Evolution towards Virulence in a Burkholderia Two-Component System

Abstract

Bacteria in the Burkholderia cepacia complex (BCC) are significant pathogens for people with cystic fibrosis (CF) and are often extensively antibiotic resistant. Here, we assess the impacts of clinically observed mutations in fixL, which encodes the sensor histidine kinase FixL. FixL along with FixJ compose a two-component system that regulates multiple phenotypes. Mutations in fixL across two species, B. dolosa and B. multivorans, have shown evidence of positive selection during chronic lung infection in CF. Herein, we find that BCC carrying the conserved, ancestral fixL sequence have lower survival in macrophages and in murine pneumonia models than mutants carrying evolved fixL sequences associated with clinical decline in CF patients. In vitro phosphotransfer experiments found that one evolved FixL protein, W439S, has a reduced ability to autophosphorylate and phosphorylate FixJ, while LacZ reporter experiments demonstrate that B. dolosa carrying evolved fixL alleles has reduced fix pathway activity. Interestingly, B. dolosa carrying evolved fixL alleles was less fit in a soil assay than those strains carrying the ancestral allele, demonstrating that increased survival of these variants in macrophages and the murine lung comes at a potential expense in their environmental reservoir. Thus, modulation of the two-component system encoded by fixLJ by point mutations is one mechanism that allows BCC to adapt to the host infection environment. IMPORTANCE Infections caused by members of the Burkholderia cepacia complex (BCC) are a serious concern for patients with cystic fibrosis (CF) as these bacteria are often resistant to many antibiotics. During long-term infection of CF patients with BCC, mutations in genes encoding the FixLJ system often become prevalent, suggesting that these changes may benefit the bacteria during infection. The system encoded by fixLJ is involved in sensing oxygen and regulating many genes in response and is required for full virulence of the bacteria in a murine pneumonia model. Evolved fixL mutations seen later in infection improve bacterial persistence within macrophages and enhance infection within mice. However, these adaptations are short sighted because they reduce bacterial fitness within their natural habitat, soil.

Keywords: Burkholderia; evolution; two-component regulatory systems; virulence regulation.

FIG 1

Mutations in the predicted sensory domain of FixL in B. dolosa are associated with decline in lung function in patients with cystic fibrosis. (a) Domains predicted by SMART (68) and mutations coded by observed single nucleotide polymorphisms (SNPs) in B. dolosa in black (17, 18) and B. multivorans in red (19). Domain abbreviations: TM, transmembrane; PAC, motif C-terminal to PAS motif; HisKA, histidine kinase; HATPase, histidine kinase-associated ATPase. (b) A modest decline was observed in percent predicted FEV1 (ppFEV1) in patient M who did not have a detectable fixL mutation. Rapid declines in ppFEV1 were seen in CF patients L (c) and J (d) after detection of mutations predicted to be in the sensory domain (PAS) of FixL; purple arrow, ancestral fixL allele; red arrow, fixL mutation in sensory domain; green arrow, fixL mutation not in sensory domain; black arrow, fixJ mutation. In c, a P value of 0.05 by linear regression comparing ppFEV1 slopes before and after detection of fixL mutations in the sensory domain is indicated by red arrows. In d, the slope was significantly different from 0 after detection of fixL mutations in the sensory domain (P < 0.0001 by linear regression), while the slope was not significantly different from 0 before detection of fixL mutations in the sensory domain.

FIG 2

Burkholderia cepacia complex strains with evolved fixL alleles are more virulent in human macrophages and a murine pneumonia model but are less able to survive in soil. PMA-treated THP-1 human macrophages were infected with ∼2 × 10⁶ CFU/well (MOI of ∼10:1) of (a) B. dolosa strain AU0158 or (b) B. multivorans strain VC7102 otherwise isogenic mutants carrying different fixL alleles for 2 h, after which the percentage of internalized bacterial relative to the total bacterial growth was determined by killing extracellular bacteria with kanamycin (1 mg/ml). Means from 2 to 3 separate experiments with three replicates per experiment are plotted with error bars representing one standard deviation; *, P < 0.05 by analysis of variance (ANOVA) with Tukey’s multiple-comparison test compared to (a) ΔfixLJ + empty vector and ΔfixLJ + FixL/FixJ or (b) strain VC7102. (c) THP-1-derived macrophages were infected with ∼2 × 10⁶ CFU/well of B. dolosa for 2 h, after which the extracellular bacteria were treated with kanamycin (1 mg/ml) for various amounts of time. The, the percentage of internalized bacteria relative to the total bacterial growth within the initial 2-h infection was determined; *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test compared to ΔfixLJ + FixL/FixJ. (d) THP-1-derived macrophages were infected with ∼2 × 10⁶ CFU/well of B. dolosa for various amounts of time (15 min to 2 h), after which the number of internalized bacteria was determined by killing extracellular bacteria with kanamycin (1 mg/ml); *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test compared to ΔfixLJ + empty vector; ^#, P < 0.05 by ANOVA with Tukey’s multiple-comparison test compared to ΔfixLJ + FixL/FixJ. C57BL/6 mice were intranasally challenged with ∼4 × 10⁸ CFU/mouse of (e) B. dolosa strain AU0158 or (f) B. multivorans strain VC7102 mutants carrying fixL alleles. Bacterial loads were measured in the lungs and spleen 7 days after infection. Data are representative of two separate experiments with 7 to 8 mice per group. Each point represents one mouse, and bars represent medians; *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test. (g) B. dolosa strain AU0158 ΔfixLJ mutants complemented with fixL alleles or empty vector were inoculated into 1 g of sterile soil (1 to 6 × 10⁶ CFU in minimal medium) and incubated for 10 days. Bacterial load was measured and plotted relative to the inoculum used; *, P < 0.05 compared to ΔfixLJ + FixL/FixJ by ANOVA with Tukey’s multiple-comparison test.

FIG 3

Burkholderia cepacia complex strains with evolved fixL alleles were more motile, made less biofilm, and had altered gene expression. Biofilm formation of (a) B. dolosa strain AU0158 or (b) B. multivorans strain VC7102 mutants carrying fixL alleles on PVC plates as measured by crystal violet staining at 48 h. Means from three separate experiments with 5 to 6 replicates per experiment are plotted with error bars representing one standard deviation; *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test to ΔfixLJ+ FixL/FixJ or VC7102. Motility of (c) B. dolosa strain AU0158 or (d) B. multivorans strain VC7102 mutants carrying fixL alleles on low-density (0.3%) LB agar and swimming distance were measured after incubation for 48 h. Means from three separate experiments with 3 to 4 replicates per experiment are plotted with error bars representing one standard deviation. *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test compared to construct carrying ancestral fixL allele; ^#, P < 0.05 by ANOVA with Tukey’s multiple-comparison test. (e) Volcano plot depicting the differential regulation of genes. Green dots signify genes with expression 2-fold lower in the B. dolosa strain AU0158 mutant carrying the evolved fixL allele (encoding FixL W439S) than in a mutant carrying the ancestral fixL allele, with a q value of <0.05. Red dots signify genes with expression 2-fold higher in the mutant carrying the evolved fixL allele (encoding FixL W439S) than in a mutant carrying the ancestral fixL allele, with a q value of <0.05. (f) GO terms that were enriched with an adjusted P value of <0.05 among genes that were statistically upregulated (q value of <0.05, at least 2-fold) in B. dolosa carrying the evolved fixL allele (encoding FixL W439S) relative to B. dolosa carrying the ancestral fixL allele.

FIG 4

Evolved fixL alleles downregulate fix pathway activity. (a) Representative plots of autophosphorylation of B. dolosa FixL proteins. (b) Density measurements from three independent experiments were normalized relative to FixL at the same time point. (c) Representative plots of phosphotransfer to FixJ of B. dolosa FixL proteins. (d) Density measurements from three independent experiments were normalized based on the level of FixL phosphorylation at time zero for each construct. (e) B. dolosa strain AU0158 mutants carrying fixL alleles or an empty vector carrying a pfixK-lacZ reporter plasmid (21) grown in ambient or low (<5%) oxygen conditions. Bars represent the means of triplicate biological replicates, and error bars represent one standard deviation (representative of three independent experiments); *, P < 0.05 by ANOVA with Tukey’s multiple-comparison test for growth in ambient oxygen compared to growth under low oxygen; ^#, P < 0.05 by ANOVA with Tukey’s multiple-comparison test to ΔfixLJ+ FixL/FixJ in both ambient and low oxygen.