On the evolution of pesticide resistance in Phytophthora infestans : an experimental evolution approach

Resistance to antimicrobial agents is a serious problem for both medicine and agriculture. The initial success of such toxins is due to absence of resistant genotypes in pathogen populations before treatment. The initial low frequency of resistance may be explained by negative pleiotropic effects of the resistance mutations on fitness in the absence of the toxin. However, when resistant and sensitive natural isolates are compared, no cost of resistance may often be seen. This high fitness of resistant genotypes is explained by the existence of compensatory mutations, i.e. mutations that ameliorate the negative effect of the initial resistance mutation. The accumulation of these compensatory mutations can only occur at a sufficient rate if the population of resistant genotypes is large, which in part depends on the pesticide use. As a result, resistance may be stably maintained in a population after pesticide treatment, whereas it was absent before pesticide treatment. Resistance to the phenylamide pesticide metalaxyl is common in populations of Phytophthora infestans, the oomycete that causes the late blight disease in potato. This resistance evolved within a few years after metalaxyl became commercially available in 1977. Such fast evolution of pesticide resistance may present a serious threat for potato production. However, fast evolution of resistance to other pesticides has not been described in P. infestans, whereas resistance to metalaxyl and other pesticides are common in other oomycetes. In the absence of metalaxyl, resistant isolates of P. infestans have a higher fitness during epidemics compared to sensitive isolates. This observation has led to the assumption that metalaxyl resistance has a direct positive effect on epidemic fitness. A reduction of the survival during the winter has been put forward as an explanation for the absence of resistance before the commercial release of metalaxyl. However, given a high mutation rate towards metalaxyl resistance, a direct positive effect of metalaxyl resistance would always lead to a high frequency of resistance, even in the absence of metalaxyl. In this thesis, we report on an experimental evolution approach to investigate the fast evolution and apparent stability of metalaxyl resistance in P. infestans. We used an experimental evolution procedure because it has several advantages. First, it enables a detailed inspection of the trajectory of evolutionary changes (e.g. in resistance phenotype). Second, it enables comparisons of fitness components between near-isogenic genotypes that only differ in their sensitivity to metalaxyl, whereas conclusions based on comparisons of field isolates and offspring from sexual crosses are hampered by differences in the genetic background. Two experiments were performed to investigate the rate of evolution of resistance to a number of pesticides. First, we used a fluctuation test to estimate the mutation rate towards resistance against metalaxyl, fluazinam and cyazofamid. This fluctuation test, however, was not suitable to estimate the mutation rate in P. infestans, due to background growth of sensitive genotypes for metalaxyl and due to the low amount of zoospores that can be obtained from a single colony for the other two pesticides. Second, we analyzed the response of selection on low pesticide concentrations. Serial transfer of zoospores on a constant low concentration of metalaxyl resulted in full resistance. The same procedure on a low concentration of cyazofamid with a similar effect on growth rate, however, did not result in full resistance. The genotypes that emerged after the selection procedure were compared with the sensitive ancestor with respect to a number of fitness components in the absence of metalaxyl. This approach did not reveal a cost of resistance. On the contrary, resistant genotypes showed a higher in vitro fitness compared to their sensitive ancestor. However, the results from a population survey during a nation-wide ban of metalaxyl containing pesticides in the Netherlands were suggestive of a negative effect of resistance on winter survival and on epidemic fitness. These results suggest that the relative high frequency of metalaxyl resistant isolates on untreated fields can only be explained by assuming that metalaxyl resistance has an initial cost that is followed by compensatory mutations that only restores epidemic fitness. A new selection procedure that included serial transfer of zoospores, revealed a small initial cost on in vitro performance by increasing the lag phase before the start of growth. Subsequent compensatory evolution under the selection conditions could then explain the absence of a cost of resistance in previous experiments. However, no such compensatory evolution was directly observed in this latter selection procedure. The results presented in this thesis suggest that metalaxyl resistance is selected on a wide range of concentrations, whereas this is not the case for (some) other pesticides. This may explain the relative fast evolution of resistance to metalaxyl relative to that for other pesticides. Additionally, the absence of resistance before the commercial release of metalaxyl can be explained by a negative effect due to delaying growth. Finally, the methodology of experimental evolution may be a valuable addition to the multitude of methods used to study the population dynamics of this important pathogen