Competition and facilitation between parasites as drivers of virulence evolution – EVOLVIR

We did experiments in the laboratory with two different species of spider mite, Tetranychus urticae and T. evansi, and a plant virus Tomato spotted wilt virus on mini tomato plants.

Our first series of experiments assessed how TSWV impactsed spider mite life-history traits (number of adult offspring, offspring sex ratio). We did this by comparing traits on plants infected with TSWV or on virus-free plants. We also investigated how spider mites affected TSWV load and how order of exposure to T. evansi and TSWV impacted facilitation. Another experiment investigated if T. evansi were more attracted to plants infected with TSWV.

Next, we evolved the spider mite Tetranychus evansi in single infections or coinfection with TSWV, which facilitates T. evansi, making the host easier to exploit, or in competition with T. urticae, for 33 generations on whole tomato plants. Each generation we transferred 255 T. evansi onto new plants. In the coinfection with TSWV, T. evansi were transferred onto plants infected with naive TSWV each generation. In the coinfection with T. urticae, T. evansi were transferred onto plants with 145 naive T. urticae adult females each generation. In all treatments females were left to lay eggs, offspring develop and after 2 weeks we placed another tomato plant beside the first. T. evansi were allowed to walk onto this second plant during a period of one week and then 255 adult T. evansi were collected from this second plant and transferred onto a new plant.

After evolution we measured the virulence, growth and transmission of all lines in single infections and coinfections with TSWV and T. urticae. We also measured the size of evolved lines, egg size and the ability of evolved lines to impact TSWV growth.

We established that the virus has a positive impact on both species of spider mites, increasing the number of offspring they have. In contrast, spider mites have no effect on viral load. Both spider mites and virus infections reduce plant height, but when they infect the plant together in coinfection plant height is reduced even more. We found out that the facilitation still occurs, regardless the order of exposure to the different parasites, but the relative degree of facilitation is higher when TSWV infects first.

After evolving in coinfection it seems that spider mites evolved with competitors have higher levels of transmission. It also seems that mites evolved with competitor mites or with the virus can maintain some transmission from more damaged hosts. Spider mites that evolved in single infections did not have any transmission when hosts were really sick. These results still need to be confirmed.

Currently we are trying to get a better understanding of mites and viruses circulating on tomato plants in agricultural settings. Some of the species I work with were discovered in Australia and Europe in parallel. I am in Australia at the moment and would like to establish whether communities are similar in Australia and France.

I would like to establish more complex communities of mites and viruses across multiple plants in laboratory settings to see how interactions among species affects epidemics.

Theory predicts that coinfections, when a host harbours more than one parasite, can drive epidemics and virulence evolution. Competition between parasites for shared host resources, or mediated via the immune system, are predicted to select for higher virulence. In contrast, facilitation between parasites may select for lower virulence. Thus, coinfection may have wide-reaching consequences for public health, agriculture and the basic understanding of processes shaping natural communities. Despite this, there are no empirical tests of the above predictions on parasite virulence evolution.

This project uses experimental evolution to address the interplay between coinfection driven selection in the within-host environment, and multiple infections across a population of hosts (different parasite strains or species co-circulating) for virulence evolution. It is important to consider the relationship between parasite traits in the within- and between-host environment because higher virulence favoured in a coinfected host may not be selected across the parasite population (if it cannot transmit to new hosts).

The positive and negative effects of coinfection on virulence evolution will be investigated using different parasite species (the spider mites Tetranychus urticae and T. evansi, and tomato spotted wilt virus, TSWV) infecting the tomato host (Solanum lycopersicum). T. evansi and T. urticae compete for shared host resources and interact via the host immune system; TSWV positively affects T. urticae via negative immune cross-talk and/or the release of free amino acids.

Objective 1 will investigate whether competition (between T. urticae and T. evansi) and facilitation (between T. urticae and TSWV) can drive virulence evolution in the within-host environment when parasites are constrained to be in coinfection each generation with natural transmission to new hosts. Experimental evolution will be coupled with mathematical models to address how these interactions impact virulence evolution. This will reveal whether interactions in the within-host environment are important for virulence evolution, but it is unlikely that sequential coinfection at each generation reflects epidemiological dynamics in natural populations.

Objective 2 will address this by allowing parasites to evolve in semi-natural host populations harbouring multiple infections; different parasite species co-occurring in a host population, not necessarily in the same host. Thus parasites may be in a coinfection one generation and a single infection the next. Here interactions in the within- host environment will only drive virulence evolution if there is a high frequency of coinfected hosts. Objective 2 will also identify natural tomato populations with multiple infections of T. evansi and T. urticae, and TSWV and T. urticae, to see if dynamics in the laboratory parallel epidemics in the field.

This project builds on previous studies of parasite evolution in both content and approach. Spider mite parasites are unique in that individual life-history traits can be linked to population measures of virulence. Theoretical and laboratory experiments will be employed to reveal the impact of different interaction types, and for the first time the relative importance of the within- and between-host environment, for virulence evolution in coinfections investigated. Empirical tests of theoretical predictions will provide proof of concept showing that coinfection driven parasite evolution can occur over relatively short time-scales. This approach to studying parasite evolution in individual plants, reinforced by the epidemiological component, and following evolution in laboratory and natural populations, will make headway towards a deeper understanding of the evolutionary consequences of coinfections.