**4.3** *H. pylori* **RecN**

8 DNA Repair

strains), *recA* mutant bacteria were never recovered, while some *addA* or *addB* mutant

In addition to the synapsis protein RecA, the genes for post-synapsis proteins (RuvABC and RecG) are also well conserved among bacteria [72]. Genes for RuvABC proteins are present in *H. pylori*, thus *H. pylori* seems to be able to restore Holliday Junctions in a similar way to *E. coli*. RuvC is a Holliday junction endonuclease that resolves recombinant joints into nicked duplex products. A *ruvC* mutant of *H. pylori* was more sensitive (compared to the wild type) to oxidative stress and other DNA damaging agents including UV light, mitomycin C, levofloxacin and metronidazole [73]. As Macrophage cells are known to produce an oxidative burst to kill bacterial pathogens, the survival of *H. pylori ruvC* mutant within macrophages was shown to be 100-fold lower than that of the wild type strain [73]. Furthermore, mouse model experiments revealed that the 50% infective dose of the *ruvC* mutant was approximately 100-fold higher than that of the wild-type strain. Although the *ruvC* mutant was able to establish colonization at early time points, infection was spontaneously cleared from the murine gastric mucosa over long periods (36 to 67 days) [73]. This was the first experimental evidence that DNA recombination processes are important for establishing and maintaining long-term *H. pylori* infection. Further studies suggested that RuvC function and, by inference, recombination facilitate bacterial immune evasion by altering the adaptive immune response [74], although the underlying

RuvAB proteins are involved in the branch migration of Holliday junctions. The annotated *H. pylori* RuvB (HP1059) showed extensive homology (52% sequence identity) to *E. coli* RuvB, particularly within the helicase domains. However, unlike in *E. coli, ruvA, ruvB*, and *ruvC* are located in separate regions of the *H. pylori* chromosome, which may predict possible functional differences. In contrast to *E. coli ruvB* mutants, which have moderate susceptibility to DNA damage, the *H. pylori ruvB* mutant has intense susceptibility to UV, similar to that of a *recA* mutant [75]. Similarly, the *H. pylori ruvB* mutant has a significantly diminished MIC (minimal inhibitory concentration) for ciprofloxacin, an agent that blocks DNA replication fork progression, to the same extent as the *recA* mutant. In agreement with these repair phenotypes, the *ruvB* mutant has almost completely lost the ability of natural transformation of exogenous DNA (frequency of <10−8), similar to the *recA* mutant. In an assay measuring the intra-genomic recombination (deletion frequency between direct repeats), the *ruvB* mutants displayed significantly (four- to sevenfold) lower deletion frequencies than the background level. All four phenotypes of the *ruvB* mutant suggested that *H. pylori* RuvAB is the predominant pathway for branch migration in DNA

In *E. coli*, an alternative pathway processing branch migration of Holliday junctions is the RecG helicase. In marked contrast to *E. coli, H. pylori recG* mutants do not have defective DNA repair, as measured by UV-light sensitivity and ciprofloxacin susceptibility [76]. Furthermore, *H. pylori recG* mutants have increased frequencies of intergenomic recombination and deletion, suggesting that branch migration and Holliday junction resolution are more efficient in the absence of RecG function [75, 76]. Thus, the effect of *H. pylori* RecG seems to be opposite to that of the RuvAB helicase. In the RuvABC pathway, the RuvC endonuclease nicks DNA, catalyzing Holliday junction resolution into doublestranded DNA. Although the resolvase in the RecG pathway has not been completely

bacteria were recovered from mouse stomachs.

mechanisms remain obscure.

recombinational repair [75].

**4.2 Post-synapsis proteins RuvABC and RecG** 

Bacterial RecN is related to the SMC (structure maintenance of chromosome) family of proteins in eukaryotes, which are key players in a variety of chromosome dynamics, from chromosome condensation and cohesion to transcriptional repression and DNA repair [78]. SMC family proteins have a structural characteristic of an extensive coiled-coil domain located between globular domains at the N- and C-termini that bring together Walker A and B motifs associated with ATP-binding [79]. *E. coli* RecN is strongly induced during the SOS response and was shown to be involved in RecA-mediated recombinational repair of DSBs [64]. In *Bacillus subtilis*, RecN was shown to be recruited to DSBs at an early time point during repair [63, 80, 81]. In vitro, RecN was shown to bind and protect 3' ssDNA ends in the presence of ATP [82].

In the published *H. pylori* genome sequence [12], HP1393 was annotated as a *recN* gene homolog. The *H. pylori recN* mutant is much more sensitive to mitomycin C, an agent that predominantly causes DNA DSBs, indicating RecN plays an important role in DSB repair in *H. pylori* [83]. In normal laboratory growth conditions, an *H. pylori recN* mutant does not show a growth defect, but its survival is greatly reduced under oxidative stress which resembles the *in vivo* stress condition. While very little fragmented DNA was observed in either wild type or *recN* mutant strain when cells were cultured under normal microaerobic conditions; after oxidative stress treatment the *recN* mutant cells had a significantly higher proportion of the DNA as fragmented DNA than did the wild type [83]. Similar roles of RecN in protection against oxidative damage have been demonstrated in *Neisseria gonorrhoeae* [84, 85]. In addition, the *H. pylori recN* mutant is much more sensitive to low pH than the wild type strain, suggesting that RecN is also involved in repair of acid-induced DNA damage [83]. This could be relevant to its physiological condition, as *H. pylori* appears to colonize an acidic niche on the gastric surface [41].

As mentioned in the sections above, loss of *H. pylori* RecA, RuvB or RuvC functions results in a great decrease of DNA recombination frequency. Similarly, the *H. pylori recN* mutant has a significant decrease of DNA recombination frequency, suggesting that RecN is a critical factor in DNA recombinational repair [83]. In contrast, loss of UvrD or MutS2 in *H. pylori* resulted in an increase of DNA recombination frequency [46, 48]. Suppression of DNA recombination by UvrD or MutS2, and facilitation of DNA recombination by RecN, may play a role in coordinating DNA repair pathways. Recombinational repair could be mutagenic due to homeologous recombination or cause rearrangement due to recombination with direct repeat sequences. In addition, recombinational repair systems are much more complex and require more energy to operate, compared to nucleotide excision repair (NER) and base excision repair (BER) systems. Thus UvrD, as a component of NER, and MutS2 as a likely component of a BER (8-oxoG glycosylase) system [49], both suppress DNA recombination. Both NER and BER systems would be expected to continuously function in low stress conditions. Under a severe stress condition when large amounts of

A Recombination Puzzle Solved: Role for New

single-strain infections [69, 92].

*H. pylori* transformation frequency [27].

**4.5** *H. pylori* **RecRO pathway** 

interesting question.

DNA Repair Systems in *Helicobacter pylori* Diversity/Persistence 11

level of sensitivity was similar to that seen for a *recA* mutant, but more severe than for the *recN* mutant. It is thus concluded that AddAB plays a major role in the repair of DNA ds breaks [69, 92]. On the other hand, the *addA* and *addB* mutants were markedly less sensitive to UV irradiation than a *recA* mutant, suggesting that AddAB does not play a major role in repair of UV damage in *H. pylori* [69]. AddA was shown to be important for *H. pylori* protection against oxidative stress-induced damage, as the *addA* mutant cells were significantly more sensitive to oxidative stress and contained a large amount of fragmented DNA [92]. Furthermore, loss of AddA resulted in reduced frequencies of apparent gene conversion between homologous genes encoding outer membrane proteins (*babA* to *babB*) [69]. Finally, it was shown that the *addA* and *addB* mutant strains display a significantly attenuated ability to colonize mouse stomachs, in both competition experiments and during

While *addA* and *addB* are adjacent in the chromosome in most bacteria, including other epsilon Proteobacteria, this is not the case in *H. pylori*. However, the phenotypes of *H. pylori addA* and *addB* mutants are indistinguishable. Thus, it was proposed [69] that the AddA and AddB act together in a complex, as do the RecBCD polypeptides and AddAB polypeptides of other bacteria. If so, the control of the unlinked *H. pylori addA* and *addB* genes to maintain the proper stoichiometry of the two polypeptides remains an

Regarding the role of *H. pylori* AddA in DNA recombination during natural transformation, conflicting results were reported from different studies. The *addA* (note: it was named *recB* in certain references) mutant showed enhanced [68, 70], decreased [20, 71, 92], or no change [27, 69] in transformation frequency. Indeed, a high degree of variability (>100-fold) in transformation frequency in *H. pylori* was observed between different strains and different experiments. The use of different assay systems may partly explain the discrepancy in transformation results. For example, the total genomic DNA from antibiotic-resistant strain was used for the transformation assay in certain studies, while in others the defined linear DNA fragments of small size [92]. Use of the transformation frequency as an indicator of DNA recombination frequency is based on the assumption that the wild type *H. pylori* and its isogenic *rec* strains are equally competent for DNA uptake. However, it is now known that this assumption is not valid because DNA damage triggers genetic exchange in *H. pylori* [55]. *H. pylori addA* mutant cells suffered more DNA damage [92], and have an enhanced competence for DNA uptake [55]. Thus, the accumulation of unrepaired DNA damage and subsequent poor growth, as well as unknown strain differences, could be the main cause of the high degree of variability in

RecFOR is a highly conserved DNA recombination pathway in bacteria, and is mainly used for ssDNA gap repair [72]. In the published *H. pylori* genome sequences, only the *recR* gene was annotated [12, 13]. Although RecF historically served as a reference for RecFOR pathway, it is absent from genomes of many bacteria including *H. pylori* [72]. By bioinformatics analysis, Marsin et al [68] identified HP0951 as a novel RecO orthologue, although its sequence identity with the *E. coli* protein is lower than 15%. Recent studies in *E. coli* indicated that RecOR in the absence of RecF can perform recombination by loading RecA [95, 96]. Whereas the RecO protein can displace ssDNA-binding protein (SSB) and

DSBs are formed, RecN perhaps recognizes DSBs and recruits proteins required for initiation of DNA recombination.

The role of *H. pylori* RecN *in vivo* has been demonstrated, as the *recN*-disrupted *H. pylori* cells are less able to colonize hosts than wild type cells [83]. However, the mouse colonization phenotype of the *recN* strain seems to be less severe than those observed for the *recA* or *ruvC* mutants. In contrast to RecA or RuvC which are major components of DNA recombination machinery, RecN is a protein specific for repairing DSBs by linking DSB recognition and DNA recombination initiation. It was proposed that the attenuated ability to colonize mouse stomachs by *recN* cells was mainly due to the strain's failure to repair DSBs through a DNA recombinational repair pathway.

#### **4.4 AddAB helicase-nuclease**

DNA helicases play key roles in many cellular processes by promoting unwinding of the DNA double helix [86]. Bacterial genomes encode a set of helicases of the DExx family that fulfill several, sometimes overlapping functions. Based on the sequence homology, bacterial RecB, UvrD, Rep, and PcrA were classified as superfamily I (SF1) helicases [86- 88]. In the well-studied *E. coli*, RecBCD form a multi-functional enzyme complex that processes DNA ends resulting from a double-strand break. RecBCD is a bipolar helicase that splits the duplex into its component strands and digests them until encountering a recombinational hotspot (Chi site). The nuclease activity is then attenuated and RecBCD loads RecA onto the 3' tail of the DNA [89]. Another bacterial enzyme complex AddAB, extensively studied in *Bacillus subtilis*, has both nuclease and helicase activities similar to those of RecBCD enzyme [90, 91].

The genes for RecBCD or AddAB were missing in the published *H. pylori* genome [12, 13]. However, HP1553 from strain 26695 was annotated as a gene encoding a putative helicase [12], and the corresponding gene from strain J99 was annotated as *pcrA* [13]. Amino acid sequence alignment of HP1553 to *E. coli* RecB (or to *B. subtilis* AddA) revealed 24% identity (to both heterologous systems) at the N-terminal half (helicase domain), and no significant homology at the C-terminal half (including nuclease domain). Thus, HP1553 could be a RecB (or AddA)-like helicase [69, 92]. Furthermore, by using the highly conserved AddB nuclease motif "GRIDRID" in BLAST search, HP1089 was identified as the putative AddB homolog [69]. Now it is accepted that HP1553 and HP1089 are termed *addA* and *addB* respectively in *H. pylori* with a reminder that previous *recB* [20, 68, 70, 92] was the equivalent of *addA* [69, 71, 93]. Both genes *addA* and *addB* are present in 56 *H. pylori* clinical isolates from around the world [94]; thus they are considered core genes that are not strain variable.

The biochemical activities of *H. pylori* AddAB helicase-nuclease have been demonstrated [69]. Cytosolic extracts from wild-type *H. pylori* showed detectable ATP-dependent nuclease activity with ds DNA substrate, while the *addA* and *addB* mutants lack this activity. Cloned *H. pylori addA* and *addB* genes express ATP-dependent exonuclease in *E. coli* cells. These genes also conferred ATP-dependent DNA unwinding (helicase) activity to an *E. coli recBCD* deletion mutant, indicating that they are the structural genes for this enzyme [69]. The roles of individual (helicase, exonuclease) activity of the AddA and AddB in DNA repair, recombination, and mouse infection have been further studied by site-directed mutagenesis approach [93].

*H. pylori addA* and *addB* mutant strains showed heightened sensitivity to mitomycin C and the DNA gyrase inhibitor ciprofloxacin, both of which lead to DNA ds breaks [69, 92]. The

DSBs are formed, RecN perhaps recognizes DSBs and recruits proteins required for

The role of *H. pylori* RecN *in vivo* has been demonstrated, as the *recN*-disrupted *H. pylori* cells are less able to colonize hosts than wild type cells [83]. However, the mouse colonization phenotype of the *recN* strain seems to be less severe than those observed for the *recA* or *ruvC* mutants. In contrast to RecA or RuvC which are major components of DNA recombination machinery, RecN is a protein specific for repairing DSBs by linking DSB recognition and DNA recombination initiation. It was proposed that the attenuated ability to colonize mouse stomachs by *recN* cells was mainly due to the strain's failure to repair

DNA helicases play key roles in many cellular processes by promoting unwinding of the DNA double helix [86]. Bacterial genomes encode a set of helicases of the DExx family that fulfill several, sometimes overlapping functions. Based on the sequence homology, bacterial RecB, UvrD, Rep, and PcrA were classified as superfamily I (SF1) helicases [86- 88]. In the well-studied *E. coli*, RecBCD form a multi-functional enzyme complex that processes DNA ends resulting from a double-strand break. RecBCD is a bipolar helicase that splits the duplex into its component strands and digests them until encountering a recombinational hotspot (Chi site). The nuclease activity is then attenuated and RecBCD loads RecA onto the 3' tail of the DNA [89]. Another bacterial enzyme complex AddAB, extensively studied in *Bacillus subtilis*, has both nuclease and helicase activities similar to

The genes for RecBCD or AddAB were missing in the published *H. pylori* genome [12, 13]. However, HP1553 from strain 26695 was annotated as a gene encoding a putative helicase [12], and the corresponding gene from strain J99 was annotated as *pcrA* [13]. Amino acid sequence alignment of HP1553 to *E. coli* RecB (or to *B. subtilis* AddA) revealed 24% identity (to both heterologous systems) at the N-terminal half (helicase domain), and no significant homology at the C-terminal half (including nuclease domain). Thus, HP1553 could be a RecB (or AddA)-like helicase [69, 92]. Furthermore, by using the highly conserved AddB nuclease motif "GRIDRID" in BLAST search, HP1089 was identified as the putative AddB homolog [69]. Now it is accepted that HP1553 and HP1089 are termed *addA* and *addB* respectively in *H. pylori* with a reminder that previous *recB* [20, 68, 70, 92] was the equivalent of *addA* [69, 71, 93]. Both genes *addA* and *addB* are present in 56 *H. pylori* clinical isolates from around the world [94]; thus they are considered core genes that

The biochemical activities of *H. pylori* AddAB helicase-nuclease have been demonstrated [69]. Cytosolic extracts from wild-type *H. pylori* showed detectable ATP-dependent nuclease activity with ds DNA substrate, while the *addA* and *addB* mutants lack this activity. Cloned *H. pylori addA* and *addB* genes express ATP-dependent exonuclease in *E. coli* cells. These genes also conferred ATP-dependent DNA unwinding (helicase) activity to an *E. coli recBCD* deletion mutant, indicating that they are the structural genes for this enzyme [69]. The roles of individual (helicase, exonuclease) activity of the AddA and AddB in DNA repair, recombination, and mouse infection have been further studied by site-directed mutagenesis

*H. pylori addA* and *addB* mutant strains showed heightened sensitivity to mitomycin C and the DNA gyrase inhibitor ciprofloxacin, both of which lead to DNA ds breaks [69, 92]. The

initiation of DNA recombination.

**4.4 AddAB helicase-nuclease** 

those of RecBCD enzyme [90, 91].

are not strain variable.

approach [93].

DSBs through a DNA recombinational repair pathway.

level of sensitivity was similar to that seen for a *recA* mutant, but more severe than for the *recN* mutant. It is thus concluded that AddAB plays a major role in the repair of DNA ds breaks [69, 92]. On the other hand, the *addA* and *addB* mutants were markedly less sensitive to UV irradiation than a *recA* mutant, suggesting that AddAB does not play a major role in repair of UV damage in *H. pylori* [69]. AddA was shown to be important for *H. pylori* protection against oxidative stress-induced damage, as the *addA* mutant cells were significantly more sensitive to oxidative stress and contained a large amount of fragmented DNA [92]. Furthermore, loss of AddA resulted in reduced frequencies of apparent gene conversion between homologous genes encoding outer membrane proteins (*babA* to *babB*) [69]. Finally, it was shown that the *addA* and *addB* mutant strains display a significantly attenuated ability to colonize mouse stomachs, in both competition experiments and during single-strain infections [69, 92].

While *addA* and *addB* are adjacent in the chromosome in most bacteria, including other epsilon Proteobacteria, this is not the case in *H. pylori*. However, the phenotypes of *H. pylori addA* and *addB* mutants are indistinguishable. Thus, it was proposed [69] that the AddA and AddB act together in a complex, as do the RecBCD polypeptides and AddAB polypeptides of other bacteria. If so, the control of the unlinked *H. pylori addA* and *addB* genes to maintain the proper stoichiometry of the two polypeptides remains an interesting question.

Regarding the role of *H. pylori* AddA in DNA recombination during natural transformation, conflicting results were reported from different studies. The *addA* (note: it was named *recB* in certain references) mutant showed enhanced [68, 70], decreased [20, 71, 92], or no change [27, 69] in transformation frequency. Indeed, a high degree of variability (>100-fold) in transformation frequency in *H. pylori* was observed between different strains and different experiments. The use of different assay systems may partly explain the discrepancy in transformation results. For example, the total genomic DNA from antibiotic-resistant strain was used for the transformation assay in certain studies, while in others the defined linear DNA fragments of small size [92]. Use of the transformation frequency as an indicator of DNA recombination frequency is based on the assumption that the wild type *H. pylori* and its isogenic *rec* strains are equally competent for DNA uptake. However, it is now known that this assumption is not valid because DNA damage triggers genetic exchange in *H. pylori* [55]. *H. pylori addA* mutant cells suffered more DNA damage [92], and have an enhanced competence for DNA uptake [55]. Thus, the accumulation of unrepaired DNA damage and subsequent poor growth, as well as unknown strain differences, could be the main cause of the high degree of variability in *H. pylori* transformation frequency [27].

#### **4.5** *H. pylori* **RecRO pathway**

RecFOR is a highly conserved DNA recombination pathway in bacteria, and is mainly used for ssDNA gap repair [72]. In the published *H. pylori* genome sequences, only the *recR* gene was annotated [12, 13]. Although RecF historically served as a reference for RecFOR pathway, it is absent from genomes of many bacteria including *H. pylori* [72]. By bioinformatics analysis, Marsin et al [68] identified HP0951 as a novel RecO orthologue, although its sequence identity with the *E. coli* protein is lower than 15%. Recent studies in *E. coli* indicated that RecOR in the absence of RecF can perform recombination by loading RecA [95, 96]. Whereas the RecO protein can displace ssDNA-binding protein (SSB) and

A Recombination Puzzle Solved: Role for New

an enormous genetic diversity.

oxidative and acid stress.

**6. Acknowledgements** 

future biochemical and genetic approaches.

R21AI076569 and by the University of Georgia Foundation.

**5. Concluding remarks and perspectives** 

DNA Repair Systems in *Helicobacter pylori* Diversity/Persistence 13

Severe *Helicobacter pylori*-mediated gastric diseases are associated with the bacterium's persistence in the host and its adaptability to host differences, which in turn is associated with its remarkable genetic variability. DNA recombination is an extraordinarily frequent event in *H. pylori*, and this manifests itself into a bacterium with unusual flexibility in stresscombating enzymes, repair mechanisms, and other adaptability characteristics. Nearly every *H. pylori* recombination-related gene studied thus far by a gene directed mutant analysis approach has documented they are individually important in stomach colonization ability; this underscores the importance of these recombination repair processes in bacterial survival in the host. It is well recognized that homologous DNA recombination is a special system in bacteria for repairing stalled replication forks and double strand breaks, while generating genetic diversity as an advantageous byproduct [102]. *H. pylori* may be an especially fruitful organism in which to learn the ultimate boundaries in roles of recombination repair enzymes, as *H. pylori* is subject to intense and prolonged host mediated stress and it displays

Substantial progress has been made recently in unraveling the complex systems of DNA recombinational repair in *H. pylori*. As expected, whole genome sequencing has been a powerful tool to aid in identifying recombination-related proteins in *H. pylori*. For example, *recA, recR, recN*, and *ruvABC* were identified and confirmed to play important roles in *H. pylori* as could be expected from results for other bacteria. Some recombination-related proteins (e.g. MutS2, RecG), however, play unique roles in *H. pylori*. Most of the genes for the major components of the two pre-synapsis pathways (RecBCD and RecFOR) were not annotated from *H. pylori* genome sequences, which drove researchers' interest to search for additional novel systems required for *H. pylori* DNA recombinational repair. Recent studies revealed the existence of both pathways, AddAB and RecRO, in *H. pylori*. Although they display a limited level of sequence homology to the known recombination enzymes, both AddAB and RecRO were shown to play important roles in *H. pylori* DNA recombinational repair, conferring resistance to

The major components of DNA recombinational repair machinery in *H. pylori* are listed in Table 1. *H. pylori* RecN protein may recognize DNA double strand breaks and recruits AddAB helicase-nuclease complex for further processing. While not being involved in repair of DNA double strand breaks, *H. pylori* RecRO proteins play a major role in intra-genomic recombination at repeat sequences. Both pre-synapsis pathways (AddAB and RecRO) require RecA for catalyzing DNA strand exchange (synapsis) and *H. pylori* RuvABC is the predominant pathway for DNA branch migration and Holliday Junction resolution (postsynapsis). Although the major functions of these components are similar to those observed in model bacteria, some novel attributes of these components have been discovered, which may be related to the highly-specific lifestyle of *H. pylori*. Additional new components that work synergistically with these pathways could be found in this unique bacterium via

The work on *H. pylori* DNA repair in our laboratory was supported by NIH grant

bind to ssDNA, RecR is the key component for loading RecA onto ssDNA [95, 97]. Likely, the RecRO pathway (with no RecF) is present in *H. pylori*.

The *recR* and *recO* mutants showed marked sensitivity to DNA damaging agents metronidazole and UV light, indicating roles of RecR and RecO in DNA repair. Unlike the *addA* (*recB*) mutant, the *recR* and *recO* mutants did not show significant sensitivity to ionizing radiation (IR) and to mitomycin C [68, 71], suggesting that RecRO pathway is not responsible for repairing DNA damage induced by these agents, most likely double strand breaks. This is in contrast to *E. coli* where the RecFOR pathway sometimes substitutes for the RecBCD pathway and in *Deinococcus radiodurance* where the RecFOR pathway plays a major role in double strand break repair [98, 99]. On the other hand, *H. pylori recR* and *recO* mutants were shown to be much more sensitive to oxidative stress and to acid stress than the wild type strain [71], indicating that *H. pylori* RecRO pathway is involved in repairing DNA damage induced by these stress conditions. The *addA recO* double mutant (deficient in both AddAB and RecRO pathways) was significantly more sensitive to atmospheric oxygen than the *recO* single mutant, indicating that both RecRO and AddAB pathways are important for survival of oxidative damage. Similar roles of the RecBCD and the RecFOR pathways for survival of oxidative damage were also observed in *E. coli* [57, 100] and in *Neisseria gonorrhoeae* [84]. In those bacteria, however, the RecBCD appeared to be the predominant (over the RecFOR) repair pathway for oxidative damage. Our results suggest that the two pathways in *H. pylori* play similarly important roles in repairing oxidative stress-derived DNA damage [71]. In accordance with the sensitivity to oxidative and acid stress in vitro, *H. pylori recR* and *recO* mutants were shown to be less able to colonize mouse stomachs [71]. Furthermore, the mouse colonization ability of the *addA recO* double mutant was significantly lower than that of the *addA* or *recO* single mutant. Therefore, both AddAB- and RecRO-mediated DNA recombinational repair in *H. pylori* play an important role in bacterial survival and persistent colonization in the host.

Although differing results regarding the effect of *addA* gene on transformation frequency were reported by different research groups, it was agreed that the RecRO-pathway is not involved in recombination of exogenous DNA into the *H. pylori* genome in the process of transformation [68, 71]. The RecRO pathway is known to have a major role in intragenomic recombination at repeat sequences [101]. Using an assay to assess the deletion frequency resulting from recombination on direct repeat sequences (358 bp long), Marsin et al [68] showed that the *recR* and *recO* mutants exhibited a statistically significantly lower deletion frequency than the wild type strain, suggesting a role of RecRO in intragenomic recombination. Recently we adopted a similar assay using DNA constructs (deletion cassettes) that contain identical repeat sequences of different length (IDS100 and IDS350) [71]. The results indicated that the intra-genomic recombination of 100 bp-long direct repeat sequences in *H. pylori* is partially dependent on RecR and RecA, yet a large portion of the recombination event is RecA-independent. This is basically in agreement (with small variance) with the results of Aras et al [35] who reported that the repeat sequences of 100 bp or shorter recombined through a RecA-independent pathway. For the deletion cassette containing repeat sequences of 350 bp in length, inactivation of *recR* or *recA* resulted in a significant 4-fold or 35-fold decrease respectively in deletion frequency, indicating that RecR plays a significant role in recombination of IDS350, while this recombination was highly dependent on RecA.
