Fig 1.
Hydrogen peroxide killing curves of E. coli MG1655.
Panel A shows bacterial sensitivity to 1 mM H2O2 with a strong dose-response effect (p = 2.1x10-16, DRC model fitted based on maximum likelihood). Panel B shows the priming effect or improved response when cells are treated in advance with a non-lethal dose (0.1 mM). Curves represent the mean of five independent cultures per time point. Asterisks represent significant differences (Welch’s test, one asterisk for p<0.05 and two asterisks for p<0.01). Only comparisons between primed and non-primed groups are shown.
Fig 2.
Bacterial survival from naïve and primed populations triggering at 30 minute intervals after the priming stimulus (memory of the priming response).
During memory decline experiment, the viability of the cells remains unaltered before the addition of the trigger (A). The priming memory declined over time after induction of priming response with 0.1 mM H2O2 (trigger) (B). Asterisks indicate significant differences between each time-point pair (Welch’s test, one asterisk for p<0.05 and two asterisks for p<0.01).
Fig 3.
Heatmap of relative protein expression based on label-free quantification by liquid chromatography/mass spectrometry (LC-MS).
Hierarchical clustering of the intensities was performed using Euclidean distances between means. Rows indicate the fluctuation of protein (by gene names) level at 5 minutes after addition of 0.1 and 1 mM H2O2. Intensity ranges of the log2 fold-changes are given from highest intensity (green) to lowest (red). Only the 50 most statistically significant up-regulated proteins are shown, taking as a reference the 0.1 mM concentration.
Fig 4.
Heatmap of relative protein expression based on quantitative label-free quantification by liquid chromatography/mass spectrometry (LC-MS).
Hierarchical clustering of the intensities was performed by using Euclidean distances between means. Rows indicate the fluctuation of protein (by gene names) level at different time points. Intensity ranges of the log2 fold-changes are given from highest intensity (green) to lowest (red). Only the most 50 statistically significant up-regulated proteins at time point T1 (30 minutes after priming) are represented to follow their fluctuation in the other time points with 30 minutes between them.
Fig 5.
Projection of differentially translated proteins 30 minutes after addition of H2O2 on protein-protein interaction network of E. coli K12.
Pale green nodes indicate up-regulated proteins while pale red ones represent down-regulated ones. Note that the most critical nodes (DnaK and GloL) of the protein-protein interaction network are affected. The interaction among nodes shows the proteome-wide impact of H2O2.
Fig 6.
Priming experiments of different mutants of E. coli MG1655 in selected H2O2-responsive genes.
Asterisks represent significant differences (Welch’s test, one asterisk for p<0.05 and two asterisks for p<0.01). Only comparison between primed and non-primed groups are shown.
Fig 7.
Proportion of population extinctions (20 populations per treatment) due to exposure to increasing concentrations of H2O2 during experimental evolution of 40 individual populations.
Note that every passage was carried out every 24 hours although time in the x-axis is represented as continuous. Both, priming and triggering doses were increased twofold daily up to 32 mM, where total extinction occurred. The extinction was perceived by negative growth in the next passage and by the absence of growth in LB plates during contamination controls. Non-evolving population control (grey line, 20 populations) is also shown. Evolvability differs between the two treatments, naïve (red line) and primed (blue line) populations (Log-rank test, p< 0.01). Differences with non-evolved populations were not determined.
Fig 8.
Chromosomal mapping of mutations in H2O2-evolved populations.
The positions indicate the approximate locations of mutations in the E. coli MG1655 chromosome relative to the wild-type. Different types of mutations that were found in both evolving regimes are indicated in the left panel highlighted with different colours (non-synonymous amino acid changes in yellow, intergenic mutations in green and frameshift in red). The right panel indicates mutations that were present for both regimes (grey) and those that were exclusive to the priming regime (blue). To see full reports of mutation for each population see supplementary material.
Fig 9.
Representative fluorescent micrographs of attachment to a glass slide by E. coli MG1655 and its derivative mutant fimE (Δ1 bp, position 248 out of 597 nucleotides).
The image was taken after vital staining (LIVE/DEAD BacLight Bacterial Viability Kit). Top panel shows greater attachment and reduced sensitivity to H2O2 MIC associated with fimE inactivation. Complementation restores both H2O2 sensitivity and the low-attachment phenotype of the wild-type strain (lower panel). In both panels control cells (transformed with the cloning vector pCA24N) can be also observed.
Fig 10.
Motility test of E. coli MG1655 and its derivative mutants in the intergenic region between insB1 and flhD (insB1→flhD, positions 1978493 nt, Δ10bp and 1978504 nt, Δ1 bp).
Panel A shows the initial phenotypes of insB1→flhD mutants, low motility and resistance to 4 mM H2O2 while the parental strain is motile with a MIC of 4 mM H2O2. A complementation experiment showed identical phenotypes when the strains were transformed with the cloning vector (panel B) and recovery of motility and original resistance to 1 mM H2O2 when both mutants are transformed with the plasmid overexpressing the operon flhDC [80] (panel C).
Fig 11.
Frequencies of different types of mutations among evolved populations of both primed and non-primed conditions.
Mutational spectra. The number of mutations (nucleotide substitutions and indels) is plotted against respective nucleotide positions within the gene fragment. In this analysis, the mutations were taken by type regardless of the targets by generating two frequency datasets to compare frequency by type of mutation of evolved populations under priming and non-priming regimes. The mutational spectra are significantly different between the two conditions according to the spectrum analysis software iMARS [46].