Vibrio cholerae requires oxidative respiration through the bd-I and cbb3 oxidases for intestinal proliferation

Abstract

Vibrio cholerae respires both aerobically and anaerobically and, while oxygen may be available to it during infection, other terminal electron acceptors are proposed for population expansion during infection. Unlike gastrointestinal pathogens that stimulate significant inflammation leading to elevated levels of oxygen or alternative terminal electron acceptors, V. cholerae infections are not understood to induce a notable inflammatory response. To ascertain the respiration requirements of V. cholerae during infection, we used Multiplex Genome Editing by Natural Transformation (MuGENT) to create V. cholerae strains lacking aerobic or anaerobic respiration. V. cholerae strains lacking aerobic respiration were attenuated in infant mice 105-fold relative to wild type, while strains lacking anaerobic respiration had no colonization defect, contrary to earlier work suggesting a role for anaerobic respiration during infection. Using several approaches, including one we developed for this work termed Comparative Multiplex PCR Amplicon Sequencing (CoMPAS), we determined that the bd-I and cbb3 oxidases are essential for small intestinal colonization of V. cholerae in the infant mouse. The bd-I oxidase was also determined as the primary oxidase during growth outside the host, making V. cholerae the only example of a Gram-negative bacterial pathogen in which a bd-type oxidase is the primary oxidase for energy acquisition inside and outside of a host.

Fig 1. Verification of MuGENT generated mutant strains.

(a) Chromosomal map of V. cholerae terminal electron reducing complex loci. (b) Multiplex allele-specific PCR (MASC-PCR) of V. cholerae terminal electron reducing complexes MuGENT mutants. Lanes are labelled with the strain name where a strain preceded by a ‘^Mu+’ (Lanes 3–6) indicates the oxidase complex as the sole remaining functional oxidase in that strain and strains preceded by a ‘^Mu’ (Lanes 7–10) indicates that the specified locus is the targeted knock out. Targeted gene loci are labelled to the right of each gel image. The presence of a band indicates a targeted knockout in the gene locus whereas the absence of a band indicates the wild type gene is present.

Fig 2. Terminal oxidases support aerobic growth in V. cholerae.

Growth characteristics of the terminal oxidases in V. cholerae. (a-b) Single terminal oxidase mutants, both MuGENT and isogenic deletion, growth in LB. Inoculums were prepared anaerobically and subsequently grown in aerobic and anaerobic conditions, respectively. (c-d) Single terminal oxidase isogenic deletion strain in vitro LB competition assays in both aerobic and anaerobic conditions, respectively. Competitive index scores were calculated as a ratio of output versus input [(Target_Output/ΔlacZ_Output) / (Target_Input/ΔlacZ_Input)], where a ΔlacZ strain served as a psuedo-wild type to determine relative fitness via blue-white screening. Red dots indicate the limit of detection where no target strain CFUs were recovered for these trials. (e-f) Single terminal oxidase isogenic deletion mutant growth in LB where inoculums were prepared aerobically and subsequently grown in aerobic and anaerobic conditions, respectively. (g) In vitro expression of terminal oxidases in anaerobic, microaerobic, and aerobic growth conditions.

Fig 3. Individually, oxidases cbb₃, bd-I, and bd-III support aerobic growth in V. cholerae.

(a-b) Combinatorial terminal oxidase deletion mutant growth in LB. Inoculums were prepared anaerobically and subsequently grown in aerobic and anaerobic conditions, respectively. Triple deletion mutant strains have a ‘+’ with an oxidase name (e.g. +cbb₃), indicating the sole remaining oxidase, with the other three oxidases disrupted by mutation. (c-d) Combinatorial terminal oxidase deletion mutant growth in LB. Inoculums were prepared aerobically and subsequently grown in aerobic and anaerobic conditions, respectively. (e) In vitro aerobic and anaerobic competition assay between Aero7 and wild type V. cholerae with competitive index scores calculated as [(Aero7_Output/WT_Output) / (Aero7_Input/WT_Input)]. Growth curves are an average of three biological replicates where error bars represent the standard error of the mean. Bars for in vitro competitions and expression data represent the arithmetic mean where error bars represent the standard error of the mean.

Fig 4. Terminal reductase mutants are reduced for anaerobic growth in the presence of cognate electron acceptor molecules.

Growth characteristics of terminal reductases of V. cholerae. MuGENT generated terminal reductase mutants grown in LB in the presence and absence of alternative electron acceptors (a) 50mM fumarate, (b) 50mM trimethylamine-N-oxide (TMAO), (c) 50mM nitrate, and (d) 50mM dimethyl sulfoxide (DMSO). Inoculums were prepared aerobically and subsequently grown in anaerobic conditions. Growth curves are an average of three biological replicates where error bars represent the standard error of the mean.

Fig 5. Aerobic respiration, but not anaerobic respiration, is required for growth and colonization of the infant mouse small intestine.

Aero7 and Ana4 small intestine colonization in single strain and competition infections. (a) Single strain infection of strain Aero7. (b) Competition infection of strain Aero7. (c) Single strain infection of strain Ana4. (d) Competition infection of strain Ana4. Bars represent the geometric mean. Horizontal dashed lines indicate the limit of detection (LOD) and red dots indicate recovered CFUs were below the LOD. Competitive index scores were calculated as [(Mutant_Output/WT_Output) / (Mutant_Input/WT_Input)]. Statistical analysis was performed using GraphPad PRISM. *, P < 0.05. A Mann-Whitney U-test was used in the determination of significance between WT and Aero7. A Student’s t test was performed on log transformed data in the determination of significance between WT and Ana4.

Fig 6. bd-I oxidase is critical in competitive infection of the infant mouse small intestine, yet colonization is supported by functional redundancy of terminal oxidases in single strain infections.

Single oxidase in vivo colonization dynamics in the small intestine. (a) Comparative Multiplex PCR Amplification Sequencing (CoMPAS) sequence analysis. Comparative index scores were calculated as [(Output Pool_{Target Reads} / Output Pool_{toxT Reads}) / (Input Pool_{Target Reads} / Input Pool_{toxT Reads})]. Vertical red dashed lines indicate a 2-fold change in output to input sequence ratios. Sequence coverage for each input pool (IP) and associated mouse output pool (MP) are shown in the bar plots. (b) Individual oxidase deletion single strain infections. Bars represent the geometric mean. Horizontal dashed lines indicate the limit of detection (LOD) and red dots indicate recovered CFUs were below the LOD. Statistical analysis was performed using GraphPad PRISM. *, P < 0.05. A Mann-Whitney U-test was used in the determination of significance between WT and Aero7 whereas an Analysis of Variance with post hoc Dunnett’s multiple comparisons test was conducted on log transformed CFU/g intestine for all other strain comparisons.

Fig 7. bd-I oxidase alone supports wild type levels of colonization in the infant mouse small intestine with cbb₃ supporting colonization to a lesser extent.

(a) Combinatorial oxidase deletion in vivo colonization dynamics in the small intestine. (b) Colonization of aerobically prepared wild type and +cbb₃ oxidase inoculums in the small intestine. Triple deletion mutant strains have a ‘+’ with an oxidase name (e.g. +cbb₃), indicating the sole remaining oxidase, with the other three oxidases disrupted by mutation. Bars represent the geometric mean. Horizontal dashed lines indicate the limit of detection (LOD) and red dots indicate recovered CFUs were below the LOD. Statistical analysis was performed using GraphPad PRISM. *, P < 0.05. A Mann-Whitney U-test was used in the determination of significance between WT and +bd-II, +bd-III, and Δcbb₃Δbd-I whereas an Analysis of Variance with post hoc Dunnett’s multiple comparisons test was conducted on log transformed CFU/g intestine for all other strain comparisons. *, P < 0.05.