Printer Friendly

Host-parasite coevolution: evidence for rare advantage and time-lagged selection in a natural population.

The Red Queen hypothesis relies on time-lagged selection by parasites against common host genotypes. Such selection potentially sets up oscillations in genotype frequencies in both the host and the parasite and can lead to selection for cross fertilization as a strategy for the genotypic diversification of offspring (Jaenike 1978; Hamilton 1980; Bell 1982; Hamilton et al. 1990). The potential for oscillatory dynamics is well known from computer simulations (e.g., Seger 1988; Seger and Hamilton 1988), but the idea has been difficult to test in nature. One reason for this difficulty is that it is virtually impossible to know the genotypes for sexual hosts and parasites at the loci that govern the interaction. Another difficulty is that multigenerational (long-term) studies are needed on natural populations where the prevalence of infection can be easily determined.

Here we report the results of a five-year study of dynamics in a clonal population of freshwater snails (Potamopyrgus antipodarum) and its trematode parasites, focusing on the dominant parasite, Microphallus sp. We took advantage of the fact that interaction loci in the clonal host are linked to the allozyme loci that we use to distinguish among clones. We tracked the frequency of four common snail clones over the five-year period in a natural lake population on the South Island of New Zealand. If the clones are oscillating over time, we expected to find regular changes in the frequencies of the snail clones over the course of the study. In addition, if parasites respond to host dynamics, we expected to find that common clones become heavily infected and that these clones would be driven down in frequency in future samples. Finally, if parasites provide a rare advantage, then recently common host clones should be more infectible than rare clones, which we tested in a laboratory infection experiment at the end of the field study. The results were consistent with these general expectations and with the predictions of computer simulations specifically designed to mimic the study system.

Natural History and Background Information

Potamopyrgus antipodarum is a small (4-6 mm) prosobranch snail that is native to the lakes and streams of New Zealand. Populations of the snail are typically comprised entirely of parthenogenetic females, but some populations contain mixtures of sexual and asexual individuals, making it an ideal system for contrasting the alternative models for the maintenance of sex in eukaryotes. Parthenogenetic females are triploid, and egg production is apomictic; sexual females, in contrast, are diploid (Dybdahl and Lively 1995b). Both reproductive modes appear to be fixed (there is no evidence for switching between sexual and asexual reproduction in our laboratory stocks). Generation time in the laboratory (18 [degrees] C) is approximately four months.

The snail serves as the first intermediate host to at least 14 species of digenetic trematode parasites (Winterbourn 1974; MacArthur and Featherston 1976), but we were specifically interested in an undescribed species of digenetic trematode in the genus Microphallus, which is responsible for most of the trematode infections in lake populations of the snail (Lively 1987, 1992). This trematode produces metacercarial cysts in the snail in three to four months under laboratory conditions, at which time it is competent to infect the final hosts (a variety of waterfowl and wading birds). The cysts "hatch" following ingestion by the final host, and the resulting worms produce cross-fertilized eggs within several days, which then pass into the environment. Snails may become infected after ingesting the eggs; a snail typically becomes infected by only one egg (Dybdahl and Lively 1996). The infection of snails by Microphallus is thought to be genetically based because clonal genotypes vary in their levels of Microphallus infection (Dybdahl and Lively 1995a; Jokela et al. 1997) and Microphallus populations are adapted to infect local snail populations (Lively 1989).

We began a long-term study of the Potamopyrgus-Microphallus system in 1992 in an effort to determine whether there was any evidence for time-lagged selection by parasites. We chose Lake Poerua on New Zealand's South Island for this study because it contains a multiclonal, asexual population of snails (Dybdahl and Lively 1995a). Lake Poerua is a small (2.15 [km.sup.2], 4.2 km long), low-elevation (124 m) glacial lake, located in a maritime climate on the west side of the Southern Alps (Livingston et al. 1986). The advantage of using a clonal host for this study is that allozyme markers are linked to loci that determine the genotypes for defense against parasites. Therefore, genetic changes in the host population could be equated with changes in frequency of the different allozyme genotypes (clones). The parasites on the other hand are sexual and outcrossed (Dybdahl and Lively 1996), and their interaction loci are not likely to be linked to visible genetic markers. However, we were able to indirectly track genetic changes in the parasite population by examining samples of snails that were infected by trematodes. The frequency of a particular genotype in the parasite population is indirectly revealed by the frequency of individual clones among infected snails.


We constructed a model of parasite-host coevolution to generate predictions for the special case of clonal hosts interacting with sexual parasites. We were specifically interested in how parasite virulence and infection rates affected the time lag for the parasite response. We also wanted to determine the sensitivity of the results to the number of host generations between samples. In our field study, we sampled once per year (instead of once per generation); we therefore wanted to know how the number of generations per year affected the parasite's response in the next sample.

The model is based on two alleles at each of two haploid loci that determined infection. Hosts were clonal with four possible genotypes, and parasites were sexual with four possible genotypes and free recombination between loci. A successful infection resulted from a match of alleles at both loci between parasite and host. The frequency of each clonal host genotype ([H[prime].sub.ij]) in generation t + 1 was calculated as:

[Mathematical Expression Omitted], (1a)

[Mathematical Expression Omitted], (1b)

[Mathematical Expression Omitted], (1c)

[Mathematical Expression Omitted]. (1d)

In these equations, T is the probability of encountering a parasite and V, the parasite virulence, gives the relative fitness of infected hosts (e.g., if V = 0.8, then relative fitness is 1 - 0.8 = 0.2). The frequency of allele 1 at the first locus is p and q is the frequency of allele 1 at the second locus. Hence, the probability of infection of a clonal host having genotype 11 ([H.sub.11]) by a randomly drawn parasite is equal to the product of p and q. Because T gives the probability of encounter with a parasite, the overall probability of infection for a host having genotype [H.sub.11] is Tpq; we refer to this probability as [P.sub.11]. Similarly [P.sub.12] = Tp(1 - q); [P.sub.21] = T(1 - p)q; and [P.sub.22] = T(1 - p)(1 - q). Finally, [Mathematical Expression Omitted] is the sum of the numerators on the right-hand side in equations (1a-d).

We calculated the frequency of different host clones among infected individuals ([I.sub.ij]) each generation as:

[Mathematical Expression Omitted], (2a)

[Mathematical Expression Omitted], (2b)

[Mathematical Expression Omitted], (2c)

[Mathematical Expression Omitted], (2d)

where [Mathematical Expression Omitted] is the sum of the numerators on the right-hand side. This model allowed us to obtain "data" from the simulation ([P.sub.ij] and [I.sub.ij]) in the same way as we collected it in the field.

The frequencies of alleles in the parasite population after selection were calculated as:

[Mathematical Expression Omitted] (3a)


[Mathematical Expression Omitted], (3b)


[Mathematical Expression Omitted]. (4)

The frequency of parasite alleles in the next generation was calculated as

p[prime] = [p.sup.*] - [p.sup.*]m + (1 - [p.sup.*])m (5a)


q[prime] = [q.sup.*] - [q.sup.*]m + (1 - [q.sup.*])m, (5b)

where m is the mutation rate or migration rate. Mutation or migration is often included in models of parasite-host coevolution to prevent fixation during cycling (e.g., Hamilton et al. 1990; Howard and Lively 1994). We used a mutation/migration rate of m = 0.003.

This coupled set of deterministic equations exhibits both regular and chaotic cycles (cf. May and Anderson 1983; Seger 1988) depending on parasite virulence and encounter rates [ILLUSTRATION FOR FIGURE 1 OMITTED]. Increasing either virulence or encounter rates decreases the cycle length and increases the infection rate of the population (averaged across genotypes over time). Under a wide range of conditions, the parasite response lags behind changes in the host clone (as shown by Hutson and Law 1981; Bell 1982). For example, the frequency of a parasite genotype (not shown) and the frequency of a host clone in the infected sample ([I.sub.ij]) lags behind and overshoots the frequency of a clone in the population [ILLUSTRATION FOR FIGURE 1 OMITTED]. The probability of infection of a host clone ([P.sub.ij]) also lags behind the clone's frequency in the population at large, but the overshoot varies with encounter rates [ILLUSTRATION FOR FIGURE 1 OMITTED]. During the time lag, a common clone is decreasing in the population at large while increasing in the infected population as the parasite alleles that match the clonal genotype overshoot. Because of this time-lagged response, changes in the frequency of a host clone between generations should be strongly correlated with the parasite response after the time lag, when compared to the contemporaneous parasite response.

We wanted to know whether this prediction about the timelagged correlated response was robust to different numbers of snail generations per yearly sampling interval and different infection rates of the population. Because infection by Microphallus has large fitness consequences, we were especially interested in the predictions when virulence was high. Using output from simulations of the model with different values of T and V, we calculated the changes in the frequency of each host clone in the population at large,

[Delta][H.sub.ij] = [H.sub.ij,t+n] - [H.sup.ij,t], (6a)

where n = 1, 2 . . . 5 is the number of generations over which change in frequency was calculated. We also calculated the response in the clone-specific infection rate,

[Delta][P.sub.ij] = [P.sub.ij,t+n] - [P.sup.ij,t+[Tau]], (6b)

where [Tau] = 0, 1 . . . 8 is the interval in generations between the changes and the response. We then examined the correlation between [Delta][H.sub.ij] and [Delta][P.sub.ij] using different values for [Tau]. These correlations were calculated for all four clones in the model across 100 generations and then averaged. We also examined the correlations between [Delta][H.sub.ij] and [Delta][I.sub.ij] (the response in clone frequencies in the infected sample) and obtained similar results. We observed the same general relationship between the correlation and [Tau], regardless of the number of generations over which change in frequency within a sample was calculated (n). Figure 2 shows the correlation between [Delta][H.sub.ij] and [Delta][P.sub.ij] when changes are calculated over three generations (n = 3), the interval probably covered by our yearly sampling of the Lake Poerua population. The magnitude of the correlation always peaks when the response in the clone's infection rate is calculated after a lag of three-to-six generations (3 [less than] [Tau] [less than] 6), and is lower for the contemporaneous response ([Tau] = 0). This pattern holds for a wide range of population infection rates, as determined by encounter rates and virulences [ILLUSTRATION FOR FIGURE 2 OMITTED]. This pattern is also robust to a wide range of starting conditions for mutation rate and initial genotype frequencies, as long as oscillations exist. Thus, the response of the infection rate of the host clone after a one-year lag in sampling ([Tau] = about three snail generations) should be more strongly correlated with changes in host clone frequency (compared to the contemporaneous response) if time-lagged dynamics are occurring.


Dynamics in a Natural Population

To monitor host-parasite interactions in a natural population, we collected snails from rocks along the same interval of shoreline in Lake Poerua each February (1992-1996). This five-year period probably represents about 10-15 snail generations. We obtained two separate samples from each collection: a random sample and a sample of infected snails. Random samples were comprised of 70-150 randomly selected adult individuals (shell length greater than 3 mm). Infected samples were comprised of adult individuals that were infected with Microphallus (Microphallus-infected), or with any of the other trematode parasites (non-Microphallusinfected). Both random and infected samples were assayed using cellulose acetate electrophoresis to identify five-locus allozyme genotypes that we refer to as different clones (details in Dybdahl and Lively 1995b).

We compared the clonal representation in random and infected samples using hierarchical log-linear analyses and likelihood ratio [[Chi].sup.2] statistics, which can be constructed to evaluate the independence of the frequencies of clones in the two samples. If dynamic coevolution is occurring in Lake Poerua, we expected clone frequencies to change over time in random samples (population at large) and in infected samples (reflecting a response in the parasite population). We first examined the year-by-clone interaction for each sample (random and infected) using five clonal groupings, the four common clones and a group containing all rare clones. Because this term was significant, we examined each clone individually to determine which clones changed in frequency in the random sample and in the infected samples.

Because we were interested in whether parasites tracked host clones with a time lag, we examined whether clones become overinfected in the year of their peak in abundance or in the following year using the clone-by-sample interaction term. A clone would be overinfected if its proportion in the infected sample exceeds its proportion in the random sample (Dybdahl and Lively 1995a). For those common clones that were overinfected, we examined whether their frequencies were driven down by testing the year-by-clone interaction term in years following overinfection.

These predictions for over- and underinfection of clones and correlated changes over time can be compared to the predictions for two alternative models of parasitism. First, if there is no genetic basis to parasitism and parasites infect different host genotypes randomly, clones would rarely be significantly over- or underrepresented among infected individuals. Second, if parasite effects on fitness are too weak, parasite-host coevolution would not drive demographic changes in host clones. Thus, the parasite population would be at a frequency-dependent equilibrium where the most common host genotype would always be most parasitized.

Tracking and overinfection of common clones by parasites should lead to high rates of clone-specific infection by both Microphallus and non-Microphallus parasites, and to clonal dynamics in the random sample. We calculated clone-specific infection rates for each year based on a clone's frequency in the random samples, and in the infected samples, and on the population-wide infection rate by each parasite group. Thus, infection rate for each clone was calculated as (I/H)L, where I is the frequency of a clone in the infected sample for each parasite group (Microphallus or non-Microphallus), H is the frequency of a clone in the random sample, and L is the population prevalence of that parasite group (Table A2). Infection rate for all parasites is the sum of the infection rates of both parasite groups. If the parasite response is timelagged, then changes in host clone frequencies should be more strongly correlated with the responses in the infection rates after a time lag, rather than with contemporaneous responses [ILLUSTRATION FOR FIGURE 2 OMITTED]. For each of the four common clones, we calculated the changes in clone frequencies from one year (y) to the next (y + 1) in the random sample ([H.sub.ij,y+1] - [H.sub.ij,y]), and the response in the estimated infection rate for the same set of years ([P.sub.ij,y+1] - [P.sub.ij,y]) and for the set of years after a one-year time lag ([P.sub.ij,y+2] - [P.sub.ij,y+1]). The correlations between clone frequency changes in the random sample and contemporaneous and lagged responses in the infection rates were evaluated using Spearman rank-correlation coefficients (Norusis 1990) and compared to the predictions of the model.

A Laboratory Test for Rare Advantage

We expected that the Microphallus population would have evolved to attack the clones that were common during our field study. To test this, we examined the infectibility of common versus rare clones by Microphallus in a laboratory experiment using mice as the final host (for general methods, see Lively 1989; Lively and Jokela 1996).

To obtain both snails and parasites for the experiment, we collected thousands of snails from the shoreline rocks in Lake Poerua in February 1996. These snails were transported to the Edward Percival Field Station in Kaikoura, New Zealand, where the infection experiment was conducted. Snails from the collection were dissected and the metacercariae from a total of 25 infected snails were fed to each of four laboratory mice. Mouse fecal pellets were collected between days 2 and [TABULAR DATA FOR TABLE 1 OMITTED] 6 after the mice ingested metacercariae. The fecal pellets were distributed evenly among four 1-L containers of fresh water. Water was changed in these containers several times daily to wash away most of the fecal material and concentrate the Microphallus eggs, which sink to the bottom of the container. At the same time, we set up four replicate containers with 150 snails from the Lake Poerua shoreline collection. The concentrated eggs from each container were then added to each of the experimental snail populations daily after repeated washings. Snails were kept in the containers with the concentrated parasite eggs for 24 days, with water changed twice per day. The snails were then packed in damp paper towels and transported to Indiana University where they were held in 4 L of water with regular water and food change.

After about 90 days from initial exposure to parasite eggs, 75 snails were sampled from each replicate. We dissected each snail to determine its infection status and to preserve snail tissue samples (head and foot) for electrophoretic analysis of its five-locus allozyme genotype. We recognized three asexually reproduced larval stages within the snail that appear sequentially over a period of approximately 70-100 days in the laboratory: early germinal, blastocercarial, and metacercarial stages. In the blastocercarial and metacercarial stages, most of the snail tissue in the shell apex has been replaced and the snails are sterilized because females no longer carry brooded eggs or embryos. We scored each infection according to five categories related to developmental stage of the parasite: (1) germinal; (2) predominantly blastocercaria; (3) mixed blastocercarial and metacercaria; (4) predominantly metacercaria; and (5) all metacercaria. Infections in stage 4 or less were considered to be experimental infections, and Microphallus infections of stage 5 or of other parasite species were considered to have occurred in the field (background infection). Thus, background infections could easily be distinguished from experimental infections. Consistent with this, background infection rates by Microphallus sp. determined at the end of the experiment (0.049 [+ or -] 0.011 SE, n = 287) matched closely and did not exceed the infection rates by Microphallus sp. in Lake Poerua at the time the sample was collected (0.083 [+ or -] 0.014 SE, n = 420). Because the infection rate by metacercariae in the experiment was less than that in the field population, a few metacercariae-infected individuals may have died during the experiment.

We focused our analysis on those snails that were infected in the experiment. We classified the snails by their clonal identity as one of the four recently common clones (12, 19, 22, 63), which accounted for 47% of the sample, or as a fifth group called "rare" clones, which contained 40 different clones (53% of the sample). We used a hierarchical log-linear analysis to examine how clonal identity and replicate affected the likelihood of infection. Significant clone-by-infection terms would suggest that Microphallus differentially infected clonal groups. We report likelihood ratio [[Chi].sup.2] statistics from a backward model selection routine in SPSS to determine the significance of the interaction terms (Norusis 1990).


Dynamics in a Natural Population

Infections of P. antipodarum in Lake Poerua during the study were comprised of trematodes from eight groups, but Microphallus accounted for between 50% and 66% of all parasite infections except in 1994 when both Microphallus and Stegodexamene anguilla were prevalent (Table A1). Infection rates of adult snails by Microphallus and by nonMicrophallus parasites also varied throughout the study (Table A2). Rates of infection of adult snails by all parasites were lowest in 1995 (9.0%) and highest in 1996 (16.9%). Levels of infection by Microphallus varied by more than twofold, from about 3% to 11%.

We discovered a total of 112 different clones over the five years, but only four clones were abundant in the samples in any year (Table 1). Clonal diversity in Lake Poerua was relatively high, with an average of 0.29 clones/individual in the random samples over the five years. However, only four clones (12, 19, 22, 63) individually exceeded 15% of the random sample and ranked either first or second in frequency in at least one of the years. The four clones together comprised over 50% of the random sample in four of the five years and 60% of the random sample in a fifth year. Because the four clones exist with a background of rare clones, their frequencies can vary independently.

Overall, the frequency of different clones in the random and infected samples changed across years (year-by-cloneby-sample interaction, [[Chi].sup.2] = 59.14, df = 32, P = 0.002). Considering the random sample alone, the frequencies of four common clones changed significantly over the five years (Table 2, [ILLUSTRATION FOR FIGURE 3 OMITTED]). Clones 19, 22, and 63 changed significantly across years when tested individually, and changes in the frequency of clone 12 were nearly significant. Clone 12 was the most common in the first year, but less common in later years. Clone 22 peaked in abundance in the second year. [TABULAR DATA FOR TABLE 2 OMITTED] Clone 19 was the most common in the third year, and declined thereafter; and clone 63 increased sharply in the fourth year.

Changes in the Microphallus population were indicated by significant changes in the frequency of different clones in the Microphallus-infected sample over the five years (Table 2). The sample of Microphallus-infected snails was dominated by clone 12 in 1992, clone 22 in 1993 and 1994, clone 19 in 1995, and clones 12 and 19 in 1996. In the sample of snails infected by non-Microphallus parasites, only clones 22 and 63 changed significantly over time (Table 2).

Significant overinfection of a clone following its peak in abundance suggests the occurrence of time-lagged overshoot. Three different clones (12, 22, and 19) were disproportionately parasitized by Microphallus over the period of the study (Table 3). Clone 12 was significantly overinfected in 1992 when it was almost twice as common among Microphallusinfected individuals compared to its frequency in the population (see also Dybdahl and Lively 1995a). The overinfection in clone 22 and 19 occurred with a time delay following their peak abundance (Table 3). In the year of its peak abundance in 1993, clone 22 was not disproportionately represented in the Microphallus-infected sample. However, clone 22 was over three times more common among Microphallusinfected individuals in 1994 (31%) than in the population at large (9%). Clone 19 became common in 1994 (after being rare in 1993), and was significantly underinfected in 1994. Underinfection of clone 19 when common may account for the low prevalence by Microphallus in 1994. However, clone 19 was overinfected in 1995 and 1996, and this overinfection was significant in 1995 (Table 3).

Clone 63 did not become overinfected by Microphallus in 1996, following its peak in 1995. However, there was some evidence for tracking by non-Microphallus parasites of clones 22 and 63, but no others (Table 3). Both clones were relatively common in the non-Microphallus infected sample in the year after their peak, but neither was significantly overinfected. Conversely, clone 19 was significantly underinfected by nonMicrophallus parasites during and after its peak in the random sample.

Tracking and overinfection of common clones by Microphallus, along with high rates of clone-specific infection by both Microphallus and non-Microphallus parasites, could have led to the observed clonal dynamics in the random sample [ILLUSTRATION FOR FIGURE 4 OMITTED]. Overinfection of clone 12 and clone 22 were associated with high rates of infection and significant changes in clonal abundance in the following years. Although changes in the frequency of clone 12 were marginally nonsignificant across all four years (Table 2), Clone 12 decreased significantly between 1992 and 1993 (year-by-clone interaction, [[Chi].sup.2] = 4.27, df = 1, P = 0.039). Clone 22 declined significantly in frequency between 1993 and 1994 ([[Chi].sup.2] = 4.38, df = 1, P = 0.036), and also between 1994 and 1995 ([[Chi].sup.2] = 4.64, df = 1, P = 0.031). It is interesting to note that clone 22, which had the strongest dynamics, was overrepresented in both Microphallus and non-Microphallus infected samples, and was infected at a high rate in 1994. A significant decline in clone 63 ([[Chi].sup.2] = 4.55, df = 1, P = 0.033) was associated with high rates of total infection by parasites. However, the trend of decline in abundance between 1995 and 1996 for clone 19 was not significant ([[Chi].sup.2] = 0.834, df = 1, P = 0.364) despite tracking and high rates of infection by Microphallus.

The strong correlation between changes in clone frequencies and the time-lagged response in infection rates further suggests that parasite infection rates evolved in response to clonal dynamics. Changes in clone frequencies in the random sample were correlated with their time-lagged, but not with their contemporaneous, changes in clone-specific rates of Microphallus infection [ILLUSTRATION FOR FIGURE 4 OMITTED]. The weak correlation between clonal dynamics and the contemporaneous responses in parasite infection rates are due to time-lagged overshoot (Spearman [Rho] = 0.091, n = 12, P = 0.779 for Microphallus infection rates; [Rho] = -0.009, n = 12, P = 0.829 for total infection rate). For example, between 1993 and 1994 clone 22 declined in the random sample but the infection rates in clone 22 increased. Similarly, clone 19 increased in the random sample between 1993 and 1994, but did not change in the infected sample. However, the correlation between changes in random sample dynamics and the responses one year later in infection rates were positive, significant, and matched the expectations based on the model ([Rho] = 0.748, n = 12, P = 0.026 for Microphallus infection rates; [Rho] = 0.811, n = 12, P = 0.001 for total infection rate; [ILLUSTRATION FOR FIGURE 4 OMITTED]). Random sample changes in clone 22 and 19 were matched by responses in the infection rate in the following year. Consequently, the changes in clone frequencies were more strongly correlated with the timelagged responses than the contemporaneous responses in the infection rate.

A Laboratory Test for Rare Advantage

The experimental infection of Lake Poerua snails resulted in high and consistent levels of new Microphallus infection in each of the replicate containers (average prevalence of experimental infection = 69.2%, range among replicates was 64-72%). There was no significant difference among replicates in the prevalence (replicate-by-infection interaction, [[Chi].sup.2] = 1.369, df = 3, P = 0.713) or among replicates in the experimental prevalence in different clones (replicate-by-infection-by-clone interaction [[Chi].sup.2] = 5.471, df = 12, P = 0.94). Therefore, we grouped all replicates together in order to examine clone-specific prevalences.

The five clonal groups differed significantly in the prevalence of experimental infections (infection-by-clone interaction, [[Chi].sup.2] = 64.336, df = 4, P [less than] 0.0001; [ILLUSTRATION FOR FIGURE 5 OMITTED]). Rare clones were the least infected group and significantly less infected than the common clones as a group ([[Chi].sup.2] = 41.391, df = 1, P [less than] 0.0001). Among the four common clones, there was significant heterogeneity in experimental infection rate ([[Chi].sup.2] = 22.943, df = 3, P [less than] 0.0001). Three were heavily parasitized in the experiment (clones 12, 22, and 63; [ILLUSTRATION FOR FIGURE 5 OMITTED]). The experimental prevalence in clone 19 (65%) was very close to the average prevalence in the experiment.


We tracked the frequencies of four clonal genotypes of the snail P. antipodarum over a five-year period to determine whether selection by parasites is operating in a time-lagged manner. The overall pattern of host-clone frequencies over time gave evidence of host cycling [ILLUSTRATION FOR FIGURE 3 OMITTED] as well as timelagged correlated responses by parasites [ILLUSTRATION FOR FIGURE 4 OMITTED]. In addition, experimental exposures to parasites showed that recently common clones were more infectible than rare clones, indicating that rare genotypes have an advantage under parasite attack [ILLUSTRATION FOR FIGURE 5 OMITTED]. These results are consistent with Red Queen models of host-parasite interactions (see Jaenike 1978; Hamilton 1980; Hutson and Law 1981; Bell 1982; Seger 1988; Seger and Hamilton 1988; Hamilton et al. 1990).

Time-lagged selection by Microphallus was indicated by overinfection of clones that were common during the field study. One was overinfected at the beginning of the study (clone 12; whether this response occurred after a time lag is unknown), and two became sequentially overinfected in the field, following a one-year time lag, during the course of the study (clones 22 and 19). In addition, the high infection rate of the fourth clone (clone 63) in the 1996 laboratory experiment is consistent with a time-lagged response [ILLUSTRATION FOR FIGURE 5 OMITTED]. All four clones declined in frequency in association with their disproportionately high levels of infection in the field [ILLUSTRATION FOR FIGURE 3 OMITTED], and two of these four declined significantly. Curiously, clone 12 did not decline to the same low level as that observed in the other two clones [ILLUSTRATION FOR FIGURE 3 OMITTED]. This result may indicate the presence of a second, rare clone with the same multilocus genotype or it may be the consequence of the complicated, uneven dynamics that can occur in interactions with highly virulent parasites [ILLUSTRATION FOR FIGURE 1 OMITTED]. In any case, even when the dynamics are very complicated, the parasites are expected to show time-lagged correlated responses to host genotypes [ILLUSTRATION FOR FIGURE 2 OMITTED], which we observed in the natural population [ILLUSTRATION FOR FIGURE 4 OMITTED]. Furthermore, clone 22 seems to have completed an entire cycle during the field study.

This "tracking" of common host genotypes by Microphallus could help maintain host genotype diversity because Microphallus is highly virulent, and thus provides a strong source of selection. Specifically, high levels of infection in the field could help account for clonal diversity in Lake Poerua, [TABULAR DATA FOR TABLE 3 OMITTED] which is high compared to lakes with either all clonal or mixed clonal and sexual populations (Dybdahl and Lively 1995a,b; Fox et al. 1996). Unfortunately, we cannot determine from these results whether the trematode parasites are sufficient to maintain sexual hosts in mixed (sexual and asexual) populations, but some of the necessary conditions would seem to be met. This later point is also illustrated by the advantage that rare clones had in the laboratory infection experiment. The group of four clones that were common at some point between 1992 and 1996 were significantly more infected by the field-derived Microphallus than the assemblage of 40 rare clonal genotypes [ILLUSTRATION FOR FIGURE 5 OMITTED]. We can not rule out the possibility that some of these rare clones may increase in future samples, as clones 22 and 63 have done during the field study. In any case, these rare clones mimic the diversity of rare genotypes that might be expected in a sexual population and show that there is an advantage to rarity independent of the mode of reproduction.

There is one reasonable alternative explanation for a timelagged overinfection of a clone that relies on a response by the parasite to host age, rather than a genetic response. Because the cumulative probability of infection by trematodes increases with host age (Jokela and Lively 1995), the infection rate in a growing population of a clone would be low, but increase as cohorts of younger individuals live another year. We measured the length of snails in the random samples in 1995 and 1996 and we can use this size as a correlate of age to examine this explanation. Contrary to the prediction that an overinfected clone would be comprised of older individuals, shell length did not differ significantly among the common clones in the random sample (12, 19, 63, and all others) in 1995 or in 1996 (one-way ANOVAs; 1995: n = 95, F = 2.408, df = 3, P = 0.072; 1996: n = 155, F = 1.873, df = 3, P = 0.137). In 1995, when clone 19 was significantly overinfected and clone 63 was underinfected, clone 19 was slightly smaller than clone 63, contrary to the prediction from the demographic explanation. In any case, no pairwise differences in length were significant in a posthoc test (Sheffe's test, df = 91, P [greater than] 0.147 for all comparisons). In addition, parasitism in Microphallus is known to have a genetic basis, and the infection experiment suggests that the Microphallus population has responded genetically to clone frequency [ILLUSTRATION FOR FIGURE 5 OMITTED].

It is difficult to be certain that the dynamics observed here are actually caused by parasites, but alternative explanations for clonal dynamics would require complicated or indirect ecological mechanisms. For example, dynamics in the quality of ecological niches of different clones could result in the exchange of dominance between different clones over time, with parasites responding to changes in clone frequencies but not driving them. We know from other studies that clones are often restricted to specific depth-stratified vegetation zones (Fox et al. 1996), but the snails in the present study were all collected from the same shallow-water habitiat in Lake Poerua. During the field study of Lake Poerua, declines in common host clones were associated with clone-specific overinfection by Microphallus and high trematode infection rates. To explain this association and the time-lagged correlated response, niche qualities would have to change just after overinfection of clones occurred. Such patterns of timelagged oscillations are more likely explained by antagonistic species than by time-lagged intraspecific interactions (Bell 1982). Furthermore, the observed time-delayed selection by highly virulent trematode parasites seems more likely to drive host dynamics than changes in niche qualities. The levels of parasite infection in our field study (Table A2, [ILLUSTRATION FOR FIGURE 3 OMITTED]) were concordant with those that produced such dynamics in our model [ILLUSTRATION FOR FIGURES 1, 2 OMITTED].

The apparent time-lagged overinfection of common clonal genotypes in P. antipodarum over five years should in theory cause both host and parasite to recycle the same types over and over (Clarke 1976; Hutson and Law 1981; Bell 1982; Seger and Hamilton 1988), but cycles may lead to the random loss of parasite alleles especially if cycles are chaotic (e.g., May and Anderson 1983; Seger and Hamilton 1988). Such cycles might be sustainable by Microphallus because the high gene flow among lake populations should restore genetic variation and genetic (allozyme) diversity within populations is relatively high (Dybdahl and Lively 1996). For the host, we would like to know whether clones recycle (e.g., clone 22) in a multiclonal host population like the one in Lake Poerua, or whether new clones arise. There are theoretical reasons to believe that some clones driven through cycles will accumulate mutations during bottlenecks, thus impacting their fitness compared to younger clones (Howard and Lively 1994; Lively and Howard 1994). Whether such cyclical dynamics are sustained will require additional years of tracking host genotype changes in Lake Poerua.

In summary, there is evidence from the PotamopyrgusMicrophallus system that suggests that parasites play an important role in genotype-frequency dynamics. Microphallus infects different host clones nonrandomly, shows a timelagged response to common clones, and depresses common clones after periods of overinfection. Although we used a clonal population to study Red Queen dynamics in nature, our infection experiment confirmed that the parasite has evolved to infect common genotypes, producing an advantage to rare genotypes. Because Microphallus seems to be maintaining a dynamic polymorphism in P. antipodarum and creating a selective advantage to rare genotypes, some essential features of the parasite theory for the maintenance of sexual reproduction have been demonstrated.


We thank J. Jaenike, D. Ebert, and an anonymous reviewer for comments on the manuscript and J. Jaenike for suggesting the demographic alternative and that we include a transmission parameter in the model. We also thank members of the Department of Zoology at the University of Canterbury for logistical support, especially J. McKenzie, J. Van Berkel, F. Sin, and M. Winterbourn. Our work was supported by U.S. National Science Foundation grants BSR-9008848, DEB9317924, and DEB-9629849, and by the Marsden Fund of New Zealand (contract LL 0501).


BELL, G. 1982. The masterpiece of nature: the evolution and genetics of sexuality. Univ. of California Press, Berkeley, CA.

BELL, G., AND J. MAYNARD SMITH. 1987. Short-term selection for recombination among mutually antagonistic species. Nature 328: 66-68.

CLARKE, B. 1976. The ecological genetics of host-parasite relationships. Pp. 87-103 in A. E. R. Taylor and R. Muller, ds. Genetic aspects of host-parasite relationships. Vol. 14. Blackwell Scientific, Oxford.

DYBDAHL, M. F., AND C. M. LIVELY. 1995a. Host-parasite interactions: infection of common clones in natural populations of a freshwater snail (Potamopyrgus antipodarum). Proc. R. Soc. Lond. B Biol. Sci. 260:99-103.

-----. 1995b. Diverse, endemic and polyphyletic clones in mixed populations of the freshwater snail Potamopyrgus antipodarum. J. Evol. Biol. 8:385-398.

-----. 1996. The geography of coevolution: comparative population structures for a snail and its trematode parasite. Evolution 50:2264-2275.

FOX, J. A., M. F. DYBDAHL, J. JOKELA, AND C. M. LIVELY. 1996. Genetic structure of coexisting sexual and clonal subpopulations in a freshwater snail (Potamopyrgus antipodarum). Evolution 50: 1541-1548.

HAMILTON, W. D. 1980. Sex versus non-sex versus parasite. Oikos 35:282-290.

HAMILTON, W. D., R. AXELROD, AND R. TANESE. 1990. Sexual reproduction as an adaptation to resist parasites: a review Proc. Natl. Acad. Sci. USA 87:3566-3573.

HOWARD, R. S., AND C. M. LIVELY. 1994. Parasitism, mutation accumulation and the maintenance of sex. Nature 367:554-557.

HUTSON, V., AND R. LAW. 1981. Evolution of recombination in populations experiencing frequency-dependent selection with time delay Proc. R. Soc., Lond. B Biol. Sci. 213:345-359.

JAENIKE, J. 1978. An hypothesis to account for the maintenance of sex within populations. Evol. Theor. 3:191-194.

JOKELA, J., AND C. M. LIVELY. 1995. Spatial variation for infection by digenetic trematodes in a population of freshwater snails (Potamopyrgus antipodarum). Oecologia 103:509-517.

JOKELA, J., C. M. LIVELY, J. A. Fox, AND M. F. DYBDAHL. 1997. Flat reaction norms and "frozen" phenotypic variation in clonal snails (Potamopyrgus antipodarum). Evolution 51:1120-1129.

LIVELY, C. M. 1987. Evidence from a New Zealand snail for the maintenance of sex by parasitism. Nature 328:519-521.

-----. 1989. Adaptation by a parasitic trematode to local populations of its host Evolution 46:1663-1671.

-----. 1992. Parthenogenesis in a freshwater snail: reproductive assurance versus parasitic release. Evolution 46:907-913.

LIVELY, C. M., AND R. S. HOWARD. 1994. Selection by parasites for clonal diversity and mixed mating. Philo. Trans. R. Soc. B Biol. Sci. 346:271-281.

LIVELY, C. M, AND J. JOKELA. 1996. Clinal variation for local adaptation in a host-parasite interaction. Proc. R. Soc. Lond. B Biol. Sci. 263:891-897.

LIVINGSTON, M. E., B. J. BIGGS, AND J. S. GIFFORD. 1986. Inventory of New Zealand lakes. Part II. South Island. Water and Soil Miscellaneous Publication no. 81. National Water and Soil Conservaton Authority, Wellington, New Zealand.

MACARTHUR, C. P., AND D. W. FEATHERSTON. 1976. Suppression of egg production in Potamopyrgus antipodarum (Gastropoda: Hydrobiidae) by larval trematodes. N.Z.J. Zool. 3:35-38.

MAGURRAN, A. E. 1988. Ecological diversity and its measurement. Princeton Univ. Press, Princeton, NJ.

MAY, R. M., AND R. M. ANDERSON. 1983. Epidemiology and genetics in the coevolution of parasites and hosts. Proc. R. Soc. Lond. B Biol. Sci. 219:281-313.

NORUSIS, M. J. 1990. SPSS advanced statistics user's guide. SPSS Inc., Chicago.

PARKER, E. D., JR. 1979. Ecological implications of clonal diversity in parthenogenetic morphospecies. Am. Zool. 19:753-762.

SEGER, J. 1988. Dynamics of some simple host-parasite models with more than two genotypes in each species. Philo. Trans. R. Soc. Lond. B Biol. Sci. 319:541-555.

SEGER, J., AND W. D. HAMILTON. 1988. Parasites and sex. Pp. 176-193 in R. E. Michod and B. R. Levin, eds. The evolution of sex. Sinauer, Sunderland MA.

WINTERBOURN, M. J. 1974. Larval Trematoda parasitizing the New Zealand species of Potamopyrgus (Gastropoda: Hydrobiidae). Mauri Ora 2:17-30.
COPYRIGHT 1998 Society for the Study of Evolution
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Dybdahl, Mark F.; Lively, Curtis M.
Date:Aug 1, 1998
Previous Article:Does Bateman's principle apply to broadcast-spawning organisms? Egg traits influence in situ fertilization rates among congeneric sea urchins.
Next Article:Cytological characterization of premeiotic versus postmeiotic defects producing hybrid male sterility among sibling species of the Drosophila...

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |