Knowledge about the distribution of mutational fitness effects (DMFE) is essential for many evolutionary models. In recent years, the properties of the DMFE have been carefully described for some microorganisms. In most cases, however, this information has been obtained only for a single environment, and very few studies have explored the effect that environmental variation may have on the DMFE. Environmental effects are particularly relevant for the evolution of multi-host parasites and thus for the emergence of new pathogens. Here we characterize the DMFE for a collection of twenty single-nucleotide substitution mutants of Tobacco etch potyvirus (TEV) across a set of eight host environments. Five of these host species were naturally infected by TEV, all belonging to family Solanaceae, whereas the other three were partially susceptible hosts belonging to three other plant families. First, we found a significant virus genotype-by-host species interaction, which was sustained by differences in genetic variance for fitness and the pleiotropic effect of mutations among hosts. Second, we found that the DMFEs were markedly different between Solanaceae and non- Solanaceae hosts. Exposure of TEV genotypes to non- Solanaceae hosts led to a large reduction of mean viral fitness, while the variance remained constant and skewness increased towards the right tail. Within the Solanaceae hosts, the distribution contained an excess of deleterious mutations, whereas for the non- Solanaceae the fraction of beneficial mutations was significantly larger. All together, this result suggests that TEV may easily broaden its host range and improve fitness in new hosts, and that knowledge about the DMFE in the natural host does not allow for making predictions about its properties in an alternative host.
Mutations are the raw material on which natural selection operates to optimize the fitness of populations. The occurrence of selection and its strength depend on the effect that mutations may have on the survival and reproduction of individuals: mutations can be lethal, deleterious, neutral, or beneficial. Thus, determining how many mutations belong to each of these categories is of importance for predicting the evolutionary fate of a population. For emerging infectious diseases, this distribution determines the likelihood that a pathogen crosses the species barrier and successfully infects a new host. We characterized such distributions across a panel of alternative hosts for a plant virus and found that fitness effects of individual mutations varied across hosts in an unpredictable way and that many mutations considered deleterious in the natural host may turn out to be beneficial in other hosts.