј] in another mode, and the fork is unwound. Consistent with this prediction, DNase I footprinting indicated the loss of specific interaction of PriA with A-fork [3Ј-P] around the branch point that was observed with A-fork [3Ј] (data not shown).
Although the PriA protein can bind to various arrested fork structures, whether it can mobilize the helicase activity for fork unwinding depends on its mode of binding. On A-fork [3Ј], the presence of a 3Ј terminus at the fork permits only the helicase orientation, which is not capable of unreplicated duplex unwinding (Fig. 6A, stable binding). In the absence of a 3Ј terminus, however, PriA binds to the substrate in another orientation as well, albeit with reduced affinity, and it now promotes unwinding of the unreplicated duplex, i.e. advances the replication fork, potentially causing replication fork destabilization. This would explain the UV light sensitivity and slow growth phenotypes displayed by the strains carrying 3Ј-pocket mutants of PriA (35). Thus, the presence of a 3Ј-end at the branch point freezes the PriA, through its 3Ј terminus binding pocket, in configurations that do not permit the unwinding of the unreplicated duplex, but are proficient for primosome assembly.
PriA and RecG Collaborate to Stabilize an Arrested Replication Fork-RecG, another conserved DEXH-type helicase (45,46), is known to act on the arrested fork with a nascent lagging strand at the branch point to reverse the fork (Fig. 6B, pathway I) (12,21,47). PriA would act on this structure to unwind the unreplicated duplex segment (Fig. 6B, pathway II), potentially causing fork destabilization. We have shown that RecG competes out the PriA protein on this structure and suppresses the fork unwinding activity of the latter protein.
We have then devised a novel arrest-fork-substrate and analyzed how RecG suppresses the potentially hazardous fork unwinding activity of PriA and facilitates stable binding of PriA. Our results indicated that by closed and open arrowheads on each gel, respectively. In the assays containing both proteins, two proteins were mixed prior to addition to the reaction mixtures. The structures of substrates used are schematically drawn at the top and the left of the each panel. The protections with PriA and RecG are indicated by gray and open ellipses, respectively, along the drawings of the substrates. For quantification of the protection, the intensity of each band was quantified and was normalized to that in the absence of PriA or RecG protein, which was taken as 100%. The nucleotides are numbered from the branch point with ϩ on the replicated arm and Ϫ on the unreplicated arm, as is indicated in the inset. B, gel shift assays were conducted with indicated amounts of PriA (K230D, helicase dead) and/or RecG proteins on an arrest-fork-substrate, pcA-fork [3Ј, 5Ј] 49 L10, in the absence (lanes 1-7) and presence (lanes 8 -14) of ATP (0.05 mM). The arrowhead indicates the ternary complex containing both PriA and RecG, which is observed in the absence of ATP but disappears in the presence of ATP. Similar results are obtained with the wild-type PriA protein (data not shown). In pathway II, PriA advances the replication fork, resulting in its destabilization. However, this is normally prevented by RecG protein, which regresses the fork and generates a chicken foot structure that is now recognized and bound by PriA. Duplex unwinding activity of PriA is suppressed under this condition, and the fork is stabilized (pathway I). PriA is bound near the 3Ј-end of the arrested leading strand prior to RecG-mediated fork reversal. This binding may prevent further unwinding by RecG (which would generate a Holliday Junction structure) and may facilitate the coupled action of RecG and PriA on this arrested fork structure. See text for details. FEBRUARY 10, 2006 • VOLUME 281 • NUMBER 6

Stabilization of a Stalled Replication Fork
RecG preferentially binds to the branch point and promotes the fork reversal pathway, generating a structure (carrying a 3Ј terminus at the branch point), which is now preferentially bound by PriA for fork stabilization. We also discovered that PriA and RecG bind to the model arrest-fork-structure simultaneously, and upon mobilization of RecG helicase by addition of ATP, RecG reverses the fork, leaves the complex, and PriA captures the fork of the chicken foot structure. Footprinting data indicate that the PriA specifically and efficiently binds to the 3Ј-end of the leading strand at the edge of the gap, and this may explain why the presence of a gap between the branch point and the 3Ј-end of the leading strand significantly inhibits the displacement of the lagging strand by PriA (19. 29). RecG also may inhibit PriA-mediated loading of DnaB onto a lagging strand template on this substrate, which would also lead to uncoupled DNA synthesis because of the absence of a 3Ј-end of the leading strand at the junction (19).
Our model can also explain suppression of a recG mutation by helicase-defective mutants of PriA as well as increased damage sensitivity of a recG mutant upon overexpression of helicase-proficient PriA but not that of helicase-dead PriA (41). The abortive fork unwinding by PriA in the absence of sufficient RecG activity is expected to increase fork instability and cause genetic damages (Fig. 6B, pathway II). Once RecG captures the fork and generates a "chicken foot" structure, this could now be stably bound by PriA with high affinity because of the presence of a 3Ј terminus at the branch point, resulting in the stabilization of the fork and reassembly of a replisome (Fig. 6B, pathway I). It was proposed before that RecG promotes Holliday Junction formation, which may facilitate reannealing of a damaged site on the leading strand and its repair, followed by restoration of a full fork and PriA-mediated reassembly of a primosome. The Holliday junction could also be processed by recombination machinery, followed by PriA-mediated primosome assembly on a D-loop structure (12,19,21). Our model proposes another pathway in which replication fork would be stabilized by highly coordinated actions of RecG and PriA without involving the Holliday Junction formation. This coordination is likely to be achieved by simultaneous binding of RecG and PriA on a stalled replication fork, which was demonstrated in our in vitro studies. It should be noted, however, that functional assays based on primosome assembly and DNA synthesis would be needed to further characterize RecG-and PriA-mediated fork stabilization and replication restart on arrested forks.
The mechanism described in this report would provide important insight into how stalled replication forks could be stabilized in eukaryotes. Helicases and other factors have been implicated in processing of eukaryotic replication forks as well (48 -53). It would be interesting to search for the presence of a 3Ј terminus binding pocket in the eukaryotic fork processing factors and examine whether similar fork stabilizing mechanisms operate in eukaryotes as well.
A remaining question is what is the biological significance of the helicase activity of PriA. The strict conservation of the helicase motifs of PriA across the various eubacteria species strongly indicates the conserved and important roles of the ATPase/helicase activity of PriA protein. However, the model in Fig. 6 suggests that helicase activity of PriA contributes to increase of fork instability rather than its stabilization. Indeed, the helicase-defective mutant of PriA is functional in most of the functions including primosome assembly in vitro, cell survival, and UV resistance (27,28,40,54,55). It was proposed that displacement of the lagging strand is required for loading of DnaB helicase at an arrested fork (29). priA deletion mutant is also partially defective in RecA-dependent DNA replication (56). Recombination-dependent replication requires extensive fork unwinding initiated from initial strand invasion (40). This process may need to be assisted by the helicase action of PriA protein.