Fig 1.
Lysis plaques produced by phages T5 and T7 on E. coli MG1655 are enlarged in the presence of sublethal doses of antibiotics.
(a, b) Aspect of the lysis plaques of phages T5 and T7 at the maximum sub-inhibitory concentration of the tested antibiotics. (c, d, e, f, g, h, I, j). Boxplots of lysis plaque radius measurements at increasing concentrations of: ciprofloxacin (Cp), ceftazidime (Cz), mecillinam (Mc), chloramphenicol (Cm) and kanamycin (Km) for phages T5 and T7. The gradient of concentration used was right below the inhibitory concentration. N = for each condition, between 18 and 414 plaques were measured. Whiskers represent the range of the data within 1.5 times the interquartile range from the lower quartile and upper quartile, mean values are shown as crosses. The asterisks represent significant pairwise comparisons, with significance determined by a linear model (ANOVA followed by post-hoc pairwise comparisons). Three asterisks represent a p-value of less than 0.001, one asterisk represents a p-value of less than 0.05; whereas “n.s.” represents p-values higher than 0.05 in the same test. P-values are listed in S1 Table.
Fig 2.
Morphological effects of sublethal antibiotics in E. coli MG1655.
(a) Phase-contrast microscopy of E. coli microcolonies trapped in a bidimensional LB-agarose 1% pad, in the presence of the same antibiotic concentrations tested in Fig 1. (b) Scatter plot showing the mean bacterial cell width and length of E. coli populations treated with increasing concentrations of ciprofloxacin (Cp), ceftazidime (Cz), mecillinam (Mc), chloramphenicol (Cm) and kanamycin (Km). N = for each condition, between 171 and 2684 individual bacteria were measured. Whiskers represent the standard error of the mean.
Fig 3.
dCas9/sgRNA-driven morphological changes in bacteria produce larger lysis plaques.
(a) Effect of dCas9/sgRNA repression of key regulators of bacterial morphology and its impact on plaque size of phages T5 and T7. First column: phase-contrast microscopy of E. coli LC-E75 microcolonies under an agarose pad. Each image showcases a microcolony of the LC-E75 strain carrying a sgRNA that either does not target any E. coli locus (psgRNA), targets ftsZ open reading frame (psgRNA-ftsZ) or target mreB open reading frame (psgRNA-mreB), in the presence of aTc at maximum sublethal dose. Second and third columns, images of lysis plaques of phages T5 and T7 on LCE75 carrying the same sgRNAs and under the same induction concentrations. (b) Scatter plot showing the mean bacterial width and length of E. coli LC-E75 in the presence of dCas9 and the sgRNAs mentioned above, at increasing inducer (aTc) concentration. N = For each condition, between 369 and 4,286 individual bacteria were measured, a total of 26,978 cells analyzed. Whiskers represent the standard error of the mean. (c, d, e) Measurements of phage T5 lysis plaque radii on E. coli LC-E75 with either a non-targeting sgRNA, sgRNA-mreB, or sgRNA-fstZ. (f, g, h) Measurements of phage T7 on E. coli LC-E75 with either a non-targeting sgRNA, sgRNA-mreB, or sgRNA-fstZ. N = For each condition, between 21 and 352 plaques were measured. Whiskers represent the range of the data within 1.5 times the interquartile range from the lower quartile and upper quartile, mean values are shown as crosses. The asterisks represent significant pairwise comparisons, with significance determined by a linear model (ANOVA followed by post-hoc pairwise comparisons). Three asterisks represent a p-value of less than 0.001, one asterisk represents a p-value of less than 0.05; whereas “n.s.” represents p-values higher than 0.05 in the same test. P-values are listed in S1 Table.
Fig 4.
Effects of dCas9-mediated morphological changes on key parameters influencing epidemic spread.
(a, b) One step growth curve of phage T5, and phage T7 in E. coli LC-E75 displaying different morphological changes. Time = 0 corresponds to the time of phage-bacteria mixing. Induction of the morphological change by dCas9 transcription inhibition was carried out for 2 hours prior to the beginning of the infection. Morphological changes at the time of the phage addition were confirmed by phase-contrast microscopy. Concentrations of the inducer were the maximum sublethal concentrations tolerated by the strain. N = 3 curves per condition. Error bars represent the standard error of the mean. (c) One step growth curve of phage T7 in E. coli MG1655 in the presence of the antibiotics mecillinam and ciprofloxacin at their maximum subinhibitory concentrations. Time = 0 corresponds to the time of phage-bacteria mixing. (d) Phase contrast microscopy images of time lapse growth of bacteria in the presence of mecillinam in E. coli MG1655 or after repression of the mreB gene by the dCas9/sgRNA in strain LCE-75. Cells were cultured in liquid medium and pre-treated with the antibiotic or the inducer one-hour prior sampling and microscopy imaging. (e) Percentage of a given surface area (900 µm²) occupied by microcolonies of E. coli under different conditions over 105 minutes of growth at 37°C. N = between 30 to 43 microcolonies were followed and measured per condition. Error bars represent the standard error of the mean. Three asterisks represent a p-value of less than 0.001 for an upper-tailed t-test comparing values at time = 105 minutes.
Fig 5.
Relative squared speeds of expanding plaques as a function of changes in morphological parameters of cell in mecillinam and ciprofloxacin treated cells.
The theoretical model leads to an analytical expression of relative squared speeds as a function of morphological parameters (length and width) and the phage physiological parameters as given in Equation 2. For mecillinam-treated cells, bacterial length remains unchanged while width increases, resulting in a linear relationship between squared speed and width (Eq. 3), as shown in Fig 5a. The dots represent experimentally measured speeds as a function of cell width, which align well with the model’s predictions. Similarly, in ciprofloxacin-treated cells, bacterial width remains constant while length changes (Eq. 4). In this case, the experimentally determined relative squared speed (dots) is plotted against cell length, and the model’s fit is shown in Fig 5b.