, 2008). Deletion of the intervening sequence by recombination between the repeats yields a functional kanamycin-resistance gene. With this construct, 90% of the deletion events occurring spontaneously are dependent on a functional RecA (Table 1 and Marsin et al., 2008). As shown in Table 2, inactivation of addB resulted in a 40% reduction in recombination rates. This value is comparable to the one obtained in the single addA mutant (Marsin et al., 2008), suggesting that AddA and AddB are epistatic. In order to evaluate the relative contributions of the two pathways to intrachromosomal Metformin cell line recombination, we introduced the recombination substrate into the recR gene, disrupting it (recR∷KDA). The recombination

rate in this case is slightly higher (Table 1) than the one obtained when the substrate was located in rdx (Table 1) probably due to sequence context. Inactivation of recO did not affect the rate obtained in the single recR mutant, again confirming the notion that recO and recR are likely to act as a complex in H. pylori. Conversely, the inactivation

of addB reduced the rate of intrachromosomal recombination of the recR mutant by an E7080 additional 60% (Table 1). This result indicates that during spontaneous recombination of direct chromosomal repeats, both RecOR- and AddAB-dependent presynaptic pathways can act, but they do so in an additive way. It is tempting to speculate that the initial event, i.e. the formation of a gap or a ds break, will determine which presynaptic complex initiates recombination. During natural transformation, H. pylori can integrate exogenous DNA into its chromosome by HR. This process is

dependent on a functional RecA (Schmitt et al., 1995); however, in strain 26695, the absence of either HR initiation complexes does not impair the integration process (Amundsen et al., 2008; Marsin et al., 2008). Consistently, Table 2 shows that disruption of addB did not reduce the frequency of transformation with chromosomal DNA carrying a mutation conferring resistance to streptomycin. Moreover, similar to what we have reported for the addA mutant, the transformation frequency in the addB mutant was fivefold higher than that in the HSP90 wild-type strain. The double addAB mutant also had an elevated transformation frequency (Table 2), indicating that the AddAB complex might act as a suppressor of transformation. This adds the AddAB complex to RecG (Kang et al., 2004), UvrD (Kang & Blaser, 2006) and MutS2 (Pinto et al., 2005) in the list of DNA metabolism proteins suppressing transformation in H. pylori. While inactivation of RexAB, the functional homologue of AddAB in Streptococcus pneumoniae, did not significantly affect chromosomal transformation (Halpern et al., 2004), no data are available on mutants defective in the other presynaptic pathway. In the other transformation model system, B.