Recruitment of Bacillus subtilis RecN to DNA Double-Strand Breaks in the Absence of DNA End Processing
http://www.100md.com
细菌学杂志 2006年第1期
Department Microbial Biotechnology, Centro Nacional de Biotecnologia, CSIC, C/Darwin 3, Campus Universidad Autonoma de Madrid, 28049 Madrid, Spain,Biochemie, Fachbereich Chemie, Hans-Meerwein-Strae, Philipps-Universitt Marburg, 35032 Marburg, Germany
ABSTRACT
The recognition and processing of double-strand breaks (DSBs) to a 3' single-stranded DNA (ssDNA) overhang structure in Bacillus subtilis is poorly understood. Mutations in addA and addB or null mutations in recJ (recJ), recQ (recQ), or recS (recS) genes, when present in otherwise-Rec+ cells, render cells moderately sensitive to the killing action of different DNA-damaging agents. Inactivation of a RecQ-like helicase (recQ or recS) in addAB cells showed an additive effect; however, when recJ was combined with addAB, a strong synergistic effect was observed with a survival rate similar to that of recA cells. RecF was nonepistatic with RecJ or AddAB. After induction of DSBs, RecN-yellow fluorescent protein (YFP) foci were formed in addAB recJ cells. AddAB and RecJ were required for the formation of a single RecN focus, because in their absence multiple RecN-YFP foci accumulated within the cells. Green fluorescent protein-RecA failed to form filamentous structures (termed threads) in addAB recJ cells. We propose that RecN is one of the first recombination proteins detected as a discrete focus in live cells in response to DSBs and that either AddAB or RecQ(S)-RecJ are required for the generation of a duplex with a 3'-ssDNA tail needed for filament formation of RecA.
INTRODUCTION
Structural aberrations in the DNA template, including endogenous lesions that hinder replication fork progression or the generation of double-strand breaks (DSBs), can result in the arrest or collapse of replication forks, whose restoration relies on replication primed by homologous recombination (8, 9, 20, 22, 25). Recombinational repair of DSBs in bacteria involves five general steps: (i) damage recognition, (ii) end processing, (iii) recombinase loading, (iv) homologous pairing, and (v) DNA heteroduplex extension and resolution. Very little is known about the process of DNA damage recognition. Genetic and biochemical studies performed with wild-type (wt) Escherichia coli cells revealed that both end processing (step ii) and loading of RecA (step iii) are mainly performed by the RecBCD enzyme (21, 34). In the absence of RecBCD in recBC sbcB sbcC mutant cells, the RecF pathway is activated (14). In the absence of the major component of the DSB repair machinery (RecBCD), end processing requires RecQ, which unwinds a duplex or a duplex with 3'overhangs, and RecJ, which degrades single-stranded DNA (ssDNA) with 5'-to-3' polarity. In concert, these proteins generate a duplex with a 3'-ssDNA tail to which the single-stranded binding protein binds (8, 13, 24, 36). The RecFOR complex promotes and modulates loading of RecA by displacing single-stranded binding protein from the exposed 3' ssDNA tails of gapped duplex DNA (26, 32, 35, 37). Recently, the existence of hybrid pathways, with interchangeable parts of E. coli RecBCD and RecF recombination machines, was also documented (4). Here, RecB1080CD and RecJ, acting in concert, degrade the 5' ssDNA tail (step ii), and RecFOR helps to load RecA onto the 3' ssDNA tail of the duplex (step iii) (15).
The recombinational repair steps of DSBs are less well defined in Bacillus subtilis. Null mutations in the recN, recO, or recU genes in an otherwise Rec+ strain render cells severely impaired in DNA repair (11, 19), whereas mutations in addA5 addB72 (collectively termed addAB) or null mutations in the recS (1, 12, 31), recJ, or recQ gene (this work) render cells moderately sensitive to the killing action of DNA-damaging agents. Cytological studies have revealed that recombination proteins assemble at the site of DNA damage, in a discrete temporal order (18, 19). Upon induction of random or defined DSBs, the nucleoids fuse and RecN localizes as a discrete focus on the nucleoids in a majority of the cells (15 to 30 min after induction of DSBs), whereas two to three foci are rarely observed (19). RecN forms foci in addAB, recO, recF, recJ, recU, recG, and recA strains (19). The RecO or RecA proteins are recruited to the RecN focus about 15 min later (30 to 45 min after the induction of DSBs) (18, 19). RecA forms highly dynamic filamentous structures (termed threads) that emanate from the RecN-promoted repair centers towards the opposite cell half (18). RecF is loaded at 60 to 90 min, followed by RecU, to random or defined DSBs. The highest number of RecU foci was observed 120 min after induction of DSBs, colocalizing with RecN (19, 31). Growth resumed 180 min after mitomycin C (MMC) treatment (19). Recently, it was shown that RecN exhibits ssDNA-dependent ATP hydrolysis and ssDNA binding activity in the presence of Mg2+; it binds ssDNA independently of nucleotide cofactors but also binds specifically to 3' ssDNA extensions in a nucleotide-dependent manner (30).
The functions required for the processing of DNA ends in B. subtilis are poorly defined. There are some obvious differences between E. coli end-processing proteins and their B. subtilis counterparts. Unlike E. coli RecQ (RecQEco) or budding and fission yeast (Sgs1 and Rqh1, respectively) (17), firmicutes possess two RecQ-like (RecQ and RecS) homologs (12). RecS contains the DExH box helicase motif and part of the RecQ conserved C-terminal (designated RQC) domain, whereas RecQ contains the DExH, RQC, and the helicase and RNase D C-terminal (designated HRDC) domains (27). AddAB, which composed of two polypeptides and is the functional analog of RecBCDEco, contains one (rather than two) helicase and two (rather than one) nuclease motifs and recognizes a different, shorter hot spot (Chi) sequence than the RecBCDEco enzyme (6, 7). It is currently unknown whether AddAB loads RecA onto the processed ends.
To probe the validity of the E. coli paradigm (see above) in end processing in B. subtilis and to learn whether RecN acts prior or after end processing, we have constructed recJ, recQ, or recS mutations in different combinations and with addAB. These strains were exposed to the killing action of different DNA-damaging agents, and their viability was compared to that of recA cells (impaired in homologous pairing). The inactivation of recQ or recS in the addAB background had an additive effect, and inactivation of recJ in addAB cells showed a strong synergistic effect, with survival matching that of recA cells. Furthermore, we show that RecF was nonepistatic with RecJ or AddAB. We determined whether RecN formed foci upon induction of DSBs in addAB recJ cells and used the loading of RecA as control. Cytological studies revealed that the AddAB, RecQ, and RecJ proteins are not required for loading of RecN onto DSBs but that both AddAB and RecJ nucleases are essential for the formation of RecA threads (RecA loading and nucleoprotein filament formation). The results presented provide the first evidence for the following. (i) addAB or recJ cells survive the killing of different DNA-damaging agents with relative similar frequency in B. subtilis cells. (ii) In the absence of the AddAB and RecJ nucleases, RecA loading is impaired. (iii) End processing is essential for RecA loading and homologous pairing. (iv) The structural maintenance of chromosome-like RecN, which might maintain DNA integrity, may be one of the first proteins recognizing a DSB in live bacteria.
MATERIALS AND METHODS
Bacterial strains. All B. subtilis strains used in this study are isogenic and are listed in Table 1. A 2-kb six-cat-six cassette containing two directly repeated copies of the site-specific recombinase target site (six) surrounding the chloramphenicol acetyltransferase gene (cat) was introduced within the coding sequence of recJ and recQ. The recJ:six-cat-six and recQ:six-cat-six disruptions were then transferred into the wt strain by a double-crossover event as previously described (1). The site-mediated site-specific recombination between the two six sites leads to deletion of one six site and the cat gene, leaving an intact copy of one six site. The chloramphenicol-sensitive rec-deficient strains were transformed with recJ:six-cat-six, recQ:six-cat-six, or recS:cat derivatives, and the double or triple mutants were generated by a double-crossover event as previously described (1).
Survival studies. Exponentially growing cells were obtained by inoculation of overnight cultures in fresh LB medium and were grown to an optical density at 560 nm of 0.4 at 37°C. Serial dilutions of exponential-phase cells listed in Table 1 were spotted on plates containing the indicated amount of methyl methane sulfonate (MMS), 4-nitroquinoline-1-oxide (4NQO), or MMC and incubated overnight at 37°C. MMS reacts with single reactive groups in adenine (N3-alkyladenine) and guanine (N7-alkylguanine) and can result in the stalling and subsequent collapse of the replication fork. 4NQO is a potent mutagen that induces two main guanine adducts at positions C8 and N2. 4NQO-damaged double-stranded DNA or ssDNA can result in the stall or collapse of the replication fork. MMC promotes an intrastrand cross-link that can result in a DSB without a preexisting nick.
To perform viability tests, different B. subtilis recombination-deficient strains were plated and incubated in LB medium overnight. At least six independent colonies from each strain were resuspended in fresh LB medium and shaken for 30 min to minimize aggregates. Appropriate dilutions were plated, and CFU were counted or stained with membrane-permeative SYTO 9 and membrane-impermeative propidium iodide and subjected to a conventional direct count of total cells. SYTO 9, which labels live and dead bacteria with green fluorescence, and propidium iodide, which stains membrane-compromised bacteria with red fluorescence, were purchased from Molecular Probes (Leiden, The Netherlands).
Protein comparison analysis. Using standard BLASTP, we searched all 115 nonredundant chromosomes for the presence of RecJ, RecBCD/AddAB, and RecQ proteins as previously described (3). The cutoff E value used in this analysis was 10–4. Identical results were observed when the PSI-BLAST program was used.
Construction of fluorescence-tagged strains and image acquisition. The construction of RecN-green fluorescent protein (GFP), RecN-yellow fluorescent protein (YFP), and GFP-RecA was previously reported, and the stopless recQ gene was fused to the gfp gene as described previously (18, 19). The N-terminal GFP fusion to RecA (GFP-RecA) retained almost full activity (data not shown). The fused genes were moved by transformation, as previously described, into the strains listed in Table 1 (19).
Fluorescence microscopy was performed with an Olympus AX70 microscope. Cells were mounted on agarose pads containing S750 growth medium on object slides. Images were acquired with a digital MicroMax charge-coupled device camera; signal intensities and cell length were measured using the MetaMorph program, version 4.6. DNA was stained with 4',6-diamidino-2-phenylindole (DAPI; final concentration, 0.2 ng/ml), and membranes were stained with FM4-64 (final concentration, 1 nM).
RESULTS
AddAB and RecJ are required for DNA repair. The mechanism by which end processing (step ii; see the introduction) occurs in B. subtilis is largely unknown. This is partly due to the fact that mutations in both subunits of the helicase/nuclease enzyme, addA5 addB72 (collectively termed addAB), render cells only moderately sensitive to the killing action of DNA-damaging agents (1). To probe the validity of the E. coli paradigm, we compared the impact of mutations in B. subtilis addAB and of null mutations in the recQ (recQ), recS (recS), and recJ (recJ) genes and compared them with recA cells, which are impaired in homologous pairing (step iv; see the introduction).
In the absence of any DNA damage, the number of viable cells per colony of addAB, recQ, recS, recJ, addAB recQ, or addAB recS strains was only marginally affected (<1.5 fold), whereas the number of viable cells per colony of recA and addAB recJ cells was 4-fold and 10-fold reduced, respectively, compared to the wt strain (data not shown) (Fig. 1A). This was consistent with the observation that the proportion of recA or addAB recJ cells stained with propidium iodide (an indicator of "membrane-compromised" bacteria) increased 4-fold and 10-fold, respectively, compared to wt cells (reference 31 and data not shown).
The mutants studied in this work (Table 1) were exposed to different concentrations of different DNA-damaging agents, and their phenotypes were recorded. Strains deficient in recombination, when present in an otherwise Rec+ background, could be grouped into different classes depending on their sensitivity to the DNA-damaging agents (MMS, 4NQO, and MMC). With few exceptions, Fig. 1 shows the maximal concentration of the DNA-damaging agents that did not seem to affect the plating efficiency of wt or mutant cells. The growth of the wt strain was unaffected when it was exposed to 250 μg/ml MMS (Fig. 1B), to 24 μg/ml 4NQO (Fig. 1C), or to 150 ng/ml MMC (Fig. 1D) compared to the condition without any drug added (Fig. 1A), and it defined class a (minimal effect). However, a growth defect in the wt strain was observed at higher drug concentrations (300 μg/ml MMS, 35 μg/ml 4NQO, or 200 ng/ml MMC) (data not shown).
The null recQ, recS, or recJ mutant strains did not show any detectable growth impairment even at 200-μg/ml MMS (class b) (Fig. 1B) but were slightly sensitive to the presence of 250-μg/ml MMS (Table 2). The addAB, recQ recS, recQ recJ, or recS recJ cells showed a slight growth defect in the presence of 80-μg/ml MMS (class c) (Fig. 1B) but were moderately sensitive in the presence of 150 μg/ml MMS and very sensitive in the presence of 200 μg/ml MMS (Table 2). The addAB recQ and addAB recS strains were marginally affected by the addition of 40 μg/ml MMS (class d) (Fig. 1B) and very sensitive in the presence of 80 μg/ml MMS (Table 2). The addAB recJ strain was extremely sensitive to the killing action of 5 μg/ml MMS, with survival matching that of a recA strain (class e) (Fig. 1B; Table 2).
addAB, recJ, recQ, recS, recQ recS, recQ recJ, or recS recJ strains did not show growth defects with 12 μg/ml 4NQO or 100 ng/ml MMC (class b) (Fig. 1C and D). Unlike the wt (class a), these mutant strains showed a defect in colony formation in the presence of 24 μg/ml 4NQO or 150 ng/ml MMC (Table 2). The addAB recQ or addAB recS strains were very sensitive to the killing action of 12 μg/ml 4NQO or to 100-ng/ml MMC (Table 2), but their growth was only slightly affected in the presence of 3 μg/ml 4NQO or 12 ng/ml MMC (class c) (Fig. 1C and D). The addAB recJ or recA strains were extremely sensitive to the killing action of 4NQO or MMC; they showed a growth defect even in the presence of 0.75 μg/ml 4NQO or 12 ng/ml of MMC (class d) (Fig. 1C and D).
From these results, we conclude that the inactivation of a RecQ-like helicase (recQ or recS) and recJ has a mild effect and that they are epistatic to one another, whereas recQ or recS in combination with addAB has an additive effect (Table 2). However, when recJ and addAB mutations were combined, a strong synergistic effect was observed (Table 2).
DNA repair in recF mutant cells is dependent on AddAB and RecJ nucleases. RecBCD is a major component of the DSB repair machinery in E. coli. In the absence of RecBCD, repair occurs via the activated RecF pathway, which is detectable only after inactivation of exonuclease I (SbcB) and nuclease SbcCD (for reviews, see references 4, 20, and 21). In contrast to E. coli cells (see above), B. subtilis addAB and recF cells are moderately and very sensitive to the killing action of DNA-damaging agents, respectively (1, 2). However, a strong synergistic effect was observed when addAB and recF mutations were combined, with survival reduced to the level seen in recA cells (1). The absence of RecS marginally increases the sensitivity of recF cells to DNA-damaging agents and markedly reduces genetic recombination (12). No information is available on the effect of the absence of B. subtilis RecJ or RecQ functions in recF cells. We examined the impact of recQ or recJ in recF mutant cells exposed to the killing action of MMS, 4NQO, or MMC and compared these strains with addAB recF mutant cells.
The growth of recF mutant cells, when present in an otherwise-Rec+ background, was similar in the presence of 5 μg/ml MMS (Fig. 2B), 0.75 μg/ml 4NQO (Fig. 2C), or 3 ng/ml MMC (Fig. 2D) or in the absence of additions (Fig. 2A). A recF mutation rendered cells extremely sensitive to the DNA-damaging agents and generated a growth defect in the presence of 12 μg/ml MMS, 3 μg/ml 4NQO, or 12 ng/ml MMC (data not shown). Figure 2 shows that the recJ deletion or the addAB mutation increased the sensitivity of recF cells to a level comparable to that of recA cells and reduced the colony size of recF mutant cells (1). The recQ deletion, as was shown to be the case for recS (12), slightly increased the sensitivity of recF mutant cells. It is likely that recF is nonepistatic with recQ, recJ, recS (12), or addAB (1) genes.
RecQ localizes throughout the nucleoids. Cells transiently engage distinct proteins of the DNA repair machinery, many of which accumulate into focal assemblies at the site of a DSB (see the introduction). Previously, it was shown that upon induction of random or of defined DSBs (i) RecN localizes on the nucleoids as a discrete focus in a majority of the cells, while RecO and RecA localize later (18, 19), and (ii) the AddA-GFP or AddB-YFP proteins, expressed as the sole source within the cells, localize throughout the cell (J. Mascarenhas, H. Sanchez, S. Tadesse, D. Kidane, J. C. Alonso, and P. L. Graumann, submitted for publication). In a previous section, it was shown that RecJ and RecQ or RecJ and RecS were involved in concert in end processing. To investigate the localization of the RecQ protein, the 3' end of the recQ gene was fused to the cfp (for cyan fluorescent protein) gene as described in Materials and Methods. RecQ-CFP, expressed as the sole source of RecQ within cells, supported DSB repair to a level similar to that of wt cells (data not shown). RecQ-CFP localized throughout the nucleoids in the presence or absence of DSBs (Fig. 3), suggesting that the RecQ helicase is constitutively associated with chromosomal DNA. Unlike the B. subtilis RecQ protein (Fig. 3), RecQEco colocalizes with the replisome in the absence of an external source of DNA damage, and the focus disappears when rounds of replication are completed (33). It is currently unknown whether RecS has any specific localization.
Our results show that not all proteins involved in DSB repair assemble into discrete foci after DNA damage; therefore, it is unlikely that the formation of foci is an artifact due to the tagging with a fluorescent protein (GFP, YFP, or CFP).
RecN assembly occurs in the absence of AddAB and RecJ. Previously, it was shown that RecN is recruited to a defined or to a random DSB at an early time point during repair, followed by RecO, RecA, and RecF, which colocalize with the induced RecN focus (18, 19). To investigate whether DNA end-processing functions might affect the localization of RecN, we moved the recN-yfp fusion into mutant strains deficient in recQ, recS, recJ, or addAB or in the recJ and addAB genes. RecN-YFP formed inducible foci that were indistinguishable from those in wt cells in the absence of RecQ, RecS, or RecJ (Table 3 and data not shown), but the foci were much fainter in addAB or in recA cells than in the wt (19). The RecN-YFP foci were also fainter in the addAB recJ cells than in wt cells (Fig. 4D and E; Table 3). This can be explained by our finding that the SOS response is strongly reduced in addAB cells and blocked in recA cells, resulting in lower levels of RecN (19).
In the absence of an external source of DNA damage, RecN-YFP was mainly distributed throughout addAB recJ cells. However, 2% of the exponentially growing cells formed RecN-YFP foci (Fig. 4C), a higher percentage than that observed with wt cells (0.05%) (19). It is likely that spontaneous DSBs accumulated at a higher frequency in addAB recJ mutant cells. Additionally, the mutant cells showed an abnormal nucleoid morphology and a high degree of nonsegregated nucleoids, leading to the formation of cells that were much longer than wt cells (compare Fig. 4C and 4A).
Seventy-eight percent of the addAB recJ cells contained fluorescent foci on the nucleoids after the addition of MMC. Unlike wt cells, which generally contain a single RecN-YFP focus (19), a large fraction of addAB recJ cells contained two foci (52%) rather than one (40%) (Fig. 4D). Only 2 to 3% of wt cells contained two or more RecN-YFP foci after the induction of DSBs, while 8% of the addAB recJ mutant cells contained three or more foci (Fig. 4E). Earlier experiments have suggested that several DSBs could be repaired within a single repair center (19). Possibly, AddAB and RecJ play a role in combining DSBs to a single repair center. These experiments show that although RecN-YFP can form foci at DSBs in the absence of AddAB and RecJ proteins, these nucleases regulate RecN-YFP focus formation.
AddAB and RecJ are required for the formation of RecA threads. RecA is one of the most important proteins for DNA repair by homologous recombination (8, 9, 20, 22, 25). When GFP-RecA was expressed from an ectopic site on the chromosome and the original recA gene was deleted, the cells grew indistinguishably from wt cells and were as sensitive to MMC as wt cells (18). GFP-RecA formed threads after induction of DSBs in addAB, recQ, recS, or recJ cells (Fig. 4F and data not shown); however, it failed to form threads in addAB recJ cells (18). This was in good agreement with our genetic data showing that repair relied on the AddAB or RecJ nucleases in an otherwise rec+ background and that these proteins acted at an early step during DSB repair before RecA formed filaments. We ruled out that the absence of GFP-RecA threads was the result of lower levels of RecA because GFP-RecA forms threads in addAB cells but fails to form them in the addAB recJ cells, and unlike in addAB cells, the absence of RecJ does not seem to affect the SOS response.
DISCUSSION
In this work, we provide evidence that different pathways are operative in B. subtilis cells during DSB repair. Strains carrying mutations in the two subunits of the AddAB enzyme or a loss of function of the recQ, recS, or recJ genes displayed only a modest sensitivity to MMS, 4NQO, or MMC, but recF15 and recA cells are very and extremely sensitive to DNA-damaging agents, respectively (1, 11, 12). Recombinational repair was highly impaired in addAB recQ or addAB recS cells but blocked in the addAB recJ strain. The addAB recJ strain was as sensitive to DNA-damaging agents as the recA strain. To rationalize our data we propose the following. (i) Inactivation of SbcC is not required to detect the effect of RecJ, RecQ, or RecS in DNA repair. (ii) The homologous RecQ and RecS proteins might have some overlapping activities. (iii) AddAB and/or RecJ provides two complementary avenues for the processing of DNA ends. (iv) AddAB and/or (in concert) RecJ and RecQ or RecJ and RecS, collectively termed RecQ(S)-RecJ, process DNA ends with relatively similar efficiency in an otherwise-Rec+ strain. (v) The RecQ(S)-RecJ repair avenue is operative even in the presence of the AddAB enzyme. RecF was nonepistatic with RecJ or with AddAB. Unlike B. subtilis cells (this work), the repair of collapsed or stalled replication forks in E. coli is highly impaired in recBCD cells and only moderately impaired in recJ, recQ, or recB1067 recF cells (16, 28, 29). Therefore, it is likely that the E. coli paradigm is not valid for all other bacteria (31).
An evolutionary study of the bacterial end-processing machinery revealed that among the 115 deposited nonredundant sequenced genomes, the RecQ(S)-RecJ avenue is more widely spread (81%) than the AddAB (RecBCD) helicase/nuclease (48%). Only 5 of the 115 genomes contain an AddAB (RecBCD) helicase/nuclease and lack the RecQ-RecJ helicase/nuclease. Both the RecQ-RecJ and AddAB (RecBCD) functions are missing in only those species that are obligate intracellular parasites or obligate endosymbionts (15% of total nonredundant sequenced genomes).
DNA replication is the main route to convert DNA cleavages in one or in both DNA strands into DSBs (9, 20, 22). In B. subtilis wt cells, RecN relocalizes from a diffuse distribution throughout the cells to a distinct focus on the nucleoids within 15 to 30 min after induction of defined or of random DSBs (18, 19). The number of RecN foci per cell does not increase with higher number of DSBs (19), indicating that several breaks might be repaired within a single RecN focus. The RecO and RecA proteins are recruited to RecN foci within the first 30 to 45 min after induction of DSBs. Even in the absence of both the AddAB and RecJ DNA end-processing nucleases, RecN forms a distinct focus on the nucleoid, suggesting that RecN binds directly to the DNA ends, and thus appears to be the earliest sensor of a DSB (Fig. 4; Table 3). This is consistent with the observation that RecN binds to ssDNA, specifically to 3' ssDNA ends, in the presence of Mg2+ and subsequently forms large protein-DNA networks in the presence of ATP (30). In eukaryotes, the Mre11/Rad50/Xrs2 (Nbs1) complex binds directly to the DNA ends and appears to be the earliest sensor of a DSB, but the mechanism by which DNA end processing takes place is largely unknown (23). Interestingly, both Rad50 and RecN are structural maintenance of chromosome-like proteins and may serve similar functions in the detection of DSBs. In any event, it has become clear that the induction of DSBs in both prokaryotes and eukaryotes leads to the activation and integration of a diverse network of functions, crucial for maintaining viability in the face of genotoxic insult (19, 23).
In the absence of an external source of DNA damage, RecN-YFP foci were present in about 35% of recA cells (19); under similar conditions, 2% of the exponentially growing addAB recJ cells formed RecN-YFP foci. It is likely that due to the lack of proper end processing, the number of spontaneous DSBs in addAB recJ mutant cells is underestimated, compared with wt cells. RecN forms multiple foci in the addAB recJ strain, suggesting that end processing by the AddAB or RecJ nucleases is required for the formation of a single RecN focus within the cell. Furthermore, either one of the end-processing avenues—AddAB or RecJ and recQ(S)—is required for the function of RecA (18; this work).
If a lesion (nick) occurs on the lagging strand template, a 3' ssDNA region is generated, whereas if a lesion arises on the leading-strand template, a 5' ssDNA segment may accumulate. However, the RecBCD (AddAB) enzyme cannot load onto DNA ends that are not blunt or nearly blunt (8, 9, 20, 22, 25). We propose that RecN binds to the 3' ssDNA tails of duplex DNA of the lagging-strand template. Then, the region of 3' ssDNA is enlarged on the lagging-strand template by the concerted action of RecQ-RecJ or RecS-RecJ helicase ssDNA exonuclease, whereas a 3' ssDNA of a DSB on a leading-strand template is de novo generated at blunt or nearly blunt ends by the AddAB helicase/nuclease. A model implying the different recombinational requirements at each end of the DNA DSB was proposed earlier (10). Further assembly of RecN onto the products of the processed break may facilitate the gathering of different DNA ends to form a discrete focus or repair center on the nucleoid (18, 19, 23). Next, RecO and RecA are successively recruited to the defined RecN focus, and RecA threads are formed. Alternatively, AddAB may load RecA onto the product of the break-processing reaction. These views are consistent with the following observations. (i) addAB recJ cells (this work) and addAB recF cells (1) are highly impaired in DNA repair. (ii) RecF was nonepistatic with RecJ or AddAB (this work). (iii) RecA is not loaded onto the DNA substrate in the absence of end processing (addAB recJ cells) (this work). (iv) RecN is the first recombination protein observed to be recruited to defined or random DSBs in live cells (this work). (v) In vitro, RecN binds and protects 3' ssDNA ends in the presence of ATP (30).
ACKNOWLEDGMENTS
This work was partially supported by grants BMC2003-00150 and BIO2001-4342-E from Direccion General de Investigacion and GR/SAL/0668/2004 from Comunidad de Madrid to J.C.A. and from the Deutsche Forschungsgemeinschaft to P.L.G.
We are very grateful to D. Camerini-Otero and S. C. Kowalczykowski for critical reading of the manuscript.
Present address: Institut für Mikrobiologie, Universitt Freiburg, Verfügungsgebude, Stefan-Meier-Str. 17, 79104 Freiburg, Germany.
REFERENCES
Alonso, J. C., A. C. Stiege, and G. Luder. 1993. Genetic recombination in Bacillus subtilis 168: effect of recN, recF, recH and addAB mutations on DNA repair and recombination. Mol. Gen. Genet. 239:129-136.
Alonso, J. C., R. H. Tailor, and G. Luder. 1988. Characterization of recombination-deficient mutants of Bacillus subtilis. J. Bacteriol. 170:3001-3007.
Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Amundsen, S. K., and G. R. Smith. 2003. Interchangeable parts of the Escherichia coli recombination machinery. Cell 112:741-744.
Ceglowski, P., G. Luder, and J. C. Alonso. 1990. Genetic analysis of recE activities in Bacillus subtilis. Mol. Gen. Genet. 222:441-445.
Chedin, F., and S. C. Kowalczykowski. 2002. A novel family of regulated helicases/nucleases from gram-positive bacteria: insights into the initiation of DNA recombination. Mol. Microbiol. 43:823-834.
Chedin, F., P. Noirot, V. Biaudet, and S. D. Ehrlich. 1998. A five-nucleotide sequence protects DNA from exonucleolytic degradation by AddAB, the RecBCD analogue of Bacillus subtilis. Mol. Microbiol. 29:1369-1377.
Courcelle, J., and P. C. Hanawalt. 2003. RecA-dependent recovery of arrested DNA replication forks. Annu. Rev. Genet. 37:611-646.
Cox, M. M., M. F. Goodman, K. N. Kreuzer, D. J. Sherratt, S. J. Sandler, and K. J. Marians. 2000. The importance of repairing stalled replication forks. Nature 404:37-41.
Cromie, G. A., C. B. Millar, K. H. Schmidt, and D. R. Leach. 2000. Palindromes as substrates for multiple pathways of recombination in Escherichia coli. Genetics 154:513-522.
Fernandez, S., S. Ayora, and J. C. Alonso. 2000. Bacillus subtilis homologous recombination: genes and products. Res. Microbiol. 151:481-486.
Fernandez, S., A. Sorokin, and J. C. Alonso. 1998. Genetic recombination in Bacillus subtilis 168: effects of recU and recS mutations on DNA repair and homologous recombination. J. Bacteriol. 180:3405-3409.
Harmon, F. G., and S. C. Kowalczykowski. 1998. RecQ helicase, in concert with RecA and SSB proteins, initiates and disrupts DNA recombination. Genes Dev. 12:1134-1144.
Horii, Z., and A. J. Clark. 1973. Genetic analysis of the recF pathway to genetic recombination in Escherichia coli K12: isolation and characterization of mutants. J. Mol. Biol. 80:327-344.
Ivancic-Bace, I., P. Peharec, S. Moslavac, N. Skrobot, E. Salaj-Smic, and K. Brcic-Kostic. 2003. RecFOR function is required for DNA repair and recombination in a RecA loading-deficient recB mutant of Escherichia coli. Genetics 163:485-494.
Ivancic-Bace, I., E. Salaj-Smic, and K. Brcic-Kostic. 2005. Effects of recJ, recQ, and recFOR mutations on recombination in nuclease-deficient recB recD double mutants of Escherichia coli. J. Bacteriol. 187:1350-1356.
Khakhar, R. R., J. A. Cobb, L. Bjergbaek, I. D. Hickson, and S. M. Gasser. 2003. RecQ helicases: multiple roles in genome maintenance. Trends Cell Biol. 13:493-501.
Kidane, D., and P. L. Graumann. 2005. Dynamic formation of RecA filaments at DNA double strand break repair centers in live cells. J. Cell Biol. 170:357-366.
Kidane, D., H. Sanchez, J. C. Alonso, and P. L. Graumann. 2004. Visualization of DNA double-strand break repair in live bacteria reveals dynamic recruitment of Bacillus subtilis RecF, RecO and RecN proteins to distinct sites on the nucleoids. Mol. Microbiol. 52:1627-1639.
Kowalczykowski, S. C. 2000. Initiation of genetic recombination and recombination-dependent replication. Trends Biochem. Sci. 25:156-165.
Kowalczykowski, S. C., D. A. Dixon, A. K. Eggleston, S. D. Lauder, and W. M. Rehrauer. 1994. Biochemistry of homologous recombination in Escherichia coli. Microbiol. Rev. 58:401-465.
Kuzminov, A. 1999. Recombinational repair of DNA damage in Escherichia coli and bacteriophage lambda. Microbiol. Mol. Biol. Rev. 63:751-813.
Lisby, M., and R. Rothstein. 2004. DNA damage checkpoint and repair centers. Curr. Opin. Cell Biol. 16:328-334.
Lovett, S. T., and R. D. Kolodner. 1989. Identification and purification of a single-stranded-DNA-specific exonuclease encoded by the recJ gene of Escherichia coli. Proc. Natl. Acad. Sci. USA 86:2627-2631.
Michel, B., G. Grompone, M. J. Flores, and V. Bidnenko. 2004. Multiple pathways process stalled replication forks. Proc. Natl. Acad. Sci. USA 101:12783-12788.
Morimatsu, K., and S. C. Kowalczykowski. 2003. RecFOR proteins load RecA protein onto gapped DNA to accelerate DNA strand exchange: a universal step of recombinational repair. Mol. Cell 11:1337-1347.
Morozov, V., A. R. Mushegian, E. V. Koonin, and P. Bork. 1997. A putative nucleic acid-binding domain in Bloom's and Werner's syndrome helicases. Trends Biochem. Sci. 22:417-418.
Nakayama, H., K. Nakayama, R. Nakayama, N. Irino, Y. Nakayama, and P. C. Hanawalt. 1984. Isolation and genetic characterization of a thymineless death-resistant mutant of Escherichia coli K12: identification of a new mutation (recQ1) that blocks the RecF recombination pathway. Mol. Gen. Genet. 195:474-480.
Ryder, L., M. C. Whitby, and R. G. Lloyd. 1994. Mutation of recF, recJ, recO, recQ, or recR improves Hfr recombination in resolvase-deficient ruv recG strains of Escherichia coli. J. Bacteriol. 176:1570-1577.
Sanchez, H., and J. C. Alonso. 2005. Bacillus subtilis RecN binds and protects 3'-single-stranded DNA extensions in the presence of ATP. Nucleic Acids Res. 33:2343-2350.
Sanchez, H., D. Kidane, P. Reed, F. A. Curtis, M. C. Cozar, P. L. Graumann, G. J. Sharples, and J. C. Alonso. The RuvAB branch migration translocase and RecU Holliday junction resolvase are required for double-stranded DNA break repair in Bacillus subtilis. Genetics, in press. (Online.) doi:10.1534/genetics.105.045906.
Shan, Q., J. M. Bork, B. L. Webb, R. B. Inman, and M. M. Cox. 1997. RecA protein filaments: end-dependent dissociation from ssDNA and stabilization by RecO and RecR proteins. J. Mol. Biol. 265:519-540.
Sherratt, D. J., B. Soballe, F. X. Barre, S. Filipe, I. Lau, T. Massey, and J. Yates. 2004. Recombination and chromosome segregation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359:61-69.
Taylor, A. F., and G. R. Smith. 1992. RecBCD enzyme is altered upon cutting DNA at a chi recombination hotspot. Proc. Natl. Acad. Sci. USA 89:5226-5230.
Umezu, K., N. W. Chi, and R. D. Kolodner. 1993. Biochemical interaction of the Escherichia coli RecF, RecO, and RecR proteins with RecA protein and single-stranded DNA binding protein. Proc. Natl. Acad. Sci. USA 90:3875-3879.
Umezu, K., K. Nakayama, and H. Nakayama. 1990. Escherichia coli RecQ protein is a DNA helicase. Proc. Natl. Acad. Sci. USA 87:5363-5367.
Webb, B. L., M. M. Cox, and R. B. Inman. 1997. Recombinational DNA repair: the RecF and RecR proteins limit the extension of RecA filaments beyond single-strand DNA gaps. Cell 91:347-356.(Humberto Sanchez, Dawit K)
ABSTRACT
The recognition and processing of double-strand breaks (DSBs) to a 3' single-stranded DNA (ssDNA) overhang structure in Bacillus subtilis is poorly understood. Mutations in addA and addB or null mutations in recJ (recJ), recQ (recQ), or recS (recS) genes, when present in otherwise-Rec+ cells, render cells moderately sensitive to the killing action of different DNA-damaging agents. Inactivation of a RecQ-like helicase (recQ or recS) in addAB cells showed an additive effect; however, when recJ was combined with addAB, a strong synergistic effect was observed with a survival rate similar to that of recA cells. RecF was nonepistatic with RecJ or AddAB. After induction of DSBs, RecN-yellow fluorescent protein (YFP) foci were formed in addAB recJ cells. AddAB and RecJ were required for the formation of a single RecN focus, because in their absence multiple RecN-YFP foci accumulated within the cells. Green fluorescent protein-RecA failed to form filamentous structures (termed threads) in addAB recJ cells. We propose that RecN is one of the first recombination proteins detected as a discrete focus in live cells in response to DSBs and that either AddAB or RecQ(S)-RecJ are required for the generation of a duplex with a 3'-ssDNA tail needed for filament formation of RecA.
INTRODUCTION
Structural aberrations in the DNA template, including endogenous lesions that hinder replication fork progression or the generation of double-strand breaks (DSBs), can result in the arrest or collapse of replication forks, whose restoration relies on replication primed by homologous recombination (8, 9, 20, 22, 25). Recombinational repair of DSBs in bacteria involves five general steps: (i) damage recognition, (ii) end processing, (iii) recombinase loading, (iv) homologous pairing, and (v) DNA heteroduplex extension and resolution. Very little is known about the process of DNA damage recognition. Genetic and biochemical studies performed with wild-type (wt) Escherichia coli cells revealed that both end processing (step ii) and loading of RecA (step iii) are mainly performed by the RecBCD enzyme (21, 34). In the absence of RecBCD in recBC sbcB sbcC mutant cells, the RecF pathway is activated (14). In the absence of the major component of the DSB repair machinery (RecBCD), end processing requires RecQ, which unwinds a duplex or a duplex with 3'overhangs, and RecJ, which degrades single-stranded DNA (ssDNA) with 5'-to-3' polarity. In concert, these proteins generate a duplex with a 3'-ssDNA tail to which the single-stranded binding protein binds (8, 13, 24, 36). The RecFOR complex promotes and modulates loading of RecA by displacing single-stranded binding protein from the exposed 3' ssDNA tails of gapped duplex DNA (26, 32, 35, 37). Recently, the existence of hybrid pathways, with interchangeable parts of E. coli RecBCD and RecF recombination machines, was also documented (4). Here, RecB1080CD and RecJ, acting in concert, degrade the 5' ssDNA tail (step ii), and RecFOR helps to load RecA onto the 3' ssDNA tail of the duplex (step iii) (15).
The recombinational repair steps of DSBs are less well defined in Bacillus subtilis. Null mutations in the recN, recO, or recU genes in an otherwise Rec+ strain render cells severely impaired in DNA repair (11, 19), whereas mutations in addA5 addB72 (collectively termed addAB) or null mutations in the recS (1, 12, 31), recJ, or recQ gene (this work) render cells moderately sensitive to the killing action of DNA-damaging agents. Cytological studies have revealed that recombination proteins assemble at the site of DNA damage, in a discrete temporal order (18, 19). Upon induction of random or defined DSBs, the nucleoids fuse and RecN localizes as a discrete focus on the nucleoids in a majority of the cells (15 to 30 min after induction of DSBs), whereas two to three foci are rarely observed (19). RecN forms foci in addAB, recO, recF, recJ, recU, recG, and recA strains (19). The RecO or RecA proteins are recruited to the RecN focus about 15 min later (30 to 45 min after the induction of DSBs) (18, 19). RecA forms highly dynamic filamentous structures (termed threads) that emanate from the RecN-promoted repair centers towards the opposite cell half (18). RecF is loaded at 60 to 90 min, followed by RecU, to random or defined DSBs. The highest number of RecU foci was observed 120 min after induction of DSBs, colocalizing with RecN (19, 31). Growth resumed 180 min after mitomycin C (MMC) treatment (19). Recently, it was shown that RecN exhibits ssDNA-dependent ATP hydrolysis and ssDNA binding activity in the presence of Mg2+; it binds ssDNA independently of nucleotide cofactors but also binds specifically to 3' ssDNA extensions in a nucleotide-dependent manner (30).
The functions required for the processing of DNA ends in B. subtilis are poorly defined. There are some obvious differences between E. coli end-processing proteins and their B. subtilis counterparts. Unlike E. coli RecQ (RecQEco) or budding and fission yeast (Sgs1 and Rqh1, respectively) (17), firmicutes possess two RecQ-like (RecQ and RecS) homologs (12). RecS contains the DExH box helicase motif and part of the RecQ conserved C-terminal (designated RQC) domain, whereas RecQ contains the DExH, RQC, and the helicase and RNase D C-terminal (designated HRDC) domains (27). AddAB, which composed of two polypeptides and is the functional analog of RecBCDEco, contains one (rather than two) helicase and two (rather than one) nuclease motifs and recognizes a different, shorter hot spot (Chi) sequence than the RecBCDEco enzyme (6, 7). It is currently unknown whether AddAB loads RecA onto the processed ends.
To probe the validity of the E. coli paradigm (see above) in end processing in B. subtilis and to learn whether RecN acts prior or after end processing, we have constructed recJ, recQ, or recS mutations in different combinations and with addAB. These strains were exposed to the killing action of different DNA-damaging agents, and their viability was compared to that of recA cells (impaired in homologous pairing). The inactivation of recQ or recS in the addAB background had an additive effect, and inactivation of recJ in addAB cells showed a strong synergistic effect, with survival matching that of recA cells. Furthermore, we show that RecF was nonepistatic with RecJ or AddAB. We determined whether RecN formed foci upon induction of DSBs in addAB recJ cells and used the loading of RecA as control. Cytological studies revealed that the AddAB, RecQ, and RecJ proteins are not required for loading of RecN onto DSBs but that both AddAB and RecJ nucleases are essential for the formation of RecA threads (RecA loading and nucleoprotein filament formation). The results presented provide the first evidence for the following. (i) addAB or recJ cells survive the killing of different DNA-damaging agents with relative similar frequency in B. subtilis cells. (ii) In the absence of the AddAB and RecJ nucleases, RecA loading is impaired. (iii) End processing is essential for RecA loading and homologous pairing. (iv) The structural maintenance of chromosome-like RecN, which might maintain DNA integrity, may be one of the first proteins recognizing a DSB in live bacteria.
MATERIALS AND METHODS
Bacterial strains. All B. subtilis strains used in this study are isogenic and are listed in Table 1. A 2-kb six-cat-six cassette containing two directly repeated copies of the site-specific recombinase target site (six) surrounding the chloramphenicol acetyltransferase gene (cat) was introduced within the coding sequence of recJ and recQ. The recJ:six-cat-six and recQ:six-cat-six disruptions were then transferred into the wt strain by a double-crossover event as previously described (1). The site-mediated site-specific recombination between the two six sites leads to deletion of one six site and the cat gene, leaving an intact copy of one six site. The chloramphenicol-sensitive rec-deficient strains were transformed with recJ:six-cat-six, recQ:six-cat-six, or recS:cat derivatives, and the double or triple mutants were generated by a double-crossover event as previously described (1).
Survival studies. Exponentially growing cells were obtained by inoculation of overnight cultures in fresh LB medium and were grown to an optical density at 560 nm of 0.4 at 37°C. Serial dilutions of exponential-phase cells listed in Table 1 were spotted on plates containing the indicated amount of methyl methane sulfonate (MMS), 4-nitroquinoline-1-oxide (4NQO), or MMC and incubated overnight at 37°C. MMS reacts with single reactive groups in adenine (N3-alkyladenine) and guanine (N7-alkylguanine) and can result in the stalling and subsequent collapse of the replication fork. 4NQO is a potent mutagen that induces two main guanine adducts at positions C8 and N2. 4NQO-damaged double-stranded DNA or ssDNA can result in the stall or collapse of the replication fork. MMC promotes an intrastrand cross-link that can result in a DSB without a preexisting nick.
To perform viability tests, different B. subtilis recombination-deficient strains were plated and incubated in LB medium overnight. At least six independent colonies from each strain were resuspended in fresh LB medium and shaken for 30 min to minimize aggregates. Appropriate dilutions were plated, and CFU were counted or stained with membrane-permeative SYTO 9 and membrane-impermeative propidium iodide and subjected to a conventional direct count of total cells. SYTO 9, which labels live and dead bacteria with green fluorescence, and propidium iodide, which stains membrane-compromised bacteria with red fluorescence, were purchased from Molecular Probes (Leiden, The Netherlands).
Protein comparison analysis. Using standard BLASTP, we searched all 115 nonredundant chromosomes for the presence of RecJ, RecBCD/AddAB, and RecQ proteins as previously described (3). The cutoff E value used in this analysis was 10–4. Identical results were observed when the PSI-BLAST program was used.
Construction of fluorescence-tagged strains and image acquisition. The construction of RecN-green fluorescent protein (GFP), RecN-yellow fluorescent protein (YFP), and GFP-RecA was previously reported, and the stopless recQ gene was fused to the gfp gene as described previously (18, 19). The N-terminal GFP fusion to RecA (GFP-RecA) retained almost full activity (data not shown). The fused genes were moved by transformation, as previously described, into the strains listed in Table 1 (19).
Fluorescence microscopy was performed with an Olympus AX70 microscope. Cells were mounted on agarose pads containing S750 growth medium on object slides. Images were acquired with a digital MicroMax charge-coupled device camera; signal intensities and cell length were measured using the MetaMorph program, version 4.6. DNA was stained with 4',6-diamidino-2-phenylindole (DAPI; final concentration, 0.2 ng/ml), and membranes were stained with FM4-64 (final concentration, 1 nM).
RESULTS
AddAB and RecJ are required for DNA repair. The mechanism by which end processing (step ii; see the introduction) occurs in B. subtilis is largely unknown. This is partly due to the fact that mutations in both subunits of the helicase/nuclease enzyme, addA5 addB72 (collectively termed addAB), render cells only moderately sensitive to the killing action of DNA-damaging agents (1). To probe the validity of the E. coli paradigm, we compared the impact of mutations in B. subtilis addAB and of null mutations in the recQ (recQ), recS (recS), and recJ (recJ) genes and compared them with recA cells, which are impaired in homologous pairing (step iv; see the introduction).
In the absence of any DNA damage, the number of viable cells per colony of addAB, recQ, recS, recJ, addAB recQ, or addAB recS strains was only marginally affected (<1.5 fold), whereas the number of viable cells per colony of recA and addAB recJ cells was 4-fold and 10-fold reduced, respectively, compared to the wt strain (data not shown) (Fig. 1A). This was consistent with the observation that the proportion of recA or addAB recJ cells stained with propidium iodide (an indicator of "membrane-compromised" bacteria) increased 4-fold and 10-fold, respectively, compared to wt cells (reference 31 and data not shown).
The mutants studied in this work (Table 1) were exposed to different concentrations of different DNA-damaging agents, and their phenotypes were recorded. Strains deficient in recombination, when present in an otherwise Rec+ background, could be grouped into different classes depending on their sensitivity to the DNA-damaging agents (MMS, 4NQO, and MMC). With few exceptions, Fig. 1 shows the maximal concentration of the DNA-damaging agents that did not seem to affect the plating efficiency of wt or mutant cells. The growth of the wt strain was unaffected when it was exposed to 250 μg/ml MMS (Fig. 1B), to 24 μg/ml 4NQO (Fig. 1C), or to 150 ng/ml MMC (Fig. 1D) compared to the condition without any drug added (Fig. 1A), and it defined class a (minimal effect). However, a growth defect in the wt strain was observed at higher drug concentrations (300 μg/ml MMS, 35 μg/ml 4NQO, or 200 ng/ml MMC) (data not shown).
The null recQ, recS, or recJ mutant strains did not show any detectable growth impairment even at 200-μg/ml MMS (class b) (Fig. 1B) but were slightly sensitive to the presence of 250-μg/ml MMS (Table 2). The addAB, recQ recS, recQ recJ, or recS recJ cells showed a slight growth defect in the presence of 80-μg/ml MMS (class c) (Fig. 1B) but were moderately sensitive in the presence of 150 μg/ml MMS and very sensitive in the presence of 200 μg/ml MMS (Table 2). The addAB recQ and addAB recS strains were marginally affected by the addition of 40 μg/ml MMS (class d) (Fig. 1B) and very sensitive in the presence of 80 μg/ml MMS (Table 2). The addAB recJ strain was extremely sensitive to the killing action of 5 μg/ml MMS, with survival matching that of a recA strain (class e) (Fig. 1B; Table 2).
addAB, recJ, recQ, recS, recQ recS, recQ recJ, or recS recJ strains did not show growth defects with 12 μg/ml 4NQO or 100 ng/ml MMC (class b) (Fig. 1C and D). Unlike the wt (class a), these mutant strains showed a defect in colony formation in the presence of 24 μg/ml 4NQO or 150 ng/ml MMC (Table 2). The addAB recQ or addAB recS strains were very sensitive to the killing action of 12 μg/ml 4NQO or to 100-ng/ml MMC (Table 2), but their growth was only slightly affected in the presence of 3 μg/ml 4NQO or 12 ng/ml MMC (class c) (Fig. 1C and D). The addAB recJ or recA strains were extremely sensitive to the killing action of 4NQO or MMC; they showed a growth defect even in the presence of 0.75 μg/ml 4NQO or 12 ng/ml of MMC (class d) (Fig. 1C and D).
From these results, we conclude that the inactivation of a RecQ-like helicase (recQ or recS) and recJ has a mild effect and that they are epistatic to one another, whereas recQ or recS in combination with addAB has an additive effect (Table 2). However, when recJ and addAB mutations were combined, a strong synergistic effect was observed (Table 2).
DNA repair in recF mutant cells is dependent on AddAB and RecJ nucleases. RecBCD is a major component of the DSB repair machinery in E. coli. In the absence of RecBCD, repair occurs via the activated RecF pathway, which is detectable only after inactivation of exonuclease I (SbcB) and nuclease SbcCD (for reviews, see references 4, 20, and 21). In contrast to E. coli cells (see above), B. subtilis addAB and recF cells are moderately and very sensitive to the killing action of DNA-damaging agents, respectively (1, 2). However, a strong synergistic effect was observed when addAB and recF mutations were combined, with survival reduced to the level seen in recA cells (1). The absence of RecS marginally increases the sensitivity of recF cells to DNA-damaging agents and markedly reduces genetic recombination (12). No information is available on the effect of the absence of B. subtilis RecJ or RecQ functions in recF cells. We examined the impact of recQ or recJ in recF mutant cells exposed to the killing action of MMS, 4NQO, or MMC and compared these strains with addAB recF mutant cells.
The growth of recF mutant cells, when present in an otherwise-Rec+ background, was similar in the presence of 5 μg/ml MMS (Fig. 2B), 0.75 μg/ml 4NQO (Fig. 2C), or 3 ng/ml MMC (Fig. 2D) or in the absence of additions (Fig. 2A). A recF mutation rendered cells extremely sensitive to the DNA-damaging agents and generated a growth defect in the presence of 12 μg/ml MMS, 3 μg/ml 4NQO, or 12 ng/ml MMC (data not shown). Figure 2 shows that the recJ deletion or the addAB mutation increased the sensitivity of recF cells to a level comparable to that of recA cells and reduced the colony size of recF mutant cells (1). The recQ deletion, as was shown to be the case for recS (12), slightly increased the sensitivity of recF mutant cells. It is likely that recF is nonepistatic with recQ, recJ, recS (12), or addAB (1) genes.
RecQ localizes throughout the nucleoids. Cells transiently engage distinct proteins of the DNA repair machinery, many of which accumulate into focal assemblies at the site of a DSB (see the introduction). Previously, it was shown that upon induction of random or of defined DSBs (i) RecN localizes on the nucleoids as a discrete focus in a majority of the cells, while RecO and RecA localize later (18, 19), and (ii) the AddA-GFP or AddB-YFP proteins, expressed as the sole source within the cells, localize throughout the cell (J. Mascarenhas, H. Sanchez, S. Tadesse, D. Kidane, J. C. Alonso, and P. L. Graumann, submitted for publication). In a previous section, it was shown that RecJ and RecQ or RecJ and RecS were involved in concert in end processing. To investigate the localization of the RecQ protein, the 3' end of the recQ gene was fused to the cfp (for cyan fluorescent protein) gene as described in Materials and Methods. RecQ-CFP, expressed as the sole source of RecQ within cells, supported DSB repair to a level similar to that of wt cells (data not shown). RecQ-CFP localized throughout the nucleoids in the presence or absence of DSBs (Fig. 3), suggesting that the RecQ helicase is constitutively associated with chromosomal DNA. Unlike the B. subtilis RecQ protein (Fig. 3), RecQEco colocalizes with the replisome in the absence of an external source of DNA damage, and the focus disappears when rounds of replication are completed (33). It is currently unknown whether RecS has any specific localization.
Our results show that not all proteins involved in DSB repair assemble into discrete foci after DNA damage; therefore, it is unlikely that the formation of foci is an artifact due to the tagging with a fluorescent protein (GFP, YFP, or CFP).
RecN assembly occurs in the absence of AddAB and RecJ. Previously, it was shown that RecN is recruited to a defined or to a random DSB at an early time point during repair, followed by RecO, RecA, and RecF, which colocalize with the induced RecN focus (18, 19). To investigate whether DNA end-processing functions might affect the localization of RecN, we moved the recN-yfp fusion into mutant strains deficient in recQ, recS, recJ, or addAB or in the recJ and addAB genes. RecN-YFP formed inducible foci that were indistinguishable from those in wt cells in the absence of RecQ, RecS, or RecJ (Table 3 and data not shown), but the foci were much fainter in addAB or in recA cells than in the wt (19). The RecN-YFP foci were also fainter in the addAB recJ cells than in wt cells (Fig. 4D and E; Table 3). This can be explained by our finding that the SOS response is strongly reduced in addAB cells and blocked in recA cells, resulting in lower levels of RecN (19).
In the absence of an external source of DNA damage, RecN-YFP was mainly distributed throughout addAB recJ cells. However, 2% of the exponentially growing cells formed RecN-YFP foci (Fig. 4C), a higher percentage than that observed with wt cells (0.05%) (19). It is likely that spontaneous DSBs accumulated at a higher frequency in addAB recJ mutant cells. Additionally, the mutant cells showed an abnormal nucleoid morphology and a high degree of nonsegregated nucleoids, leading to the formation of cells that were much longer than wt cells (compare Fig. 4C and 4A).
Seventy-eight percent of the addAB recJ cells contained fluorescent foci on the nucleoids after the addition of MMC. Unlike wt cells, which generally contain a single RecN-YFP focus (19), a large fraction of addAB recJ cells contained two foci (52%) rather than one (40%) (Fig. 4D). Only 2 to 3% of wt cells contained two or more RecN-YFP foci after the induction of DSBs, while 8% of the addAB recJ mutant cells contained three or more foci (Fig. 4E). Earlier experiments have suggested that several DSBs could be repaired within a single repair center (19). Possibly, AddAB and RecJ play a role in combining DSBs to a single repair center. These experiments show that although RecN-YFP can form foci at DSBs in the absence of AddAB and RecJ proteins, these nucleases regulate RecN-YFP focus formation.
AddAB and RecJ are required for the formation of RecA threads. RecA is one of the most important proteins for DNA repair by homologous recombination (8, 9, 20, 22, 25). When GFP-RecA was expressed from an ectopic site on the chromosome and the original recA gene was deleted, the cells grew indistinguishably from wt cells and were as sensitive to MMC as wt cells (18). GFP-RecA formed threads after induction of DSBs in addAB, recQ, recS, or recJ cells (Fig. 4F and data not shown); however, it failed to form threads in addAB recJ cells (18). This was in good agreement with our genetic data showing that repair relied on the AddAB or RecJ nucleases in an otherwise rec+ background and that these proteins acted at an early step during DSB repair before RecA formed filaments. We ruled out that the absence of GFP-RecA threads was the result of lower levels of RecA because GFP-RecA forms threads in addAB cells but fails to form them in the addAB recJ cells, and unlike in addAB cells, the absence of RecJ does not seem to affect the SOS response.
DISCUSSION
In this work, we provide evidence that different pathways are operative in B. subtilis cells during DSB repair. Strains carrying mutations in the two subunits of the AddAB enzyme or a loss of function of the recQ, recS, or recJ genes displayed only a modest sensitivity to MMS, 4NQO, or MMC, but recF15 and recA cells are very and extremely sensitive to DNA-damaging agents, respectively (1, 11, 12). Recombinational repair was highly impaired in addAB recQ or addAB recS cells but blocked in the addAB recJ strain. The addAB recJ strain was as sensitive to DNA-damaging agents as the recA strain. To rationalize our data we propose the following. (i) Inactivation of SbcC is not required to detect the effect of RecJ, RecQ, or RecS in DNA repair. (ii) The homologous RecQ and RecS proteins might have some overlapping activities. (iii) AddAB and/or RecJ provides two complementary avenues for the processing of DNA ends. (iv) AddAB and/or (in concert) RecJ and RecQ or RecJ and RecS, collectively termed RecQ(S)-RecJ, process DNA ends with relatively similar efficiency in an otherwise-Rec+ strain. (v) The RecQ(S)-RecJ repair avenue is operative even in the presence of the AddAB enzyme. RecF was nonepistatic with RecJ or with AddAB. Unlike B. subtilis cells (this work), the repair of collapsed or stalled replication forks in E. coli is highly impaired in recBCD cells and only moderately impaired in recJ, recQ, or recB1067 recF cells (16, 28, 29). Therefore, it is likely that the E. coli paradigm is not valid for all other bacteria (31).
An evolutionary study of the bacterial end-processing machinery revealed that among the 115 deposited nonredundant sequenced genomes, the RecQ(S)-RecJ avenue is more widely spread (81%) than the AddAB (RecBCD) helicase/nuclease (48%). Only 5 of the 115 genomes contain an AddAB (RecBCD) helicase/nuclease and lack the RecQ-RecJ helicase/nuclease. Both the RecQ-RecJ and AddAB (RecBCD) functions are missing in only those species that are obligate intracellular parasites or obligate endosymbionts (15% of total nonredundant sequenced genomes).
DNA replication is the main route to convert DNA cleavages in one or in both DNA strands into DSBs (9, 20, 22). In B. subtilis wt cells, RecN relocalizes from a diffuse distribution throughout the cells to a distinct focus on the nucleoids within 15 to 30 min after induction of defined or of random DSBs (18, 19). The number of RecN foci per cell does not increase with higher number of DSBs (19), indicating that several breaks might be repaired within a single RecN focus. The RecO and RecA proteins are recruited to RecN foci within the first 30 to 45 min after induction of DSBs. Even in the absence of both the AddAB and RecJ DNA end-processing nucleases, RecN forms a distinct focus on the nucleoid, suggesting that RecN binds directly to the DNA ends, and thus appears to be the earliest sensor of a DSB (Fig. 4; Table 3). This is consistent with the observation that RecN binds to ssDNA, specifically to 3' ssDNA ends, in the presence of Mg2+ and subsequently forms large protein-DNA networks in the presence of ATP (30). In eukaryotes, the Mre11/Rad50/Xrs2 (Nbs1) complex binds directly to the DNA ends and appears to be the earliest sensor of a DSB, but the mechanism by which DNA end processing takes place is largely unknown (23). Interestingly, both Rad50 and RecN are structural maintenance of chromosome-like proteins and may serve similar functions in the detection of DSBs. In any event, it has become clear that the induction of DSBs in both prokaryotes and eukaryotes leads to the activation and integration of a diverse network of functions, crucial for maintaining viability in the face of genotoxic insult (19, 23).
In the absence of an external source of DNA damage, RecN-YFP foci were present in about 35% of recA cells (19); under similar conditions, 2% of the exponentially growing addAB recJ cells formed RecN-YFP foci. It is likely that due to the lack of proper end processing, the number of spontaneous DSBs in addAB recJ mutant cells is underestimated, compared with wt cells. RecN forms multiple foci in the addAB recJ strain, suggesting that end processing by the AddAB or RecJ nucleases is required for the formation of a single RecN focus within the cell. Furthermore, either one of the end-processing avenues—AddAB or RecJ and recQ(S)—is required for the function of RecA (18; this work).
If a lesion (nick) occurs on the lagging strand template, a 3' ssDNA region is generated, whereas if a lesion arises on the leading-strand template, a 5' ssDNA segment may accumulate. However, the RecBCD (AddAB) enzyme cannot load onto DNA ends that are not blunt or nearly blunt (8, 9, 20, 22, 25). We propose that RecN binds to the 3' ssDNA tails of duplex DNA of the lagging-strand template. Then, the region of 3' ssDNA is enlarged on the lagging-strand template by the concerted action of RecQ-RecJ or RecS-RecJ helicase ssDNA exonuclease, whereas a 3' ssDNA of a DSB on a leading-strand template is de novo generated at blunt or nearly blunt ends by the AddAB helicase/nuclease. A model implying the different recombinational requirements at each end of the DNA DSB was proposed earlier (10). Further assembly of RecN onto the products of the processed break may facilitate the gathering of different DNA ends to form a discrete focus or repair center on the nucleoid (18, 19, 23). Next, RecO and RecA are successively recruited to the defined RecN focus, and RecA threads are formed. Alternatively, AddAB may load RecA onto the product of the break-processing reaction. These views are consistent with the following observations. (i) addAB recJ cells (this work) and addAB recF cells (1) are highly impaired in DNA repair. (ii) RecF was nonepistatic with RecJ or AddAB (this work). (iii) RecA is not loaded onto the DNA substrate in the absence of end processing (addAB recJ cells) (this work). (iv) RecN is the first recombination protein observed to be recruited to defined or random DSBs in live cells (this work). (v) In vitro, RecN binds and protects 3' ssDNA ends in the presence of ATP (30).
ACKNOWLEDGMENTS
This work was partially supported by grants BMC2003-00150 and BIO2001-4342-E from Direccion General de Investigacion and GR/SAL/0668/2004 from Comunidad de Madrid to J.C.A. and from the Deutsche Forschungsgemeinschaft to P.L.G.
We are very grateful to D. Camerini-Otero and S. C. Kowalczykowski for critical reading of the manuscript.
Present address: Institut für Mikrobiologie, Universitt Freiburg, Verfügungsgebude, Stefan-Meier-Str. 17, 79104 Freiburg, Germany.
REFERENCES
Alonso, J. C., A. C. Stiege, and G. Luder. 1993. Genetic recombination in Bacillus subtilis 168: effect of recN, recF, recH and addAB mutations on DNA repair and recombination. Mol. Gen. Genet. 239:129-136.
Alonso, J. C., R. H. Tailor, and G. Luder. 1988. Characterization of recombination-deficient mutants of Bacillus subtilis. J. Bacteriol. 170:3001-3007.
Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller, and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402.
Amundsen, S. K., and G. R. Smith. 2003. Interchangeable parts of the Escherichia coli recombination machinery. Cell 112:741-744.
Ceglowski, P., G. Luder, and J. C. Alonso. 1990. Genetic analysis of recE activities in Bacillus subtilis. Mol. Gen. Genet. 222:441-445.
Chedin, F., and S. C. Kowalczykowski. 2002. A novel family of regulated helicases/nucleases from gram-positive bacteria: insights into the initiation of DNA recombination. Mol. Microbiol. 43:823-834.
Chedin, F., P. Noirot, V. Biaudet, and S. D. Ehrlich. 1998. A five-nucleotide sequence protects DNA from exonucleolytic degradation by AddAB, the RecBCD analogue of Bacillus subtilis. Mol. Microbiol. 29:1369-1377.
Courcelle, J., and P. C. Hanawalt. 2003. RecA-dependent recovery of arrested DNA replication forks. Annu. Rev. Genet. 37:611-646.
Cox, M. M., M. F. Goodman, K. N. Kreuzer, D. J. Sherratt, S. J. Sandler, and K. J. Marians. 2000. The importance of repairing stalled replication forks. Nature 404:37-41.
Cromie, G. A., C. B. Millar, K. H. Schmidt, and D. R. Leach. 2000. Palindromes as substrates for multiple pathways of recombination in Escherichia coli. Genetics 154:513-522.
Fernandez, S., S. Ayora, and J. C. Alonso. 2000. Bacillus subtilis homologous recombination: genes and products. Res. Microbiol. 151:481-486.
Fernandez, S., A. Sorokin, and J. C. Alonso. 1998. Genetic recombination in Bacillus subtilis 168: effects of recU and recS mutations on DNA repair and homologous recombination. J. Bacteriol. 180:3405-3409.
Harmon, F. G., and S. C. Kowalczykowski. 1998. RecQ helicase, in concert with RecA and SSB proteins, initiates and disrupts DNA recombination. Genes Dev. 12:1134-1144.
Horii, Z., and A. J. Clark. 1973. Genetic analysis of the recF pathway to genetic recombination in Escherichia coli K12: isolation and characterization of mutants. J. Mol. Biol. 80:327-344.
Ivancic-Bace, I., P. Peharec, S. Moslavac, N. Skrobot, E. Salaj-Smic, and K. Brcic-Kostic. 2003. RecFOR function is required for DNA repair and recombination in a RecA loading-deficient recB mutant of Escherichia coli. Genetics 163:485-494.
Ivancic-Bace, I., E. Salaj-Smic, and K. Brcic-Kostic. 2005. Effects of recJ, recQ, and recFOR mutations on recombination in nuclease-deficient recB recD double mutants of Escherichia coli. J. Bacteriol. 187:1350-1356.
Khakhar, R. R., J. A. Cobb, L. Bjergbaek, I. D. Hickson, and S. M. Gasser. 2003. RecQ helicases: multiple roles in genome maintenance. Trends Cell Biol. 13:493-501.
Kidane, D., and P. L. Graumann. 2005. Dynamic formation of RecA filaments at DNA double strand break repair centers in live cells. J. Cell Biol. 170:357-366.
Kidane, D., H. Sanchez, J. C. Alonso, and P. L. Graumann. 2004. Visualization of DNA double-strand break repair in live bacteria reveals dynamic recruitment of Bacillus subtilis RecF, RecO and RecN proteins to distinct sites on the nucleoids. Mol. Microbiol. 52:1627-1639.
Kowalczykowski, S. C. 2000. Initiation of genetic recombination and recombination-dependent replication. Trends Biochem. Sci. 25:156-165.
Kowalczykowski, S. C., D. A. Dixon, A. K. Eggleston, S. D. Lauder, and W. M. Rehrauer. 1994. Biochemistry of homologous recombination in Escherichia coli. Microbiol. Rev. 58:401-465.
Kuzminov, A. 1999. Recombinational repair of DNA damage in Escherichia coli and bacteriophage lambda. Microbiol. Mol. Biol. Rev. 63:751-813.
Lisby, M., and R. Rothstein. 2004. DNA damage checkpoint and repair centers. Curr. Opin. Cell Biol. 16:328-334.
Lovett, S. T., and R. D. Kolodner. 1989. Identification and purification of a single-stranded-DNA-specific exonuclease encoded by the recJ gene of Escherichia coli. Proc. Natl. Acad. Sci. USA 86:2627-2631.
Michel, B., G. Grompone, M. J. Flores, and V. Bidnenko. 2004. Multiple pathways process stalled replication forks. Proc. Natl. Acad. Sci. USA 101:12783-12788.
Morimatsu, K., and S. C. Kowalczykowski. 2003. RecFOR proteins load RecA protein onto gapped DNA to accelerate DNA strand exchange: a universal step of recombinational repair. Mol. Cell 11:1337-1347.
Morozov, V., A. R. Mushegian, E. V. Koonin, and P. Bork. 1997. A putative nucleic acid-binding domain in Bloom's and Werner's syndrome helicases. Trends Biochem. Sci. 22:417-418.
Nakayama, H., K. Nakayama, R. Nakayama, N. Irino, Y. Nakayama, and P. C. Hanawalt. 1984. Isolation and genetic characterization of a thymineless death-resistant mutant of Escherichia coli K12: identification of a new mutation (recQ1) that blocks the RecF recombination pathway. Mol. Gen. Genet. 195:474-480.
Ryder, L., M. C. Whitby, and R. G. Lloyd. 1994. Mutation of recF, recJ, recO, recQ, or recR improves Hfr recombination in resolvase-deficient ruv recG strains of Escherichia coli. J. Bacteriol. 176:1570-1577.
Sanchez, H., and J. C. Alonso. 2005. Bacillus subtilis RecN binds and protects 3'-single-stranded DNA extensions in the presence of ATP. Nucleic Acids Res. 33:2343-2350.
Sanchez, H., D. Kidane, P. Reed, F. A. Curtis, M. C. Cozar, P. L. Graumann, G. J. Sharples, and J. C. Alonso. The RuvAB branch migration translocase and RecU Holliday junction resolvase are required for double-stranded DNA break repair in Bacillus subtilis. Genetics, in press. (Online.) doi:10.1534/genetics.105.045906.
Shan, Q., J. M. Bork, B. L. Webb, R. B. Inman, and M. M. Cox. 1997. RecA protein filaments: end-dependent dissociation from ssDNA and stabilization by RecO and RecR proteins. J. Mol. Biol. 265:519-540.
Sherratt, D. J., B. Soballe, F. X. Barre, S. Filipe, I. Lau, T. Massey, and J. Yates. 2004. Recombination and chromosome segregation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359:61-69.
Taylor, A. F., and G. R. Smith. 1992. RecBCD enzyme is altered upon cutting DNA at a chi recombination hotspot. Proc. Natl. Acad. Sci. USA 89:5226-5230.
Umezu, K., N. W. Chi, and R. D. Kolodner. 1993. Biochemical interaction of the Escherichia coli RecF, RecO, and RecR proteins with RecA protein and single-stranded DNA binding protein. Proc. Natl. Acad. Sci. USA 90:3875-3879.
Umezu, K., K. Nakayama, and H. Nakayama. 1990. Escherichia coli RecQ protein is a DNA helicase. Proc. Natl. Acad. Sci. USA 87:5363-5367.
Webb, B. L., M. M. Cox, and R. B. Inman. 1997. Recombinational DNA repair: the RecF and RecR proteins limit the extension of RecA filaments beyond single-strand DNA gaps. Cell 91:347-356.(Humberto Sanchez, Dawit K)