Experimental evolution of a biocontrol agent: monitoring the adaptation of a bacteriophage (Tequintavirus) to different genotypes of Salmonella enterica

Following the exponential development of multi-resistant pathogenic bacteria in recent years, it has become urgent to find alternatives treatments. The use of bacteriophages (or phages) as biocontrol agents is a good approach. Indeed, these are viruses that have only infected bacteria for more than 3 billion years. They are generally specific to a bacterial species, or even a particular ST ("Sequence Type"), which gives them "specialist" status. In an ecological niche containing a bacterial population, environmental changes and pressure from predators like phages constitute selection pressures that generate bacterial genotypic diversity and resistance. Subject to co-evolution with their hosts, phages are also able to adapt in changing their genotypic frequencies. It is thus possible through experimental evolution to adapt a phage to several bacterial genotypes, and thus make it more "generalist". Given that the expansion of the host range of viral parasites is often associated with a fitness cost, is it possible to generate a more generalist organism without an associated fitness cost? We followed the adaptive path of a virulent phage (Tequintavirus) evolving in a spatially variable environment, composed of four susceptible and four resistant bacterial isolates (Salmonella enterica serotype Tennessee, ST5018 and ST319 respectively). We used at different times the Appelmans' protocol as experimental evolution, through serial passages on non-co-evolving bacterial isolates. We thus obtained several independent populations of evolved phages, with an expanded host range and increased virulence. Sequencing of evolved phage populations revealed multiple mutations occurring at the same loci (parallel mutations) in all populations, including exo- and endo-nuclease, dUTPase, and tail proteins. Two parallel mutations present in the Long Tail Fiber gene were fixed in the populations before the other parallel mutations. The introduction of these two mutations by reverse genetics into the ancestral genome broadened the host range but only substantially increased virulence, highlighting the effect of compensatory mutations.

Since the observed phenotypes generated by these mutations can vary depending on the environmental pressures exerted, we used a complex environmental model adapted for the study of phage/bacteria interaction: the Galleria mellonella animal model, in the larval state. The fitness of S. enterica Tennessee ST319 and ST5018, infected either by the ancestral phage or by the evolved phages, was monitored by bioluminescence following the addition of an iLux plasmid by cloning. Bioluminescence, a non-invasive method for monitoring bacterial kinetics, allowed us to obtain more precise quantitative data than the observation of survival/mortality.