Nuanced differences in adenylate cyclase toxin production, acylation, and secretion may contribute to the evolution of virulence in Bordetella species

Abstract

Bordetella pertussis, which causes the acute human disease whooping cough, evolved from Bordetella bronchiseptica, which causes long-term, chronic infections in a broad range of mammals. Both B. pertussis and B. bronchiseptica produce adenylate cyclase toxin (ACT), a bifunctional protein with adenylate cyclase-mediated cell intoxication and pore-forming activities. CyaC-mediated acylation of ACT is important for cell intoxication and required for pore-forming activity in vitro, but its role in vivo was unknown. Our comparative analysis showed that although ACT secreted by B. pertussis is fully acylated at residues K860 and K983, only a fraction of ACT secreted by B. bronchiseptica is modified. We showed that B. bronchiseptica produces more ACT than B. pertussis and is more efficient at ACT-dependent intoxication of macrophages in vitro than B. pertussis, but for both organisms, acylation of ACT greatly enhances intoxication. Acylation also enhances ACT secretion. Using a natural-host model, we determined that non-acylated ACT is functional during the initial stage of B. bronchiseptica infection, but not at later time points, and that acylation of K860, but not K983, is required for persistence in the lower respiratory tract. These data indicate a role for both acylated and non-acylated ACT during infection. Acylation of ACT was similarly not absolutely required for B. pertussis persistence in the murine lower respiratory tract. Overall, our data revealed nuanced differences in production, acylation, and secretion of ACT between B. pertussis and B. bronchiseptica that may correlate with their different virulence strategies.IMPORTANCEBordetella pertussis causes the acute disease whooping cough and survives only in the human respiratory tract, while Bordetella bronchiseptica causes long-term, chronic infections in a broad range of mammals and can also survive in extra-host environments. These bacteria produce a nearly identical set of virulence factors, including adenylate cyclase toxin (ACT), a protein that is modified by the addition of acyl chains. Acylation is required for ACT to cause hemolysis and for efficient intoxication of host cells in vitro. We found that ACT acylation is also important, but not absolutely required, during infection. We also discovered differences in ACT production, acylation, and secretion between B. bronchiseptica and B. pertussis that may contribute to the different virulence strategies of these species. This study highlights the advantage of conducting comparative analyses between closely related species to better understand the evolution of virulence.

Keywords: Bordetella; Bordetella bronchiseptica; Bordetella pertussis; acylation; adenylate cyclase toxin; bacterial respiratory infection; post-translational modification.

Fig 1

Acylation is required for B. bronchiseptica and B. pertussis ACT-dependent hemolysis on blood agar. Growth after 72 hours (B. bronchiseptica strains) or 5 days (B. pertussis strains) on Bordet-Gengou agar containing 12.5% defibrinated sheep’s blood. WT, wild type.

Fig 2

ACT production, secretion, and surface association are greater in B. bronchiseptica than in B. pertussis, and acylation enhances secretion. Western blot analyses of whole cell lysates (A) or supernatants (B) collected from B. bronchiseptica and B. pertussis strains grown in SS medium supplemented to 2 mM CaCl₂. Samples were taken after 16 hours of growth and normalized to culture OD₆₀₀. Blots were probed with monoclonal α-ACT antibody. Relative abundance (RA) values under each lane correspond to quantified ACT by normalizing ACT signal to the signal of a protein of relative equal abundance identified by Revert 700 Total Protein Stain (LI-COR) present in whole cell lysates across all strains. (C) Dot blot analyses of intact bacteria that were cultured, normalized, and probed as described for panels A and B. WT, wild type.

Fig 3

ACT acylation enhances B. bronchiseptica and B. pertussis intoxication of J774A.1 murine macrophage-like cells. J774A.1 cells were infected with B. bronchiseptica and B. pertussis wild-type (WT) or mutant strains at an MOI of 100 for 30 minutes. Intracellular cAMP concentrations were determined by ELISA, and the concentration of cAMP for the uninfected control was subtracted from the experimental samples. The data shown are compiled from four biologically independent experiments for B. bronchiseptica WT and ∆cyaC strains and two biologically independent experiments for all other B. bronchiseptica and B. pertussis strains.

Fig 4

Only a fraction of ACT secreted by B. bronchiseptica is modified at K860 and K983. Linear representations of the peptide sequences following residues K860, K983, and K783 of ACT. Dashed lines indicate trypsin digestion sites. Sequences in blue between dashed lines represent the peptides that were quantified by LC-MS/MS. Open circles represent no peptide detected in the sample and plotted at the limit of detection. Relative abundances (LFQ intensities) were corrected using global normalization to compare across samples. WT, wild type.

Fig 5

Acylated and non-acylated ACT are functional during B. bronchiseptica infection. (A) Bacterial burden over time in the trachea (top) and right lung (bottom) of mice inoculated with wild-type (WT) B. bronchiseptica or mutant strains. Data were compiled from four independent experiments, each including WT. Dashed line represents the limit of detection. Only statistically significant differences in bacterial burden compared to WT (above; black lines) or ∆cyaA (below; pink lines) are reported as determined using unpaired Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001. (B) Representative images of hematoxylin-and-eosin-stained 5 µm left lung sections from day 3 mice inoculated with WT or mutant strains in panel A.

Fig 6

Acylation of ACT residue K860 is important for B. bronchiseptica persistence in the lower respiratory tract. (A) Growth after 72 hours on Bordet-Gengou agar containing 6% defibrinated sheep’s blood. (B) Bacterial burden over time in the trachea (top) and right lung (bottom) of mice inoculated with B. bronchiseptica wild-type (WT) or mutant strains. Data were compiled from two independent experiments. Dashed line represents the limit of detection. Only statistically significant differences in bacterial burden compared to WT (above; black lines) or ∆cyaA (below; pink lines) are reported as determined using unpaired Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001.

Fig 7

Acylation is only partially required for B. pertussis persistence in mice. (A) Bacterial burden over time in the trachea (top) and right lung (bottom) of mice inoculated with B. pertussis wild-type (WT) or mutant strains. Data were compiled from two independent experiments. Dashed line represents the limit of detection. Only statistically significant differences in bacterial burden compared to WT (above; black lines) or ∆cyaA (below; pink lines) are reported as determined using unpaired Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001. (B) Representative images of H&E-stained 5 µm left lung sections from day 3 mice inoculated with PBS vehicle control, WT, or mutant strains in panel A.