Reply to “dupA ₁ Is Associated with Duodenal Ulcer and High Interleukin-8 Secretion from the Gastric Mucosa”

REPLY

We appreciate the comments from Hussein et al. DupA has been implicated in the risk of developing duodenal ulcers; however, the mechanisms remain unclear. We hypothesize that the dupA cluster forms a novel functional type IV secretion system (T4SS), although direct confirmatory evidence is still lacking. Several vir gene homologues (virB8, virB9, virB10, virB11, virD4, and virD2) are present before and after the dupA locus, which has been proposed as a novel putative T4SS (tfs3a) (6). If so, strains with both dupA and a complete tfs3a may have increased pathogenicity related to the presence of the T4SSs (6). Our study revealed many mosaic patterns of vir gene homologues near dupA and suggested that these different patterns may be responsible for different degrees of virulence. We first showed that the presence of a complete dupA cluster increased the duodenal ulcer risk compared to that with the presence of incomplete clusters and that gastric mucosal IL-8 levels were the highest among patients infected with H. pylori possessing a complete dupA cluster. We proposed that the presence of the complete dupA cluster, rather than dupA alone, was possibly involved in the development of duodenal ulcers and increased IL-8 productions. By analogy, a complete dupA cluster, like an intact cag pathogenicity island, is needed for full expression of the virulence factor (1, 2, 4).

Several fully sequenced strains (e.g., Shi470, Cuz20, and SouthAfrica7) show a complete dupA cluster (tfs3a). However, strain G27 contains a tfs3a that is unlikely to be functional in that dupA has a frameshift mutation. Furthermore, recent fully sequenced data of H. pylori have shown that the length of dupA varies between strains (i.e., strains Shi470 and Cuz20 have an open reading frame of approximately 2,500 bp that is approximately 600 bp longer than that of strain J99 due to an addition to the N-terminal region of dupA) (5). These data suggest that dupA may have two genotypes according to the signal sequence of the N-terminal region. None of the previous studies took this N-terminal region into account. We believe that a functional dupA should rely on detecting intact DupA protein by using immunoblotting techniques described in our Discussion (3). We propose that only the complete dupA cluster with an intact, long dupA is likely to express full virulence. Further study is necessary to elucidate how dupA might play a role in the development of duodenal ulcer disease.