Obtaining reliable estimates of mutation fitness effects and characterizing directly the dynamics of mutation appearance was impossible previously since mutations have never been witnessed in action in single cells. Such estimates have wide application and are critical for most evolutionary studies. We accomplished this in Escherichia coli by employing microfluidics, time-lapse imaging, and by visualizing mutations in single cells by using a fluorescent tag of the Mismatch Repair System.
Use microfluidics, time-lapse microscopy, and Escherichia coli to study for the first time directly at the level of single cells the dynamics of appearance of spontaneous mutations and their effects on fitness. Characterize using this approach and non-parametrically the whole distribution of spontaneous mutation fitness effects (without any assumption of form of this distribution). Determine the mutation rate variability in cells in the absence of stress or in the presence of some exogenous stress. Determine the origin of mutation rate variability in the absence of stress, in particular examine whether the cell cycle or proteome quality (translation or transcription fidelity) affects the mutation rate variability in unstressed cells. Determine the fitness consequences of increased variability in mutation rates.
To visualize mutations we use a method that we developed previously, where the fluorescent mismatch repair protein MutL forms fluorescent foci on replication errors that are detected in single cells by fluorescent microscopy. We developed an automated image analysis procedure to detect and track MutL foci, which allows precise quantification. To follow the dynamics of MutL foci appearance we grow cells in the mother machine chip containing a series of separate microchannels that are closed on one side and where cells grow in a single row. This specific geometry allows retaining and following by microscopy the cell abutting the dead end in each microchannel through hundreds of consecutive divisions. For fitness measurements we imaged cells growing in >1000 microchannels of mother machine by phase contrast microscopy in parallel every 4 minutes during 3 days . This allowed measuring two fitness components, growth rate and survival, at single cell level with high temporal resolution and on large timescales. The specific geometry of microchanneles allows realizing a perfect bottleneck of one individual at each generation, thus measuring fitness for the first time in the complete absence of natural selection.
Our study shows that spontaneous mutations in unstressed cells do not occur by puffs but following a Poisson process. We also show that the rate of mutations is not constant but varies by 3 times in unstressed cells due to the cell cycle. We found that the average effect of a mutation is very small, 0.31%, well below the previous estimates. We characterized in a non-parametric way the whole distribution of mutation effects of on fitness, which is dominated by almost neutral mutations and with very rare deleterious mutations. Our study shows in addition that deleterious mutations appear with a constant rate suggesting a limited macroscopic epistasis between mutations having a strong effect and those with a very weak effect on fitness. Finally, we show that 1% of spontaneous mutations are lethal in Escherichia coli.
Future work will aim to analyze the effect of environment as well as exogenous stress on the dynamics of mutations and their fitness effects. We will also examine the mechanism by which exogenous stress generates mutation rate variability in Escherichia coli cells as well as the fitness consequences of increased variability in mutation rates during exogenous stress. This work is relevant to many human health phenomena such as bacterial virulence, antibiotic resistance, and cancer. All of these processes are driven by mutations and all can be accelerated if some cells in the population were transiently or permanently mutating more than the rest of the population.
1. «Mutation dynamics and fitness effets followed in single cells«, manuscript in revision at Science, submitted on March 1, 2017 2. «Method for detecting and studying genome-wide mutations in single living cells in real-time«, manuscript accepted at Spr
Understanding how mutations arise is critical for understanding both evolution and pathological cell physiology. Due to their rarity, mutations were difficult to detect directly. For the most of the history of genetics they were inferred by detecting different phenotypes in populations but this approach allows only non-neutral mutations to be scored. In recent years, whole genome sequencing made it possible to detect all non-lethal mutations that separate two lineages but this approach has limited temporal resolution. Furthermore, all mutation detection methods used until now are based on the study of populations to infer how mutation appear and therefore lack information on the single-cell variability of the mutagenesis process. Therefore, measuring precisely the rate at which mutations occur by the presently available mutation detection methods is complex.
We address, for the first time, directly, on the single-cell level, how mutations appear and how they affect the cell fitness, i.e. death, growth, and reproduction. For this, we combine an unprecedented way of investigating mutagenesis via genetic tools and the imaging of live Escherichia coli cells and novel microfluidic technologies allowing long-term experiments on single cells under controlled conditions. The methodology we use allows detecting whole-genome mutations, is direct, non-invasive, high-throughput (million of cells), and on large timescales (hundreds of generations).
Our approach to investigating mutagenesis allows several fundamental questions to be uncovered. First, it allows investigating if mutation rates are constant in populations during unperturbed growth, under stress, and during aging of single cells. Second, our system is unique as it detects mutations prior to selection. This allows, for the first time, scoring even lethal mutations and therefore establishing, in an unbiased way, the distribution of mutation fitness effects, i.e. rates of neutral, deleterious, lethal and beneficial mutations. Third, it allows investigating how the cell-to-cell phenotypic variability, resulting from stochasticity of expression of genes important for DNA repair, affects the cell-to-cell mutation rate variability. Finally, it allows addressing evolutionary consequences of the among-cell mutation rate variability in populations.
All these questions have attracted great attention for decades but were impossible to address previously. They are relevant to several evolutionary phenomena. Estimates of rates of different mutation types are key parameters in many evolutionary theories such as those trying to explain the evolution of recombination, aging, frequency of mutator strains in microbial populations and extinction of small population. However, few empirical data exist on these rates. Mutation rate heterogeneity in populations is relevant to the evolution of antibiotic resistance, microbial pathogenesis strategies, radiation resistance, chemotherapy resistance and tumor progression. In fact, the acquisition of beneficial mutations could benefit from the transient increases in mutation rates in some cells in the population. In the case of pathogenic bacteria, this could concern phase variation allowing rapid host immune system evasion. In the case of cancer, five different mutations at least, are necessary for acquiring a fully malignant phenotype. To date it is still not clear if these mutations accumulate gradually or if they are acquired in a single step. The last scenario is incompatible with the ordinary rates of mutations but is achievable if a subpopulation of cells mutate substantially more than the rest of the population.
Madame MARINA ELEZ (LABORATOIRE JEAN PERRIN)
The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.
LJP LABORATOIRE JEAN PERRIN
Help of the ANR 347,335 euros
Beginning and duration of the scientific project: September 2014 - 48 Months