Bringing together evolution on serpentine and polyploidy: spatiotemporal history of the diploid-tetraploid complex of Knautia arvensis (Dipsacaceae)

Abstract

Polyploidization is one of the leading forces in the evolution of land plants, providing opportunities for instant speciation and rapid gain of evolutionary novelties. Highly selective conditions of serpentine environments act as an important evolutionary trigger that can be involved in various speciation processes. Whereas the significance of both edaphic speciation on serpentine and polyploidy is widely acknowledged in plant evolution, the links between polyploid evolution and serpentine differentiation have not yet been examined. To fill this gap, we investigated the evolutionary history of the perennial herb Knautia arvensis (Dipsacaceae), a diploid-tetraploid complex that exhibits an intriguing pattern of eco-geographic differentiation. Using plastid DNA sequencing and AFLP genotyping of 336 previously cytotyped individuals from 40 populations from central Europe, we unravelled the patterns of genetic variation among the cytotypes and the edaphic types. Diploids showed the highest levels of genetic differentiation, likely as a result of long term persistence of several lineages in ecologically distinct refugia and/or independent immigration. Recurrent polyploidization, recorded in one serpentine island, seems to have opened new possibilities for the local serpentine genotype. Unlike diploids, the serpentine tetraploids were able to escape from the serpentine refugium and spread further; this was also attributable to hybridization with the neighbouring non-serpentine tetraploid lineages. The spatiotemporal history of K. arvensis allows tracing the interplay of polyploid evolution and ecological divergence on serpentine, resulting in a complex evolutionary pattern. Isolated serpentine outcrops can act as evolutionary capacitors, preserving distinct karyological and genetic diversity. The serpentine lineages, however, may not represent evolutionary 'dead-ends' but rather dynamic systems with a potential to further influence the surrounding populations, e.g., via independent polyplodization and hybridization. The complex eco-geographical pattern together with the incidence of both primary and secondary diploid-tetraploid contact zones makes K. arvensis a unique system for addressing general questions of polyploid research.

Figure 1. Ploidy level, genome size and habitat differentiation of the examined populations of Knautia arvensis agg.

Light grey circles – diploids from ‘non-relict’ genome size group, black circles – diploids from ‘relict’ genome size group, squares – tetraploids, white ovals – relict limestone habitats (open pine forests or subalpine grasslands), grey ovals – relict serpentine pine forests; the remaining populations inhabit semiruderal grasslands (ploidy levels according to ref. 31). The map covers the region of eastern part of central Europe, the inset displays the situation in the diploid-tetraploid serpentine area in the Slavkovský les Mts.

Figure 2. Phylogeographical grouping of 40 analyzed populations of Knautia arvensis agg. in central Europe.

Grouping is according to the nonhierarchical K-means clustering of AFLP phenotypes. Pie charts represent the proportion of individuals belonging to each of the seven detected groups (K1–K7). The size of the pie chart reflects the sample size. The inset displays the situation in the Slavkovský les serpentine area. White ovals denote populations from relict limestone habitats (open pine forests or subalpine grasslands), grey ovals populations from relict serpentine pine forests. Note the presence of several relict diploid populations in the western part of the area (P03, P04, and P05) with the genetic composition highly similar to the surrounding tetraploids. The distribution of chloroplast haplotypes is indicated (A–M).

Figure 3. Principal coordinate analysis based on Jaccard similarity among AFLP multilocus phenotypes of Knautia arvensis agg.

(a) entire data set; (b) excluding the most divergent group K1 (i.e., non-relict diploids). The different colours represent the groups identified by nonhierarchical K-means clustering (same as in Fig. 2). The centroid of each group and its connection with other points are displayed as well as an ellipse reflecting the variance of the group and the covariance on the axes.

Figure 4. Network of 13 plastid DNA haplotypes found within 77 examined individuals of Knautia arvensis agg.

The size of the circles is proportional to the number of individuals, while their shading indicates the ploidy level and monoploid genome size of the samples (black – relict 2× only, dark grey – relict 2×+4×, light grey – all 2×+4×, white – unique for a single non-relict 2× – haplotypes J and M – or 4× – haplotypes C and K – population). The double line indicates an insertion-deletion. For more detailed information, see Table 1.

Figure 5. Serpentine outcrop covered by open pine forest near Borovsko, central Czech Republic

(A.) This locality probably served as a Holocene refugium for several rare plant taxa. Morphologically distinct ‘relict diploid’ cytotype of Knautia arvensis (B, population P02 in this study) also occurs at this site.