To detect genes that were less abundant or absent in gastric cancer-associated H. pylori, PCR products

of the L library inserts were arrayed on nylon membranes and hybridized with DIG-labeled L301 or B975 digested DNAs (data not shown). Nine positive clones of superficial gastritis-specific DNAs were selected and sequenced. Homology analysis reveals that the less abundant cancer-specific genes belong to several functional groups (Table 2). These include (1) nucleotide transport and metabolism (clone 67), (2) cellular processing and signaling (clones 86, 128 and 140), (3) metabolism (clone 24), (4) information storage and processing (clones 99 and 117) and (5) function-unknown (clones 5 and 74). To further confirm that the positive genes as shown in Table 1 were gastric caner-specific, we screened the genes in 64 H. pylori strains that were isolated PD-0332991 solubility dmso either from gastric cancer patients (22 strains) or from superficial gastritis patients (42 strains). Among the 12 positive high-copy genes, we focused on clone 35 PPIase, because PPIases has been characterized as a virulence factor of L. pneumophila and T. cruzi (Fischer et al., 1992; Pereira et al., 2002), and it seems that PPIase plays HCS assay an important role in H. pylori-induced epithelial cell damage (Basak et al., 2005). PCR-based screening analysis showed that 11 out of 22 gastric cancer-associated strains were positive for PPIase. In contrast,

only 10 out of 42 superficial gastritis-associated strains were positive for PPIase (Table 3). Among the other high-copy genes, clone 88 encoding tyrosyl-tRNA synthetase was

also statistically associated with gastric cancer-associated strains. PCR-based screening analysis showed that 17 out of 22 gastric cancer-associated strains were positive for clone 88. In contrast, only Glutathione peroxidase 21 out of 42 superficial gastritis-associated strains were positive for clone 88 (Table 3). The absence of clones 86 and 128 encoding flagellar hook protein (see Table 2) in the gastric cancer-associated strain as detected by dot blot analysis was interesting and suggested that loss of flagellar genes may be a feature of gastric cancer-associated strains. To test this idea, H. pylori strains isolated either from superficial gastritis or from gastric cancer patients were screened to detect clones 86 and 128 genes using PCR-based screening analysis. The screening results revealed that although the percentage of flagellar gene-positive superficial gastritis strains was higher than that of gastric cancer-associated strains for both clones, the difference in the absence of the flagellar genes between gastric cancer-associated and nongastric cancer-associated H. pylori strains did not reach statistical significance. Thus, our result does not support the idea that loss of flagellar genes is a feature of the gastric cancer-associated strain.