Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Struct Mol Biol. Author manuscript; available in PMC 2013 December 1.
Published in final edited form as:
PMCID: PMC3685420

Preferential D-loop Extension by a Translesion DNA Polymerase Underlies Error-Prone Recombination


Although homologous recombination (HR) is considered an accurate form of DNA repair, genetics suggest that Escherichia coli (E. coli) translesion DNA polymerase (pol) IV (DinB) promotes error-prone recombination during stress which allows cells to overcome adverse conditions. How pol IV functions and is regulated during recombination under stress, however, is unknown. We show that pol IV is highly proficient in error-prone recombination, and is preferentially recruited to D-loops at stress-induced concentrations in vitro. Unexpectedly, we find that high-fidelity pol II switches to exonuclease mode at D-loops which is stimulated by topological stress and reduced deoxy-ribonucleotide pools observed during stationary-phase. The exonuclease activity of pol II enables it to compete with pol IV which likely suppresses error-prone recombination. These findings indicate that preferential D-loop extension by pol IV facilitates error-prone recombination and explain how pol II reduces such errors in vivo.

Homologous recombination (HR) repairs double-strand breaks (DSBs) by directing replication to copy sequence information from a homologous donor (Fig. 1a)14. For example, following formation of a DSB, nucleases resect the DNA resulting in a 3’ single-strand DNA (ssDNA) tail. RecA type recombinases form a filament along the tail which facilitates strand invasion within a homologous donor DNA resulting in a displacement loop (D-loop). DNA polymerase (pol) then extends the 3’ end of the invading strand by using the complementary strand within the donor DNA as a template—a process called recombination-directed replication (RDR) or D-loop extension. Recombination intermediates are then further processed to form Holliday junctions which are resolved by endonucleases.

Fig. 1
Pol IV is highly proficient and error-prone in recombination-directed replication

Importantly, the accuracy of HR, which is widely considered to be high, is dependent on the fidelity of RDR. Although high-fidelity replicative pols predominantly perform RDR58, mounting evidence indicates that low-fidelity translesion pols also carry out this function which suggests that HR is error-prone. For example, previous studies in eukaryotes including humans, yeast, chickens, frogs and flies indicate that low-fidelity translesion pols η, ν, θ, and ζ contribute to RDR which probably promotes mutagenesis915. Moreover, recent evidence from Drosophila suggests that translesion pols even compete with replicative pols during HR10. Although translesion pols are widely known to promote replication past lesions in DNA, it is becoming clear that these low-fidelity enzymes function during HR in various organisms and therefore may have been selected to perform RDR in all domains of life.

In E. coli, several years of genetic studies have suggested that Y-family translesion pol IV (DinB) promotes error-prone HR in the form of mutations specifically during stress1620. Such error-prone recombination allows E. coli to rapidly evolve and overcome stressful conditions including nutritional starvation and exposure to antibiotics16,19. Pol IV induced recombination errors are therefore also referred to as stress-induced or adaptive mutations. Our current knowledge of pol IV involvement in error-prone HR is mostly based on genetic data. For example, previous genetic studies have demonstrated that pol IV induced mutations are targeted to regions of DSBs and require SOS-induced levels of pol IV (~2,500 molecules per cell), recombination factors (i.e. RecA, RecBCD, RuvABC), and the RpoS stress response which further upregulates pol IV (~100%) and downregulates mismatch repair16,17,19,21,22. Current models based on these findings propose that pol IV promotes mutations near D-loops during HR under stress17,20. Yet clear evidence for pol IV RDR activity, which is outside its normal role in translesion synthesis, has never been demonstrated. Furthermore, how pol IV is recruited, regulated, and competes with other pols during HR under stress remains unclear. Here we sought to provide mechanistic insight into the activity and regulation of pol IV during HR under stress.


Pol IV is proficient and error-prone in RDR

We used a biochemical approach to investigate the activity and regulation of pol IV in RDR (D-loop extension)(Fig. 1b). The results demonstrated that pol IV promotes RDR (Fig. 1c, lanes 1–4), which requires the polymerase and RecA (Fig. 1d). In contrast, the related Y-family pol V (UmuD’2C) failed to perform RDR under identical conditions (Fig. 1c, lanes 5–8). The inability of pol V to promote RDR was surprising since its activity requires RecA filaments in trans which are present in excess as indicated by the free ssDNA (Fig. 1c, lower band)23. Nevertheless, we repeated the reaction with pol V, but added more than twice the amount of RecA along with increasing concentrations of unlabeled heterologous trans ssDNA which can not form a D-loop. Still, pol V failed to promote RDR (Fig. 1e). As a positive control, we demonstrated that the same amount of pol V extends a simple primer-template which requires ssDNA and RecA in trans as shown previously (Fig. 1f)23. Considering pol V plays a small role in promoting stress-induced mutations during recombination, additional factors may be needed to stimulate its activity at D-loops24.

We next examined whether β, which confers processivity onto pols, is required for pol IV RDR activity. Here a relatively high polymerase to D-loop ratio was used which we based on approximate conditions in SOS-induced cells (Table 1). For example, pol IV is highly upregulated (~101) by the SOS response to ~2,500 molecules per cell, making it the most abundant DNA polymerase in stressed cells (Table 1)25,26. Since the average number of DSBs leading to D-loops in stressed cells is unknown, it is difficult to model in vitro. We therefore used an approximation of four DSBs per cell since previous estimates suggest that each chromosome incurs a break and E. coli contains four chromosomes in rich medium2,27,28. Since DSBs result in two DNA ends capable of forming D-loops, this results in a polymerase to DNA end ratio of 312.5:1 for pol IV (Table 1). Considering SOS-induced cells may contain fewer than four DSBs, this ratio may under represent the amount of pol IV relative to D-loops. Nevertheless, using these relative amounts which take into account the approximate concentration of D-loops formed in our assay (Supplementary Fig. 1), we showed that pol IV is able to promote RDR in the absence of β (Fig. 1g, left), even under conditions of high ionic strength (Fig. 1g, right). This result was unexpected and is in contrast to previous biochemical studies of yeast proteins which have demonstrated that PCNA, the eukaryotic equivalent of β, is required for RDR by replicative pol δ and translesion pol η11. Our results suggest that pol IV may not require β for its involvement in RDR due to its abundance in stressed cells.

Table 1
Relative amounts of translesion pols in SOS-induced cells

Next, we investigated the ability of pol IV to promote mutations at D-loops which is thought to be the central mechanism of error-prone recombination. To our knowledge, the fidelity of pols on D-loops has not previously been investigated in vitro. As a control, we first examined the fidelity of pol IV on a primer-template which resembles the DNA substrate used during translesion synthesis. The results showed that pol IV strongly discriminates against incorporating incorrect nucleotides on a primer-template (Fig. 1h, left). In contrast, pol IV exhibited a surprisingly high efficiency of nucleotide misincorporation on a D-loop under similar conditions and sequence context as in the previous experiment with the primer-template (Fig. 1h, compare right and left panels). Pol IV also appears to be more prone to mismatch extension on the D-loop which is indicated by the upper bands in lanes two and four (Fig. 1h, right). These data suggest that RDR may be inherently error-prone. Alternatively, since RecA interacts with pol IV and collaborates with UmuD to modulate the fidelity of the polymerase, the recombinase may reduce the accuracy of pol IV on the D-loop29.

Considering genetic data strongly implicate pol IV in promoting mutations during DSB repair under stress, our observation of pol IV mutagenic activity on a D-loop provides the molecular basis for its involvement in stress-induced mutagenesis. Comparison of pol IV activity on D-loops with different sequences suggests that the polymerase preferentially misincorporates the nucleotide that is complementary to the +2 template base which is located two positions downstream from the 3’ end of the invading ssDNA (compare Fig. 1h, right and 1i). For example, in the right panel of figure 1h pol IV preferentially misincorporated dCMP which is complementary to the +2 template base guanosine. In contrast, dGMP is preferentially misincorporated in figure 1i where cytosine is the +2 template base. Previous biochemical studies have clearly shown that pol IV is able to ‘skip over’ the correct template base which facilitates a −1 frameshift mutation30. The ability of pol IV to promote −1 frameshifts has been widely used to detect its activity in error-prone recombination16,18,19. Consistent with these previous studies, we showed that pol IV is capable of promoting a −1 frameshift on a D-loop (Fig. 1j). Importantly, pol IV induced mutations are likely to go unrepaired during stationary-phase since mismatch repair is deficient16,19. The data presented in figure 1 demonstrate that pol IV is highly proficient in error-prone RDR which provides mechanistic insight into its role in stress-induced mutagenesis18.

Mechanism of pol IV recruitment to D-loops

Given that the replisome performs RDR in unstressed proliferating cells, an important consideration is how pol IV gains access to D-loops during stress5. Since SOS concentrations of pol IV are required for its role in error-prone recombination, we reasoned that upregulation of the polymerase might facilitate its recruitment to D-loops31. For example, polymerase access to D-loops in proliferating cells is prevented by primosomal protein PriA which binds tightly to D-loops where it recruits other primosomal proteins to assemble the replisome during RDR (Fig. 2a, left)32. Consistent with previous studies, PriA inhibited D-loop extension by replicative pol III (Fig. 2a, right)32. Similarly, concentrations of PriA corresponding to known amounts in the cell (~70 molecules per cell33) suppressed RDR by pol IV at levels corresponding to non-stressed cells (~250 molecules per cell25,26)(Fig. 2b, left panel). In contrast, PriA only slightly inhibited D-loop extension by pol IV when the polymerase was added at levels comparable to stress-induced concentrations (~2,500 molecules per cell; Fig. 2b, right panel). This indicates that pol IV outcompetes PriA at D-loops during stress (Fig. 2b, right schematic). Similar preferential RDR activity by pol IV was observed in the presence of RuvAB which also binds tightly to D-loops and promotes stress-induced mutations (Supplementary Fig. 2)16. We further found that pol IV levels comparable to SOS-induced cells promote RDR in the presence of all the primosomal and replisome components and supersede D-loop dependent replisome activity (Supplementary Fig. 3). Together, these results indicate that pol IV gains access to D-loops and overrides the primosome dependent RDR machinery during stress.

Fig. 2
High levels of pol IV comparable to SOS-induced cells facilitate its recruitment to D-loops

D-loop dependent stimulation of pol II exonuclease activity

Genetic data suggest that translesion pol II also plays a role in RDR during growth-limiting conditions20,34,35. Similar to pol IV, pol II is upregulated during stress (~350 molecules per cell, Table 136) and is considered a translesion pol due to its ability to promote replication past certain lesions26. However, unlike pol IV which is a Y-family DNA polymerase that exhibits low-fidelity DNA synthesis, pol II is among the B-family of pols which exhibit high-fidelity DNA synthesis and exonuclease activities. Pol II therefore possesses similar characteristics to eukaryotic replicative pols δ and ε which are also B-family members. Interestingly, genetic studies indicate that the exonuclease domain of pol II suppresses stress-induced mutations35. Considering pol IV promotes most stress-induced mutations, this suggests that pol II may regulate pol IV activity during RDR by competing with and proofreading pol IV errors.

We investigated whether pol II performs RDR in figure 3. Remarkably, we found that although pol II initially extends the D-loop, the reaction is subsequently reversed, presumably due to the enzyme’s 3’–5’ exonuclease activity (Fig. 3a, left). Indeed, the reverse reaction was not performed by a previously characterized mutant version of pol II (D155A, E157A) that is deficient in exonuclease activity (Fig. 3a, right)37. Such constant exonuclease activity by a DNA polymerase is unprecedented considering the reaction was performed with a saturating deoxy-ribonucleotide (dNTP) concentration (50 µM) that exceeds the enzyme’s Km for dNTPs (3.7 µM)38. Given that dNTP pools are reduced to similar concentrations (~35–50 µM) during stationary-phase39, and concentrations of pol II corresponding to SOS conditions gain access to D-loops in the presence of PriA (Supplementary Fig. 4), the observed phenomenon is likely relevant to pol II activity in stressed cells and thus warranted further investigation.

Fig. 3
Pol II switches to an active exonuclease mode at D-loops

We hypothesized that the exonuclease domain of pol II is stimulated due to inhibition of forward movement by topological constraint in the DNA generated by positive supercoils during D-loop extension. Consistent with this notion, we demonstrated that wild-type and exonuclease deficient pol II act similar on a linear double-strand DNA template and a circular primer-template which indicates that the observed exonuclease activity is specific to D-loops (Fig. 3b). Our hypothesis was further supported by the effects of gyrase which removes positive supercoils in DNA and suppressed the reverse reaction (Fig. 3c), which appears to be due to a slight delay in pol II exonuclease activity (Supplementary Fig. 5). These results support a model whereby superhelical tension generated during D-loop extension promotes the reverse translocation and exonuclease function of pol II which facilitates a switch to a highly active exonuclease mode (Fig. 3d). This model is supported by biophysical studies that have demonstrated that mechanical tension on the DNA template stimulates proofreading by bacteriophage phi29 pol40. We further found that reducing the dNTP pool to 10 µM facilitates the reverse reaction (compare Fig. 3e with Fig. 3a, left), whereas increasing the dNTP pool to 100 µM—which reflects the conditions in proliferating cells—either during (Fig. 3f) or prior (Fig. 3g) to the reaction prevents pol II from switching to exonuclease mode during the same timecourse39. These findings suggest that the enzyme’s Km for dNTPs is increased under conditions of opposing force (i.e. superhelical tension).

We next examined whether this exonuclease activity is specific to pol II by comparing RDR by wildtype and exonuclease deficient pol III. Here, similar conditions were used as with pol II (i.e. 50 µM dNTPs), however, the γ-subunit of the clamp-loader was substituted with τwhich specifically binds to pol III and facilitates its recruitment to DNA; τrepresents the full-length version of γ which is truncated due to a translational frameshift41. The results showed that exonuclease deficient pol III extends D-loops further than wild-type pol III (Fig. 3h), but acts the same as wild-type on a primer-template (Fig. 3i) and within the replisome (Fig. 3j). Hence, the exonuclease domain of pol III was also surprisingly activated at D-loops. However, only pol II exhibited the unique ability to reverse the D-loop extension reaction which we attribute to its highly active exonuclease domain compared to pol III (Fig. 3k).

Pol II requires exonuclease activity to compete with pol IV

We next examined competition between pol II and pol IV by performing RDR in the presence of both enzymes using their relative concentrations observed in SOS-induced cells (see Table 1). Since each polymerase produced a distinct product, we were able to determine which enzyme acts dominantly. The results showed that pol IV competes with pol II during RDR as indicated by a net increase in D-loop extension by pol IV versus D-loop resection by pol II (Fig. 4a, left; compare lanes 7–9 to lanes 1–3 and 4–6). Further reduction of the dNTP pool (10 µM) enhanced pol IV activity compared to pol II (Fig. 4a, right). Since the exonuclease function of pol II reduces (~83%) stress-induced mutations in vivo35, we wondered whether the observed exonuclease activity affects the ability of pol II to compete with pol IV. To test this idea, we examined competition between exonuclease deficient pol II and pol IV during RDR. The different D-loop extension products indicated that exonuclease deficient pol II fails to effectively compete with pol IV (Fig. 4b, left and right panels; compare lanes 7–9 to lanes 1–3 and 4–6). Considering similar products were generated at 50 µM dNTPs (left), we further resolved the polymerases’ products which unequivocally showed that exonuclease deficient pol II is unable to effectively compete with pol IV (Fig. 4c). These results were in contrast to those observed for wild-type pol II which competes with pol IV (Fig. 4a). Thus, the data in figure 4 demonstrate the unexpected finding that pol II requires a functional exonuclease domain to compete with pol IV and thus regulate error-prone RDR.

Fig. 4
Pol II requires a functional exonuclease domain to compete with pol IV at D-loops

Importantly, the concentration of pol IV is further upregulated approximately 100 percent (to ~5,000 molecules per cell) by the RpoS general stress response in stationary-phase cells17,20. Whether this increase in pol IV expression contributes to its activity at D-loops, however, is unknown. We found that increasing the relative levels of pol IV by 100 percent enabled pol IV to further compete with pol II as indicated by a marked reduction in net exonuclease activity at D-loops (compare Fig. 5 with Fig. 4a). Considering pol IV outcompeted pol II, the pol III replisome (Supplementary Fig. 3) and pol I (Supplementary Fig. 6) which is the second most abundant pol in stressed cells (~400 molecules per cell42), our data suggest that pol IV plays a major role in RDR during stress which likely facilitates adaptive evolution.

Fig. 5
Levels of pol IV corresponding to RpoS-induced cells allow it to limit pol II exonuclease activity at D-loops


Models based on genetics have long proposed a role for pol IV in error-prone RDR in growth-limited cells as the mechanism of stress-induced mutagenesis, also known as error-prone recombination16,17,19. Genetics also implicate pol IV in RDR during replication restart43. In vitro evidence supporting pol IV RDR activity, however, has been lacking. This report verifies the ability of pol IV to promote RDR in a reconstituted assay which, in conjunction with previous genetics, establishes a new DNA repair function for this enzyme. Our data further show that pol IV is efficient in misincorporation and mismatch extension during D-loop extension which is consistent with its ability to generate mutations during HR in stressed cells. The high proficiency of pol IV in error-prone RDR and its abundance in growth-limited cells (~5,000 molecules per cell17,20) provide an explanation of why this enzyme is responsible for most (85%) stress-induced mutations18.

Intriguingly, we found that pol IV is considerably more error-prone on a D-loop compared to a primer-template (Fig. 1h). We suspect that the DNA structure or RecA which are unique to the D-loop extension reaction contribute to the reduced fidelity of pol IV. Previous studies showed that RecA interacts with pol IV and collaborates with UmuD to modulate the fidelity of the polymerase29. Thus, RecA binding to pol IV could conceivably reduce its ability to discriminate against incorrect nucleotides during RDR. Alternatively, the dynamic structure of the D-loop may affect the fidelity of pol IV. For example, RecA-mediated D-loops are unstable due to the ability of the recombinase to promote dissociation of the invading strand44. Thus, the inherent instability of the D-loop may compromise the fidelity of all pols. This would indicate that RDR is inherently error-prone. However, considering error-prone recombination has not been observed in proliferating E. coli cells, other factors may be needed to promote high-fidelity RDR. For example, mismatch repair proteins which ensure accurate recombination and replication may prevent error-prone RDR in the absence of stress. Consistent with this, suppression of mismatch repair during stationary-phase is thought to contribute to errorprone recombination18. Further studies are needed to determine what reduces the accuracy of pol IV at D-loops, whether this phenomenon is universal for all pols, and whether other factors promote high-fidelity RDR.

Genetic studies have shown that SOS-induced levels of pol IV are necessary for its involvement in errorprone recombination31. This has suggested that upregulation of pol IV promotes its RDR activity. Inline with the previous genetics, we have demonstrated that SOS-induced concentrations of pol IV facilitate its recruitment to D-loops (Fig. 2). For example, at concentrations corresponding to non-stressed conditions, PriA which facilitates replisome assembly at D-loops mostly blocked pol IV RDR activity. Yet at concentrations corresponding to stressed-induced conditions, pol IV outcompeted PriA at D-loops. We further found that pol IV outcompeted pol I (Supplementary Fig. 6), pol II (Fig. 5) and the pol III replisome (Supplementary Fig. 3) at polymerase concentrations relevant to stressed-induced cells. Although previous genetics have implied competition between pol IV and these other pols20, our data suggest that pol IV is preferentially recruited to D-loops under stress-induced conditions which likely facilitates error-prone recombination.

Previous in vivo studies have shown that the exonuclease domain of pol II reduces (~83%) stress-induced mutations by an unknown mechanism35. In an effort to elucidate this regulatory function of pol II, we examined its activity and ability to compete with pol IV during RDR. Remarkably, we found that the exonuclease activity of pol II is dramatically stimulated during D-loop extension even at saturating dNTP concentrations (~50 µM)(Fig. 3). This D-loop dependent stimulation causes pol II to switch to an active exonuclease mode during RDR which is promoted by topological stress (i.e. positive supercoils) and reduced dNTP pools observed during stationary-phase (Fig. 3). Although gyrase reduces positive supercoils in DNA, we found that the topoisomerase only delays the exonuclease mode of pol II in our assay (Supplementary Fig. 5). This suggests that the exonuclease activity of pol II may be similarly stimulated at D-loops in vivo.

We also surprisingly found that the exonuclease activity of pol II enables it to compete with pol IV at D-loops (Fig. 4). Although pol II is capable of promoting a small fraction of stress-induced mutations20,34, it performs high-fidelity DNA synthesis compared to pol IV. Thus, the ability of exonuclease proficient pol II to compete with pol IV and presumably correct pol IV errors provides an explanation of why the exonuclease domain of pol II suppresses stress-induced mutations in vivo.

Lastly, cellular studies have demonstrated that both the SOS and RpoS stress responses are necessary for pol IV involvement in error-prone recombination17. Upregulation of pol IV (to ~2,500 molecules per cell) by the SOS response is thought to be the sole role of this stress response in error-prone recombination31. Consistent with this, we have demonstrated that levels of pol IV corresponding to SOS-induced cells facilitate its recruitment to D-loops (Fig. 2). How the RpoS stress response which is activated during growth-limiting conditions contributes to pol IV RDR activity, however, has remained unclear. We showed that an additional 100 percent increase in pol IV levels which is observed during the RpoS stress response enables pol IV to outcompete pol II which results in a net reduction in D-loop resection by pol II (Fig. 5)17. Thus, our findings indicate that upregulation of pol IV by the RpoS response contributes to its role in error-prone RDR. The RpoS stress response may, however, elicit other factors or processes that facilitate pol IV involvement in RDR. For example, the possibility exists that pol IV may undergo post-translational modifications that promote its RDR activity.

In summary, this report provides unexpected insight into the mechanism and regulation of error-prone recombination during stress and supports the model illustrated in figure 6. Our data suggest that pol IV plays a dominant role in RDR during stress which is due to its upregulation by the SOS and RpoS responses. Preferential D-loop extension by pol IV during stress likely facilitates error-prone RDR which allows E. coli to rapidly evolve and overcome adverse environmental conditions such as exposure to antibiotics16,19. We propose that pol IV RDR errors are suppressed by pol II through its active exonuclease domain which enables it to compete with pol IV and delete pol IV errors by partially resecting the extended D-loop. However, since pol IV is highly abundant in stressed cells, it likely regains access to the D-loop by displacing pol II from DNA. This model supports a dynamic interplay among translesion pols at D-loops during stress and explains how the exonuclease domain of pol II reduces stress-induced mutations in vivo35.

Fig. 6
Model of translesion DNA polymerase activity at D-loops during stress

Online Methods

Recombination-directed replication

16 µM (in nucleotides) 32P-5’ labeled ssDNA was incubated with 5.2 µM RecA, 0.5 mM ATP, 40 mM phosphocreatine, 1 µg creatine phosphokinase, and 200 µM dNTPs in a total volume of 5 µl of buffer A (25 mM TrisHCl pH 7.5, 0.1 mg/ml BSA, 1 mM DTT, 15 mM MgCl) for 5 min. The reaction was then mixed with 5 µl of buffer A containing 740 µM (in nucleotides) supercoiled pRP27, 0.5 mM ATP, 40 mM phosphocreatine and 1 µg creatine phosphokinase for a further 1.5 min. Next, the reaction was mixed with 10 µl of buffer A containing 0.5 mM ATP, 0.2 µM γ-complex, 2.6 µM SSB and 1 µM β2 for 1 min. Pol was then added for the times indicated. In figures 2, ,3c,3c, and supplementary figures 2, 4 and 5, pols were added along with γ-complex, SSB, β and other factors (PriA, RuvAB, Gyrase). γ-complex was replaced with τ-complex in figures 2a, ,3h,3h, and supplementary figure 2. Reactions were terminated by the addition of 20 mM EDTA, 2 mg/ml proteinase K and 0.6% SDS and incubated for a further 15–30 min. Reaction products were resolved in non-denaturing agarose gels except for those in figures 1h–j which were purified twice through microspin S-400 HR columns (GE Healthcare) then resolved in denaturing urea polyacrylamide gels. Reaction products were analyzed by phosphorimager. Reactions in figure 1e included 3.3 µM final concentration of RecA and the indicated concentrations of trans ssDNA (RP158). Reactions were performed at 37 °C. Concentrations of pols used were as follows except where indicated in figures: pol III (50 nM); pol I (1 µM); pol II (875 nM); pol IV (SOS concentrations, 6.25 µM; non-SOS concentrations, 625 nM). RP192 ssDNA was used in all experiments except for figure 1h (RP219A), and figures 1i,j (RP235A). ssDNA was 5’-end labeled with 32P-γ-ATP using T4 polynucleotide kinase (New England Biolabs).

Primer-template extension assays. (Fig. 1f)

Primer extension by pol V with RecA and trans ssDNA was performed as described23, however, a different primer-template (RP158/RP25) was used and the concentration of pol V was 500 nM. Products were resolved in a denaturing urea polyacrylamide gel and visualized by phosphorimager. The primer-template was assembled by mixing equimolar concentrations of RP25 and RP158 followed by heating to 90–100 °C then slowly cooling to room temp. (Fig. 1h) Primer extension was performed by incubating 500 nM β, 100 nM γ-coplex, and 1.3 µM SSB with 40 nM primer-template (RP312/313) and 50 µM of the indicated dNTP in buffer A at 37° C for 1 min followed by the addition of 500 nM pol IV for an additional 5 min. Assembly of the primer-template and resolution of the reaction products were performed as above. Primers were 5’-end labeled with 32P-γ-ATP using T4 polynucleotide kinase (New England Biolabs).

Pol II replication of double-strand DNA (Figure 3b, left)

10 nM linear double-strand DNA containing a replication fork at the 5’ end was mixed with 100 nM γ-complex, 500 nM β, 50 µM dNTPs, 2 mM ATP, 3 µCi 32P-α-dATP and 1.3 µM SSB for 1 min at 37° C in buffer A. 875 nM wild-type or exonuclease deficient pol II was then added and aliquots of the reaction were terminated at the indicated times by the addition of 20 mM EDTA and 0.5% SDS. DNA products were resolved in alkaline agarose gels and analyzed by phosphorimager.

Pol II replication of a primer-template (Figure 3b, right)

1.2 nM primed M13mp18 ssDNA was mixed with 4 nM γ-complex, 15 nM β, 50 µM dNTPs, 2 mM ATP, 3 µCi 32P-α-dATP and 1.68 µM SSB for 2 min at 37° C in buffer A. 4 nM wild-type or exonuclease deficient pol II was then added and aliquots of the reaction were terminated at the indicated times by the addition of 20 mM EDTA and 0.5% SDS. DNA products were resolved in alkaline agarose gels and analyzed by phosphorimager.

Replisome replication assay (Fig. 3j)

Reactions containing 100 fmol 5’-biotinylated 100mer rolling circle DNA, 60 µM dCTP and dGTP, 50 µM γ-S-ATP, 4 pmol DnaB6, 2.5 pmol β2, 0.5 pmol pol III* (pol III WT or pol III exonuclease minus containing an ε mutant (D12A and E14A) mixed with τ-complex) of a total volume of 25 µl of buffer B (20 mM TrisHCl (pH 7.5), 0.5 mM EDTA, 5 mM DTT, 10% glycerol, 8 mM MgCl) were incubated for 5 min at 37°C then immobilization to streptavidin beads (Invitrogen). Beads were washed three times in 500 µl buffer B containing 12 pmol β2 and 50 µM γ-S-ATP, 60 µM dCTP, 60 µM dGTP. Replication was initiated by adding 2.5 pmol β2, 0.5 mM ATP, 480 nM SSB4, 60 µM dATP, 10 µM dTTP and 3 µCi α32PdTTP in a total volume of 25 µl of buffer B. Reactions were terminated at the indicated times by addition of 25 µl of stop buffer (40 mM HEPES (pH 7.5), 40 mM EDTA, 6% SDS). DNA was released from beads by incubating at 95 °C for 3 min then resolved in 0.8% denaturing alkaline agarose gels.

Replisome recombination-directed replication (Supplementary Fig. 3)

20 µM (in nucleotides) RP192 ssDNA was incubated with 6.5 µM RecA, 0.5 mM ATP, 40 mM phosphocreatine, 1 µg creatine phosphokinase, and 250 µM dNTPs in a total volume of 5 µl of buffer B (25 mM TrisHCl pH 7.5, 0.1 mg/ml BSA, 10 mM DTT, 15 mM MgCl) for 5 min. The reaction was then mixed with 5 µl of buffer B containing 925 µM (in nucleotides) supercoiled pRP27, 0.5 mM ATP, 40 mM phosphocreatine, 1 µg creatine phosphokinase, 1 µCi 32P-α-dATP and 7 µM SSB for a further 1.5 min. Next, the reaction was mixed with 15 µl of buffer B containing 0.5 mM ATP, 85 nM Pol III*, 0.85 µM β, 17 or 34 nM of PriA/B/C, 612 nM DnaB, 714 nM DnaC, 680 nM DnaG, 306 nM DnaT, 25.5 nM gyrase, 340 µM NTPs and the indicated final concentrations of pol IV for 40 min. Reactions were terminated by the addition of 20 mM EDTA, 2 mg/ml proteinase K and 0.6% SDS and incubated for a further 15–30 min. Reaction products were resolved in non-denaturing agarose gels (and in an alkaline agarose gel where indicated) and analyzed by phosphorimager. Reactions were performed at 37° C.

Pol III replication of a primer-template (Fig. 3i)

Reactions contained 37.5 fmol primed M13mp18 ssDNA, 60 µM dCTP and dGTP, 0.5 mM ATP, 0.9 µg SSB, 350 fmol β2 and 100 fmol Pol III* or Pol III* containing an ε mutant (D12A and E14A) that eliminates the 3'–5' exonuclease activity, in 22 µl replication buffer (20 mM Tris-HC1 (pH 7.5), 4% glycerol, 0.1 mM EDTA, 40 µg/ml BSA, 5 mM DTT, 10 mM MgOAc2). Reactions were incubated at 37° C for 5 min, before adding 60 µM dATP, 20 µM dTTP and 1 µCi α32P-dTTP were added to initiate replication. Total reaction volume was 25 µl. DNA synthesis was quenched after the indicated time points by adding 25 µl of 40 mM EDTA and 1% SDS. Quenched reactions were resolved in an alkaline agarose gel and analyzed by phosphorimager.

Exonuclease activity assay (Fig. 3k)

10 nM 32P-labeled primer-template (25/10 or 25G/10 (contains mismatch)) was mixed with 8 mM MgCl, 200 nM β, 60 nM τ-complex, 1 mM ATP and 1 µM SSB in buffer C (20 mM Tris-HCl pH 7.5, 0.5 mM EDTA, 5 mM DTT, 10% glycerol) for 2 min. 30 nM pol II or pol III was added as indicated and aliquots of the reactions were terminated at the indicated time intervals by the addition of 25 mM EDTA and 45% formamide. Reactions were performed at 37° C. Products were resolved in a urea polyacrylamide gel and analyzed by phosphorimager.


Pol III (WT and Exo-), pol IV, pol V, β, and reconstituted γ-complex were purified as described23,45,46. τ-complex was reconstituted from pure proteins and purified in an identical fashion as γ-complex, however, γ was replaced with τ His-tagged versions of PriA/B/C and DnaT were purified by standard Ni2+ affinity chromatography methods. DnaG/B/C were purified as described 46,47. Wild-type and exonuclease deficient pol II were purified as described37,48. RuvA/B were purified as described49. RecA, Pol I (Klenow fragment and Klenow fragment 3’–5’ exonuclease minus) and Gyrase were purchased from New England Biolabs.



Supplementary Material


We thank M. Skangalis, D. Zhang and O. Yurieva for technical support. This work was supported by the National Institutes of Health (grants to M.O.D. (GM38839), R.T.P. (K99CA160648) and M.F.G. (GM21422, ES012259)) and the Howard Hughes Medical Institute.


Author Contributions:

R.T.P. conceived the idea for the study, wrote the manuscript, and performed and interpreted all experiments with the exception of those in Fig. 3i,j which were performed by I.K. M.O.D. and M.F.G. provided editorial input.


1. Cox MM. Recombinational DNA repair of damaged replication forks in Escherichia coli: questions. Annu Rev Genet. 2001;35:53–82. [PubMed]
2. Cox MM, et al. The importance of repairing stalled replication forks. Nature. 2000;404:37–41. [PubMed]
3. Mazon G, Mimitou EP, Symington LS. SnapShot: Homologous recombination in DNA double-strand break repair. Cell. 2010;142:646. 646 e1. [PubMed]
4. Li X, Heyer WD. Homologous recombination in DNA repair and DNA damage tolerance. Cell Res. 2008;18:99–113. [PMC free article] [PubMed]
5. Heller RC, Marians KJ. Replisome assembly and the direct restart of stalled replication forks. Nat Rev Mol Cell Biol. 2006;7:932–943. [PubMed]
6. Zahradka K, et al. Reassembly of shattered chromosomes in Deinococcus radiodurans. Nature. 2006;443:569–573. [PubMed]
7. Motamedi MR, Szigety SK, Rosenberg SM. Double-strand-break repair recombination in Escherichia coli: physical evidence for a DNA replication mechanism in vivo. Genes Dev. 1999;13:2889–2903. [PubMed]
8. Lydeard JR, et al. Break-induced replication requires all essential DNA replication factors except those specific for pre-RC assembly. Genes Dev. 2010;24:1133–1144. [PubMed]
9. Kawamoto T, et al. Dual roles for DNA polymerase eta in homologous DNA recombination and translesion DNA synthesis. Mol Cell. 2005;20:793–799. [PubMed]
10. Kane DP, Shusterman M, Rong Y, McVey M. Competition between replicative and translesion polymerases during homologous recombination repair in Drosophila. PLoS Genet. 2012;8:e1002659. [PMC free article] [PubMed]
11. Li X, Stith CM, Burgers PM, Heyer WD. PCNA is required for initiation of recombination-associated DNA synthesis by DNA polymerase delta. Mol Cell. 2009;36:704–713. [PMC free article] [PubMed]
12. Hashimoto Y, Puddu F, Costanzo V. RAD51- and MRE11-dependent reassembly of uncoupled CMG helicase complex at collapsed replication forks. Nat Struct Mol Biol. 2012;19:17–24. [PMC free article] [PubMed]
13. Moldovan GL, et al. DNA polymerase POLN participates in cross-link repair and homologous recombination. Mol Cell Biol. 2010;30:1088–1096. [PMC free article] [PubMed]
14. McIlwraith MJ, et al. Human DNA polymerase eta promotes DNA synthesis from strand invasion intermediates of homologous recombination. Mol Cell. 2005;20:783–792. [PubMed]
15. Kohzaki M, et al. DNA polymerases nu and theta are required for efficient immunoglobulin V gene diversification in chicken. J Cell Biol. 2010;189:1117–1127. [PMC free article] [PubMed]
16. Rosenberg SM. Evolving responsively: adaptive mutation. Nat Rev Genet. 2001;2:504–515. [PubMed]
17. Ponder RG, Fonville NC, Rosenberg SM. A switch from high-fidelity to error-prone DNA double-strand break repair underlies stress-induced mutation. Mol Cell. 2005;19:791–804. [PubMed]
18. McKenzie GJ, Lee PL, Lombardo MJ, Hastings PJ, Rosenberg SM. SOS mutator DNA polymerase IV functions in adaptive mutation and not adaptive amplification. Mol Cell. 2001;7:571–579. [PubMed]
19. Foster PL. Stress-induced mutagenesis in bacteria. Crit Rev Biochem Mol Biol. 2007;42:373–397. [PMC free article] [PubMed]
20. Hastings PJ, et al. Competition of Escherichia coli DNA polymerases I II and III with DNA Pol IV in stressed cells. PLoS One. 2010;5:e10862. [PMC free article] [PubMed]
21. Shee C, Gibson JL, Rosenberg SM. Two mechanisms produce mutation hotspots at DNA breaks in Escherichia coli. Cell Rep. 2012;2:714–721. [PMC free article] [PubMed]
22. Harris RS, Bull HJ, Rosenberg SM. A direct role for DNA polymerase III in adaptive reversion of a frameshift mutation in Escherichia coli. Mutat Res. 1997;375:19–24. [PubMed]
23. Schlacher K, Cox MM, Woodgate R, Goodman MF. RecA acts in trans to allow replication of damaged DNA by DNA polymerase V. Nature. 2006;442:883–887. [PubMed]
24. Petrosino JF, Galhardo RS, Morales LD, Rosenberg SM. Stress-induced beta-lactam antibiotic resistance mutation and sequences of stationary-phase mutations in the Escherichia coli chromosome. J Bacteriol. 2009;191:5881–5889. [PMC free article] [PubMed]
25. Kim SR, Matsui K, Yamada M, Gruz P, Nohmi T. Roles of chromosomal and episomal dinB genes encoding DNA pol IV in targeted and untargeted mutagenesis in Escherichia coli. Mol Genet Genomics. 2001;266:207–215. [PubMed]
26. Tippin B, Pham P, Goodman MF. Error-prone replication for better or worse. Trends Microbiol. 2004;12:288–295. [PubMed]
27. Akerlund T, Nordstrom K, Bernander R. Analysis of cell size and DNA content in exponentially growing and stationary-phase batch cultures of Escherichia coli. J Bacteriol. 1995;177:6791–6797. [PMC free article] [PubMed]
28. Withers HL, Bernander R. Characterization of dnaC2 and dnaC28 mutants by flow cytometry. J Bacteriol. 1998;180:1624–1631. [PMC free article] [PubMed]
29. Godoy VG, et al. UmuD and RecA directly modulate the mutagenic potential of the Y family DNA polymerase DinB. Mol Cell. 2007;28:1058–1070. [PMC free article] [PubMed]
30. Kobayashi S, Valentine MR, Pham P, O'Donnell M, Goodman MF. Fidelity of Escherichia coli DNA polymerase IV. Preferential generation of small deletion mutations by dNTP-stabilized misalignment. J Biol Chem. 2002;277:34198–34207. [PubMed]
31. Galhardo RS, et al. DinB upregulation is the sole role of the SOS response in stress-induced mutagenesis in Escherichia coli. Genetics. 2009;182:55–68. [PubMed]
32. Xu L, Marians KJ. PriA mediates DNA replication pathway choice at recombination intermediates. Mol Cell. 2003;11:817–826. [PubMed]
33. Shlomai J, Kornberg A. A prepriming DNA replication enzyme of Escherichia coli. I. Purification of protein n': a sequence-specific, DNA-dependent ATPase. J Biol Chem. 1980;255:6789–6793. [PubMed]
34. Frisch RL, et al. Separate DNA Pol II- and Pol IV-dependent pathways of stress-induced mutation during double-strand-break repair in Escherichia coli are controlled by RpoS. J Bacteriol. 2010;192:4694–4700. [PMC free article] [PubMed]
35. Foster PL, Gudmundsson G, Trimarchi JM, Cai H, Goodman MF. Proofreading-defective DNA polymerase II increases adaptive mutation in Escherichia coli. Proc Natl Acad Sci U S A. 1995;92:7951–7955. [PubMed]
36. Bonner CA, et al. Purification and characterization of an inducible Escherichia coli DNA polymerase capable of insertion and bypass at abasic lesions in DNA. J Biol Chem. 1988;263:18946–18952. [PubMed]
37. Cai H, Yu H, McEntee K, Kunkel TA, Goodman MF. Purification and properties of wild-type and exonuclease-deficient DNA polymerase II from Escherichia coli. J Biol Chem. 1995;270:15327–15335. [PubMed]
38. Goodman MF, Petruska J, Boosalis MS, Bonner C, Randall SK, Sowers LC, Mendelman L. DNA Replication and Mutagenesis. Washington, D.C.: American Society for Microbiology; 1988.
39. Buckstein MH, He J, Rubin H. Characterization of nucleotide pools as a function of physiological state in Escherichia coli. J Bacteriol. 2008;190:718–726. [PMC free article] [PubMed]
40. Ibarra B, et al. Proofreading dynamics of a processive DNA polymerase. EMBO J. 2009;28:2794–2802. [PubMed]
41. Downey CD, McHenry CS. Chaperoning of a replicative polymerase onto a newly assembled DNA-bound sliding clamp by the clamp loader. Mol Cell. 2010;37:481–491. [PMC free article] [PubMed]
42. Baker T, Kornberg A. DNA Replication. 2nd Edition. U.S.: Uinversity Science Books; 1992.
43. Lovett ST. Replication arrest-stimulated recombination: Dependence on the RecA paralog, RadA/Sms and translesion polymerase, DinB. DNA Repair (Amst) 2006;5:1421–1427. [PubMed]
44. Register JC, 3rd, Griffith J. Direct visualization of RecA protein binding to and unwinding duplex DNA following the D-loop cycle. J Biol Chem. 1988;263:11029–11032. [PubMed]
45. Indiani C, McInerney P, Georgescu R, Goodman MF, O'Donnell M. A sliding-clamp toolbelt binds high- and low-fidelity DNA polymerases simultaneously. Mol Cell. 2005;19:805–815. [PubMed]
46. Georgescu RE, et al. Mechanism of polymerase collision release from sliding clamps on the lagging strand. Embo J. 2009;28:2981–2991. [PubMed]
47. Davey MJ, Fang L, McInerney P, Georgescu RE, O'Donnell M. The DnaC helicase loader is a dual ATP/ADP switch protein. EMBO J. 2002;21:3148–3159. [PubMed]
48. Indiani C, Langston LD, Yurieva O, Goodman MF, O'Donnell M. Translesion DNA polymerases remodel the replisome and alter the speed of the replicative helicase. Proc Natl Acad Sci U S A. 2009;106:6031–6038. [PubMed]
49. Kaplan DL, O'Donnell M. RuvA is a sliding collar that protects Holliday junctions from unwinding while promoting branch migration. J Mol Biol. 2006;355:473–490. [PubMed]
50. McInerney P, O'Donnell M. Functional uncoupling of twin polymerases: mechanism of polymerase dissociation from a lagging-strand block. J Biol Chem. 2004;279:21543–21551. [PubMed]
51. Pomerantz RT, O'Donnell M. The replisome uses mRNA as a primer after colliding with RNA polymerase. Nature. 2008;456:762–766. [PMC free article] [PubMed]