Polyploid evolution in arctic-alpine Draba (Brassicaceae)

The mainly arctic-alpine genus Draba is well known for its complex morphological and chromosomal vanation. This paper reviews a larger study of Nordic Draba, aimed to provide insights into evolutionary processes that confound taxonomic relationships in the genus. The populations analyzed were referred to 16 currently recognized species of the sections Draba (petals white), Chrysodraba (petals yellow), andDrabella (petals yellow), and investigated using enzyme electrophoresis, restriction site analysis of cpDNA and rDNA, and analysis of chromosome numbers, artificial and natural hybrids, reproductive biology, habitat differentiation, and morphology. Section Draba comprises three diploids and seven polyploids (4x-10x) based on x = 8, sect. Chrysodraba comprises five polyploids (4x-16x) based on x = 8, and sect. Drabella comprises D. crassifolia, which probably is octoploid based on x = 5. Artificial F₁ hybrids were obtained in 19 interspecific combinations. Later-generation hybrids were obtained in seven of these combinations. The genetic data suggest that 1) all polyploids are genetic allopolyploids, i.e., they show disomic inheritance and are highly fixed-heterozygous; 2) several of the polyploids have originated recurrently, some of them even polyphyletically; 3) some of the alloploid populations may have originated from cross-incompatible, sibling species that all belong to a single diploid taxonomic species; 4) interspecific gene flow across chromosome number barriers is possible and probably occurs in natural situations; 5) each of three of the polyploids represents an independent alloploid lineage, whereas sect. Draba and two species of sect. Chrysodraba form an intricate phylogenetic network; 6) some of the polyploids have originated locally, others have migrated repeatedly into the Nordic area; and 7) the phenotypic expression of genes encoding taxonomically important morphological characters does not follow consistent patterns in hybrids; this result may explain the discrepancies between genetic and taxonomic relationships in this highly reticulate genus. All species are sexual autogams, but there was large variation in autogamous seed set and traits promoting cross-pollination. The diploids are genetically depauperate, extremely inbreeding stress-tolerators occupying restricted ecological niches, whereas most of the polyploids have high levels of genetic variation, occupy a wide range of niches, and are either stress-tolerant competitors with a mixed mating system or primarily inbreeding ruderals. Mixed mating appears advantageous in the polyploids occurring in competitive habitats, although their fixed heterozygosity buffers the effect of selfing with respect to loss of variability. The ecological amplitude, heterozygosity, and biochemical diversity in the species were positively correlated and increased significantly with ploidal level. Two hypotheses are advanced to explain these correlations: 1) the general-purpose genotype hypothesis, which suggests that a high level of fixed heterozygosity in an allopolyploid genotype per se allows for exploitation of several different niches; and 2) the special-purpose genotype hypothesis, which suggests that repeated alloploidizations involving genetically divergent progenitors result in different fixed-heterozygous genotypes, each of which may exploit a particular niche. The evidence for multiple polyploid origins, differentiation into sibling diploids, and interploidal gene flow · in Draba add to a growing data base suggesting that polyploid complexes represent considerably more dynamic genetic systems than previously envisioned. In Draba, the principal evolutionary importance of these processes is probably that they serve as escapes from genetic and ecological depauperation caused by uniparental inbreeding at the diploid level. These processes inevitably result, however, in incongruities between taxonomic and evolutionary entities in the genus, supporting the use of a wide species concept.