The ArcA regulon and oxidative stress resistance in Haemophilus influenzae

Abstract

Haemophilus influenzae transits between niches within the human host that are predicted to differ in oxygen levels. The ArcAB two-component signal transduction system controls gene expression in response to respiratory conditions of growth and has been implicated in bacterial pathogenesis, yet the mechanism is not understood. We undertook a genome-scale study to identify genes of the H. influenzae ArcA regulon. Deletion of arcA resulted in increased anaerobic expression of genes of the respiratory chain and of H. influenzae's partial tricarboxylic acid cycle, and decreased anaerobic expression levels of genes of polyamine metabolism, and iron sequestration. Deletion of arcA also conferred a susceptibility to transient exposure to hydrogen peroxide that was greater following anaerobic growth than after aerobic growth. Array data revealed that the dps gene, not previously assigned to the ArcA modulon in bacteria, exhibited decreased expression in the arcA mutant. Deletion of dps resulted in hydrogen peroxide sensitivity and complementation restored resistance, providing insight into the previously uncharacterized mechanism of arcA-mediated H(2)O(2) resistance. The results indicate a role for H. influenzae arcA and dps in pre-emptive defence against transitions from growth in low oxygen environments to aerobic exposure to hydrogen peroxide, an antibacterial oxidant produced by phagocytes during infection.

Fig. 1

Differential transcript abundance in the H. influenzae arcA mutant versus parental strain. Northern blots containing 6 μg of total RNA from anaerobically grown Rd parental strain, RdV, arcA deletion mutant, RAA6V (ΔarcA) and complemented strain, RAA6C hybridized with probes corresponding to genes increased (A) or decreased (B) in expression in the ΔarcA mutant (Tables 1 and 2 respectively). Ethidium bromide stained gel is directly below each blot. 16S and 23S are the ribosomal RNAs. C. Schematic representation of the genomic organization of genes modulated in the ΔarcA mutant (black open reading frames) of which the genes in bold were evaluated on Northern blots in panels A and B. Numbers in parentheses near each gene represent the fold change in differential gene expression in the ΔarcA mutant versus parent comparison (Tables 1 and 2).

Fig. 2

Sensitivity of the H. influenzae arcA mutant to hydrogen peroxide. The parental strain, RdV, arcA deletion mutant, RAA6V (ΔarcA) and complemented strain, RAA6C (ΔarcA, arcA+) from anaerobically (A) and aerobically (B) grown cultures were treated with 0.5 mM H₂O₂ for 10 min (Experimental procedures). The survival ratios are plotted as the percentage of cfu obtained from H₂O₂ treated/untreated samples. Values represent the mean of three independent cultures of each strain tested, and the error bars represent the standard deviations. Statistically significant differences between the arcA mutant and parent (P < 0.01) and between the arcA mutant and complemented strain (P < 0.05) were observed (one-way anova with Bonferroni's multiple comparison test) for cultures grown anaerobically. The arcA mutant exhibits a 15-fold and 1.5-fold increase in sensitivity to H₂O₂ challenge compared with parent when grown anaerobically and aerobically respectively.

Fig. 3

Sensitivity of the H. influenzae HI1349 (dps) mutant to hydrogen peroxide. H₂O₂ sensitivity assays with the parental strain, RdV, dps deletion mutant, RdpsV (Δdps) and complemented strain, RdpsC (Δdps, dps+), from anaerobically (A) and aerobically (B) grown cultures were performed as described in Fig. 2. Statistical differences: parent versus dps mutant (P < 0.001 for anaerobic growth, P > 0.05 for aerobic growth); complemented strain versus dps mutant (P < 0.001 for anaerobic and aerobic growth) (one-way anova with Bonferroni's multiple comparison test). The dps mutant exhibits a 9.4-fold and twofold increase in sensitivity to H₂O₂ challenge compared with parent when grown anaerobically and aerobically respectively.

Fig. 4

Persistence defect of H. influenzae mutants in a mouse model of bacteraemia. H. influenzae Rd wild-type (wt), Rcp5 (lpsA) and Rcp19 (galU), and arcA deletion mutant (ΔarcA) were inoculated IP into five C57BL/6 mice per strain and blood was sampled daily for cfu determination (Experimental procedures). Bars represent the average cfu μl⁻¹ (224 and 28 for wt and ΔarcA mutant, respectively, at 24 h post inoculation). Asterisks denote differences from wild type that were statistically significant (P < 0.01). lower limit of detection (LLD) = 2 cfu μl⁻¹.