- Top of page
- Materials and Methods
- Conflict of Interest
- Data Accessibility
Studies on life-history traits let us expect that many of the phenotypic differences observed among individuals have a genetic basis (Roff 1992; Stearns 1992). There are no reasons to assume that parasite life-history traits would be an exception of this pattern (Poulin 1996), although evolution of parasite life-history traits is less studied using the concepts of life-history theory. This may be surprising because understanding the processes that maintain genetic variation in transmission-related traits is fundamental for our understanding of how parasite transmission evolves. Transmission, on the other hand, is a key factor underlying evolution of parasite virulence (defined here as harm to the host) as well as epidemiological dynamics of infection and impact of parasites on host populations.
Parasites that reproduce clonally at some stage of their life cycle provide interesting opportunities to test how much of the phenotypic variation has a genetic basis. Clonal stages allow testing of how much of the variation in phenotypic traits among the same genotype results from environmental effects (Falconer and Mackay 1996). The comparison of genetically identical individuals has classically been used in twin studies, in experiments with clonal plants and animals, and more recently also with clonal parasite stages (see e.g., Anderson et al. 2010; Koehler et al. 2011). Variation among genotypes reflects the opportunity for natural selection to operate on life-history trait expression, whereas a high variability within a genotype reflects phenotypic plasticity or classical bet-hedging strategies that are beneficial in stochastic environments (Fenton and Hudson 2002; Kussell and Leibler 2005; Donaldson-Matasci et al. 2008; Reece et al. 2009).
One potential mechanism maintaining variation in life-history traits of parasites is genotype-by-environment interactions (G × E). The speciality of complex (indirect) life cycle parasites is that in some cases genetic variation among parasite genotypes includes the effects of the interaction with unique host genotypes. Such an intimate interaction is particularly common in parasites with complex life cycles where propagules produced by sexual reproduction (egg or miracidia larva) infect host genotypes that also originate from sexual events. These unique parasite–host genotype combinations are not replicated and both the genetic and phenotypic effects of the host genotype become part of the parasite genotype's extended phenotype. Nevertheless, it is valuable to evaluate how much of the variation in trait expression among parasite genotypes is due to genetic effects, even if the contribution of host genotype cannot be separated from the direct effects of the parasite genotype. The magnitude of genetic variance among parasite genotypes allows to evaluate how much of the fitness differences among parasite genotypes are caused by external environment. In evolutionary ecology such genetic variation that includes both heritable and nonheritable variation is coined in measures of “broad-sense” genetic variability (see e.g., Falconer and Mackay 1996).
The special features that the intimate interaction of parasites with their host and their external environment creates motivate studies of their evolutionary ecology. Developmental conditions (e.g., individual quality, immune status, and number of conspecific parasites) within the hosts tend to vary substantially, and often unpredictably, among different host individuals (Poulin 2007; Reece et al. 2009). Therefore, genetically identical parasites are not necessarily phenotypically similar as they have to deal with different host environments (Poulin 1996). Conditions for parasite development have important implications for initial infection success, within-host infection dynamics, and transmission success to the next host. On one hand, it is generally assumed that hosts in good condition are hostile environments for parasites as they are effective in resisting infections. On the other hand, hosts in bad condition might provide fewer resources for parasite growth and reproduction (Ebert 2000; Bedhomme et al. 2004; Ebert et al. 2004), and the resulting shorter duration of infection (due to eventual death of the host) could lead to reduced parasite fitness. The effect of host condition on host susceptibility (Krist et al. 2004) and the effect of environmental stress on parasite virulence (Jokela et al. 1999, 2005; Pulkkinen and Ebert 2004) have been studied intensively, but studies concerning the effect of host condition on parasite traits have received less attention (but see Logan et al. 2005; Tschirren et al. 2007). For example, in snail–trematode interactions host individuals supplied with low-quality food release fewer larval stages compared with hosts provided with high-quality diets (Keas and Esch 1997; Sandland and Minchella 2003; Seppälä et al. 2008). This suggests that host exploitation by parasites is dependent on host condition, which again could determine the quality of larval stages (Seppälä et al. 2007, 2008).
In this study we focus on the variation in life-history traits of trematodes, a group of parasites with a complex life cycle typically including three transmission steps, an endothermic final host, a mollusc first intermediate host, and an invertebrate or vertebrate second intermediate host. Trematodes show distinctive behavioral (reviewed by Combes et al. 1994; Haas 1994), physiological, and morphological traits (Koehler et al. 2011), which are potentially important during transmission between the hosts. Transmission-related traits of larval stages (e.g., cercarial output, activity, survival, and infection success) are good candidates for characters that determine transmission success and subsequent parasite fitness as they affect the likelihood that the parasite reaches its final host. However, maximizing all traits simultaneously is practically impossible due to physiological and ecological constraints (Partridge and Harvey 1988). For example, natural selection could favor parasite genotypes that produce high numbers of clonally produced larval transmission stages in the first intermediate host, as this could increase the probability that the genotype will successfully encounter the second intermediate host. On the contrary, a high investment in asexual reproduction could be constrained by low quality of the larvae or increased host mortality (Davies et al. 2001) and decrease parasite fitness. For free-swimming parasite stages such as trematode cercariae, active movement is an important determinant of transmission success as the parasites must localize and enter the next host. However, increased energy (glycogen) consumption through movement shortens their life span (Ginetsinskaya 1988) as these larval stages do not feed. The longevity of the parasite larvae is important as longer survival increases the time that the parasite genotype can use to localize the subsequent host. Moreover, a successful parasite genotype should be infective, but infectivity may trade-off with other traits as shown in a snail–schistosome system, where cercarial production in the snail host was negatively correlated with infectivity to a mouse definitive host (Davies et al. 2001).
In this study, we measured genotype-level variation in transmission-related traits of Diplostomum parasites and studied experimentally how host (freshwater snail, Lymnaea stagnalis) nutritional status affects these traits. We measured traits from parasites originating from field-collected snails that were manipulated for their nutritional status by dividing them into two food treatments (“ad libitum food” and “no food”) for 2 weeks. The food manipulation was applied to examine how differences in host condition affect parasite traits and to identify possible trade-offs between the traits as allocation theory predicts that trade-offs should be most evident when resources are limiting (e.g., Bazzaz et al. 1987).
- Top of page
- Materials and Methods
- Conflict of Interest
- Data Accessibility
Transmission-related traits are important determinants of parasite fitness. In addition to genetic differences among parasite genotypes, environmental factors such as the condition of the previous host may affect infection success of the parasite in the next host. In this study we examined genotype-specific transmission characteristics of clonal parasite stages (D. pseudospathaceum, Trematoda) and tested if food deprivation of the snail intermediate host affected these traits.
We found considerable variation in all the measured transmission traits among parasite genotypes that originated from naturally infected snails. Depending on the trait, 51–95% of the variance was attributable to parasite genotype in the broad sense that we use it in this study. In addition to parasite genotype, experiment week and genotype-by-week interaction were the most important effects in the statistical tests. These results are in line with those from a marine trematode (Maritrema novaezealandensis) that show significant differences among cercarial clones in morphology, photoreactive behavior, and survivorship (Koehler et al. 2011), as well as with those from malaria parasites (Plasmodium sp.) for which substantial intraspecific life-history variation is commonly found (e.g., Mackinnon and Read 1999; Eisen 2000; Eisen and Schall 2000). Moreover, a recent study by Rieger et al. (2013) found large differences among four different clonal lines of D. pseudospathaceum in traits such as overall hatching success, time of hatching, and infection rates in snails and three-spined sticklebacks. Taken together, these results suggest that genotype-level variation in transmission traits is common in many parasite species.
Although we observed relatively high (ranging between 0.27 and 0.67) and stable H2 estimates of cercarial activity and infection success, the genotypes showed different mean trait values on different weeks, indicating that genotypes switched ranks during the experiment in both treatment groups. As H2 estimates include the genetic and phenotypic variation introduced by the host individuals, the estimates are literally clonal repeatabilities including variation that usually would be estimated under maternal and epistatic effects (see below). Therefore, our data give evidence for high clonal repeatability within the clones measured at a given time point, but the phenotypic effects of the snail host may also contribute to the large week-to-week phenotypic variation within the parasite genotypes. Thus, variation in transmission traits within a genotype may be adaptive or nonadaptive for the parasite, but requires further study.
The variance of a trait may be itself subjected to selection and parasites may exhibit different levels of variability in life-history traits as a result of selection in hosts or habitats that differ in predictability (Poulin 2007). D. pseudospathaceum parasites infect several different freshwater fish species (Rellstab et al. 2011) and have highly motile birds as final hosts which ensure high levels of parasite gene flow among snail populations (Louhi et al. 2010). In such a situation intraclonal variability could assure that at least a small proportion of the phenotypes always match to the available hosts (Cosseau et al. 2010). This idea is known as bet-hedging strategy (reviewed by Hopper 1999) and could be useful in variable environments, where it could maximize the likelihood that some clones will survive and transmit successfully (Fenton and Hudson 2002; Kussell and Leibler 2005; Donaldson-Matasci et al. 2008).
Another possibility is that the observed variation is nonadaptive for the parasite and mainly reflects temporal host–parasite dynamics in the snail. Although we minimized phenotypic snail effects by keeping all snails on an ad libitum diet for considerable time period before the experiment and by using snails infected with only one parasite genotype, it was not possible to control for all snail effects. Thus, variation among the parasite genotypes also includes the genetic and phenotypic effects of the snail host (Eisen and Schall 2000). For example, the age of infection in the snail host could affect parasite traits, if the amount of resources that are available for the parasite changes when the infection spreads in host tissue (Seppälä et al. 2008). However, if the amount of resources declined in some of the experimental snails, we would have expected to see a gradual decline in the parasite trait values in such snails and no significant weekly variation in trait values as observed here. Therefore, it is more likely that temporal dynamics related to cercarial production in the snail could have affected the parasite traits. Such dynamics could be caused by differences in the infection phase of the snails as trematode infections typically alternate between cercarial output and production of new daughter sporocysts within the snails (Ward et al. 1988; Kechemir and Théron 1989; Touassem and Théron 1989). For example, cercarial performance could be determined by the level of synchrony in cercarial release and host energy level (see Karvonen et al. 2004a). Such temporal variation stresses the challenge of measuring parasite fitness–related traits in practice as a measurement taken in one time point might or might not correlate with the lifetime performance of that particular clone. If such “maternal effects” of the snail explain the differences then parasite genotypes should show substantial phenotypic plasticity in response to these temporal dynamics in the snail. Unfortunately, it was not possible to separate effect of the genetic and “snail maternal effects” as it would require an experimental design where several snail individuals are infected with the same parasite genotype. However, the advantage of our approach is that the broad-sense genetic variation that we measured gives a more realistic estimate of the variation present in natural parasite populations.
The importance of selection on traits that contribute to transmission success is difficult to estimate without detailed understanding how particular traits interact (Agnew and Koella 1999). For example, we found a trade-off between cercarial output and activity (at the beginning of the experiment) and a positive correlation between activity and infection success (“ad libitum” group, week 1). The trade-off between activity and cercarial output indicates that parasite quality may be constrained when the parasite invests resources into reproduction. The positive correlation observed between activity and infectivity, on the other hand, suggests that those parasite genotypes that are actively swimming have also higher infection success in the next host. However, the instability or inconsistency of the genetic correlations indicates that the positive and negative genetic correlations may vary over time due to environmental effects (see Gutteling et al. 2007), such as the temporal dynamics in the snail. Alternatively, it may suggest that trade-offs do not constrain evolution of these traits. Interestingly, no correlations or trade-offs were observed in the “no-food” group. This is somewhat contradictory as trade-offs should become more likely under unfavorable conditions (e.g., Bazzaz et al. 1987). In this case the scarcity of genetic correlations in weeks 1 and 2 might depend on the smaller sample size (n = 15–16 vs. 32 on week 0) or on weak treatment effects as discussed below.
The manipulation of food availability for the snail (L. stagnalis) host of D. pseudospathaceum explained only a small fraction of the variation observed in parasite life-history traits. The treatment groups did not differ in mortality of the snails, which might indicate that the 2-week food manipulation was not enough to cause severe starvation in the snails collected from Lake Vuojärvi. However, in our previous experiment that used snails from another lake, L. Huumojärvi, similar food deprivation increased the mortality of the snails and decreased the number of released cercariae from the “starved” snails (Seppälä et al. 2008). The difference between these results may be explained by the larger snail size in Vuojärvi compared with Huumojärvi (shell length 53 mm vs. 44 mm), when snails might have more resources to resist short periods of food stress. In addition, in this study, the snails were kept in lower ambient temperature (15°C in this study vs. 20°C in Seppälä et al. 2008), which might have reduced snail metabolism and led to weaker treatment effects.
In summary, these data reveal that transmission-related life-history traits of D. pseudospathaceum are highly variable within a naturally infected snail population. This is in line with the high levels of neutral genetic diversity reported for this trematode species (Louhi et al. 2010). The variation in parasite traits was not affected by a 2-week starvation treatment of the snail hosts. Instead, we found significant differences among the genotypes in all four transmission traits examined as well as significant temporal variation in the traits within each genotype. This may be indicative of an adaptive bet-hedging strategy, where temporal within-genotype variation increases the likelihood of some parasite clones surviving and reaching the next host. Alternatively, such variation may result from phenotypic plasticity in response to temporal dynamics within the snail hosts.