coli both constitutively and in response to H2O2 treatment (Figure 4 and Table 2). Our further analysis on the messenger RNA level of fliC indicates that the RNA levels are higher in the ΔarcA mutant E. coli and corresponded mTOR inhibitor to the ISRIB protein levels, suggesting that the regulation is likely on the transcriptional or post-transcriptional level (Figure 5). Oshima et al. did not detect a significant alteration in the expression of fliC in their microarray analysis, although flagellar synthesis was identified as a system that was affected in the ΔarcA mutant but not the ΔarcB mutant E. coli [23]. The discrepancy is possibly due to the differences in experimental conditions (shaking

bacterial cultures at 120 rpm vs. 225 rpm) and detection methods (microarray vs. Real-Time Reverse Transcriptase PCR and 2-D gel electrophoresis). Since we detected an elevation of both

mRNA and protein levels ALK inhibitor of flagellin in the ΔarcA mutant E. coli (Figures 4 and 5), we believe that our observation is valid. The regulation of ArcA on flagellin is likely to be indirect, as we did not detect specific binding of recombinant ArcA protein to the upstream sequence of fliC (data not shown). Given that the ArcAB system regulates a large number of genes in E. coli, its role in the ROS resistance is likely to be complex. We have demonstrated that mutation of ArcA or ArcB did not alter the H2O2 scavenging ability of E. coli (Figure 2), however, the precise molecular mechanism on how ArcA regulates ROS resistance in E. coli is yet to be elucidated. ArcA was reported to be necessary for the ROS resistance of Haemophilus influenzae due to its regulation of Dps, a ferritin-like small protein that was previously reported to be involved in ROS resistance of Salmonella [39, 47]. The mechanism

of the ROS resistance mediated by ArcA is likely to be different in E. coli, since dps is expressed close to the wild type level in the ΔarcA or ΔarcB mutant (84% and 99% respectively), and our preliminary microarray analysis with Salmonella ΔarcA mutant indicated that dps responded old normally to H2O2 in the ΔarcA mutant (unpublished results). One possible clue on the mechanism of how ArcAB contributes to the ROS resistance of E. coli came from our proteomic analysis that showed altered expression of flagellin, GltI and OppA between the wild type and ΔarcA mutant E. coli (Table 2). The constitutive GltI and OppA levels are higher in the ΔarcA mutant than in the wild type E. coli, suggesting that the mutant may have a higher need for amino acid transport. In contrast to the GltI and OppA levels in the wild type E. coli that increased 6- and 24-fold respectively in response to H2O2 exposure (possibly due to a higher need for amino acid transport under ROS stress), the level of neither protein in the ΔarcA mutant increased under the same condition (Table 2).