The persistence of sexual reproduction in natural populations remains as one of the major unsolved problems for evolutionary biology. The problem stems from the fact that male-producing, sexual populations are subject to invasion and rapid replacement by clonal females. In fact, a population of one million sexual individuals would be replaced in less than 50 generations by a clone beginning with a single asexual female (explanation). Why, then, is there sex? A similar paradox exists for recombination. The effect of recombination is to break up the non-random associations between alleles at different loci. Why would this be favored by selection, especially if these non-random associations (linkage disequilibria) were generated by selection in the previous generations? Why break up a good thing? Over 20 hypotheses have been suggested to explain the evolutionary stability of sexual reproduction, most of which focus on the possible advantages gained by the production of variable offspring. Many of these ideas also apply to recombination. The solution seems far from certain, but most of the effort at present is focused on two general areas: the advantages of sex in clearing deleterious mutations, and the advantages of sex in variable environments. We have been engaged in both empirical and theoretical studies of these hypotheses (link to theoretical studies). Empirical studies. We work on a New Zealand snail in order to contrast the predictions of the leading theories. The snail is useful for this purpose because sexual and asexual females exist, and often coexist in the same (mixed) population (more about the snail). We have taken a strong-inference approach of forcing the different theories to make different falsifiable predictions, which can be evaluated using biogeographic and/or experimental evidence. Our work to date has been inconsistent with many of the leading hypotheses, but we have been unable to falisfy the idea that coevolution with parasites provides an advantage to cross-fertilization in hosts, and vice versa (the Red Queen hypothesis). The basic idea behind the Red Queen hypothesis is that parasites will be under strong selection to infect the most common host genotypes. This kind of frequency-dependent selection will favor rare host genotypes, which should increase in frequency over time, eventually becoming common. The parasite is then under selection to infect these previously rare, but now common host genotypes. This sort of coevolutionary interaction easily leads to oscillations in genotypic frequencies in both the host and the parasite. Under the Red Queen hypothesis, sexual reproduction is selected over asexual reproduction as a mechanism to produce variable progeny, some of which may have relatively rare genotypes, which escape infection. The gist of the idea can be seen, perhaps, by asking this question: if I have a garden, and my family depends solely on the production of this garden for its survival, should I plant a monoculture of the highest yielding wheat, or should I plant a mixture of genotypes? There are clear advantages of planting the high-yield monoculture, but there is also one huge disadvantage: any disease or herbivore that can successfully infect (or attack) this genotype might potentially wipe it out. Your family suffers catastrophic consequences as a result. There is also one obvious disadvantages of the mixture; lower yield. But low sustained yield may be better in the long run. This is classic bet-hedging. Sex may be a bet-hedging strategy, if the Red Queen is correct: a lower arithmetic mean, but a higher geometric mean over time. In other words, for the family, it may be better to have consistent, but low resources, rather than extremes in boom and bust. The cost of a bust year is too high. Similarly, the cost of a clonal genotype may become too high (extinction) once the clone becomes common. The Red Queen hypothesis makes several straightforward predictions. 1. Sexual individuals should be favored where the risk of infection is high. We have found support for this expectation both within and between lakes. In general, we find more sexual females in areas where the frequency of infection by trematode worms is high (Lively 1987, 1992; Jokela & Lively 1995a,b) [more about the parasite]. This result suggests that asexual females have displaced sexuals where the risk of infection is low. Recently we have also found that infection levels and male frequency are highly correlated over a 10-15 year period, suggesting that selection pressures exerted by parasites and levels of sexual reproduction in the hosts are stable over time (Lively & Jokela, 2002). [more on Jokela] 2. Host genotypes should oscillate over time. Mark Dybdahl and I followed the dynamics of an all clonal population in L. Poerua, New Zealand over a six-year period. We found that changes in the frequencies of clones were correlated with later changes in the prevalence of infection in the clones, exactly as expected under the Red Queen hypothesis (Dybdahl & Lively 1998). Specifically, clones oscillated over time. When they first became common they were under-infected, but the prevalence of infection increased over time (years), until the clones were over-infected. The clones then decreased in frequency. There are, of course, alternative ways to explain these coupled changes over time, but they require odd assumptions. (Link to Mark Dybdahl's home page.) 3. There should be selection against genotypes that were common in the recent past. We conducted laboratory infection experiments to test the idea that clonal genotypes that were common in the recent past were also more susceptible to infection by local populations of parasites. In our first experiment, we found striking support for the idea: all four of the recently common clones from L. Poerua were more infected than the aggregate of rare clonal genotypes (Dybdahl & Lively 1998). We then repeated the experiment, with a twist. We wondered if the common clonal genotypes were more infected because they were being tracked by the local parasites (the Red Queen hypothesis), or whether there were more infected because they trade-off parasite resistance with rapid growth and reproduction (the trade-off hypothesis). To contrast the Red Queen with the trade-off hypothesis, we exposed snails from L. Poerua to two sources of parasites: one from L. Poerua (the local source), and one from another lake (L. Ianthe) 70 km to the South (the remote source). Under the Red Queen hypothesis, the common clones should only be more susceptible to the local source of parasites. In contrast, under the trade-off hypothesis, the common clones should be more susceptible to both sources of parasites. The results were consistent with the Red Queen, and inconsistent with the trade-off hypothesis (Lively & Dybdahl 2000). Specifically, the four common clones were all more susceptible than rare clones to the local source of parasites, but no such pattern was observed for the remote source of parasites. The trade-off hypothesis is thus rejected by the data. 4. Parasites should become adapted to infecting local populations of their hosts. If parasites are tracking locally common host genotypes, they should also become better (at least periodically) at infecting hosts from their local population than hosts from remote populations. We have found strong support for this idea in three different experiments, involving seven different populations (Lively 1989; Lively & Dybdahl 2000) (show me the data). Recent theoretical work has suggested that this kind of local adaptation does not require that parasites have faster generation times than their hosts (as commonly asserted), and that it can persist in the face of substantial gene flow between parasite populations (Lively 1999). In addition, we have found (in two separate experiments) a cline in susceptibility of snails along a depth gradient. Specifically, snails from shallow water are more susceptible to infection than snails from deep water (Lively & Jokela 1995; Krist et al. 2000). There are two possible explanations for this striking pattern. One is that shallow-water snails are physiologically compromised for some unknown reason, which renders them more susceptible to infection. Another possibility is that coevolutionary interactions are restricted to shallow water, which makes the shallow-water snails more tractable by the parasites. This later idea might work as follows. The final host for these parasites are ducks, which tend to forage along the shallow-water margins of the lake. Thus they are more likely to pick up infections from snails living in shallow water than deep water; and, hence, coevolution may be a stronger force in shallow water (see diagram). Coevolution is required for tracking of common genotypes. Interestingly, we find more sexual females and higher infection levels in shallow water than deep water (Jokela & Lively 1995a), which is consistent with the Red Queen hypothesis for sex. We are currently engaged in tests designed to contrast the physiological and coevolutionary alternatives. In summary, much of our work points to coevolutionary interactions with parasites as a mechanism selecting for sexual reproduction over parthenogenesis. There, nonetheless, remain important theoretical (link to theory) and empirical difficulties. For example, critical details regarding the genetic basis of infection have yet to be worked out. It remains to be seen whether the Red Queen will survive future attacks by the data in this and other systems. But, even if she does not survive, the knowledge of host-parasite interactions gained in the attempt will more than repay the effort.
Cited papers*:
*This work has been supported by grants from the US National Science Foundation |
C. M. Lively, Dept. of Biology,
Indiana University
Go back to Lively's homepage.
