Evolution of cryptic gene pools in Hypericum perforatum: the influence of reproductive system and gene flow

Abstract

Background and aims: Hypericum perforatum (St. John's wort) is a widespread Eurasian perennial plant species with remarkable variation in its morphology, ploidy and breeding system, which ranges from sex to apomixis. Here, hypotheses on the evolutionary origin of St. John's wort are tested and contrasted with the subsequent history of interspecific gene flow.

Methods: Extensive field collections were analysed for quantitative morphological variation, ploidy, chromosome numbers and genetic diversity using nuclear (amplified fragment length polymorphism) and plastid (trnL-trnF) markers. The mode of reproduction was analysed by FCSS (flow cytometric seed screen).

Key results: It is demonstrated that H. perforatum is not of hybrid origin, and for the first time wild diploid populations are documented. Pseudogamous facultative apomictic reproduction is prevalent in the polyploids, whereas diploids are predominantly sexual, a phenomenon which also characterizes its sister species H. maculatum. Both molecular markers characterize identical major gene pools, distinguishing H. perforatum from H. maculatum and two genetic groups in H. perforatum. All three gene pools are in close geographical contact. Extensive gene flow and hybridization throughout Europe within and between gene pools and species is exemplified by the molecular data and confirmed by morphometric analyses.

Conclusions: Hypericum perforatum is of a single evolutionary origin and later split into two major gene pools. Subsequently, independent and recurrent polyploidization occurred in all lineages and was accompanied by substantial gene flow within and between H. perforatum and H. maculatum. These processes are highly influenced by the reproductive system in both species, with a switch to predominantly apomictic reproduction in polyploids, irrespective of their origin.

Keywords: Apomixis; Hypericum maculatum; Hypericum perforatum; St. John's wort; evolution; hybridization; reproductive mode.

Fig. 1.

Principal co-oordinate analysis (PCA) of the whole data set including 476 individuals (Supplementary Data Table S3) and comparing H. maculatum with H. perforatum. (A) Phenotypic constancy of morphological characters: evaluation of discriminating power of morphological characters comparing morphology-based characters obtained from individuals directly collected in the wild with the respective material cultivated in a common garden experiment at Heidelberg Botanical Garden. (B) Combination of the morphological differentiation with ploidy. Unknown cytotypes are also indicated. (C) Combination of the morphological differentiation and affiliation with one out of the three genetic clusters as revealed by AFLP analysis. A colour signature (blue = H. maculatum; green/red = H. perforatum) follows the results as summarized in Supplementary Data Table S5. A colour is assigned only if the respective proportion of the genetic variation (blue, red or green) contributes >80 %. The remaining individuals scored for AFLPs but with no contribution of any genetic cluster >80 % are shown by open circles.

Fig. 2.

Plastid DNA parsimony network of Hypericum perforatum and H. maculatum, and including various outgroup taxa. Hypericum tetrapterum and H. undulatum accessions shared an identical plastid DNA haplotype. The different colour codes correspond to those used for AFLP gene pool recognition (blue = H. maculatum; red and green = H. perforatum). Haplotype coding (HP01–HP10, HM01–HM04; and outgroup taxa) and respective sequence accession data are provided in Supplementary Data Table S4.

Fig. 3.

(A) Structure analysis of the AFLP data of single individuals (optimal K = 3; n = 195). The population code is given above the bars. The colour code corresponds to the plastid DNA data and distinguishes between Hypericum maculatum and two groups within H. perforatum (blue = H. maculatum; red and green = H. perforatum; Fig. 1). (B) Distribution map of populations. Pie charts combine the data from the individuals on the populational level. Asterisks (*) indicate diploid populations. Colour code follows (A) and Fig. 1 (blue = H. maculatum; red and green = H. perforatum). Additional populations analysed for morphometrics are shown in Supplementary Data Fig. S1.

Fig. 4.

Diagram summarizing the various modes of sexual and non-sexual reproduction in Hypericum maculatum and H. perforatum, and expected ratios in DNA content (embryo:endosperm) as measured here with FCSS (flow cytometric seed screen). The diagram uses a tetraploid (4n) mother plant as an example. The male gamete is indicated in red if contributing to endosperm fertilization and/or egg cell fertilization. The colour code corresponds to the categorization of our FCSS data summarized in Supplementary Data Table S6.

Fig. 5.

The reproductive system of diploid and polyploid Hypericum perforatum, H. maculatum and hybrids. White, proportion of sexual reproduction; grey, proportion of any kind of asexual and apomictic seed production (see also Fig. 4). Diploids and polyploids are separated and further differentiated according to their genotypic designation (gene pool as indicated with AFLP and plastid DNA data, Figs 1 and 3; Supplementary Data Table S5). The numbers indicate the number of individuals investigated with seed screen analysis and providing endosperm:embryo DNA ratios.