Metabolic trade-offs hide unforeseen benefits of plasmids carriage
The link between fitness and reproduction rate is a central tenet in microbiology, and indeed evolutionary biology: Mutants reproducing faster than the dominant wild-type are favoured by selection, but otherwise the mutation is lost. This link was given by Ronald Fisher in 1930 under the assumption that fitness can only change through mutations that boost or hinder growth rate, whence the use of logarithms on growth data by experimentalists. Here I show that logarithms are highly sensitive to sampling times, resulting in fitness estimates that are not constant over the growth of bacterial cultures. This variability invalidates typical selection measurements, and unfit mutants can be co-maintained if they reach their equilibrium. And this is what I observed in competition assays between two Escherichia coli constructs, one of which harbours a non-transmissible plasmid that protects against tetracycline (pGW155B), without using the antibiotic. Despite growing 40% slower than its drug-sensitive counterpart, the construct harbouring the plasmid persisted throughout the competition. And, perhaps more importantly, maintained the plasmid. My study suggests that reliance on growth rate masks that selection on plasmid carriage may be stronger than previously thought--explaining the seemingly-paradoxical abundance of plasmids in nature.