Interactions between chaperone and energy storage networks during the evolution of Legionella pneumophila under heat shock

Abstract

Waterborne transmission of the bacterium Legionella pneumophila has emerged as a major cause of severe nosocomial infections of major public health impact. The major route of transmission involves the uptake of aerosolized bacteria, often from the contaminated hot water systems of large buildings. Public health regulations aimed at controlling the mesophilic pathogen are generally concerned with acute pasteurization and maintaining high temperatures at the heating systems and throughout the plumbing of hot water systems, but L. pneumophila is often able to survive these treatments due to both bacterium-intrinsic and environmental factors. Previous work has established an experimental evolution system to model the observations of increased heat resistance in repeatedly but unsuccessfully pasteurized L. pneumophila populations. Here, we show rapid fixation of novel alleles in lineages selected for resistance to heat shock and shifts in mutational profile related to increases in the temperature of selection. Gene-level and nucleotide-level parallelisms between independently-evolving lineages show the centrality of the DnaJ/DnaK chaperone system in the heat resistance of L. pneumophila. Inference of epistatic interactions through reverse genetics shows an unexpected interaction between DnaJ/DnaK and the polyhydroxybutyrate-accumulation energy storage mechanism used by the species to survive long-term starvation in low-nutrient environments.

Keywords: Experimental evolution; Heat shock proteins; Heat shock response; Legionella pneumophila; Pasteurization; Phasin protein, bacteria; Poly(hydroxybutyrates).

Figure 1. Dynamics of mutation innovation and fixation in L. pneumophila populations undergoing experimental selection for heat shock tolerance.

(A) Allele frequencies of mutations accumulated in eight selected genes (clpB, clpX, dnaJ, dnaK, htpG, mreC, phaP, and rodA) across the three sequenced heat-adapted populations. (B) Linear accumulation of mutations in three sequenced heat-adapted (HA) lineages through 70 passages of selection. Mutation load totals fixed mutations and allele frequencies of polymorphisms annotated by Breseq 0.36.1 per-lineage. Linear fit depicts results of a linear regression of all per-lineage mutation load points with respect to passage number constrained to intercept (x, y) = (0, 0). Vertical dashed lines represent times of temperature increase.

Figure 2. Heat tolerance comparisons between mutant L. pneumophila strains constructed with evolutionarily-fixed mutations in isolation or combination.

(A–D) Comparisons of thermal kill rate at 55 °C of wild-type KS79 and constructed single, double, and triple mutants (identified by key below each graph), as well as the heat-adapted (HA) lineage produced by 70 rounds of selection (which contains all three mutated alleles, as indicated by key). Mutants are grouped by trajectories (TRA). See Results section for details. Data shown represent mean +/− SEM. Significance measurements depict two-tailed one-way ANCOVA relative to KS79, n = 3, (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Dotted horizontal line at y = 0 indicates full thermal resistance to 55 °C exposure.

Figure 3. Genetic and phenotypic characterization of wild-type and mutant alleles of phaP in L. pneumophila.

(A) Amino acid sequence of L. pneumophila str. Philadelphia-1 PhaP (above), annotated with predicted α-helices 1 and 2 and predicted amino acid sequence of the L. pneumophila phaP^A8→A7 gene product (below), showing a divergent sequence after 15 amino acids and pre-mature truncation of the translated product. Image produced using AliView 1.28 (Larsson, 2014). (B) AlphaFold-predicted (Jumper et al., 2021; Varadi et al., 2022) structure of PhaP (UniProt: Q5ZY12), showing assembly of two α-helices 1 and 2 and linker region. (C–E) Gene organization around the phaP locus including phaB₁, phaB₂, phaR, and phaP as an operon on the sense strand. D and E show the mutant construction strategy, with the mutated allele shown in the 5’ flanking region of the Kan^R cassette. Survival of starvation in Fraquil of indicated strains at (F) 37 °C or (G) 42 °C shows no significant differences between wild-type KS79 and the constructed mutants, n = 3. (H) Polyhydroxybutyrate production evaluated with Nile red staining. Fluorescence was measured with excitation at 535 nm and emission at 590 nm normalized to mean fluorescence in KS79. Significance is depicted relative to KS79, n = 3, Data shown represents mean +/− SD.

Figure 4. phaP frameshift affects the transcriptional and phenotypic heat shock response.

(A–C) Transcriptional changes in (A) dnaJ, (B) dnaK, and (C) rpoH mRNA levels following a 5 min exposure to heat shock were measured by RT-qPCR and quantified by efficiency-corrected ddCt (Yuan, Wang & Stewart, 2008). Data shown represent mean +/− SEM and significance calculations show two-tailed one-way ANOVA with Tukey’s correction for multiple comparisons, n = 3 (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). (D) Comparisons of thermal kill rate at 55 °C between wild-type KS79, single mutant, double mutant, and triple mutant identified in the below key. Data shown represent mean +/− SEM and significance depicts ANCOVA with Šidák correction for multiple comparisons (*p < 0.05; **p < 0.01; ****p < 0.0001).