Cytochrome oxidase requirements in Bordetella reveal insights into evolution towards life in the mammalian respiratory tract

Abstract

Little is known about oxygen utilization during infection by bacterial respiratory pathogens. The classical Bordetella species, including B. pertussis, the causal agent of human whooping cough, and B. bronchiseptica, which infects nearly all mammals, are obligate aerobes that use only oxygen as the terminal electron acceptor for electron transport-coupled oxidative phosphorylation. B. bronchiseptica, which occupies many niches, has eight distinct cytochrome oxidase-encoding loci, while B. pertussis, which evolved from a B. bronchiseptica-like ancestor but now survives exclusively in and between human respiratory tracts, has only three functional cytochrome oxidase-encoding loci: cydAB1, ctaCDFGE1, and cyoABCD1. To test the hypothesis that the three cytochrome oxidases encoded within the B. pertussis genome represent the minimum number and class of cytochrome oxidase required for respiratory infection, we compared B. bronchiseptica strains lacking one or more of the eight possible cytochrome oxidases in vitro and in vivo. No individual cytochrome oxidase was required for growth in ambient air, and all three of the cytochrome oxidases conserved in B. pertussis were sufficient for growth in ambient air and low oxygen. Using a high-dose, large-volume persistence model and a low-dose, small-volume establishment of infection model, we found that B. bronchiseptica producing only the three B. pertussis-conserved cytochrome oxidases was indistinguishable from the wild-type strain for infection. We also determined that CyoABCD1 is sufficient to cause the same level of bacterial burden in mice as the wild-type strain and is thus the primary cytochrome oxidase required for murine infection, and that CydAB1 and CtaCDFGE1 fulfill auxiliary roles or are important for aspects of infection we have not assessed, such as transmission. Our results shed light on the environment at the surface of the ciliated epithelium, respiration requirements for bacteria that colonize the respiratory tract, and the evolution of virulence in bacterial pathogens.

Fig 1. No cytochrome oxidase is necessary for growth in ambient air, but Cyd1, Cta1, and Cyo1 are each sufficient.

(A-B) Growth over time, measured via optical density (left) or CFU/mL (right). Growth of RB50 (WT, black) is compared to strains lacking a single cytochrome oxidase-encoding gene loci (A) or strains with only one of the cytochrome oxidase-encoding gene loci conserved in B. pertussis (cyd1⁺, cta1⁺, and cyo1⁺) or just the three Bp-conserved loci intact (Bp-conserved) (B). Cultures were started at 0.1 OD₆₀₀, which contains approximately 3.5x10⁸ CFU/mL. Points represent the mean of at least 3 biological replicates, with bars representing the standard deviation. (C) Intracellular ATP levels, as measured through luciferase activity, for the same strains as in B. (D) Proton motive force, as measured through uptake of fluorescently labeled gentamicin, for the same strains as in B. Statistical significance was determined using unpaired Student’s t-test. *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001. Raw data for Fig 1A: https://doi.org/10.15139/S3/F89WZH. Raw data for Fig 1B: https://doi.org/10.15139/S3/NKQ2BK. Raw data for Fig 1C: https://doi.org/10.15139/S3/EURIPE. Raw data for Fig 1D: https://doi.org/10.15139/S3/QSJXPT.

Fig 2. Growth under different atmospheric conditions relevant to the environment of the mammalian respiratory tract.

Growth, measured via optical density (left) or CFU/mL (right), for RB50 (WT, gray), a strain with only the cytochrome oxidase-encoding gene loci conserved in B. pertussis (Bp-conserved, maroon), a strain with only cydAB1 (cyd1⁺, orange), a strain with only ctaCDFGE1 (cta1⁺, blue), or a strain with only cyoABCD1 (cyo1⁺, green). Cultures were started at 0.1 OD₆₀₀, which contains approximately 3.5x10⁸ CFU/mL. Cultures were grown in either: ambient air for 24 hours (A), 5% O₂ for 48 hours (B), 2% O₂ for 72 hours (C), 5% CO₂ for 24 hours (D), or 5% O₂ 5% CO₂ for 48 hours (E). The data from A are identical to the 24-hour timepoint of Fig 1A and are included for comparison. Bars represent the mean of at least 3 biological replicates, with error bars representing the standard deviation. Statistical significance was determined using unpaired Student’s t-test. *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001. Raw data: https://doi.org/10.15139/S3/UTSSJM.

Fig 3. No single cytochrome oxidase is required during murine infection.

Bacterial burden over time within the nasal cavity (upper), trachea (middle), and right lung (lower) of mice infected with either RB50 (WT) or strains lacking a single cytochrome oxidase-encoding gene loci. Different symbols represent different batches of inoculations, which always include WT for comparison: circles (Δcco, Δcyd1, Δcyd3, and Δcio), upright triangles (Δcio, Δcyo1, Δcyo2, Δcta1, and ΔctaD2), upside-down triangles (Δcco, Δcyd1, and Δcyd3), and diamonds (Δcyo1, Δcyo2, Δcta1, and ΔctaD2). Each point represents data from one mouse. n = 4 for day 0, n = 6 for all other timepoints, from two independent experiments, for all mutant strains. Each point represents a single mouse. Dashed line represents the limit of detection. Statistical significance was determined using unpaired Student’s t-test; *, p < 0.01. Raw data: https://doi.org/10.15139/S3/POHLJP.

Fig 4. The three cytochrome oxidases conserved in B. pertussis are sufficient for persistence in the murine respiratory tract.

Bacterial burden over time within the nasal cavity (upper), trachea (middle), and right lung (lower) of mice infected with either wild-type bacteria (WT, grey circles) or a strain with only the cytochrome oxidase-encoding gene loci conserved in B. pertussis (Bp-conserved, maroon hexagon). n = 4 for day 0, n = 6 for all other timepoints, with each point representing a single mouse. Samples were collected from two independent experiments. Dashed line represents the limit of detection. Statistical significance was determined using unpaired Student’s t-test; *, p < 0.05. Raw data: https://doi.org/10.15139/S3/ZO4B8O.

Fig 5. Cyo1 is sufficient for WT-levels of persistence within the mammalian respiratory tract, while Cyd1 and Cta1 are not.

Bacterial burden over time within the nasal cavity (upper), trachea (middle), and right lung (lower) of mice infected with wild-type bacteria (WT, grey circles), a strain with only cydAB1 (cyd1⁺, orange upright triangles), a strain with only ctaCDFGE1 (cta1⁺, blue diamonds), or a strain with only cyoABCD1 (cyo1⁺, green upside-down triangles). These strains were tested in the same experiment but separated into separate graphs for analysis; thus, the mutant strains are compared to the same WT data. n = 4 for day 0, n = 6 for all other timepoints, with each point representing a single mouse. Samples were collected from two independent experiments. Dashed line represents the limit of detection. Statistical significance was determined using unpaired Student’s t-test; *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001. Raw data: https://doi.org/10.15139/S3/60DN0G.

Fig 6. Cyo1 is sufficient for the efficient establishment of colonization, while Cyd1 and Cta1 are not.

Bacterial burden over time within the nasal cavity of mice infected with a strain with only the cytochrome oxidase-encoding gene loci conserved in B. pertussis (Bp-conserved, maroon hexagon), a strain with only cydAB1 (cyd1⁺, orange upright triangles), a strain with only ctaCDFE1 (cta1⁺, blue diamonds), or a strain with only cyoABCD1 (cyo1⁺, green upside-down triangles). Samples from 10 mice across two independent experiments were collected at each timepoint for each strain. However, due to the natural microbiota of the nasal cavity, B. bronchiseptica could not always be enumerated due to contamination. Therefore, n = 7 for cta1⁺ day 3, n = 9 for cyd1⁺ day 7 and cta1⁺ day 7. Each point represents a single mouse. Dashed line represents the limit of detection. Statistical significance was determined using unpaired Student’s t-test; *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001. Raw data: https://doi.org/10.15139/S3/VO1E3R.

Fig 7. Regulation of cytochrome oxidase genes under different atmospheric conditions, measured using RNA sequencing.

(A) Transcript abundance, measured by counts per million per kilobase, for cytochrome oxidase-encoding genes from B. bronchiseptica grown in ambient air. fhaB and flaA, two highly-regulated and well-characterized genes, are included for comparison. flaA transcription level (lower line) is used as a reference, as flaA is negatively regulated under this growth condition and no protein product can be detected. fhaB (upper line) is highly expressed under this growth condition and is thus used as a reference for a highly transcribed gene. Numbers in the bars represent the mean number of transcripts normalized to gene length. (B-D) Comparative transcription, measured by fold change in transcript abundance, comparing 5% O₂ (B), 2% O₂ (C), or 5% CO₂ (D) to ambient air. Genes with log₂ fold changes greater than 1 or less than -1 have filled-in bars. Differential expression analysis was performed using edgeR’s glmQLFTest [73]; significance was adjusted using the Benjamini–Hochberg procedure. *, p < 0.05; **, p < 0.005; ***, p < 0.0005; ****, p < 0.0001. Raw data: GEO repository GSE268598