The occurrence of intragenomic recombination between babA and babB has been demonstrated in in vitro and in vivo experiments, implicating this mechanism may possibly assist H. pylori to adapt in the human stomach [12, 14]. In addition, mixed genotypes of babA and babB at locus A or B have been demonstrated . The clinical association of the babA and babB genotype of H. pylori strains and genetic profile with infections of the antrum and corpus of a single host are still unclear. In this study, we demonstrated that the AB AB genotype, one dominant genotype in the antrum, was associated with the precancerous lesion as IM, and correlated with gastric cancer. However, H. pylori infection by such AB AB genotype has not lead into a more dense colonization or inflammation severity in gastric histology. Our data indicate H. pylori babA and babB genotypes as AB AB should at least exert with better adaptation to gastric environment during carcinogenesis.
Colbeck et al.  found 9 genotypes (A B, AB B, A AB, A A, B B, B A, B C, C B and B AB) in their study. Nevertheless, our study only found four genotypes (A B, A AB, AB B and AB AB) in the 168 isolates from 19 patients’ antrum and corpus (Table 2). It indicates the genotype diversity of babAB in Taiwanese isolates could be obviously less complicate. Moreover, at least one babA gene at locus A existed in each of the isolates. This finding is compatible with our previous report to reveal the Taiwanese H. pylori isolates are nearly 100% babA-positive , and support the higher prevalence of babA in H. pylori strains from East Asian countries than those from western worlds . Moreover, Matteo et al.  demonstrated that 9 of 34 patients (26.5%) had bab gene variation across the antrum and corpus of a single host at a specific time point. We found that 12 of 19 patients (63.2%) infected by more than one genotype in either one or both gastric niches. The prevalence discrepancy between two studies could be due to the analysis of bab genotype from the bacterial pool or single-colony isolate.
Analysis of the sequences of babA and babB revealed that nonsynonymous substitutions of amino acids occurred between the individual strains (Figure 2, Table 3 and data not shown), but did not differ between the gastric niches. Pride et al.  also showed high allelic diversity within babA and babB in the strains from different patients. Judging by the 6 different nonsynonymous substitutions of amino acid 161 in the 6 patients’ strains, that codon was a highly variable site. This is worth further investigation, as it may be a special site responsible for adapting to differences in individual stomachs.
CT repeats in the 5’ coding region of babA and babB are more commonly found at locus B than locus A . We found that the corpus isolates had a higher frequency of changes in number of CT repeats of babB at locus B than the antrum isolates (Table 4). Among those 7 patients infected by the corpus isolates with a change of CT repeats, only one (patient no. 27) had the isolate changing CT repeats to in-frame (CT = 8) (Table 4). This data indicates that BabB expression could be tightly controlled by phase variation due to out of frame repeats in the corpus. Among 12 patients infected by isolates with more than one genotype, their isolates from antrum have a higher rate of A B and AB AB as dominant genotypes than corpus (9/12 vs. 2/12, p < 0.05). Moreover, half of those patients lacked a dominant genotype in their corpus isolates. These results suggest the environment in the corpus may favor different adaptation for the isolates with different H. pylori genetic diversities.
The presence of the AB AB genotype was higher in GC patients with older age (Table 1). In addition, the AB AB genotype is not correlated with more severe inflammation or precancerous changes in the non-cancer patients. Based on this cross-sectional clinical histological data, it suggests the AB AB strains may have a better adaptation to the cancerous environment in stomach, instead of leading into more toxicity in gastric carcinogenesis. In Figure 3A, we show that the babA gene at locus A dominantly determines BabA expression, and the mixed genotype as AB at locus A may decrease the BabA expression (Figure 3B and 3C). It is thus possible a mixed genotype as AB at locus A may make H. pylori isolates to contain a subpopulation losing BabA expression. Alternatively, the mixed genotype as AB at locus B may possibly allow H. pylori to change BabB expression and thus deserves further study.
In addition, our previous data have shown that the intensity of Lewis b become decreased in antrum atrophy, but can be preserved in corpus to mediate higher colonization of bug overthere . So, it shall be also implicative to test whether the AB AB genotype dominantly in antrum can have advantage to adapt the gastric epithelium with weak Lewis b expression in future.