«Evolutionary Ecology Research, 2012, 14: 403–423 Drift and selection entwined: asymmetric reproductive isolation in an experimental niche shift Jay ...»
Evolutionary Ecology Research, 2012, 14: 403–423
Drift and selection entwined: asymmetric reproductive
isolation in an experimental niche shift
Jay J. Falk1, Christine E. Parent1,2, Deepa Agashe3* and Daniel I. Bolnick4
Section of Integrative Biology, University of Texas at Austin, Austin, Texas, USA,
Department of Environmental Science, Policy, and Management, University of California Berkeley,
Berkeley, California, USA, 3Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, Massachusetts, USA and 4Howard Hughes Medical Institute and Section of Integrative Biology, University of Texas at Austin, Austin, Texas, USA ABSTRACT Question: Host races of phytophagous insects originate when a population adapts to a novel resource (host) while other populations remain resident on their ancestral host. The derived and ancestral host race populations will be subject to unequal selection intensity and genetic drift. Is this asymmetry responsible for asymmetric reproductive isolation observed in some populations?
Hypothesis: Unequal intensity of selection and/or genetic drift between populations may lead to asymmetric reproductive isolation.
Methods: We reared populations of Tribolium castaneum flour beetles on wheat or corn flour, which represent ancestral and suboptimal novel resources respectively. After approximately 43 generations, we assayed the fitness of wheat- and corn-evolved beetles on wheat and corn flour, and measured reciprocal pre-mating, pre-zygotic, and post-zygotic reproductive isolation between the ecotypes.
Results: Three of our four corn-evolved populations went extinct. The one surviving corn population exhibited evidence of adaptation to corn half way through the experiment, but by the final generation we found little evidence of adaptation. Instead, the corn-evolved population had lower survival than the wheat-evolved population, independent of rearing environment. This result is consistent with maladaptation due to fixation of deleterious alleles via genetic drift. Pre-mating and post-zygotic reproductive isolation were both asymmetric, favouring the higher-fitness ancestral population. Females from both wheat- and corn-evolved populations avoided mating with corn-evolved males. This bias improved female fitness, because corn-evolved males had offspring with lower survival, regardless of female genotype or rearing medium.
Conclusion: Strong selection in the derived (corn) population appears to have decreased population size to the extent that genetic drift led to the fixation of deleterious alleles that reduced corn-evolved male fitness. We posit that asymmetric pre-mating isolation arose because of this drift-induced maladaptation. Although our study is limited to a single pair of derived and ancestral populations, the results of our extensive pre- and post-mating reproductive Correspondence: J.J. Falk, Section of Integrative Biology, One University Station C0930, University of Texas at Austin, Austin, TX 78712, USA. e-mail: email@example.com *Current address: National Centre for Biological Sciences, GKVK, Bellary Road, Bengaluru, India.
Consult the copyright statement on the inside front cover for non-commercial copying policies.
© 2012 Jay J. Falk404 Falk et al.
isolation tests are consistent with this conclusion. This study highlights the importance of recognizing the fundamental entangling of drift and selection in the evolution of reproductive isolation, which complicates the frequently invoked dichotomy of selection and drift as distinct alternative forces in speciation.
Keywords: ecological speciation, founder effects, inbreeding depression, niche shift, phytophagous insects, post-zygotic isolation, pre-mating isolation, pre-zygotic isolation.
INTRODUCTIONSpeciation due to host shifts by phytophagous insects (Feder et al., 1988, 2003; Bush et al., 1989; Funk, 1998; Nosil et al., 2002), in which a few individuals may colonize a new host plant, often entails an evolutionary asymmetry in which a derived population is subjected to both intense selection and drift while another related population using the ancestral resource remains essentially unchanged. This asymmetry in selective regimes may cause asymmetric pre-mating reproductive isolation, which is common in nature (Watanabe and Kawanishi, 1979; Kaneshiro, 1980;
Bordenstein et al., 2000; Shine et al., 2002, Lin et al., 2012). Asymmetric post-mating isolation is also widely documented (Tiffin et al., 2001), but has been attributed strictly to genetic rather than adaptive mechanisms (Turelli and Moyle, 2007). We posit that asymmetric reproductive isolation may arise from asymmetric fitness in the ancestral and novel conditions. Selection may cause the new population to adapt, in the sense that it can now persist in the new habitat. However, if initial fitness in the new habitat is extremely low, genetic drift due to founder effects or strong selection may cause maladaptation by fixing deleterious alleles. Due to asymmetric selection and drift, the derived population may have lower survival or fecundity than the ancestral population. The ancestral population, being generally more fit, should discriminate against mates from the derived population. The derived population, being less fit, might actually prefer mates from the ancestral population because the outbreeding would increase fitness.
Considering the joint effect of selection and drift may help to explain patterns of asymmetric reproductive isolation, and more generally may highlight the necessity of considering these processes jointly rather than separately. The relative importance of natural selection and genetic drift in speciation has been hotly debated since the New Synthesis (Wright, 1931, 1932; Rice and Hostert, 1993; Coyne et al., 1997, 2004). Natural selection can drive the fixation of hybrid incompatibility genes (Dobzhansky, 1936; Muller, 1942; Orr, 1995), select against migration between populations (Nosil et al., 2005), or drive pre-mating isolation due to divergent mating cues (Ryan, 1990; Proctor, 1991, 1992). In principle, genetic drift may also fix incompatibility genes, or drive divergence in traits used for mate choice. Thus, a major question in evolutionary biology is whether reproductive isolation arises more frequently from divergent natural selection [e.g. ecological speciation (Schluter, 2001; Rundle and Nosil, 2005)] or genetic drift [e.g. founder effect speciation (Mayr, 1942, 1963)].
This dichotomy between selection- and drift-based speciation is a simplification, however, because the two evolutionary mechanisms interact in complex ways. Selection reduces effective population size and thereby increases the role of genetic drift in the fixation of alleles (Fig. 1). Conversely, genetic drift may drive maladaptive changes that undermine the effect of selection (Fig. 1). Under certain circumstances, this maladaptation may cause a population to shift off a local adaptive peak, thereby allowing natural selection to Drift and selection entwined 405 Fig. 1. A conceptual diagram of the fundamental entangling of selection and drift and their potentially interacting effect on reproductive isolation between populations.
drive populations towards more global optima [e.g. Wright’s shifting balance theory (Wright, 1931, 1932)].
A number of laboratory evolution experiments have attempted to experimentally (Mooers or statistically (reviewed in Rice and Hostert, 1993; Coyne and Orr, 2004; Fry, 2009) isolate et al., 1999; Rundle, 2003) the effects of selection and drift. Such experiments have found that natural selection may drive adaptive divergence between populations, pleiotropically generating reproductive isolation (‘ecological speciation’). However, as noted above, such selection can also increase the rate of genetic drift in fixing deleterious alleles. Consequently, very strong natural selection can simultaneously drive and (through bottlenecks) undermine adaptation, while also promoting founder effect speciation. It may therefore be important to consider the interacting effect of selection and drift in speciation experiments, especially when selection is strong enough to drive strong founder effects.
Here, we present the results of an experiment in which we maintained populations of Tribolium castaneum flour beetles on an ancestral resource (wheat) and a suboptimal novel resource (corn). Corn is a nutritionally poor resource (Via, 1991; Agashe, 2009), imposing 406 Falk et al.
extremely strong selection on beetle populations that caused most populations to go extinct within a few generations (Agashe et al., 2011). We assayed the effect of this strong selection on reproductive isolation between the derived corn and ancestral wheat beetles, and found asymmetric pre-mating and post-zygotic reproductive isolation. Upon measuring relative fitness on corn and wheat, we also found a corresponding asymmetry in fitness.
We speculate that these results arise from a strong maladaptive effect of genetic drift during the process of adaptation to survive on corn, combined with mate preferences for more fit genotypes.
Experimental lines The Tribolium castaneum populations described in this study were part of a previous, larger experiment to determine the impact of genetic variation on population dynamics and adaptation (Agashe, 2009; Agashe et al., 2011). Populations were derived from wild-type strains obtained from the Beeman Lab (Biological Research Unit, Grain Marketing and Product Research Center, Kansas) reared strictly on wheat flour (∼250 generations). We obtained a sample of ∼50 individuals of each strain in April 2006, and reared them under ancestral conditions (organic white wheat flour + 5% yeast mixture at 33 C and 60% relative humidity). After 5 months, we extracted adults from these stock populations and initiated 42 populations of 120 beetles each, on a suboptimal novel resource, corn flour (Agashe, 2009).
The corn populations had an average of 94% fewer individuals than populations in wheat during bimonthly censuses [mean N = 21 and 323 for corn and wheat populations, respectively (Agashe, 2009)]. All corn populations experienced an initial decline in population size, after which many (45%) went extinct within a year (Agashe et al., 2011). Most of the surviving populations had high founding genetic variation (i.e. were initiated with individuals derived from multiple strains). Only two of the surviving populations were derived from single strains (one each of strains Pak-3 and Col-2), corresponding to the low genetic variation treatment. Individuals from these populations had increased fecundity, survival, and growth rate on corn flour relative to their ancestors (Agashe et al., 2011).
Here, we focus on one of these surviving single-strain corn populations (Col-2) and its wheat-evolved counterpart (for logistical reasons, other wheat populations were terminated at the end of the original experiment described above). Of four replicate Col-2 corn populations, three went extinct during the original experiment described above. Individuals from the surviving population exhibited adaptation to corn after approximately 18 generations [median development rate was 6 weeks to reach adulthood (Agashe et al., 2011)]. The surviving corn population and the wheat population from which it was derived were then maintained for an additional 3 years (∼43 generations in total). At this point, we again assayed adaptation of corn-evolved and wheat-evolved beetles to each resource, and tested for reproductive isolation between these populations. We will refer to individuals as coming from either wheat-evolved (WE) or corn-evolved (CE) populations.
Note that because we had only a single surviving corn population, we did not have replicate CE populations and therefore could not assay whether reproductive isolation was stronger between the WE and CE populations than among either the WE or CE populations, as would typically be required in a test of ecological speciation (Coyne and Orr, 2004;
Ostevik et al., 2012). We therefore acknowledge that the results reported below cannot be Drift and selection entwined 407 specifically ascribed to divergent selection, though this seems likely. Instead, we present a very extensive analysis of pre-mating and post-mating reproductive isolation between the two populations.
Tests of (mal)adaptation If the surviving CE populations adapted to corn, we would expect that CE individuals reared on corn would have higher fecundity or survival than WE individuals reared on corn.
Conversely, if genetic drift induced maladaptation, we might observe decreased mean fitness (survival or fecundity) over time within the CE line independent of the resource they are fed. Survival rates were recorded both for the original stock population (in 2006) and at the end of this experiment. We can thus measure (mal)adaptation as the change in survival over time, or the fitness of contemporary CE versus WE beetles.
To assay fecundity, 50 adults from each population (CE, WE) were placed in 80 g of either wheat or corn flour. After 2 weeks for mating and oviposition, adults were removed from each treatment and larvae were allowed to continue development until pupation.
Pupae were placed in sex-specific containers (one pupa per gram of flour) and allowed to eclose over a period of 14 days. This ensured that all beetles used were virgins of similar age.
We refer to beetles reared on wheat or corn as WR and CR respectively, and we factorially crossed evolved type with rearing type generating four treatments (CE/CR; CE/WR;
WE/CR; WE/WR). Eclosed virgin males and females were then paired with individuals of the same ancestry and rearing environment (WE/WR, n = 20 pairs; CE/WR, n = 14;
WE/CR, n = 14; CE/CR, n = 27) in 0.35 g of flour corresponding to the female’s rearing flour (corn for CR, wheat for WR). Eggs were sifted from the flour and counted every 48 h for 8 days. We used a quasi-poisson generalized linear model (GLM) to test whether fecundity depends on evolved and reared flour types and their interaction.