Diversity of acid stress resistant variants of Listeria monocytogenes and the potential role of ribosomal protein S21 encoded by rpsU

Abstract

The dynamic response of microorganisms to environmental conditions depends on the behavior of individual cells within the population. Adverse environments can select for stable stress resistant subpopulations. In this study, we aimed to get more insight in the diversity within Listeria monocytogenes LO28 populations, and the genetic basis for the increased resistance of stable resistant fractions isolated after acid exposure. Phenotypic cluster analysis of 23 variants resulted in three clusters and four individual variants and revealed multiple-stress resistance, with both unique and overlapping features related to stress resistance, growth, motility, biofilm formation, and virulence indicators. A higher glutamate decarboxylase activity correlated with increased acid resistance. Whole genome sequencing revealed mutations in rpsU, encoding ribosomal protein S21 in the largest phenotypic cluster, while mutations in ctsR, which were previously shown to be responsible for increased resistance of heat and high hydrostatic pressure resistant variants, were not found in the acid resistant variants. This underlined that large population diversity exists within one L. monocytogenes strain and that different adverse conditions drive selection for different variants. The finding that acid stress selects for rpsU variants provides potential insights in the mechanisms underlying population diversity of L. monocytogenes.

Keywords: SNP analysis; glutamate decarboxylase; heterogeneity; phenotypic characterization; trade-off; whole genome sequencing.

FIGURE 1

Maximum specific growth rate of Listeria monocytogenes LO28 wild type (WT) and acid resistant variants in Brain Heart Infusion (BHI) at 37∘C (A) and 7∘C (B), determined by the twofold dilution (2FD) method. Dark gray bars indicate the slightly resistant variants, light gray the variable resistant variant and white the highly acid resistant variants (Metselaar et al., 2013). Error bars represent the SD of the mean for three (A) or two (B) independent replicates. Significantly different values for μ_max compared to the WT are indicated by ∗(p< 0.05).

FIGURE 2

Stress resistance of late-exponential phase cells of L. monocytogenes LO28 WT and acid resistant variants in BHI broth. Late-exponential phase cells were exposed to pH 2.5 for 3.5 min at 37∘C (adapted from Metselaar et al., 2013) (A), 55∘C for 6 min (B), 420 mM hydrogen peroxide for 9 min at 30∘C (C), and 20 mg/L benzalkonium chloride (BAC) for 5 min at 30∘C (D). Results are expressed as reduction in log₁₀ cfu/ml after exposure compared to log₁₀ cfu/ml at t = 0. Errors bars represent the SD and significant differences from the WT are indicated with ∗(p < 0.05).

FIGURE 3

GABA_i concentration of stationary phase cells exposed to pH 4.0 in BHI for 60 min. Solid bars are the GABA_i concentrations at t = 0 and dashed bars at t = 60 min. GABA_i concentrations are corrected for total amount of protein. Error bars include the SD of the GABA and protein measurements and significant differences from the WT are indicated with ∗(p < 0.05).

FIGURE 4

rpsU and upstream region sequence alignment of Sanger sequenced L. monocytogenes LO28 and acid resistant variants. The upper alignment represent the nucleotide sequence of the region where mutations were found. The black line indicates the start codon of the rpsU gene. The five sequences below the WT are the variations of rpsU found in several variants (indicated with the numbers). The lower alignment represents the amino acid sequence of the complete rpsU gene and the effect of the mutations on the amino acid sequence.

FIGURE 5

Hierarchical cluster analysis based on Euclidean distance of the phenotypes (A) and heat map of the six phenotypes for each cluster (B). The variants in black were the initial selection for cluster analysis and the gray variants were used to validate the resulting clusters. The heat map was made by taking the average of each phenotype within one cluster. Subsequently, the relative ratio of the cluster to the WT was calculated for each phenotype.