A Novel, Molybdenum-Containing Methionine Sulfoxide Reductase Supports Survival of Haemophilus influenzae in an In vivo Model of Infection

Abstract

Haemophilus influenzae is a host adapted human mucosal pathogen involved in a variety of acute and chronic respiratory tract infections, including chronic obstructive pulmonary disease and asthma, all of which rely on its ability to efficiently establish continuing interactions with the host. Here we report the characterization of a novel molybdenum enzyme, TorZ/MtsZ that supports interactions of H. influenzae with host cells during growth in oxygen-limited environments. Strains lacking TorZ/MtsZ showed a reduced ability to survive in contact with epithelial cells as shown by immunofluorescence microscopy and adherence/invasion assays. This included a reduction in the ability of the strain to invade human epithelial cells, a trait that could be linked to the persistence of H. influenzae. The observation that in a murine model of H. influenzae infection, strains lacking TorZ/MtsZ were almost undetectable after 72 h of infection, while ∼3.6 × 10³ CFU/mL of the wild type strain were measured under the same conditions is consistent with this view. To understand how TorZ/MtsZ mediates this effect we purified and characterized the enzyme, and were able to show that it is an S- and N-oxide reductase with a stereospecificity for S-sulfoxides. The enzyme converts two physiologically relevant sulfoxides, biotin sulfoxide and methionine sulfoxide (MetSO), with the kinetic parameters suggesting that MetSO is the natural substrate of this enzyme. TorZ/MtsZ was unable to repair sulfoxides in oxidized Calmodulin, suggesting that a role in cell metabolism/energy generation and not protein repair is the key function of this enzyme. Phylogenetic analyses showed that H. influenzae TorZ/MtsZ is only distantly related to the Escherichia coli TorZ TMAO reductase, but instead is a representative of a new, previously uncharacterized clade of molybdenum enzyme that is widely distributed within the Pasteurellaceae family of pathogenic bacteria. It is likely that MtsZ/TorZ has a similar role in supporting host/pathogen interactions in other members of the Pasteurellaceae, which includes both human and animal pathogens.

Keywords: DMSO reductase enzyme family; Haemophilus influenzae; host–pathogen interaction; methionine sulfoxide reductase; molybdenum enzymes.

FIGURE 1

Structure and function of the HI2019 torYZ operon. (A) Schematic representation of the HI2019 torYZ gene region. The black bar indicates the position of the PCR product shown in (B). (B) PCR based co-transcription analysis of HI2019 torYZ. ‘+’ = positive control containing gDNA; ‘-’ = no template control, C = cDNA (derived from anaerobic culture) as template. (C) Relative expression of the torZ gene in HI2019 cultures grown under aerobic (AE), microaerophilic (MA) and anaerobic (AN) growth conditions. Data were normalized using expression of the gyrA gene. (D) DMSO reductase activity in HI2019^WT Cell-free extracts derived from cultures grown under aerobic (AE), microaerophilic (MA), and anaerobic (AN) growth conditions. DMSO reductase activity in cell-free extracts is a relative measure of the activity of Mo-containing S-oxide reductases under the conditions tested. Errors are given as standard deviations of the mean, at least three repeat assay were carried out per condition.

FIGURE 2

Effects of the torZ gene inactivation on the physiology of HI2019^WT, HI2019^ΔtorZ and HI2019^ΔtorZ_comp strains. (A) S- and N-oxide reductase activity in cell-free extracts of HI2019^WT, HI2019^ΔtorZ and HI2019^ΔtorZ_comp strains. Cell extracts were prepared from cultures grown anaerobically to ensure maximal expression of torZ, enzyme assays were carried out anaerobically using sodium dithionite reduced methylviologen as the electron donor to the enzymes and different S- and N-oxides. Each assay was repeated at least three times, experimental errors are given as standard deviations of the mean. (B) Biofilm formation and in biofilm survival of HI2019^WT, HI2019^ΔtorZ and HI2019^ΔtorZ_comp strains under anaerobic conditions. Left: Crystal violet based detection of biofilm density, Right: survival of HI in anaerobic biofilms measured as CFU present per well. Bacteria were recovered from biofilms using proteinase K treatment followed by serial dilution and plating. The data shown is one of three independent experiments carried out on different days, each bar represents the combined data of three biological replicates analyzed that day with eight replicate assays/biological replicate. One Way ANOVA was used to determine statistical significance. P-values: for all instances of ^∗∗∗∗p < 0.0001, ^∗∗p = 0.0077 (right panel, comparison torZ mutant and complemented strain). (C) Susceptibility of HI2019^WT, HI2019^ΔtorZ and HI2019^ΔtorZ_comp strains to HOCl–induced oxidative stress. Bacteria were resuspended in 1xPBS (OD₆₀₀ in assay = 1.0) and exposed to different concentrations of HOCl for 60 min before serial dilution and plating. The graph shows the combined data for three experiments carried out on three different days using a total of seven biological replicates for each strains. Error bars represent 95% confidence intervals, no significant differences in strain survival were detected (2-Way ANOVA).

FIGURE 3

Deletion of the TorZ protein leads to a change in the ability of HI2019 to colonize and invade human bronchial epithelial 16HBE14 cells. (A) Immunofluorescence imaging of co-cultures of HI2019^WT and HI2019^ΔtorZ. Epithelial cells were stained using CellTracker^TM, HI cells were detected using the antibody 6E4 (Erwin et al., 2006) as set out in the method section. (B) Total adherent and intracellular HI2019 cells present in co-cultures of 16HBE14 cells with HI2019^WT or HI2019^ΔtorZ after 4 and 24 h incubation. Cell numbers were determined by serial dilution plating and are reported as CFU/ml. Data represent averages from three biological replicates, with three technical replicates per sample. Data were analyzed using ordinary one way ANOVA, using SIDAK’s multiple comparison test for comparisons between individual datasets. P-values were as follows: total adherent cells: 4 h p = 0.8640 (n.s.), 24 h ^∗∗∗∗p =< 0.0001, invasion: 4 h ^∗∗∗p = 0.0005, 24 h ^∗∗∗∗p =< 0.0001.

FIGURE 4

Survival of HI2019^WT and HI2019^ΔtorZ in the respiratory tract of BALBc mice and in contact with primary human neutrophils. (A) Survival in contact with primary human neutrophils following 2 h of incubation. Escherichia coli was used as a control. P-values determined by unpaired t-tests were HI2019^WT – E. coli p = 0.0003; HI2019^WT vs. HI2019^ΔtorZ p = 0.1072 (n.s.). (B) Survival in the respiratory tract of BALBc mice. Mice were inoculated intranasally as HI is unable to cause a stable colonization of mice, a decrease of CFU/ml over time is expected. P-values determined by unpaired t-tests were 0 h ^∗∗p = 0.00585, 24 h ^∗∗p = 0.0083, 48 h ^∗∗p = 0.00376, 72 h ^∗∗p = 0.001865. (C) Immune cell counts in BALF fluid. Total cell counts are shown. Influx of immune cells is slower for HI2019^ΔtorZ, in keeping with the reduced fitness of this strain. ^∗∗∗p = 0.0005.

FIGURE 5

Characterization of HI rTorZ. (A) 10% SDS-PAGE of purified HI rTorZ; (B) activity of HI rTorZ with different N- and S-oxide substrates in standard assays. Averages shown are from at least three assays, errors are given as standard deviations. (C) Activity of HI rTorZ with increasing amounts of racemic DL-MetSO as substrate. Three repeat assays were carried out per concentrations shown, the data were fitted with Prism 6.0 (GraphPad) using direct non-linear fitting. Kinetic parameters are shown in Table 2. (D) Repair of oxidized calmodulin by HI rTorZ. E. coli MsrP was used as the positive control. Oxidized and reduced forms of Calmodulin have altered electrophoretic properties, the data show that HI rTorZ has no calmodulin reducing activity compared to MsrP. Labels: ‘+’ = complete assay containing reductant, enzyme and Calmodulin as the substrate, ‘-‘ = negative control without reductant to ensure the reaction is due to enzymatic activity.

FIGURE 6

Phylogenetic analysis of HITorZ/MtsZ related protein sequences of the DMSO reductase enzyme family. HITorZ/MtsZ clearly forms a distinct group within this subgroup of DMSO reductase enzyme family enzymes as indicated by the branching pattern. The phylogenetic tree shown was constructed from a total of 1005 sequences retrieved from NCBI by BLAST searches (see Supplementary Data) using the neighbor-joining algorithm. The phylogenetic analyses used MEGA 6.0, bootstrapping used 500 replicates.

FIGURE 7

Current model of HIMtsZ physiological function in HI2019^WT and HI2019^ΔtorZ. MetSO, methionine sulfoxide; Met, methionine; MQ, oxidized menaquinone; MQH₂, reduced menaquinone. MQH₂ in bold (right panel) indicates accumulation of this compound due to lack of MtsZ activity.