Full text loading...
Chapter 16 : Recombinational Repair, Replication Fork Repair, and DNA Damage Tolerance
Category: Microbial Genetics and Molecular Biology
Ebook: Choose a downloadable PDF or ePub file. Chapter is a downloadable PDF file. File must be downloaded within 48 hours of purchase
This chapter briefly reviews the functions and properties of key Escherichia coli proteins involved in homologous recombination, with a particular emphasis on RecA, a versatile protein with additional roles in the induction of the SOS response and in translesion DNA synthesis and mutagenesis. Parallels can be drawn between these bacterial recombination proteins and the eukaryotic recombination proteins. Evidence supporting the existence of recombinational repair of DNA in bacteria is discussed, followed by the developing views of how recombination proteins can help cells deal with replication forks whose progress has been blocked by endogenous or exogenous DNA damage. Although many discussions of recombinational repair, replication fork repair, and other mechanisms of DNA damage tolerance consider only the formal set of simple DNA structures, it is important to keep in mind the numerous observations indicating that the DNA structures present in vivo after DNA damage may be considerably more complex. It is important to recognize that the various replication fork recovery strategies involving homologous recombination functions are not necessarily independent of translesion DNA polymerases and that such translesion DNA polymerases may participate in recombinogenic replication fork recovery strategies at times. Much remains to be learned about the relationship between homologous recombination functions, replication fork repair/recovery, and translesion synthesis.
Key Concept Ranking
- Nuclear Magnetic Resonance Spectroscopy
DNA damage can interfere with the progress of replication forks and lead to the generation of various structures. (A) DSB can be created directly by ionizing radiation and certain other agents. (B and C) A DSB can be generated by a replication fork encountering a nick in the leading-strand template (B) or by a replication fork encountering a nick in the lagging-strand template (C) (note that the DSB could have a 3’ single-strand overhang). (D and E) A single-strand gap can be generated by a replication fork encountering a lesion in the leading-strand (D) or lagging-strand (E) template. (F and G) Replication forks can regress on encountering a lesion blocking both strands (F) or just one strand (G) to form a particular fork of Holliday junction commonly referred to as a chicken foot structure.
Model of the E. coli replication fork. In E. coli the replicative helicase DnaB acts to disrupt parental DNA and expose the two individual strands. One of these strands acts as the template for synthesis of the leading strand by a DNA Pol III complex, which proceeds uninterrupted in the 5’ → 3’ direction (arrows represent the 3’ ends of DNA strands). The anchoring of the leading-strand polymerase on this template strand by the β sliding clamp allows synthesis of the leading strand to continue for many thousands of bases. By contrast, because DNA synthesis occurs in the 5 ‘ — 3 ‘ direction, synthesis of the second DNA strand proceeds in segments using RNA primers made by the primase DnaG. This allows DNA synthesis to repeatedly initiate on the lagging-strand template. Thus, the lagging-strand-DNA Pol III complex continually associates and dissociates with the lagging-strand template to extend each RNA primer and form so-called Okazaki fragments. These Okazaki fragments are 1,000 to 2,000 bp in bacteria but only 40 to 300 bp in eukaryotes. β sliding clamps that are associated with the lagging-strand polymerase are also reloaded continually onto the lagging-strand template by the γ-complex clamp loader. A single continuous strand is formed from these discontinuous lagging strands by degradation of the RNA primers and the subsequent filling in of the resultant gaps by DNA Pol I followed by ligation of the 5’ end of one Okazaki fragment with the 3’ end of the adjacent fragment. (Adapted from reference 191 .)
Model of Okazaki fragment synthesis by a replication fork encountering an AP lesion on the leading-strand template. Thin grey lines indicate the template strand DNA. Dark gold and light gold lines with arrowheads indicate leading- and lagging-strand DNA synthesis, respectively. (Adapted from reference 100 .)
Effect of UV irradiation on the rate of DNA synthesis (“induced replisome reactivation/replication restart”). An exponentially growing culture of wild-type E. coli strain AB1157 was UV irradiated (10 J/m2); 0.5 ml of culture was removed at intervals and pulse-labeled with [3H]thymidine for 2 min, and trichloroacetic acid-precipitable counts were determined. Gold line, [3H]thymidine cpm; black line, optical density at 450 nm (A450). (Adapted from reference 129 .)
Alkaline sedimentation profiles of labeled DNA showing postreplication repair and excision repair occurring in wild-type E. coli following UV irradiation. Gold lines indicate DNA uniformly labeled with [14C]thymidine; black lines indicate DNA pulse-labeled with [3H]thymidine. (A) A thymidine-requiring repair-proficient strain of E. coli K-12 was uniformly labeled with [14C]thymidine, washed, pulse-labeled with [3H]thymidine for 10 min, and then lysed for sedimentation in an alkaline sucrose gradient. (B) Same as panel A except that the 14C-labeled bacteria were exposed to 254-nm UV light at 6 J/m2 and then pulse-labeled with [3H]thymidine for 10 min. (C) Same as panel B except that the pulse-labeled cells were washed and incubated in unlabeled medium for 50 min. (D) Bacteria were uniformly labeled with [14C]thymidine, washed, exposed to UV light at 6 J/m2, incubated for 60 min, and then pulse-labeled with [3H]thymidine for 10 min. (Adapted from reference 106 .)
Comparison of structural domains in the E. coli RecA protein with the T4 UvsX, yeast (S. cerevisiae) Rad51 and Dmc1, and archaeal (S. solfataricus) RadA proteins. Core domain conservation is depicted in light gold. N-terminal domain conservation between Rad51, Dmc1, and RadA is shown in dark gold. Regions with no sequence homology include the light grey regions of the N-terminal domains and all regions C-terminal to the core, including the medium gold C-terminal domains of the E. coli RecA and T4 UvsX proteins. (Adapted from reference 174 .)
Examples of recombination reactions promoted by RecA protein. (A) Renaturation of complementary single strands; (B) asymmetric (nonreciprocal) strand exchange following the pairing of single-stranded and double-stranded DNA; (C) symmetric (reciprocal) strand exchange following the pairing of duplex DNA and partially single-stranded DNA. (Adapted from reference 236 .)
Electron micrographs of frozen-hydrated (A and B) and negatively stained (C) RecA-DNA-ATP-γS filaments. The RecA filaments in panels A and C are on double-stranded DNA, while the RecA in panel B is on a circular single-stranded DNA molecule. The white arrows in panel A show the general direction of fluid flow that occurred on the grid during blotting, prior to rapid freezing, as judged by the preferred orientation of filaments in this direction on the grid. The pitch of the filament sections between the white arrows are 114 and 108 Å (11.4 and 10.8 nm), while the pitch of the filament section that runs perpendicular to the flow direction (black arrow) is 96 A (9.6 nm). All panels are at the same magnification, and the tobacco mosaic virus particles in panel C are about 200 A (20 nm) in diameter. (Reproduced from reference 370 with permission.)
Model for the RecFOR-facilitated formation of a RecA nucleoprotein filament on gapped DNA. (A) The RecQ helicase and RecJ exonuclease might process the gapped DNA prior to recognition by the RecFOR proteins. (B and C) The SSB-coated gap (B) is first recognized by the RecFR complex (or RecF) (C). (D) The RecOR complex (or RecO) interacts with the RecFR-gapped DNA complex. (E) The RecFOR proteins serve to nucleate RecA protein filament assembly, which then extends over the single-strand region by growth in the 5’ → 3’ direction, displacing SSB. (Adapted from reference 204 .)
E. coli RecA filament. Two views of the RecA protein filament (three subunits are shown) crystallized in the absence of DNA ( 317 ). ADP binds to the central domain of RecA near the subunit interface. An N-terminal α-helix of each subunit packs against the neighboring subunit, and the C-terminal domain protrudes from the inner radius of the helical filament.
Model for homologous pairing within a RecA nucleoprotein filament. (Courtesy of S. C. West.)
General model for homologous recombination that depicts the DSB repair model suggested by Szostak et al. ( 320 ). The text in parentheses designates proteins from E. coli that function at the designated steps. Light grey lines indicate newly synthesized DNA. (Adapted from reference 139 .)
Early steps of homologous recombination in E. coli are coordinated by RecA protein, RecBCD enzyme, and χ. RecBCD enzyme binds to the end of a DSB. It unwinds the double-stranded DNA while preferentially degrading the strand that was 3’ terminal at the entry point. Interaction with a DNA sequence known as χ results in attenuation of the 3’ → 5’ nuclease activity, activation of a weaker 5’ → 3’ nuclease activity, and facilitated loading of RecA protein onto the χ-containing single-stranded DNA that was produced by continued DNA unwinding beyond x. The resulting RecA protein-single-stranded DNA filament invades homologous double-stranded DNA (dsDNA) to produce a D-loop structure. (Adapted from reference 139 .)
Resolving a chicken foot Holliday junction generates a DSB. (Adapted from reference 192 .)
RuvA and RuvB function as a branch migration motor that remodels Holliday junctions. (A) A tetramer of RuvA (two gold subunits and two white subunits) binds to one face of a four-way DNA junction (Holliday junction; colored black) in a square planar conformation ( 252 ). A second RuvA tetramer (not shown) binds to the opposite face of the DNA junction in a clam shell-like arrangement. (B) Side view of the RuvA-DNA complex shown in panel A. (C) Domain 3 of RuvA (disordered in the crystal structure shown in panel A) interacts with the RuvB helicase. RuvB has an AAA+ ATPase protein fold ( 365 ). ATP (gold) binds in a cleft between the N-and C-terminal domains of RuvB. (D) Model showing the locations of the RuvB hexamer and the RuvA octamer during branch migration of a four-way DNA junction. RuvA manages the exchange of DNA strands at the center of the junction. RuvB is the motor that translocates double-stranded DNA from two branches of the four-way junction to cause branch migration. Domain 3 (small sphere) of RuvA mediates the interaction with RuvB (shown in panel C).
RecG is a replication fork-remodeling enzyme. The crystal structure of Thermotoga maritima RecG bound to a fork-shaped DNA suggests a mechanism for reversal of a stalled replication fork ( 297 ). Domains 2 and 3 of RecG are typical of a superfamily 1 helicase, with ATP (black) binding in the interdomain cleft. Domain 1 manages the DNA substrate, interacting with two double-stranded DNA branches and inserting a “wedge” at the junction to promote strand separation during translocation of the DNA. In the proposed mechanism, the dsDNA marked “leading” would be pulled toward helicase domains 2 and 3 to cause remodeling of a stalled replication fork.
Model for daughter strand gap repair. The thick black lines represent irradiated parental DNA containing a cyclobutane dimer. The thin gold lines represent daughter DNA synthesized after irradiation. The Holliday junction can be resolved in one of two ways. The cleavages marked “a” result in an exchange event involving the parental strand without a dimer ( 265 ); the dimer stays in the parental strand. The cleavages marked “b” lead to an exchange event involving the parental strand with the dimer; the dimer is exchanged into the daughter strand (Adapted from references 82 , 330 , and 342 .)
E. coli cells that are proficient in both excision repair (uvrA+) and discontinuous DNA synthesis with gap filling (recA+) are resistant to killing by UV radiation. Cells that are defective in both of these processes (uvrA recA) are extremely UV sensitive. The survival curve of the uvrA recA double mutant is reproduced on the left with an expanded UV dose scale. From this curve, it can be estimated that one pyrimidine dimer per E. coli genome equivalent is lethal to a uvrA recA strain. (Adapted from reference 108 .)
DNA synthesized immediately after UV irradiation of E. coli has a lower molecular weight than normal. E. coli uvr cells were exposed to UV radiation and pulse-labeled briefly with [3H]thymidine. The cells were analyzed by sedimentation in alkaline sucrose gradients either immediately after being labeled (no incubation; dotted gold line) or following incubation for 70 min (solid gold line). Immediately following the pulse-label, the newly synthesized DNA is of low molecular weight and sediments near the top of the gradient. However, over time the newly synthesized DNA approaches the size of the unirradiated control (no UV; black line). (Adapted from reference 264 .)
(A) Relative yield of Lac+ colonies after an excision-defective Flac+ strain was mated to mutant lac recipients. When the recipient is defective in the recA gene, the yield of Lac+ colonies is significantly reduced relative to that observed in a wild-type recipient. However, there is no detectable difference in the yield of Lac+ colonies in recipients that are either excision repair proficient (recA uvr + ) or excision repair deficient (recA uvrA). (B) FLac+ DNA transferred from a UV-irradiated donor to an unirradiated recipient shows enhanced survival if, following transfer, the cells are exposed to photoreactivating light. This result is consistent with (but does not prove) the presence of replicative gaps opposite pyrimidine dimers. (Adapted from reference 107 .)
(A) Design and interpretation of an experiment ( 265 ) to detect exchanges between DNA strands during postreplicative gap filling. E. coli uvrA cells were grown for several generations in medium containing the heavy isotopes [13C]thymine and [15N]thymine as well as [14C]thymine. The cells were then exposed to various doses of UV light and grown for less than one generation in a medium without density markers (light medium) containing [3H]thymidine. The newly replicated DNA is of hybrid density and has the 14C label uniquely in the heavy strand and the 3H label uniquely in the light strand. (B) If the cells are not irradiated and their DNA strands are separated by heat denaturation and equilibrium CsCl density gradient centrifugation, the strands separate cleanly as 14C-labeled heavy strands and 3H-labeled light strands. (C) If the cells are irradiated, strands of intermediate density are observed after heat denaturation and equilibrium CsCl density gradient centrifugation. This observation has been interpreted as indicating that exchanges have occurred between sister duplexes in the UV-irradiated cells so that light (3H-labeled) DNA becomes covalently attached to heavy (14C-labeled) DNA. This DNA of intermediate density could be resolved into heavy and light components after shearing to a molecular weight of less than 5 × 105, suggesting that the exchanges involved segments of at least this size.
Model assigning possible proteins to steps in daughter strand gap repair. See the text for details.
Both discontinuous replication and stalled replication generate small nascent DNA molecules that subsequently are converted into high-molecular-weight ones. Thus, it can be very difficult, if not impossible, to differentiate these processes by using the sedimentation velocity of radiolabeled newly synthesized DNA. The thicker lines in the lower left diagram indicate regions of gap filling by recombinational exchange and by nonsemiconservative DNA synthesis (see Fig. 16–17 ).
Schematic depiction of possible relationships of DNA lesions to replication forks in E. coli. Replication forks A1 and A2 were established during the initiation of an initial round of DNA replication. Replication forks B1, B2, B1 ‘, and B2’ were established before the completion of this initial round of DNA replication. Lesions 1 and 2 differ with respect to their positions relative to these replication forks. See the text for details. (Adapted from reference 303 .)
Model showing how replication of DNA by a factory combined with subcellular localization of the DNA after synthesis could limit the time during which the two daughter strands could undergo recombinational repair. After DNA synthesis, the DNA becomes localized to the incipient daughter cells, so that the two daughter DNAs would not be able to undergo recombinational interactions although the two daughter strands are still with the same cell. The heavy arrow indicates a possible spatial/temporal window during which the two DNA molecules might be able to undergo such recombinational interactions.
Outline of an in vivo assay system for gap filling by recombinational repair. The gap-lesion plasmid GP21 (KanR) was introduced into E. coli cells along with a homologous partner plasmid (FGP20/Tamp; AmpR). Parallel experiments were carried out with a heterologous plasmid (pUC18). Only gap-filling repair allows GP21 to transform the cells and confer a KanR phenotype. This can be done by translesion DNA replication, which does not depend on the partner plasmid, or by a process involving recombination functions, which depends on a homologous partner plasmid. The homologous partner plasmid carries a T opposite the site corresponding to the lesion in GP21, so that GP21 filled in by recombination will have a T at that position. In contrast, GP21 filled in by DNA Pol V-dependent translesion synthesis will have primarily an A at that location ( 246 ). (Adapted from reference 14 .)
Fork regression offers an avenue for damage tolerance (and replication fork repair) that does not involve DSB or DNA strand invasion.
Possible scheme for PriA-dependent control of replication fork establishment. A broken arm of the replication fork is processed by RecBCD. RecA (white ovals) is loaded, and a D loop recombination intermediate is formed between the broken arm and the sister chromosome. If PriA does not bind, DNA synthesis from the invading 3’ end would merely extend the D loop without forming a processive replication fork. If PriA (grey hexagon) does bind to the invading 3’ end, DNA synthesis is inhibited while PriA (with PriBC DnaCT) loads DnaB helicase (grey oval) to the displaced strand of the D loop. DnaB recruits DnaG primase (gold circle), and this relieves PriA inhibition to allow both lagging- and leading-strand DNA synthesis to begin via DNA Pol III holoenzyme (gold triangles). In contrast, a broken chromosome may engage both broken arms into the D loop, thereby potentially blocking DnaB loading; repair may be accomplished with a small amount of repair synthesis without replication fork establishment. (Adapted from reference 171 .)
UV-induced DNA replication intermediates observed during the recovery of replication. (A) Diagram of the migration pattern of PvuII-digested pBR322 during two-dimensional analysis. Nonreplicating plasmids run as a linear 4.4-kb fragment. Normal replicating fragments form Y-shaped structures and migrate more slowly due to their larger size and nonlinear shape, forming an arc that extends out from the linear fragment. Double Y- or X-shaped molecules migrate in the cone region. (B) The replication intermediates persist until a time correlating with replication recovery and lesion removal. Replication recovery, lesion repair, and the relative amount of replicating fragments (gold) and cone region intermediates (black) are plotted. Replication recovery was assayed by [3H]thymine incorporation for UV-irradiated (black line) or mock-irradiated (gold line) cultures. (Adapted from reference 37 .)
Three possible mechanisms for the regression of replication forks. (A) Fork regression mediated by RecA. (B) Fork regression mediated by RecG. (C) Fork regression mediated by positive supercoils accumulated ahead of the replication fork. Plus sign represents positive supercoiling of the DNA; dark grey circle represents a lesion.
Formation of the cone region intermediates depends on the presence of UV-induced lesions, RecA, and active replication. Replication recovery was assayed by [3H]thymine incorporation, as in Fig. 16–29 , and the replication intermediates observed during the normal recovery period were monitored by two-dimensional gel analysis. (A) uvrA mutants fail to recover replication after UV-induced DNA damage, and the cone region intermediates persist and accumulate. (UV-irradiated cultures, gold; mock-irradiated cultures, black). (B) recA mutants fail to recover replication after UV-induced DNA damage and the cone region intermediates do not accumulate (UV-irradiated cultures, gold; mock-irradiated cultures, black). (C and D) y-structure (C) and cone region (D) are plotted for UV-irradiated uvrA (gold) and recA (grey) strains. (Adapted from reference 37 .)
Models for restarting DNA replication by junction cleavage. Formation of a Holliday junction from a replication fork stalled at a lesion (grey circle) might allow processing by the RuvABC helicase-endonuclease complex in one of two ways. (A) RuvABC might cleave the Holliday junction at the stalled fork before recombination from the free double-stranded DNA end has occurred, while formation of the Holliday junction might facilitate unmasking and subsequent removal of the block. The released double-stranded DNA end would then be recombined back into the homologous duplex to form a D-loop. Cleavage of the Holliday junction formed at the D-loop by RuvABC and reassembly of the replication machinery at the fork would reconstitute an active replisome. (B) The double-stranded DNA end that is spooled out from the Holliday junction might be recombined with the homologous sequences in the reannealed parental strands to form a D-loop intermediate that is linked to the original Holliday junction at the fork. A second Holliday junction would also be formed at the D-loop. Assuming that the original block could be removed, cleavage of both Holliday junctions by RuvABC would generate a forked DNA structure onto which the replication machinery could be reloaded. The 3’ ends of DNA strands are shown by arrowheads, and cleavage of Holliday junctions by RuvABC is shown by black triangles. (Adapted from reference 191 .)
Recovery from inhibition of DNA synthesis after UV irradiation: synergism in double mutants combining recA718 and umuC36. Black dashed line, recA+ umuC+; black dotted line, recA+ umuC36; gold dotted line, recA718 umuC+; gold dashed line, recA718 umuC36. The UV dose was 3 J/m2. Cells growing exponentially at 37°C were pulse-labeled with [3H]thymidine for 2 min at various times before and after UV irradiation. The optical density (OD) curve defines the range of values obtained for all strains. (Adapted from reference 362 .)
Rates of post-UV DNA synthesis in recA+ lexA+ strains carrying substitutions in polB and/or umuDC. Experiments were performed and plotted as described previously ( 238 ). (Adapted from reference 239 .)