Most plant population biologists are familiar with the classic papers titled "Evolution in closely adjacent plant populations" of Bradshaw, McNeilly, Antonovics and coworkers (e.g., McNeilly 1968; McNeilly and Antonovics 1968; Antonovics and Bradshaw 1970) that focused primarily on the evolution of metal tolerance in populations on mine tailings. Much of the work was conducted on two species of grasses, Anthoxanthum odoratum and Agrostis tenuis, and provides a classic demonstration of the power of strong natural selection in maintaining distinct sub-populations despite the potential for gene flow (McNeilly 1968; McNeilly and Bradshaw 1968). These classic papers, along with subsequent work on mine tailings also illustrate two especially intriguing features of the biology of edaphic specialists that are relevant to speciation. First, edaphic specialists seem to have a tendency to evolve tolerance of specific conditions multiple times, that is there is a tendency towards parallel establishment of edaphic specialization. Whether this represents de novo evolution of alleles conferring specialization, or parallel increases in frequency of alleles that have arisen a single time remains unclear (see below). Second, edaphic shifts often have direct or indirect effects on patterns of reproductive isolation.
In the early characterization of mine tailing populations in Agrostis and Anthoxanthum, parallel evolution was suspected because of the geographical separation of various populations on mine tailings, and the widespread occurrence of non-tolerant populations (Gregory and Bradshaw 1965). Furthermore, copper tolerance was shown to be present at low frequencies in seed samples collected in separate populations of Agrostis tenuis growing on non-contaminated soils, pointing to the likely parallel establishment of the trait in isolated localities (McNeilly and Bradshaw 1968). The hypothesis of parallel evolution of metal tolerance in Agros-tis tenuis has since been further supported by analysis of genotype x environment interaction (G x E) in copper tolerant populations from distinct localities, which reveals that copper tolerant populations display significant heterogeneity in their response to levels of copper (Nicholls and McNeilly 1982). While these patterns may reflect distinct origins of copper tolerance, they could also reflect the presence of unique modifiers of a common mechanism of tolerance that in fact had a single origin (Schat et al. 1996).
In much of the literature discussing the origins of metal tolerance, the geographical distribution of metal-tolerant populations again provides the primary indication of parallel origins. While this is a reasonable hypothesis, it is also possible that tolerant populations have been established by longdistance dispersal, perhaps aided by accidental transport by mine workers (Schat et al. 1996). Further evidence of multiple origins comes from crossing studies that examine patterns of segregation of tolerance that would indicate the action of multiple independent loci. The action of independent loci would be revealed either by a breakdown of metal tolerance in the F2 generation in crosses between homozy-gous tolerant individuals, but might also be suggested by the presence of transgressive segregation in the F1 generation. To date, such tests tend to support a common genetic basis for tolerance, though the presence of modifiers has been noted in some populations. For example, in Silene vulgaris, crosses among copper, zinc and cadmium tolerant and nontolerant populations from Germany and Ireland reveal that a single locus is responsible for tolerance to each metal, though at least one additional locus contributes to additional copper tolerance in one population (Schat et al. 1993). In this case, the modifier could have arisen subsequent to the establishment of the widespread tolerant genotype at this site. While a common genetic basis for tolerance could be regarded as evidence against parallel evolution of tolerance, studies of the genetic basis of metal tolerance suggest that constraints may limit the number of loci that can successfully mutate to yield tolerant genotypes. Therefore it is possible that crossing studies fail to show a breakdown of tolerance because mutation of the same genetic locus has given rise to tolerant genotypes independently. Finally, it should be noted that even if the same allele is responsible for tolerance in distinct populations, this allele may have been present at low frequencies in non-tolerant populations and thus may have become established in parallel at multiple localities (Schat et al. 1996). In such cases, parallel evolution of metal tolerance could be indicated by phylogenetic data that support multiple origins of tolerant populations. Where phylogenetic studies of edaphically distinct populations have been conducted, it is noteworthy that these uniformly point to parallel evolution.
Under the extreme conditions presented by severely metal-contaminated sites, the role of natural selection in establishing tolerance is understood, given that non-tolerant individuals generally do not survive the mine tailing environment. In less dramatic cases, the pattern of parallel evolution itself is suggestive of the action of natural selection, because parallel establishment of ecologically relevant traits is unlikely to occur as a result of stochastic processes (Levin 2001). For example, in previous work on Lasthenia californica, analysis of isozyme variation (Desrochers and Bohm 1998) and phylogenetic data suggested the existence of geographically-based subdivisions within the complex. Analysis of chemical characteristics of soils and plant tissue demonstrated the existence of edaphic races and suggested that divergent natural selection might play a role in diversification of the races (Rajakaruna and Bohm 1999). Combining these data has revealed a pattern of parallel evolution of an ecologically relevant suite of traits (Rajakaruna et al. 2003c), which serves as evidence for the action of natural selection in establishing racial differences. Given that relatively few systems have been explored and that many of these provide some indication of parallel evolution, this tendency would seem to be a common characteristic of edaphic specialists.
A phenomenon that may be related to parallel evolution of tolerance is the evolution of multiple tolerances within spe cies. In many cases, lineages that have evolved tolerance to one edaphic extreme have also adapted to other extreme factors. For example, tolerance of multiple populations to four elements (copper, nickel, zinc and lead) was assessed (Gregory and Bradshaw 1965) in Agrostis tenuis. Although most populations occur on pasture soils that have trace levels of heavy metals and comprise plants that are not tolerant of heavy metal contamination, mine tailing populations variously display tolerance to zinc, copper and lead, generally matching their tolerance to levels present in soils at the collection locality (Gregory and Bradshaw 1965). The authors also demonstrated that tolerance of each of these metals did not confer tolerance to the others, indicating somewhat distinct mechanisms in each case. Still, that this species has evolved tolerance to normally toxic levels of multiple ions suggest that an underlying trait, perhaps involving tolerance of low pH (Gregory and Bradshaw 1965), drought (Hughes et al. 2001) or hyper saline conditions is widespread in the progenitors of edaphic specialists and contributes to their ability to evolve tolerance to specific ionic extremes. It seems plausible that such underlying traits could facilitate either the evolution of multiple tolerances or the parallel evolution of tolerance of a single heavy metal.
The second intriguing feature of edaphic endemics is the relationship between the shift in edaphic tolerance and changes in patterns of reproductive isolation. Such changes may arise as a direct consequence of adaptive shifts, either because reproductive compatibility is a by-product of physiological adaptation to new edaphic conditions or because of linkage or pleiotropy of loci affecting ecological shifts and those affecting reproductive isolation. Alternatively, enhanced reproductive isolation between divergent edaphic specialists may reflect the action of reinforcement, i.e., selection for reduced gene flow to avoid maladaptive hybridization following a period of divergence in allopatry. It is important to note that the two alternatives are not mutually exclusive. For instance, in the classic mine tailing studies, heavy metal tolerant and non-tolerant populations of both Anthoxanthum odoratum and Agrostis tenuis were shown to have genetically-controlled differences in flowering time (McNeilly and Antonovics 1968), with tolerant populations flowering earlier in both cases. Plants closest to the mine boundary showed the greatest difference in flowering times, which McNeilly and Antonovics (1968) interpret as evidence for the action of reinforcement, though they state that a portion of the flowering time shifts also arises as a byproduct of adaptation to local conditions. They noted a relationship between flowering time and soil temperatures to support this claim. It is likely that changes in phenology can also evolve as a direct by-product of shifts in edaphic tolerances. For example, flowering time differences are associated with differences in sodium accumulation in wheat (Taeb et al. 1992). Furthermore, even if the initial divergence in flowering time is entirely environmentally-determined, this pattern can contribute to the accumulation of genetically-based flowering time differences. Stam (1983) demonstrated theoretically that spatially structured differences in flowering time can lead to spatially-coincident, genetically-based differences in flowering time. Of course as different edaphic habitats are almost by necessity spatially isolated from one another, individuals that occur in distinct edaphic environ ments are likely to experience decreased gene flow relative to individuals that occur in the same habitat (L. H. Rieseberg, pers. comm.). Another tendency that has been repeatedly observed in edaphic endemics is a shift towards increasing self-fertility. Increased selfing rates are thought to have arisen as a mechanisms to prevent gene flow between mine and non-mine populations in both Anthoxanthum odo-ratum and Agrostis tenuis (Antonovics 1968), though this explanation was not favored to explain increased self-fertility in Armeria maritima populations from zinc mines (Lefebvre 1970). Lefebvre (1970) instead interprets increased self-fertility as having provided reproductive assurance during long distance colonization of mine sites, as there are no adjacent non-mine populations in this region.
Examples of changes in post-mating reproductive isolation accompanying edaphic shifts are less common, as might be expected given that these are generally more likely to accumulate as a by-product of divergence. However, post-mating reproductive isolation can include mechanisms that act prior to zygote formation, such as pollen-pistil incompatibilities. Such post-mating, pre-zygotic isolation can be subject to reinforcement, and thus may be an indirect outcome of adaptive divergence. The most widely cited example of post-mating reproductive isolation associated with edaphic shifts comes from populations of Mimulus guttatus adapted to copper mine tailings. Macnair and co-workers have documented that the genetic locus that confers copper tolerance is closely linked to or has a pleiotropic effect on viability of hybrids between tolerant and non-tolerant individuals (Christie and Macnair 1983). Searcy and Macnair (1990) also demonstrated that copper uptake may contribute to pollen pistil incompatibilities, suggesting that crossing studies involving extremes should perhaps be conducted.
Was this article helpful?