Purification and characterization of DnaC810, a primosomal protein capable of bypassing PriA function.

Escherichia coli strains lacking PriA are severely compromised in their ability to repair UV-damaged DNA and to perform homologous recombination. These phenotypes arise because of a lack of PriA-directed replication fork assembly at recombination intermediates such as D-loops. Naturally arising suppressor mutations in dnaC restore strains carrying the priA2::kan null allele to wild-type function. We have cloned one such gene, dnaC810, and overexpressed, purified, and characterized the DnaC810 protein. DnaC810 can support a PriA-independent synthesis of phiX174 complementary strand DNA. This can be attributed to its ability, unlike wild-type DnaC, to catalyze a SSB-insensitive general priming reaction with DnaB and DnaG on any SSB-coated single-stranded DNA. Gel mobility shift analysis revealed that DnaC810 could load DnaB directly to SSB-coated single-stranded DNA as well as to D loop DNA. This explains the ability of DnaC810 to bypass the requirement for PriA, PriB, PriC, and DnaT during replication fork assembly at recombination intermediates.

PriA, a 3Ј 3 5Ј DNA helicase (1,2), was discovered because of its requirement, along with PriB, PriC, DnaT, DnaB, DnaC, and DnaG, for assembly of a primosome on X174 (X) 1 viral DNA during complementary strand DNA replication in vitro (3,4). Although the X-type primosome could provide both the DNA unwinding (via the 5Ј 3 3Ј DNA helicase activity of DnaB (5)) and Okazaki fragment priming (via the primase activity of DnaG (6)) functions at a replication fork (7,8), these biochemical analyses never revealed a role for either PriA, PriB, PriC, or DnaT during chromosomal DNA replication. The advent of strains in which priA had been disrupted soon changed this (9,10).
Based on our initial observation that the priA2::kan strain was constitutively induced for the SOS response (9) and that SOS induction could be suppressed completely by providing a mutant priA allele in trans that encoded a protein that could not act as a DNA helicase but did direct primosome assembly (11), we proposed that PriA-directed replication fork assembly was required to rescue replication forks that had formed at oriC and then stalled because of encountering endogenous DNA damage (9,12). The actual substrate for replication fork assembly became clear as detailed genetic analyses were performed with the priA null strains.
These subsequent genetic studies showed that in addition to induction of the SOS response, strains null for PriA activity had very complex phenotypes. They were poorly viable (9,10), filamented extensively (9,10), and were defective in homologous recombination (13,14), the repair of both double-stranded breaks (14) and UV-damaged DNA (10,13,14), and in the manifestation of both inducible and constitutive stable DNA replication (15). A common theme in the processes involved in the latter four phenotypes was recombination-directed DNA replication. This led to the proposal that PriA functioned by directing the assembly of replication forks at recombination intermediates such as D loops (13,16).
This proposal has been supported by recent studies. PriA binds D loops with high affinity, whereas it will not bind the corresponding bubble structure (17,18) and can direct primosome assembly on D loop DNA as well (19). Significantly, PriA can also direct replication fork assembly on double-stranded templates carrying a D loop in a reaction in vitro that also requires PriB, PriC, DnaT, DnaB, DnaC, DnaG, and the DNA polymerase III holoenzyme (Pol III HE) (20).
Strains carrying priA null mutations acquire suppressor mutations rapidly (13,14). Seventeen independent, naturally arising UV R Rec ϩ revertants of priA2::kan strains that carry extragenic suppressors have been isolated (13). All of these extragenic suppressors map to the C-terminal half of dnaC. Two of these extragenic suppressors alleles, dnaC809 and dnaC810, arise as a result of GAA 3 GGA and GAA 3 GGT nucleotide changes, respectively, at codon 176. This generates a E176G amino acid substitution in each case.
In order to understand the mechanism of suppression of ablation of PriA activity, we have cloned the dnaC810 allele and overexpressed, purified, and characterized the DnaC810 protein. This protein is capable of catalyzing a single-stranded DNA-binding protein (SSB)-independent general priming reaction with DnaB and DnaG. Direct examination by gel mobility shift analysis showed that, unlike wild-type DnaC, DnaC810 could transfer DnaB from a DnaB-DnaC810 complex to either SSB-coated single-stranded DNA (ssDNA) or D loop DNA. This gain of function of the mutant enzyme explains its ability to bypass the requirement for PriA, PriB, PriC, and DnaT during replication fork assembly on a template carrying a D loop (20). reconstituted from polymerase III* and ␤ as described by Wu et al. (8).
Construction of Plasmid pET11c-dnaC810 -A dnaC810 open reading frame was made by two-step overlapping polymerase chain reaction (PCR) (23). The N-terminal coding region of dnaC810 was PCR-amplified using plasmid pET11c-dnaC (22) as a template and two flanking primers: (i) the NdeI primer, 5Ј-TAATGCAGGCCATATGAAAAACGT-TGGCGACCTG-3Ј, which carries a NdeI site at the dnaC initiator codon, and (ii) the AgeIЈ primer, 5Ј-TCGTATTTCGAACCGGTCTG-CACG-3Ј, which carries the designed point mutation (E176G, GAA 3 GGT). The C-terminal coding region of dnaC810 was also PCR-amplified using plasmid pET11c-dnaC as a template and two different flanking primers: (i) the AgeI primer (5Ј-CGTGCAGACCGGTTCGAAAT-ACGA-3Ј), which is complementary to the AgeIЈ primer and (ii) the BamHI primer (5Ј-TTAAGCACTGGGATCCTTAATACTCTTTACCTG-TTAC-3Ј), which carries a BamHI site just downstream of the dnaC stop codon. These overlapping N-and C-terminal fragments were gelpurified after PCR and were further PCR-extended and -amplified with the two flanking NdeI and BamHI primers. The gel-purified dnaC810 open reading frame fragment was digested with NdeI and BamHI and ligated with NdeI-and BamHI-digested pET11c plasmid DNA to give pET11c-dnaC810.
Purification of DnaC810 -Because of the extreme overproduction, DnaC810 was followed during purification by SDS-polyacrylamide gel electrophoresis. BL21(DE3)pLysS carrying pET11c-dnaC810 was grown in 12 liters of L Broth (24) containing 0.4% glucose and 300 g/ml ampicillin to A 600 ϭ 0.4 and then induced in the presence of 1 mM isopropyl-1-thio-␤-D-galactopyranoside for 3 h. Cells were chilled, pelleted by centrifugation, and resuspended in 50 mM Tris-HCl (pH 8.4 at 4°C) and 10% sucrose. The cell suspension (50 ml) was adjusted to 150 mM KCl, 20 mM EDTA, 5 mM dithiothreitol, 0.02% lysozyme, and 0.1% Brij 58 and incubated at 0°C for 10 min. This suspension was centrifuged at 100,000 ϫ g for 1 h (Sorvall T865 rotor). The supernatant (fraction 1, 65 ml, 3510 mg of protein) was adjusted to 0.04% polymin P by dropwise addition of a 1% solution. The precipitate was removed by centrifugation at 47,000 ϫ g in a Sorvall SS-34 rotor for 30 min. The supernatant was further subjected to (NH 4 ) 2 SO 4 fractionation (50% saturation) by the addition of solid. The resulting protein pellet was collected by centrifugation at 47,000 ϫ g in a Sorvall SS-34 rotor for 30 min. The protein pellet was resuspended in 8 ml of buffer A (50 mM Tris-HCl (pH 7.5 at 4°C), 1 mM EDTA, 5 mM dithiothreitol, 20% glycerol, 0.01% Brij 58) plus 50 mM NaCl to give fraction 2 (13 ml, 1108 mg of protein). Fraction 2 was dialyzed against 2 liters of buffer A plus 50 mM NaCl for 12 h and then loaded onto a 100-ml DEAE-cellulose column (4 ϫ 20 cm) that had been equilibrated previously with buffer A plus 50 mM NaCl. The column was washed with 200 ml of buffer A plus 50 mM NaCl. Fractions (15 ml) of the flow-through and wash that contained protein were pooled to give fraction 3 (81 ml, 363 mg of protein). Fraction 3 was loaded directly onto a 35-ml SP-Sepharose FF column (formed in a 60-ml disposable syringe) that had been equilibrated previously with buffer A plus 50 mM NaCl. The column was washed with 200 ml of buffer A plus 50 mM NaCl, and protein was then eluted with a 350-ml linear gradient of 50 -300 mM NaCl in buffer A. DnaC810 eluted at 175 mM NaCl (fraction 4; 24 ml, 25 mg of protein). Fraction 4 was then loaded directly onto a 6-ml hydroxylapatite column (packed in a 10-ml disposable syringe) that had been equilibrated previously with buffer A plus 200 mM NaCl. The column was washed with 12 ml of equilibration buffer, and protein was eluted with a 60-ml linear gradient of 0 -400 mM (NH 4 ) 2 SO 4 in buffer A plus 200 mM NaCl. DnaC810 eluted at 150 mM (NH 4 ) 2 SO 4 to give fraction 5 (5.2 ml, 16.5 mg of protein). Fraction 5 was concentrated by dialyzing against buffer A plus 50 mM NaCl plus 30% polyethylene glycol 20,000 and loaded onto a 125-ml Superdex-200 fast protein liquid chromatography column that had been equilibrated with buffer A plus 50 mM NaCl. The column was eluted at 1 ml/min. Fractions (1 ml) containing DnaC810 were pooled to give fraction 6 (7.5 ml, 9.2 mg of protein). Fraction 6 was then loaded onto a 3-ml phosphocellulose column that had been equilibrated with buffer A plus 50 mM NaCl. The column was washed with 6 ml of equilibration buffer, and protein was eluted with a 60-ml linear gradient of 50 -400 mM NaCl in buffer A. DnaC810 eluted at 250 mM NaCl (fraction 7; 3.5 ml, 5.2 mg of protein).
Gel Mobility Shift Assay and ECL-Western Blot Analysis-Reaction mixtures (15 l) containing 50 mM Tris-HCl (pH 7.8 at 25°C), 10 mM MgOAc, 10 mM NaOAc, 0.2 mg/ml bovine serum albumin, 8 M ATP (when indicated), 1 nM 32 P-labeled DNA substrate (unless indicated otherwise), and the indicated concentrations of either SSB, DnaC, DnaC810, or DnaB were incubated at 30°C for 10 min. Samples were loaded directly onto 5% (80:1, acrylamide:bisacrylamide) gels using 6 mM Tris-HCl (pH 7.8 at 25°C), 4 mM MgOAc, 10 mM NaOAc, and 1 mM EDTA as the electrophoresis buffer. A constant voltage of 6 V/cm was applied for 4.5 h or as otherwise indicated at 4°C with constant recirculation of the buffer. Following gel electrophoresis, the gel was either dried and autoradiographed or subjected to ECL-Western blot analysis. The conditions for ECL-Western blot were as described by Ng and Marians (32).
DnaC810 Supports SSB-insensitive General Priming-The activity of the DnaC810 suppressor protein was compared with that of wild-type DnaC during X ss(c) 3 RF DNA replication. This reaction utilizes PriA, PriB, PriC, DnaB, DnaC, DnaG, DnaT, and the Pol III HE to convert SSB-coated X ss(c) DNA to the double-stranded, nicked circular form (22). In this reaction, PriA directs the assembly of a primosome at a specific region on the DNA, the primosome assembly site (PAS) (25). The primosome then can synthesize a primer that is elongated by the Pol III HE. DnaC810 was slightly more active than the wild-type DnaC protein in this assay ( Fig. 2A). This indicates that the E176G amino acid substitution present in DnaC810 does not inactivate the protein in any significant way.
Genetic studies (13) predicted that DnaC810 could direct X-type primosome assembly in the absence of PriA. Thus, we examined the activity of the mutant protein in the X ss(c) 3 RF DNA replication reaction in the absence of PriA (Fig. 2B). The activity of the mutant protein in this reaction in the absence of PriA was 40% of what it was in the presence of PriA (Fig. 2, compare A and B), whereas wild-type DnaC was completely inactive in the absence of PriA (Fig. 2B). Thus, DnaC810 had apparently gained the ability to bypass the requirement for PriA in primosome assembly. The observed difference in DnaC810 activity in the presence and absence of PriA most likely reflects the fact that in the presence of PriA, two reactions are being measured: the bypass reaction and X-type primosome assembly at the PAS. In the absence of PriA, only the bypass reaction is operative. Thus, we do not think the apparent reduced activity of DnaC810 in the absence of PriA actually reflects a lowered efficiency in the bypass reaction.
PriA-directed primosome assembly also requires PriB, PriC, and DnaT (3,4). In order to determine whether DnaC810 bypassed the requirement for these proteins as well, we assessed the ability of DnaC and DnaC810 to support X ss(c) 3 RF DNA replication in the presence of only SSB, DnaB, DnaG, and the Pol III HE (Fig. 3A). DnaC810 was still active in this assay; the wild-type protein, however, could not support the reaction. With the exception of the presence of SSB and the PAS on the DNA, the replication reaction manifest in the experiment shown in Fig. 3A resembled the general priming reaction (26). During general priming, DnaB and DnaG can cooperate to synthesize a primer on any protein-free ssDNA.
DnaC stimulates this reaction significantly by forming a complex with DnaB in solution that leads to more efficient transfer of DnaB to the DNA. We therefore investigated the requirement for a PAS sequence and the effect of SSB on the reaction.
A comparison of DNA synthesis supported by either X or f1 ss(c) DNAs in the presence of DnaB, DnaG, DnaC810, SSB, and the Pol III HE showed that the two templates were equally active (Fig. 3B). f1 viral DNA does not carry a PAS. Complementary strand synthesis during the phage life cycle is primed by RNA polymerase (27,28). Thus, the DnaC810-directed replication reaction did not require a PAS, thereby distinguishing it from the PriA-directed reaction.
The DnaC810-directed replication reaction was similar in requirements to the general priming reaction. We therefore compared the activity of the wild-type and mutant proteins in that reaction directly. As noted above for the X ss(c) 3 RF replication reaction, these two proteins displayed little difference in activity (Fig. 4A). However, the differential effect of adding enough SSB to coat the ssDNA was dramatic. The reaction supported by wild-type DnaC was inhibited by SSB, whereas the reaction supported by DnaC810 was stimulated when the ssDNA was coated with SSB ( Fig. 4B). We therefore conclude that DnaC810 can bypass the requirement for PriA, PriB, PriC, and DnaT in loading DnaB to SSB-coated DNA.
In the Presence of DnaB, DnaC810 Supports Stable Complex Formation with SSB-coated ssDNA-The primary role of DnaC is to load DnaB to ssDNA (29). In the cell, it does so in cooperation with other proteins. Afterward, DnaC is not found in the complex of replication proteins on the DNA. In order to understand the properties that allow DnaC810 to bypass PriA function, we used gel mobility shift analysis to examine its ability to form stable complexes with various DNA substrates.
Formation of the DnaB-DnaC complex in solution requires ATP but not ATP hydrolysis (30). Formation of active replication complexes such as the X-type primosome does require ATP hydrolysis (29). Although DnaB is a DNA helicase (5) and ssDNA-dependent NTPase (31), the ATP hydrolysis during loading is catalyzed by DnaC (30). Thus, binding to two oligos, 21 and 42 nt long, was assessed in both the presence and absence of ATP and SSB (Fig. 5).
Several complexes with different R F values were formed in the presence of DnaB, DnaC, DnaC810, and SSB. In no case was any complex detected with either DnaB, DnaC, or DnaC810 individually. In the absence of SSB, the combination of DnaB and DnaC gave complex i with the 21-mer (Fig. 5, A and B, lane 3) and complex iv with the 42-mer (Fig. 5, C and D,  lane 3). Formation of this complex was unaffected by the presence or absence of ATP. In the presence of SSB with the 21-mer substrate, no complex was observed in the absence of ATP (Fig.  5A, lane 9), whereas a new species, complex ii was observed in the presence of ATP (Fig. 5B, lane 9). No complexes were observed with the 42-mer substrate in the presence of SSB under any condition (Fig. 5, C and D, lane 9).
The results with DnaC810 were quite different. In the absence of SSB, the combination of DnaB and DnaC810 yielded a new complex, iii, on the 21-mer that was, like complex i, unaffected by the presence or absence of ATP but had a distinct, reduced mobility (Figs. 5, A and B, lane 4). Likewise, with the 42-mer substrate, new complexes (vi and vii), were observed that had a reduced mobility compared with complex iv (Fig. 5, C and D, lane 4). In this case, formation of complexes vi and vii was clearly stimulated by the presence of ATP. In the presence of SSB, two complexes (ii and iii) were formed with the 21-mer substrate (Fig. 5, A and B, lane 10). The yield of these complexes seemed relatively unaffected by the presence or absence of ATP. With the 42-mer substrate, ATP was required to observe three complexes (v, vi, and vii) (Fig. 5, C and D, lane 10).
Complex i was, based on its unusual U-shaped appearance, reminiscent of the type of complex we have observed previously with DnaB and DnaC on a 304-nt-long ssDNA containing a PAS (32). ECL-Western analysis demonstrated that only DnaB was present in the complex on the PAS DNA. Thus, we believe that complex i contains only DnaB. Thus, DnaC810 appears to be able to transfer DnaB to SSBcoated ssDNA. In addition, in the presence of DnaC810, DnaB, and ATP a second larger complex can be detected on SSBcoated DNA that may contain both DnaB and DnaC810. The identical mobility of complex iii in the presence and absence of SSB argues that SSB has been displaced from the DNA. The difference between complexes i and iii suggests that DnaC810 is either slower than DnaC to release DnaB to the DNA or requires an interaction with another protein to do so.
Wild-type DnaC could load DnaB to SSB-coated 21-mer but not to the 42-mer. This is most likely a result of the stability of the SSB on the DNA. The observed difference in complex formation with DnaB and DnaC is a function of the size of the oligomer. The most likely aspect of the binding reaction that can be affected by oligomer size is the number of SSB molecules bound to the DNA. The simplest way to think of this is that one SSB tetramer is bound to the 21-mer, whereas two SSB tetramers are bound to the 42-mer. This allows the SSB tetramers bound to the 42-mer to interact. The energy gain from this interaction would make it more difficult to displace SSB from the 42-mer than from the 21-mer. Alternatively, only one protomer in the SSB tetramer may be bound to the 21-mer, whereas two protomers of the tetramer may be bound to the 42-mer. This scenario would also make it more difficult for SSB to be displaced from the 42-mer than from the 21-mer.
Learn et al. (33) demonstrated that it is actually DnaC that first binds to ssDNA while transferring DnaB via a cryptic DNA-binding activity that becomes activated when bound to DnaB. The relative affinities of DnaC and SSB for the DNA thus become the crucial factor in determining whether DnaC can transfer DnaB to the DNA. Because of the potential for greater cooperative interactions on the larger DNA, less energy would be required for DnaC to displace SSB from the 21-mer than from the 42-mer. Because DnaC810-and DnaB-catalyzed general priming is SSB-insensitive (Fig. 4B), the E176D mutation in DnaC810 apparently increases its affinity for ssDNA in the presence of DnaB so that it maintains the ability to displace SSB no matter the size of the DNA. The observation that DnaB-and DnaC810-dependent complex formation on the 42mer, but not on the 21-mer, requires ATP is also consistent with this explanation.
DnaC810 Can Load DnaB to D Loop DNA in the Presence of SSB-We have shown that PriA can direct the assembly of a replication fork on a template DNA carrying a D loop in vitro. As with the X ss(c) 3 RF DNA replication reaction described above, DnaC810 could bypass the requirement for PriA, PriB, PriC, and DnaT and direct the assembly of a replication fork at D loop DNA (20). This implies that DnaC810 can load DnaB directly to SSB-coated D loop DNA. We therefore investigated this directly by gel mobility shift analysis.
We first assessed the ability of DnaC810 to bind a bubble formed from two 82-nt-long oligos. This DNA substrate has, going 5Ј 3 3Ј on the top strand, 23 nt of duplex, a 42-nt long nonhomologous bubble, and then 17 nt of duplex (18). In the absence of SSB, two complexes could be observed in the presence of DnaC810 and DnaB that were unaffected by either the presence or absence of ATP (Fig. 6, A and B, lane 4). These were similar in relative mobility to the complexes (iii and iv) observed on the oligo substrates and were expected based on the predicted ssDNA present in the nonhomologous region of the bubble substrate. Interestingly, wild-type DnaC was unable to form any similar complexes with DnaB, although it could on the oligo substrates.
Under standard reaction conditions in the presence of SSB, no binding to the bubble DNA could be detected in the presence of DnaB and either DnaC or DnaC810 either in the presence or absence of ATP (Figs. 6). This was surprising, because, as demonstrated above, DnaC810 could form a complex with DnaB on, and transfer DnaB to, SSB-coated ssDNA (Fig. 5). We investigated whether this was simply a result of a reduced affinity of the DnaC proteins for the ssDNA in the bubble. However, no binding was observed in either the presence or absence of ATP even when the concentration of DnaC and DnaC810 was increased to very high levels (data not shown).
Our previous studies have shown that PriA binds D loop DNA with high affinity but does not bind to bubble DNA at all, although it does not require a free end to bind DNA (18). Through the use of various forms of bubble DNA substrates, we concluded that this was because, although the nonhomologous region was clearly in single-stranded form (based on nuclease sensitivity), the single strands were still twisted about each other as in a tangle. This is also probably why we did not find complex formation on either the bubble substrate in the presence of DnaB and DnaC or on the SSB-coated bubble substrate in the presence of DnaB and DnaC810. Complex formation in the presence of DnaB and DnaC810 on the bubble in the absence of SSB reinforces the suggestion that the E176D mutation has increased the affinity of DnaC810 for DNA.
Binding of DnaB in the presence of either DnaC or DnaC810 to D loop DNA was investigated next. The D loop DNA substrate was formed by annealing a 42-nt-long invading strand that was homologous to the top strand of the bubble (18). In the presence of ATP, neither DnaB, DnaC, nor DnaC810 could form a stable complex on D loop DNA either in the presence or absence of SSB (data not shown).
In the absence of SSB, the combination of DnaB and DnaC gave little evidence of complex formation in the absence of ATP (Fig. 7A, lanes 2-7) and formed primarily two complexes, b and c, in the presence of ATP (Fig. 7B, lanes 3-7). Complex c exhibited the distinctive U-shaped appearance also shown by complexes i and iv that we have attributed to the presence of only DnaB on the DNA (32). Increasing the DnaC concentration significantly did little to change the distribution between complexes b and c. The combination of DnaB and DnaC810 also yielded complexes b and c. In addition, there were trace amounts of a slowly moving complex (a) (Fig. 7A, lanes 8 -12).
In the presence of ATP, complexes b and c diminished in abundance as the concentration of DnaC810 was increased, and complex a became very prominent (Fig. 7B, lanes 9 -13).
Both complexes b and a always appeared as doublets. The resolution of the dimer pair varied from gel to gel. In general, the slower moving complex in the pair was generally the most populated, but we could not associate a difference in distribution with a particular set of conditions in the reaction mixture. We assume that these dimer pairs relate either to a difference in stoichiometry of the proteins present or to a difference in conformation of one or both of the proteins present.
In the presence of SSB, no complexes were evident when DnaB and DnaC were included in the reaction mixture either in the absence (Fig. 8A, lanes 2-7) or presence of ATP (Fig. 8B,  lanes 2-7). This was consistent with the inability of this combination of proteins to form a complex on SSB-coated ssDNA that was larger than a 21-mer. In contrast, the combination of DnaB and DnaC810 formed primarily complex b at low concentrations of DnaC810 and complex a at high concentrations of DnaC810. Small amounts of another complex (cЈ) were also evident as a smeared thickening of the amount of labeled material on the sides of the lane just above the position of the SSB-coated D loop DNA (Fig. 8B, lanes 9 -12). Complex cЈ is presumably related to complex c, but we label it differently because of the presence of SSB in the reaction mixture.
The identical nature of complexes a and b formed in the absence (Fig. 7) and presence (Fig. 8) of SSB was confirmed when these reactions were analyzed side-by-side on the same gel ( Fig. 9) as described below.
Thus, the combination of DnaB and DnaC810 was clearly able to form at least two complexes on SSB-coated D loop DNA, whereas the combination of DnaB and DnaC could not. This was similar to the results obtained when SSB-coated ssDNA was used as the substrate. In that case, it appeared as if DnaC810 was capable of displacing SSB from the DNA. To assess if this were the case with the D loop DNA as well, we compared on the same gel the mobility of complexes formed in the presence and absence of SSB (Fig. 9).
In the presence of DnaB and DnaC810 and in the absence of ATP, complex b was the major product (Fig. 9A). Much higher concentrations of DnaC810 were required to form substantial amounts of complex b in the presence of SSB than in its ab-sence. Nevertheless, it is clear that the mobility of complex b formed in either the presence or absence of SSB was identical. This strongly suggests that SSB has been displaced. Note the large shift in mobility of the D loop when SSB is bound (compare lane 7 with lane 1 in Fig. 9A). As shown below, complex b contains DnaB. If only one 300-kDa hexamer of this protein were bound to the D loop in addition to SSB, there should be a significant change in mobility of the complex compared with when SSB is absent.
Similar results were obtained in the presence of ATP where complex a is the major species formed (Fig. 9B). Here as well, the mobility of complex a in the presence and absence of SSB was identical. This complex contains both DnaB and DnaC810 (see below). Thus, one would expect the difference in mobility if SSB were also bound to the D loop compared with its absence to be even greater than that for complex b. Thus, we conclude that during formation of both complexes a and b, SSB is displaced from the DNA by the action of the DnaC810 protein.
ECL-Western blotting was used to determine the distribution of DnaB and DnaC in complexes a and b. These analyses were confounded somewhat by the fact that under the conditions of electrophoresis used for gel mobility shift analysis, the free DnaB and DnaC migrated into the gel. However, the data were clear enough to allow conclusions to be made.
The mobility of free DnaB was very close to that of complex  (Fig. 10), so that under standard conditions, when DnaB was in vast excess over the DNA substrate, the smearing of the anti-DnaB antibody-reactive material obscured the position of complex b (Fig. 10B, lanes 10 -12). To circumvent this problem, D loop binding assays were performed with concentrations of the D loop substrate that were in excess of the concentration of DnaB (Fig. 10A). This is why little complex formation is evident on the autoradiogram (Fig. 10A, lanes 4 -6). However, under these conditions, the presence of DnaB in complex b could be clearly observed in the Western blot (Fig. 10A, lanes 10 -12). On the other hand, the presence of DnaB in complex a could be clearly observed under standard conditions as the concentration of DnaC810 in the reaction mixture was increased (Fig.  10B, lanes 4 -6 and 10 -12).
The situation was reversed in assessing the presence of DnaC810 in these two complexes. Free DnaC810 barely moved into the gel, obscuring the position of complex a (Fig. 11). However, as the concentration of DnaC810 was increased in the reaction mixture, a band could be observed at the trailing edge of the smear of free DnaC810 in the Western blot coincident with the position of complex a that was not present in each case when DnaB was omitted from the reaction mixture so that no complex would form (Fig. 11, lanes 2-7 and 9 -14). No DnaC810 signal was observed in the position of either complex b or c. We therefore conclude that complex a contains both DnaB and DnaC810, whereas complex b contains only DnaB.
We have no direct evidence for the composition of complex c. Thus, any conclusion made as to its identity must be considered speculative. Based on the relative mobilities of complexes a, b, and c, we suspect that it contains only DnaB. This is consistent with our previous conclusion with respect to the nature of these U-shaped bands (32). The significantly greater mobility of complex c compared with complex b suggests that complex c either has a lower ratio of DnaB to DNA or that the DnaB has assumed a different, more compact, conformation than in complex b.
We make no claim that the Western analyses can be used to determine relative stoichiometries of DnaB and DnaC810 in the complexes. Although polyclonal antibodies were used, major antigenic determinants can be buried and inaccessible in one complex but not another. Rather, these data can only help to suggest which proteins are present in which complex.
These data indicate that whereas both DnaC and DnaC810 can load DnaB to a D loop in the absence of ATP to give complex b, only DnaC810 can do so in the presence of SSB. In addition, in both the absence and presence of SSB, DnaC810 can form a  6 in A and B). ECL-Western blot analysis using anti-DnaB antiserum was performed on nitrocellulose papers after transfer from the wet gels (lanes 7-12 in A and B).
complex with DnaB on the D loop DNA (complex a). Both DnaC810-directed loading of DnaB and formation of the DnaB-DnaC810 complex result in the displacement of SSB. This ability of DnaC810 to load DnaB onto SSB-coated DNA, which represents a gain of function over the properties of the wild type, accounts for its ability to bypass the function of PriA, PriB, PriC, and DnaT during replication fork formation at a D loop (20). DISCUSSION Current models suggest that the replication forks formed at oriC often stall and collapse either at DNA damage such as nicks and noncoding lesions or by colliding with protein roadblocks on the DNA (12, 13 16, 34, 35). Correction of the lesion requires both repair and recombination proteins and continued viability of the cell requires replication fork restart. Restart is likely to occur in the majority of instances through recombination-directed DNA replication, where a recombination intermediate such as a D loop serves to provide both the primer for leading-strand synthesis (the 3Ј-OH of the invading strand) and the site at which a replisome assembles.
Because strains carrying a disruption in priA were constitutively induced for the SOS response (9) and shown to be defective in homologous recombination (13,14), the repair of both double-strand breaks (14) and UV-damaged DNA (10,13,14) and inducible stable DNA replication (15) (which requires both recA and recBCD (36)), it was proposed that PriA was involved in directing replication fork assembly at recombination intermediates (13,16). This has been supported by the demonstration that PriA binds specifically to D loop DNA (17,18) and can direct the assembly of both a X-type primosome (19), and a replication fork (20) at a D loop in vitro.
In this report, we have investigated the properties of the DnaC810 protein, encoded by dnaC810, a naturally arising, extragenic suppressor of all of the phenotypes of the priA2::kan mutation (13). Genetic analysis has indicated that dnaC810 bypasses PriA function. This suggested that DnaC810 could also direct replication fork assembly at a D loop. This proved to be the case (20). Here we show that the underlying gain of function that allows DnaC810 to bypass PriA activity is an ability to load DnaB, the replication fork DNA helicase (7,37,38), to SSB-coated DNA.
Although DnaB can bind ssDNA to act in vitro as a ssDNAdependent NTPase (31), a DNA helicase (5), and a landing pad for DnaG during general priming (26), the protein is, in general, prevented from binding indiscriminately to ssDNA in vivo by its association with DnaC in a tight stoichiometric complex (36). While in this complex, DnaC can transfer DnaB to exposed ssDNA but not to SSB-coated ssDNA. This transfer is mediated by a cryptic DNA-binding site in DnaC that is activated by binding of DnaB (33). While in the form of a DnaB-DnaC complex, DnaB cannot bind ssDNA (33). Under normal circumstances, any exposed ssDNA will be coated with SSB. Access of DnaB to the DNA is therefore effectively limited to targeting by other mechanisms that involve additional DNA replication proteins to generate a SSB-free region of ssDNA for DnaB binding. The two mechanisms that operate in vivo to do this are DnaAdirected initiation of DNA replication at oriC (37) and PriAdirected assembly of a primosome at recombination intermediates such as a D loop (20).
DnaC810 was able to load DnaB to both SSB-coated ssDNA and D loop DNA. We have previously shown by DNA footprinting studies that during assembly of the X-type primosome, DnaB is loaded to the displaced strand in the D loop (19). This is consistent with the ability of DnaC810 to form a replication fork at a D loop in the presence of SSB, DnaB, and the Pol III HE, where the invading strand is used as the leading strand primer (20). This also places DnaB on the displaced strand that becomes the lagging strand template. Thus, we believe that DnaC810 is also loading DnaB to the displaced strand of the D loop. Because this is the only region of the D loop that can bind SSB, this would also explain the inability of the wild-type DnaC to load DnaB to the D loop in the presence of SSB.
The most likely explanation for the gain of function exhibited by DnaC810 is that the affinity for ssDNA of its cryptic DNA binding activity has been increased. This would account for its ability to compete with SSB for binding to the ssDNA. In addition to having gained the ability to transfer DnaB to SSBcoated DNA, DnaC810 can also be found on the same DNA as DnaB. Although there is no direct evidence that DnaC810 and DnaB are together in a complex on the DNA, we believe that this is the case.
DnaC810 alone will not bind to ssDNA, either in the presence or absence of SSB. Thus, in order for both DnaB and DnaC810 to be on the same DNA, a DnaB-DnaC810 complex would have to have interacted with the DNA and then rearranged in such a manner that both proteins were left on the DNA. This seems unlikely, because free DnaB can be found on the DNA, transferred from the DnaB-DnaC810 complex, but free DnaC810 (or DnaC, for that matter) is never found bound to the DNA. In addition, DnaC is not present on the DNA after formation of the X-type primosome (32,39).
The nature of this DnaB-DnaC810 complex on the DNA, however, is unclear. The argument advanced above would suggest that it is bound in a stable fashion as a result of the DnaC  1-6). ECL-Western blot analysis using anti-DnaC antiserum was performed on nitrocellulose paper after transfer from the wet gel (lanes 7-12).
DNA-binding site and not that of DnaB. Presumably, transfer of DnaB to the DNA ejects DnaC from the complex. This could be seen to be consistent with the requirement for higher concentrations of DnaC810 to form this complex compared with the concentration required to transfer DnaB directly to the DNA.
Based on our biochemical analyses described here, can we account for the properties of the dnaC810 mutation in vivo?
The answer is mixed. The ability to bypass PriA function and rescue all of the phenotypes of the priA2::kan mutation can clearly be attributed to the gain of function that allows DnaC810 to load DnaB to SSB-coated D loop DNA. This allows the crucial replication fork restart downstream of recombination intermediates required for continued cell viability. Biochemically, neither PriA, PriB, PriC, nor DnaT are required for this bypass step. Thus, it is interesting that genetically, priA2::kan, dnaC810 strains appear to be partially priB-and priC-dependent.
In contrast to what was predicted based on the pathway of assembly of the X-type primosome (32), neither priB nor priC disruptions exhibit any of the phenotypes associated with the priA2::kan mutation (40). Remarkably, however, the priB, priC double mutant did display the multiple phenotypes associated with the priA2::kan mutation and was even less viable and grew more slowly (40).
This suggests that PriB and PriC have redundant functions in a pathway that involves PriA. Although the biochemical properties of DnaC810 would lead one to expect that the dnaC810 mutation should be able to completely suppress the phenotypes of the priB, priC double mutant, this is not the case. The priB, priC, dnaC810 triple mutant strain is still as recombination-deficient and induced for the SOS response as priA2::kan. A second suppressor mutation that also arises in dnaC, dnaC820, is required for complete reversion of the phenotypes. This implies, contrary to the biochemical properties demonstrated here, that DnaC810 would be partially PriB-and PriC-dependent.
We suspect that these apparent inconsistencies between the biochemical and genetic data relate to the fact that we have yet to incorporate recombination proteins into our biochemical systems. Replication fork assembly at a recombination intermediate such as a D loop in vivo presumably has to compete with the binding and action of a variety of recombination proteins. This could easily change the apparent biochemical requirements for replication fork restart.