Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Polyploidisation and Geographic Differentiation Drive Diversification in a European High Mountain Plant Group (Doronicum clusii Aggregate, Asteraceae)

  • Clemens Pachschwöll,

    Affiliation Department of Botany and Biodiversity Research, University of Vienna, Rennweg 14, A-1030 Vienna, Austria

  • Pedro Escobar García,

    Affiliations Department of Botany and Biodiversity Research, University of Vienna, Rennweg 14, A-1030 Vienna, Austria, Department of Botany, Natural History Museum, Burgring 7, A-1010 Vienna, Austria

  • Manuela Winkler,

    Affiliations Department of Botany and Biodiversity Research, University of Vienna, Rennweg 14, A-1030 Vienna, Austria, GLORIA co-ordination, University of Natural Resources and Life Sciences Vienna, Center for Global Change and Sustainability & Austrian Academy of Sciences, Institute for Interdisciplinary Mountain Research, Silbergasse 30, A-1190 Vienna, Austria

  • Gerald M. Schneeweiss ,

    gerald.schneeweiss@univie.ac.at

    Affiliation Department of Botany and Biodiversity Research, University of Vienna, Rennweg 14, A-1030 Vienna, Austria

  • Peter Schönswetter

    Affiliation Institute of Botany, University of Innsbruck, Sternwartestrasse 15, A-6020 Innsbruck, Austria

Polyploidisation and Geographic Differentiation Drive Diversification in a European High Mountain Plant Group (Doronicum clusii Aggregate, Asteraceae)

  • Clemens Pachschwöll, 
  • Pedro Escobar García, 
  • Manuela Winkler, 
  • Gerald M. Schneeweiss, 
  • Peter Schönswetter
PLOS
x

Abstract

Range shifts (especially during the Pleistocene), polyploidisation and hybridization are major factors affecting high-mountain biodiversity. A good system to study their role in the European high mountains is the Doronicum clusii aggregate (Asteraceae), whose four taxa (D. clusii s.s., D. stiriacum, D. glaciale subsp. glaciale and D. glaciale subsp. calcareum) are differentiated geographically, ecologically (basiphilous versus silicicolous) and/or via their ploidy levels (diploid versus tetraploid). Here, we use DNA sequences (three plastid and one nuclear spacer) and AFLP fingerprinting data generated for 58 populations to infer phylogenetic relationships, origin of polyploids—whose ploidy level was confirmed by chromosomally calibrated DNA ploidy level estimates—and phylogeographic history. Taxonomic conclusions were informed, among others, by a Gaussian clustering method for species delimitation using dominant multilocus data. Based on molecular data we identified three lineages: (i) silicicolous diploid D. clusii s.s. in the Alps, (ii) silicicolous tetraploid D. stiriacum in the eastern Alps (outside the range of D. clusii s.s.) and the Carpathians and (iii) the basiphilous diploids D. glaciale subsp. glaciale (eastern Alps) and D. glaciale subsp. calcareum (northeastern Alps); each taxon was identified as distinct by the Gaussian clustering, but the separation of D. glaciale subsp. calcareum and D. glaciale subsp. glaciale was not stable, supporting their taxonomic treatment as subspecies. Carpathian and Alpine populations of D. stiriacum were genetically differentiated suggesting phases of vicariance, probably during the Pleistocene. The origin (autopolyploid versus allopolyploid) of D. stiriacum remained unclear. Doronicum glaciale subsp. calcareum was genetically and morphologically weakly separated from D. glaciale subsp. glaciale but exhibited significantly higher genetic diversity and rarity. This suggests that the more widespread D. glaciale subsp. glaciale originated from D. glaciale subsp. calcareum, which is restricted to a prominent Pleistocene refugium previously identified in other alpine plant species.

Introduction

Biota of temperate European mountain ranges such as the European Alps were strongly influenced by Quaternary climatic oscillations causing range shifts and extirpations [13]. Intersecting paleo-environmental (e.g., maximum extent of glaciers and extrapolated position of the snow line) with genetic data (patterns of genetic diversity and rarity, geographical distribution of genetically delimited groups), it was shown that during Pleistocene glaciations many alpine species were forced into only locally glaciated mountain ranges at or close to the periphery of the Alps [46]. Major peripheral refugia were situated in the southwestern and the southern Alps; for silicicolous and calcicolous species, additional refugia were located in the eastern-most and in the northeastern-most Alps, respectively. Refugia on nunataks in the strongly glaciated interior of the Alps were suggested for only a few species [7,8].

During phases of warmer climate, lowland populations of alpine species got extinct while ranges expanded from the refugia towards interior areas of the Alps [9,10]. In many species phases of isolation were prolonged by incomplete range filling (temperate trees [11]; alpine plants [12]) resulting in disjunct distributions [1317], but even in species with contiguous distribution areas vicariance may still be evident in strong phylogeographic breaks and the presence of hybrid zones [1823].

Quaternary climatic oscillations did, however, not only reshuffle distribution ranges but also fostered lineage diversification and speciation [24,25]. Ecogeographic isolation, such as edaphic specialization, and polyploidisation are among the most important mechanisms driving the diversification of angiosperms in general [2629], and that of cold-adapted species in particular [3037]. This is supported by the good congruence of refugial patterns with centres of endemism and species richness [3842] or co-occurrence, partly over wider geographic scales, of intraspecific genetic groups [4346].

A good system to study the role of range shifts, polyploidisation and hybridization for diversification of mountain biota is the Eurasian genus Doronicum (Asteraceae—Senecioneae). A sound hypothesis on relationships within Doronicum is available from revisionary taxonomic and molecular phylogenetic work [4750]. In this genus polyploidy [50] and hybridization [51,52] are common phenomena and may be at least partly responsible for high morphological variation and delimitation problems [50,53,54]. Sexual reproduction and outcrossing was experimentally proven for the diploid species D. austriacum Jacq. and D. grandiflorum Lam. s.s. as well as for the tetraploid species D. stiriacum (Vill.) Dalla Torre (as D. clusii (All.) Tausch) [55,56]. In Doronicum, the cypselae possess long pappus bristles and are likely wind-dispersed [50,57].

The monophyletic D. clusii aggregate comprises four geographically, ecologically and/or karyologically differentiated subalpine to subnival forbs restricted to the European mountain systems of the Alps and the Carpathians [49,57,58,59]. They possess yellow, showy flower heads that are visited and pollinated by various Diptera, Hymenoptera and Lepidoptera [60,61]. These four taxa are grouped in two pairs [58,59] (summarized in Table 1; for more detailed descriptions see S1 Appendix): (i) D. glaciale (Wulf.) Nyman subsp. calcareum (Vierh.) Hegi from the northeastern-most Alps and the parapatric D. glaciale (Wulf.) Nyman subsp. glaciale from the eastern Alps (Fig. 1, S1 Appendix) are diploid (2n = 60) and basiphilous growing on calcareous or, in case of D. glaciale subsp. glaciale, also on base-rich siliceous substrate; (ii) D. clusii (All.) Tausch s.s. (i.e., excluding D. stiriacum (Vill.) Dalla Torre) from the Alps (except the easternmost parts) and the allopatric D. stiriacum from the easternmost Central Alps and the Carpathians (Fig. 1, S1 Appendix) are acidophilic, but differ in being diploid (2n = 60) and tetraploid (2n = 120), respectively. Whereas a geographically restricted contact zone with morphologically intermediate individuals links D. glaciale subsp. calcareum and D. glaciale subsp. glaciale (Fig. 1; S1 Appendix), putative hybrids between D. clusii s.s. and D. glaciale subsp. glaciale occur throughout their overlapping distribution ranges in the eastern Central Alps (Fig. 1; S1 Appendix). The Alpine distribution ranges of both D. glaciale subsp. calcareum and D. stiriacum overlap with Pleistocene refugia [4,5,38] suggesting that their differentiation might be connected to this period.

thumbnail
Table 1. Overview of the taxa of the Doronicum clusii aggregate and their diagnostic morphological characters (summarized from [50,58,59,113,114,116]).

https://doi.org/10.1371/journal.pone.0118197.t001

thumbnail
Fig 1. Investigated populations of the Doronicum clusii aggregate in the Alps and Carpathians.

Population numbers are as in Table 2; taxonomic entities are colour-coded: D. glaciale subsp. calcareum (yellow), D. glaciale subsp. glaciale (red) and their morphological intermediates (orange); D. stiriacum (green); D. clusii s.s. (blue); hybrids between D. clusii s.s. and D. glaciale subsp. glaciale (D. × bauhini; lavender). The insert shows Carpathian populations of D. stiriacum (turquoise); for graphical reasons the insert is in the upper left despite the Carpathians being east of the Alps. The coloured outlines are distribution ranges summarized according to various literature sources, herbarium specimens and personal observations. Scale bars: 50 km.

https://doi.org/10.1371/journal.pone.0118197.g001

thumbnail
Table 2. Populations of the Doronicum taxa investigated, sampling locations (voucher information), relative fluorescence, number of investigated individuals, AFLP-derived gene diversity and rarity, Prabclus assignments, plastid haplotype, and GenBank accession numbers.

https://doi.org/10.1371/journal.pone.0118197.t002

Here, we investigate diversification and its underlying causes in the D. clusii aggregate. To this end, we use Amplified Fragment Length Polymorphisms (AFLPs), nuclear and plastid DNA sequences, DNA ploidy level estimation by flow cytometry and chromosome counts. AFLPs are rapidly homogenizing, biparentally inherited markers well suited to resolve relationships among closely related taxa because of their ability to provide phylogenetic signal in young or rapidly evolving study systems (e.g., [62,63]). Biparentally inherited nuclear Internal Transcribed Spacer (ITS) sequences are, despite a number of potential problems [64], appropriate for addressing evolutionary questions at the species level including hybridization [65] and have been widely used in Asteraceae including Doronicum [49,52]. Maternally inherited (for Doronicum and other Asteraceae [66]) plastid DNA sequences have slower mutation rates [67,68] and were widely and successfully applied for phylogenetic and phylogeographic analyses and, together with nuclear ITS data, inference of reticulate relationships also in Doronicum [49,52]. Flow cytometry can be used to rapidly screen ploidy levels (calibrated by chromosome counts) for numerous samples [69,70].

The main goal of the present study is to determine phylogenetic relationships among the four members of the D. clusii aggregate (only D. clusii s.s. and D. glaciale subsp. glaciale have been included by [49]) and to unravel evolutionary patterns within the aggregate. We aim to answer the following questions. (1) Do the traditionally recognized taxa D. clusii s.s., D. glaciale subsp. calcareum, D. glaciale subsp. glaciale, and D. stiriacum constitute genetically distinct lineages? Are morphologically intermediate forms between D. glaciale subsp. calcareum and D. glaciale subsp. glaciale and between D. clusii s.s. and D. glaciale subsp. glaciale (i.e., D. × bauhini Saut.) of hybrid origin? (2) Are populations of D. stiriacum in the eastern Alps tetraploid, as are those in the Carpathians? Is tetraploid D. stiriacum of auto- or allotetraploid origin? (3) Do patterns of genetic diversity and/or rarity reflect putative Pleistocene refugia previously suggested for acidophilic and basiphilous species [5]?

Materials and Methods

Ethics Statement

The investigated taxa are neither endangered nor protected except at the following locations: D. stiriacum pop. 32 (collected under permit 8-NAT-259/3/2004 from the Amt der Kärntner Landesregierung, Austria); D. stiriacum pop. 36 (collected under permit from the Parcul Național Munții Rodnei, Romania); D. stiriacum pop. 37 (collected under permit no. 1762/565/04–5.1 from the Ministerstvo Životneho Prostredia, Slovakia); D. stiriacum pop. 38 (collected under permit no. 6878/2008–2.1 from the Ministerstvo Životneho Prostredia, Slovakia).

Plant Material

Sampling focused on the eastern Alps where all four taxa of the D. clusii aggregate occur (Fig. 1). Our sampling of D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and D. stiriacum is comprehensive and covers the species’ entire distribution areas. As our study did not primarily focus on intraspecific phylogeography—a wealth of data already exists for the Alps (reviewed in [5,71])—we sampled D. clusii s.s. westwards only to southern Switzerland. As outgroups, the closely related D. grandiflorum Lam. (Alps and southern European mountain ranges [50,72]) and, for DNA sequences only, the more distantly related [49] D. austriacum Jacq. and D. carpaticum (Griseb. & Schenk) Nyman (= D. columnae Ten. subsp. carpaticum (Griseb. & Schenk) Soó [54]) were included. Leaf material from three to six (median five) individuals per population was collected in the field mostly between 2008 and 2010 (Table 2) and immediately dried in silica gel. Herbarium vouchers of each population were databased and deposited at the herbarium of the University of Vienna (WU; http://herbarium.univie.ac.at/) and at the herbarium of the University of Salzburg (SZU).

DNA extraction

Total genomic DNA was extracted from c. 10 mg tissue following [73,74] with the modifications detailed in [18]. The quality of the extracted DNA was checked on 1% TAE-agarose gels and quantified with a Nanodrop Spectrophotometer ND-100 (PEQLAB, Erlangen, Germany).

Flow Cytometry and Chromosome Counts

DNA ploidy levels [75] of silica gel-dried leaf tissue were determined using 4′,6-diamidino-2-phenylindole (DAPI) flow cytometry as described in [76]. Measurements of DAPI-stained nuclei were performed for the same individuals used for AFLPs plus four additional specimens (total of 280 individuals; Table 2). In a first round, one individual per population was measured to assess putative DNA ploidy variation within a given taxon. In a second round, which was preceded by initial tests proving that a single individual with deviating ploidy could be detected if co-analysed with five individuals of another ploidy, all individuals of a population were pooled to test the possible presence of minority ploidy levels; this approach does not affect the reliability of the ploidy estimates [44,7678]. Following [76], Pisum sativum ‘Ctirad’ (1C = 4.55 pg [79]) and Vicia faba ‘Inovec’ (1C = 13.45 pg [80]) were used as internal standards, and fluorescence intensities of 3000 particles were recorded.

Actively growing root meristems were obtained from individuals that were collected in the field (along with the silica samples, but not used in the genetic analyses) and subsequently cultivated in the Botanical Garden of the University of Vienna (HBV), or from seedlings (pop. 35 only). After initial experiments, these meristems were pre-treated with 0.002 M 8-hydroxyquinoline for 2 h at room temperature and for 2 h at 4°C (putative diploids) and for 3.5 at room temperature and for 4.5 h at 4°C (putative tetraploids), respectively, in darkness; subsequently they were fixed in 3: 1 ethanol: acetic acid and stored at-20°C until use. Fixed meristems were hydrolysed in 5N HCl for 30 min at room temperature, rinsed with tap water, and stained with Schiff’s reagent (Merck, Vienna, Austria) in darkness for 60 min. Squash preparations were made in a drop of 60% acetic acid. Preparations were analysed with an Axioplan2 microscope (Carl Zeiss, Vienna, Austria). Images were acquired with a CCD camera and processed using Axiovision 4.8 (Carl Zeiss). Image quality was optimised using Adobe Photoshop CS3 (Adobe Systems, San Jose, CA, USA) with options that applied uniformly to all pixels of the image.

AFLPs

AFLP data were generated for taxa of the D. clusii aggregate and for D. grandiflorum populations from the Alps. The AFLP procedure followed [81] with the modifications described in [82] using approximately 150–500 ng of total DNA. For the selective amplification initially 24 selective primer combinations with three or four MseI selective bases were screened. The three final primer combinations were (fluorescent dye in parentheses): EcoRI (6-Fam)-ACA/MseI-CACC, EcoRI (VIC)-ACG/MseI-CAAG, and EcoRI (NED)-ACA/MseI-CAAG. In each PCR plate, the same seven individuals were replicated to calculate the error rate according to [83], and to exclude non-reproducible fragments from the analysis. Thirty-two between-plate replicates were successfully amplified, and ten individuals were extracted twice. From the restriction-ligation onwards, one blank sample per plate was included to test for systematic contamination.

DNA sequences

One individual per population (usually one that was also included in the AFLP data set) was used for generating DNA sequences. In addition, for DNA sequences up to five individuals per population were sequenced in 28 randomly selected populations to check for possible intra-population variation.

For plastid markers, the following regions were screened using eight individuals from different taxa: petL-psbE, psbJ-petA, 3′trnVUAC-ndhC, psbD-trnTGGU, atpI-atpH, trnQUUG-5′rps16, 3′rps16–5′trnKUUU, the ndhA-intron, ndhF-rpl32, rpl32-trnLUAG (all [67]), the rps16-intron [84], and the trnTUGU-trnFGAA region [85]. Of these twelve markers, the three most variable ones (ndhF-rpl32, rpl32-trnLUAG, 3′rps16–5′trnKUUU) were selected. The PCR reaction mixes (totalling 15 μL) contained 4.8 μL REDTaq Ready Mix PCR reaction mix (Sigma-Aldrich, Vienna, Austria), 7.4 μL water, 0.6 μL BSA (10 mg/mL; Promega, Vienna, Austria), 0.3 μL of each primer (10 μM; Sigma-Aldrich) and 1 μL of genomic DNA. The ndhF-rpl32 intergenic spacer region was amplified with the primers of [67], using the following PCR conditions: 1 min at 95°C; 10 cycles of 30 s at 95°C, 30 s at 47°C, 90 s at 65°C; 25 cycles of 30 s at 95°C, 30 s at 49°C, 90 s at 65°C; 8 min at 65°C. The 3′rps16–5′trnKUUU intergenic spacer was amplified with newly designed primers (rps16-F-PD2: 5′-GTGGGTTTTTATGATCCGATCAAG-3′, trnK-R-PD2: 5′-TTAAAAGCCGAGTACTCTACCGTTG-3′), using the following PCR conditions: 1 min at 95°C; 35 cycles: 30 s at 95°C, 30 s at 56 or 63°C, 90 s at 72°; 8 min at 72°C. The rpl32-trnLUAG intergenic spacer was amplified with newly designed primers (rpl32-F-PD2: 5′-AGGAAAGGATATTGGGCGGCG-3′, trnL-R-PD2: 5′-TTTCACCATAGCGGCTTGCTCG-3′), using the following PCR conditions: 1 min at 95°C, 35 cycles: 30 s at 95°C, 30 s at 63°C, 90 s at 72°; 8 min at 72°C. Samples for which the amplification of ndhF-rpl32 and rpl32-trnLUAG repeatedly failed were amplified with AmpliTaq Gold in 12.5 μL reactions containing 0.1 μL AmpliTaq Gold DNA polymerase (Applied Biosystems, Foster City, CA), 0.1 μL BSA (1mg/mL; Promega), 7.8 μL water, 1.25 μL 10×AmpliTaq Gold buffer (Applied Biosystems), 1 μL dNTPs (10 mM; Applied Biosystems), 1.25 μL MgCl2 and 0.25 μL of the primer pairs, using the same PCR conditions.

The nuclear ITS region was amplified using the primers ITS101 (17SE) and ITS102 (26SE) of [86]. PCR reactions were performed in volumes of 15 μL comprising 0.5 μL of total genomic DNA, 4.8 μL REDTaq Ready Mix PCR reaction mix (Sigma-Aldrich), 7.4 μL water, 0.6 μL BSA (1 mg/mL; New England BioLabs, Ipswich, MA, USA) and 0.3 μL (10 μM) of the primer pairs. The PCR conditions were: 1 min at 95°C, 35 cycles with 30 s at 95°C, 30 s at 52°C, 90 s at 72°C followed by 7 min at 72°C. Critical samples that repeatedly failed were amplified in 12.5 μL reactions containing 0.1 μL AmpliTaq Gold DNA polymerase (Applied Biosystems), 0.1 μL BSA (1mg/mL; Promega), 7.8 μL water, 1.25 μL 10×AmpliTaq Gold buffer (Applied Biosystems), 1 μL dNTPs (10 mM; Applied Biosystems), 1.25 μL MgCl2 and 0.25 μL of the primer pairs. The PCR conditions were 10 s at 95°C, 35 cycles with 30 s at 94°C, 1 min at 52°C, 1 min at 72°C followed by 10 s at 72°. ITS sequences of the putative hybrid D. × bauhini (pops. 25, 26) and of a subset of D. stiriacum samples (pops. 30–32, 35–38), showing a significant amount of ambiguities in direct sequencing (Table 3), were cloned. Molecular cloning was done as described in [8] with the following modifications: inserts of 3–13 clones were amplified using the universal primers M13F(–47) and M13R(–48) and 1 μL of colony DNA re-suspended in 60 μL of double distilled water. For all PCR programs a ramp temperature of 1.2°C/s was used.

thumbnail
Table 3. Polymorphic sites from the ITS region in the Doronicum clusii aggregate: nucleotides are those found without cloning and (in parentheses) after cloning.

https://doi.org/10.1371/journal.pone.0118197.t003

PCR products were purified using Exonuclease I and FastAP thermosensitive alkaline phosphatase (Fisher Scientific, St. Leon-Rot, Germany) following the manufacturer’s instructions. Cycle sequencing reactions were performed using 5 μL of purified template, 1 μL of primer (3.2 μM) and 1 μL BigDye Terminator (Applied Biosystems), cleaned with Sephadex G-50 Fine (GE Healthcare Bio-Sciences, Uppsala, Sweden) and sequenced on an ABI 3730 DNA Analyzer capillary sequencer (Applied Biosystems).

AFLP Data Analyses

Raw AFLP data were aligned with the internal size standard using ABI Prism GeneScan 3.7.1 (Applied Biosystems), and imported into Genographer 1.6.0 [87] for scoring. Each AFLP fragment was scored using the ‘thumbnail’ option, which allows the comparison of the signal of each fragment (present or absent) over all samples. Unambiguous bands in the size range of 100–500 bp were scored and exported as a presence/absence matrix. Shorter fragments were not scored due to the more frequent occurrence of non-homologous fragments [88]. Nei’s gene diversity over loci [89] termed ‘genetic diversity’ in the following, and the frequency of rare markers as frequency-down-weighted marker values ([19,90]; ‘rarity 1’ in AFLPdat), termed ‘rarity’ in the following, were calculated for each population with the R script AFLPdat ver. 20.10.2010 [91]. Briefly, each marker is down-weighted by its frequency in the entire data set. Subsequently, the rarity for an individual is calculated as the sum of the (down-weighted) markers present in this individual; population rarity values are estimated as the average of the individuals’ rarity values.

Genetic diversity and rarity were compared (1) among the four constituents of the D. clusii aggregate, (2) between Alpine populations of D. stiriacum from glacial refugia and populations from glaciated areas and (3) between D. glaciale subsp. calcareum and D. glaciale subsp. glaciale excluding the morphologically intermediate populations 6–10. Statistical comparisons were done with SPSS 21 (SPSS, Chicago, IL, USA) using two-tailed t-tests for normally distributed data and Mann-Whitney U-tests for data without normal distribution. Normal distribution was tested for with the Kolmogorov-Smirnov test. Post-hoc Tukey-Kramer tests were applied to detect differences among and within taxonomic groups at a 5% confidence level, a frequently applied confidence level in similar comparisons (e.g., [92,93]).

A NeighborNet [94] was produced with SplitsTree 4.13.1 [95] using uncorrected P-distances. NeighborNets display congruent and conflicting signals in a data set as a series of splits, where the weight of a split is visible as the length of a line or a set of parallel lines on the diagram. They are well suited to display conflicting signals, as is expected to result from hybridization or introgression. Bootstrap support values were obtained from neighbour-joining trees using Nei-Li distances and 2,000 pseudoreplicates in SplitsTree.

For a data set comprising samples of the D. clusii aggregate only (excluding D. grandiflorum) excluding the hybrid D. × bauhini (see Results), groups were delimited using Gaussian clustering with a noise component for outliers as implemented in the R library Prabclus 2.2–4 [96]. This approach is recommended for dominant multilocus genetic data and results are considered comparable or superior to model-based approaches such as STRUCTURE [97]. Briefly, nonmetric multidimensional scaling (NMDS [98]) is performed on a distance-matrix. The resulting Euclidean variables are used to determine clusters of individuals based on Gaussian clustering where the number of clusters is determined by the Bayesian Information Criterion using the R functions prabinit (default settings, distance = Jaccard) and prabclus (settings: mdsmethod = ‘kruskal’, nclus = 0:6). As the default number of nearest neighbours to determine the initial noise estimation (nnk) of 7 (i.e., the number of individuals divided by 40 and rounded to the next-largest integer: [97: p. 492]) yielded meaningless results (random non-assignments over the whole data set), this parameter was finally set to 3, that is the number of individuals from the largest of the three distinct groups (i.e., the one containing D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and their morphological intermediates) divided by 40 and rounded to the next-largest integer. The number of NMDS dimensions (mdsdim) was 3, which is the smallest mdsdim with a stress value below the arbitrary cut-value of 10% as determined by the function stressvals. Prabclus was repeated 200 times with a bootstrapped matrix (generated using seqboot from the PHYLIP package [99]) to test the support of group assignments. Bootstrap values of individuals were calculated as percentage of correct assignment to the original cluster (vs. non-assigned or assigned to another cluster). Two-dimensional visualization of the initial, non-bootstrapped matrix was performed with SigmaPlot 12.5 (Systat Software, San Jose, CA, USA).

For a data set containing D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and their morphological intermediates (data set 1) and for a second data set containing populations of D. clusii s.s., D. glaciale subsp. glaciale and their putative hybrid D. × bauhini (data set 2), population structure was inferred employing a Bayesian clustering approach based on MCMC estimations developed for dominant markers with recessive alleles (STRUCTURE 2.3.3 [100,101]). Since we assumed admixture between D. glaciale subsp. calcareum and D. glaciale subsp. glaciale and between D. clusii s.s., D. glaciale subsp. glaciale and their putative hybrid, only the admixture model with correlated (data set 1, which comprises closely related populations; [102]) or uncorrelated allele frequencies (data set 2) was used. Ten replicate runs for each K (number of groups) ranging from 1 to 10 were carried out at the Bioportal of the University of Oslo (http://www.bioportal.uio.no), using a burn-in of 105 iterations followed by 106 additional MCMC iterations. Similarity among results of different runs for the same K was calculated according to [103,104] using the R-script Structure-sum ver. 2009 (part of AFLPdat [91]). The optimal number of groups was identified as the value of K where the likelihood started to flatten out, the results of replicate runs were identical, no empty groups were encountered and the mean delta K was at the maximum. DISTRUCT 1.1 [105] was used to display the results graphically.

DNA Sequence Data Analyses

DNA sequences were assembled using SeqMan 7.0.0 (DNAStar, Madison, WI, USA) and manually edited and aligned with BioEdit 7.2.0 [106]. The alignment of the three concatenated plastid markers was analysed using statistical parsimony as implemented in TCS 1.21 [107] with the connection limit set to 95%; gaps were treated as fifth character state. For this analysis, indels longer than 1 bp and inversions were reduced to single base pair columns allowing those structural mutations to be counted as single base pair mutations only; additionally, mononucleotide repeats were removed due to their high degree of homoplasy at larger geographical scales [108]. As haplotype networks are known to be misled by ambiguous and missing data [109], ITS sequences (cloned ones in case of D. × bauhini and of D. stiriacum pops. 30–32 and 35–38) were analysed instead using maximum likelihood as implemented in RAxML 8.0.17 [110] using the GTRCAT substitution model and the fast bootstrap approach [111] with 500 replicates.

Results

Flow Cytometry and Chromosome Counts

CVs of G0/G1 peaks of the internal references standards were always below 3%. As expected, dehydrated samples showed higher CVs; the values usually did not exceed 5%, a threshold recommended by [69], but approached 7% in a few samples of the tetraploid D. stiriacum. It is known that polyploids often perform worse with DAPI in terms of quality and storability than diploids because nuclei may detoriorate faster [112]. DNA ploidy level measurements (Table 2) revealed that all individuals of D. clusii s.s., D. glaciale subsp. glaciale, D. glaciale subsp. calcareum and D. grandiflorum were DNA diploids (‘diploids’ hereafter) and all of D. stiriacum were DNA tetraploids (‘tetraploids’ hereafter). Pisum sativum was used as internal standard for 51 out of 58 populations, whereas Vicia faba was used for populations 5, 6, 28, 40, 44, 55 and 57. By mistake, population 28 was not analysed.

Chromosome numbers of diploids and tetraploids were determined to be 2n = 2x = 60 and 2n = 4x = c. 120, respectively (Fig. 2). A summary of new and published chromosome numbers is given in Table 4.

thumbnail
Fig 2. Mitotic chromosomes of members of the Doronicum clusii aggregate.

(a) D. clusii s.s. from Mt. Plose, Dolomiten, Italy (pop. 42; CP1042), 2n = 60; (b) D. stiriacum from Mt. Höchstein, Schladminger Tauern, Austria (pop. 34; CP1041), 2n = c. 120; (c) D. glaciale subsp. glaciale from Mt. Gruft, Gurktaler Alpen, Austria (pop. 15; CP1038), 2n = 60. Scale bars = 5 μm.

https://doi.org/10.1371/journal.pone.0118197.g002

thumbnail
Table 4. Published and new chromosome counts from the Doronicum clusii aggregate and from D. grandiflorum.

https://doi.org/10.1371/journal.pone.0118197.t004

AFLPs

After removing one monomorphic fragment, 47 fragments present in all but one individual as well as 41 non-reproducible fragments, the final data matrix (available from Dryad under http://dx.doi.org/10.5061/dryad.573pm) contained 319 polymorphic fragments in 276 individuals. In the AFLP profiles from replicated samples 949 differences were observed out of 27,744 phenotypic comparisons, resulting in an error rate of 3.42%.

Genetic diversity (Table 2, Fig. 3) varied approximately 22-fold among populations ranging from 0.006 in population 41 (D. clusii s.s.) to 0.137 in population 8 (morphological intermediate between D. glaciale subsp. calcareum and D. glaciale subsp. glaciale); the highest value in a non-admixed population was 0.134 in population 29 (D. stiriacum). Rarity varied sevenfold ranging from 0.48 in population 26 (D. × bauhini) to 3.20 in population 38 (D. stiriacum; Table 2, Fig. 3). After removing D. × bauhini and morphological intermediates between D. glaciale subsp. glaciale and D. glaciale subsp. calcareum, both D. clusii s.s. and D. glaciale subsp. glaciale, two taxa that occur mainly in formerly glaciated areas, exhibited significantly lower global genetic diversity and rarity (Tukey—Kramer test, α = 0.05) than both D. glaciale subsp. calcareum and D. stiriacum (Table 2, Fig. 3), two taxa that mainly occur in formerly unglaciated (refugial) areas. Alpine populations of D. stiriacum from refugia (populations 27–30, 32, 36–38) and from glaciated areas (populations 31, 33–35) did not differ with respect to genetic diversity (two-tailed t-test, t = 1.710, df = 10, P = 0.868) and rarity (two-tailed t-test, t = 1.537, df = 10, P = 0.155). Means of genetic diversity and rarity in D. glaciale subsp. calcareum and D. glaciale subsp. glaciale (excluding the morphologically intermediate populations 6–10) were significantly different (diversity: two-tailed t-test, t = 3.309, df = 17, P = 0.004; rarity: Mann—Whitney U-test, two tailed significance: Z = -3.256, P < 0.001). The number of private markers (i.e., restricted to a single taxon) was zero in D. glaciale subsp. glaciale, six in D. glaciale subsp. calcareum, 20 in D. clusii s.s., and 50 in D. stiriacum.

thumbnail
Fig 3. Patterns of (a, c) AFLP diversity and (b, d) rarity within populations of the D. clusii aggregate.

(a, b) D. glaciale subsp. calcareum (yellow), D. glaciale subsp. glaciale (red) and their morphological intermediates (orange); (c, d) D. clusii s.s. (blue) and D. stiriacum (green) from the Alps and the Carpathians (insert). Dot sizes are proportional to genetic diversities and rarities given in Table 2; five weighted categories were used. The maximum extent of the Alpine ice shield during the Last Glacial Maximum (LGM) is given as white line; hatched areas in blue and red indicate glacial refugia (i.e., areas outside the continuous glaciation during the LGM and areas within the ice shield but situated below the LGM snow line) on calcareous and siliceous bedrock, respectively (modified from [5]); the Carpathians were only locally glaciated, not exceeding 1% of their total area [125]. Scale bars: 50 km.

https://doi.org/10.1371/journal.pone.0118197.g003

The NeighbourNet (Fig. 4) showed that the D. clusii aggregate (bootstrap support BS 100) comprises three groups corresponding to D. clusii s.s. (BS 100), D. stiriacum (BS <50) and D. glaciale subsp. glaciale plus D. glaciale subsp. calcareum (BS 71). Gaussian clustering yielded four groups corresponding to D. clusii s.s., D. stiriacum, D. glaciale subsp. glaciale and D. glaciale subsp. calcareum (Table 2). Plotting of NMDS data showed three clearly distinguishable groups (Fig. 4); a weak separation of D. glaciale subsp. calcareum from D. glaciale subsp. glaciale was only evident along the third axis (not shown). Two individuals of population 5 of D. glaciale subsp. calcareum were mis-assigned to D. glaciale subsp. glaciale (Table 2, Fig. 4). Morphologically intermediate populations mostly grouped with D. glaciale subsp. calcareum (15 individuals: all from populations 7 and 9 and four and two from populations 6 and 8, respectively) whereas all individuals from the westernmost population 10 and the remaining individuals from populations 6 and 8 were assigned to D. glaciale subsp. glaciale (Table 2). The assignment of D. clusii s.s. and D. stiriacum to the respective groups was stable with high proportions of correct assignment (0.82–1.00 in D. clusii s.s. and 0.56–0.91 in D. stiriacum) and a negligible proportion of incorrect assignment (a single individual in a single bootstrap replicate was mis-assigned to D. clusii s.s.). The assignment of D. glaciale subsp. calcareum and D. glaciale subsp. glaciale to the respective clusters was less stable because the proportion of incorrect assignment of D. glaciale subsp. glaciale to the D. glaciale subsp. calcareum group was 0.23–0.39. Placement of the two mis-assigned individuals from population 5 (D. glaciale subsp. calcareum) was not stable, because the genetic assignment (to D. glaciale subsp. glaciale) was only 0.53–0.56 compared to the taxonomic assignment (to D. glaciale subsp. calcareum) of 0.29–0.30. The assignment of individuals from morphologically intermediate populations was not stable, either, and the proportion of correct assignment to the D. glaciale subsp. calcareum or D. glaciale subsp. glaciale group was 0.42–0.93 and 0.32–0.73, respectively.

thumbnail
Fig 4. Structure of AFLP variation in the Doronicum clusii aggregate.

(a) NeighborNet of the complete AFLP data set including the outgroup D. grandiflorum; splits with weight < 0.001 were omitted to aid legibility; numbers represent bootstrap values above 50% given for major groups only. (b) Non-metric multidimensional scaling of Jaccard distances using Gaussian clustering with Prabclus; only two of three dimensions are shown. The symbols represent Prabclus-groups (circles: D. clusii s.s.; squares: D. stiriacum; triangles pointing upwards: D. glaciale subsp. calcareum; triangles pointing downwards: D. glaciale subsp. glaciale; D. × bauhini was not included), the colours denote taxa as in Fig. 1; the arrows indicate two individuals of D. glaciale subsp. calcareum from population 5 that are assigned to D. glaciale subsp. glaciale. (c) Estimated population structure among D. glaciale subsp. calcareum (pops. 1–5), D. glaciale subsp. glaciale (pops. 11–24) and morphological intermediates (pops. 6–10) inferred from Bayesian clustering with STRUCTURE at K = 2. Each individual is represented by a vertical bar, black lines separate populations, and population numbers (as in Table 2) are given from west to east; the coloured bar above the barplots represents morphologically defined taxa as in Fig. 1.

https://doi.org/10.1371/journal.pone.0118197.g004

Among D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and their morphological intermediates, STRUCTURE distinguished two groups (Fig. 4). The two groups corresponded to populations 1–4 of D. glaciale subsp. calcareum plus population 9 (morphological intermediate with D. glaciale subsp. glaciale) and populations 11–24 of D. glaciale subsp. glaciale, respectively; the remaining morphologically intermediate populations were genetically admixed with either the D. glaciale subsp. calcareum gene pool (populations 5–8) or the D. glaciale subsp. glaciale gene pool (population 10) dominating.

Among D. clusii s.s., D. glaciale subsp. glaciale and their putative hybrid (D. × bauhini), STRUCTURE identified two groups. Although D. × bauhini was genomically admixed, alleles from D. glaciale subsp. glaciale dominated (at least 73% relative contribution in population 25 and at least 98% in population 26).

DNA sequences

Sequences are available from GenBank (see Table 2 for accession numbers). Sequence alignments are available from Dryad under http://dx.doi.org/10.5061/dryad.573pm. Sequence statistics are given in Table 5. Apart from excluded mononucleotide repeats, chloroplast markers showed no intra-populational variation. The statistical parsimony network, comprising 25 haplotypes from 62 sequences (Table 2, Fig. 5), shows that apart from the clearly separated outgroup species D. austriacum, D. carpaticum and D. grandiflorum, four haplotype groups are evident corresponding to D. clusii s.s., D. glaciale subsp. glaciale plus D. glaciale subsp. calcareum, D. stiriacum from the Carpathians and D. stiriacum from the Alps. Haplotype 12 of D. clusii s.s., which is common in the Central Alps, is shared by population 25 of D. × bauhini, whereas haplotype 2, the most common haplotype of D. glaciale subsp. glaciale plus D. glaciale subsp. calcareum, is found in population 26 of D. × bauhini.

thumbnail
Fig 5. Phylogenetic relationships of the Doronicum clusii aggregate based on plastid DNA sequences.

Statistical parsimony network of plastid haplotypes; non-sampled haplotypes are shown as small open circles. The diameter of circles is proportional to the number of sampled haplotypes. Colour-coding of taxa as in Fig. 1; outgroup species: D. grandiflorum (light grey), D. carpaticum (middle grey), D. austriacum (black).

https://doi.org/10.1371/journal.pone.0118197.g005

In the maximum likelihood tree of the 127 ITS sequences (-ln 2567.7; Fig. 6) the D. clusii aggregate forms a monophyletic group (BS 66). Doronicum clusii s.s., which shows no intraspecific sequence variation beyond a few ambiguous sites (Table 3), groups with cloned sequences of both D. × bauhini populations (BS 79). Doronicum stiriacum does not form a monophyletic group, but falls into two clades. An unsupported clade containing cloned sequences from the Carpathians (populations 36–38) and some, but not all populations from the Alps (populations 30–32, 35) is weakly supported sister (BS 56) to the unsupported clade containing D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and D. stiriacum. Sequences of D. stiriacum (from all populations) form a clade (BS 63) sister group to a clade (BS 55) comprising D. glaciale subsp. calcareum, D. glaciale subsp. glaciale, their morphological intermediates and cloned sequences of both D. × bauhini populations. Samples from D. glaciale subsp. calcareum, D. glaciale subsp. glaciale and their morphological intermediates intermix, and phylogenetic structure, if present, does not correspond to taxonomic boundaries.

thumbnail
Fig 6. Phylogenetic relationships of the Doronicum clusii aggregate based on nuclear DNA sequences.

Maximum likelihood tree based on nuclear ITS data; numbers above branches are bootstrap support values above 50%. The coloured bar to the right represents morphologically defined taxa as in Fig. 1.

https://doi.org/10.1371/journal.pone.0118197.g006

Discussion

Phylogeny and taxonomy of the D. clusii aggregate

The D. clusii aggregate is confirmed as a cohesive group distinct from D. grandiflorum (Figs. 5, 6). This group comprises three genetic lineages (Figs. 45), which correspond to diploid D. clusii s.s., tetraploid D. stiriacum and diploid D. glaciale subsp. glaciale plus D. glaciale subsp. calcareum (referred to as D. glaciale s.l. in the following). Álvarez Fernández et al. [49] did neither include D. glaciale subsp. calcareum nor D. stiriacum, which were considered conspecific with D. glaciale subsp. glaciale and D. clusii s.s., respectively [50], and ours are the first molecular results on both taxa. Due to overall morphological similarity D. stiriacum has often been treated as subspecies of D. clusii s.s. (e.g., [113]) or as a synonym of D. clusii s.s. (e.g., [50]). As shown here (Figs. 46), D. stiriacum is genetically clearly separated from D. clusii s.s. and, therefore, should be treated at the species level. This is also justified by differences in morphology (thicker, coarse leaves that are densely villous on both sides, sparse glands on the involucrum, and villous corolla tubes in D. stiriacum versus tender leaves that are almost glabrous on the upper side, abundant glands on the involucrum and scape, and glabrous corolla tubes in D. clusii s.s.; S1 Appendix), ploidy level (tetraploid versus diploid) and distribution range (eastern-most Alps and Carpathians versus Alps except for the eastern-most parts [59,114]).

The regionally endemic D. glaciale subsp. calcareum is weakly, but consistently differentiated from widespread D. glaciale subsp. glaciale. Plastid DNA haplotypes found in D. glaciale subsp. calcareum are shared with D. glaciale subsp. glaciale (Fig. 5) and ITS shows no divergence between the two taxa (Fig. 6). Except for the NeighborNet analysis (Fig. 4A), AFLP data are structured according to the morphological differentiation of D. glaciale subsp. calcareum and D. glaciale subsp. glaciale (Fig. 4B, C). This pattern is only blurred by population 5, which is genetically admixed and contains individuals mis-assigned by Gaussian clustering (Fig. 4B), but is not morphologically intermediate. The distinction between D. glaciale subsp. calcareum and D. glaciale subsp. glaciale implied by Gaussian clustering is, however, not stable with an average bootstrap support of only 72 for an individual’s correct assignment. As expected, morphological intermediates (pops. 6–10) show varying levels of AFLP admixture (Fig. 4C) and, consequently, are assigned to either D. glaciale subsp. calcareum or D. glaciale subsp. glaciale (Fig. 4B). As D. glaciale subsp. calcareum and D. glaciale subsp. glaciale are genetically connected in a wide contact zone mainly containing morphological intermediates, have parapatric distribution ranges (Fig. 1) and show only subtle differences in leaf and involucrum indumentum (S1 Appendix), we follow recommendations by [115] for sexually reproducing flowering plants and retain the subspecific rank for the taxon from the northeastern-most Alps originally described as separate species, D. calcareum Vierh. [58].

Our data confirm the status of D. × bauhini as hybrid between D. clusii s.s. and D. glaciale subsp. glaciale, as suggested previously based on morphology [48,51]. Based on cloned ITS sequences, the two populations of D. × bauhini group with both D. clusii s.s. and D. glaciale s.l. (Fig. 6). Similarly, population 25 possesses haplotype 12 common in D. clusii s.s., whereas population 26 possesses haplotype 2 common in D. glaciale s.l. (Fig. 5, Table 2). This suggests that both D. clusii s.s. and D. glaciale (based on geographical considerations almost certainly subsp. glaciale) acted as maternal parents. In the NeighborNet of AFLP-data D. × bauhini populations are closer to D. glaciale s.l. than to D. clusii s.s. (Fig. 4B), which agrees with the higher proportion of the D. glaciale s.l. genome in D. × bauhini inferred from STRUCTURE analyses. This likely is due to backcrossing and corresponds well to a higher morphological affinity to D. glaciale subsp. glaciale. Higher morphological similarity to one parent was also found in D. × minutilloi Peruzzi (D. columnae Ten. × D. orientale Hoffm.) from Central Italy [52]. Backcrossing is feasible, because D. × bauhini possesses regularly developed pollen as confirmed by SEM analysis (pollen from population 25; H. Halbritter and C. Pachschwöll, unpubl. data).

Origin of polyploid D. stiriacum

Flow cytometry and chromosome counts revealed constant tetraploidy of D. stiriacum over the entire distribution range including the hitherto not investigated Alps (Tables 2, 4). The molecular data indicate that tetraploid D. stiriacum is a distinct evolutionary lineage clearly separated from diploid D. clusii s.s. and D. glaciale s.l. and exhibiting a higher number of private AFLP fragments than any other member of the D. clusii aggregate. Molecular evidence is, however, inconclusive with respect to an auto- or allopolyploid origin. The mode of polyploidization (auto‐ versus allopolyploidy) was not discussed by the authors studying karyology and morphology of D. stiriacum (e.g., [55,114]). The hypothesis of an autopolyploid origin from D. clusii s.s. (or an extinct diploid lineage) finds support in overall morphological and ecological similarities between the two species [59,114] and the lack of signal for mixed ancestry in the sequence data (Figs. 5, 6). The alternative hypothesis of an allopolyploid origin involving D. glaciale s.l. and D. clusii s.s. has not been considered previously, but would be in line with morphological traits resembling D. glaciale s.l., such as the thick, coarse leaves and the rare presence of stipitate glands on the basal leaves [59,114,116], as well as the presence of hybrids in contact areas (i.e., D. × bauhini). Further evidence comes from the lack of monophyly of D. stiriacum in the ITS phylogeny (Fig. 6) that may, however, be due to divergent evolution of different 35S rDNA loci [64], the intermediate position between D. clusii s.s. and D. glaciale s.l. in the AFLP NeighborNet and the lack of long splits supporting D. stiriacum (Fig. 4A), a pattern resembling the recently evolved allopolyploid Androsace brigantiaca Jord. & Fourr. [117].

Phylogeography and spatiotemporal evolution

Species responded idiosyncratically to Pleistocene climatic oscillations [118], but generally higher values of genetic rarity—but not necessarily of genetic diversity [119]—are expected in refugia or rear edge populations than in recolonized areas [2,39,42,120,121]. Impact of climatic oscillations, likely during the Pleistocene, on genetic patterns is evident in the D. clusii aggregate and can be best seen in the elevated levels of genetic diversity and rarity (Fig. 3) in D. glaciale subsp. calcareum and D. stiriacum, whose ranges are either almost entirely restricted to peripheral refugia in the northeastern Alps or only slightly extend westwards from refugia in the easternmost Central Alps [5]. In contrast, D. clusii s.s. and D. glaciale subsp. glaciale are widely distributed mostly in previously strongly glaciated areas and possess significantly lower global genetic diversity and rarity (Fig. 3).

At the eastern margin of the distribution range of D. clusii s.s., where this species is rare and restricted to a few isolated populations (populations 39–41), genetic diversity is low. Population 39 possesses the lowest rarity of all populations (Fig. 3, Table 2) and is divergent in the NeighborNet (Fig. 4), most probably due to its genetic depauperation. Each of populations 39–41 has a unique haplotype (9, 11, 13; Fig. 5) suggesting eastwards expansion, which stopped in the Hohe Tauern (part of the Central Alps: Fig. 1), a frequently emerging break zone of alleles and species distributions in silicicolous species [22]. Otherwise, no phylogeographic structure can be discerned neither by AFLP variation nor by the uniform ITS sequences (Fig. 6). As our study focused on the eastern Alps as the only area of co-occurrence of taxa of the D. clusii aggregate, the southwestern and the middle southern Alps, which may have acted as important glacial refugia [5] also for D. clusii s.s., have remained unsampled or are underrepresented.

Vierhapper [58] postulated a recent, postglacial origin for D. stiriacum, which he deemed responsible for its restriction to a small area. However, the divergence between Alpine and Carpathian populations in all genetic data (Figs. 46) and the possession of plastid haplotypes not found in the diploid species (Fig. 5) can be better explained by an earlier preglacial origin of D. stiriacum. An earlier origin is supported by macrofossils from the northern foothills of the western Carpathians with an estimated age of c. 29,500 cal years BP [121] thus predating the Last Glacial Maximum in the western Carpathians [122124]. This fits well to the hypothesized downward displacement of alpine vegetation and is in line with the suggested predominance of herbaceous arctic‐alpine species in the Carpathian forelands during that period [125]. A plausible scenario for the evolution of D. stiriacum includes an eastern Alpine origin (based on the distribution of the species’ closest relatives) and subsequent dispersal to the Carpathians, a scenario suggested for other silicicolous species such as Jacobaea carniolica (Willd.) Schrank (syn. Senecio carniolicus Willd.; [76]), Salix herbacea L. [126] or Cardamine resedifolia L. [127]. Subsequently, divergence took place in phases of vicariance and glacial survival in both the Alps and Carpathians. As no difference in genetic rarity emerged between refugia (populations 27–30, 32, 36–38) and glaciated areas (populations 31, 33–35; Fig. 3) in the Alps, slow, broad-fronted diffusion likely led to the small, but compact current distribution. For D. stiriacum a pronounced post-glacial migration lag resulting in incomplete range filling in the eastern Alps was modelled [12], suggesting that the species’ current distribution range can be better explained with historical than with ecological factors. This strong migration lag and the limited extent of postglacial range expansion may also explain the lack of strong imprints of refugial patterns as observed in other silicicolous species of similar distribution [127129].

Doronicum glaciale s.l. exhibits a particularly clear imprint of its glacial history. Doronicum glaciale subsp. calcareum from the previously only locally glaciated northeastern Alps exhibits high levels of AFLP diversity and rarity (Fig. 3, Table 2), whereas D. glaciale subsp. glaciale, predominantly distributed in formerly glaciated areas, is strongly depauperate (Fig. 3). In none of the hitherto investigated calcicolous Alpine species [16,63,82,90] were such strong intraspecific patterns observed. A likely hypothesis is that westward leading-edge colonization [120] started from source populations on the Hochschwab Massif (the area, where the morphological intermediate populations 7–10 occur: Fig. 1), which are genetically intermediate (Fig. 4) and possess short glands on the basal leaves, a morphological feature not found in eastern populations of D. glaciale subsp. calcareum (S1 Appendix). This hypothesis identifies the Hochschwab as primary contact zone and suggests parapatric origin of D. glaciale subsp. glaciale from D. glaciale subsp. calcareum. The derived D. glaciale subsp. glaciale exhibits a wider niche (extension onto basic silicates) as observed in South American Pozoa volcanica Mathias & Constance [130]. Alternatively, D. glaciale subsp. glaciale might have undergone a founder effect within the unglaciated easternmost Central Alps, from where it subsequently expanded its range also north(east)wards to finally form a secondary contact zone at the Hochschwab Massif. This hypothesis is, however, less likely as populations in the contact zone do not show increased genetic diversity (Fig. 3, Table 2) as expected for an evolutionary melting pot [119].

Conclusion

Our study identifies range shifts, which were likely triggered by Pleistocene climatic oscillations, and polyploidisation as main forces shaping the evolutionary history and consequently the genetic structure of D. clusii and relatives. Integrating morphological, karyological, genetic and biogeographic data allows us to draw solid taxonomic conclusions with respect to controversial taxa (D. stiriacum, D. glaciale subsp. calcareum). The highly uneven distribution of genetic variation across species strongly suggests that a better understanding of the evolution of intricate alpine species groups is necessary for designing conservation strategies for endemics in the light of global warming.

Supporting Information

S1 Appendix. Systematics, morphological differentiation, ecology and geographical distribution of the constituents of the Doronicum clusii aggregate and its closest relative D. grandiflorum.

https://doi.org/10.1371/journal.pone.0118197.s001

(PDF)

Acknowledgments

This paper is dedicated to the memory of our friend and colleague Ruth Flatscher (1985–2014). The authors thank B. Frajman, C. Gilli, M. Gina, H. P. Grohmann, A. Hahnekamp, R. Hehenberger, V. Kolarčik, A. Miernik, J. Pachschwöll, M. Puşcaş, D. Reich, A. & M. Ronikier, M. Sonnleitner and A. Tribsch for collecting plant material or assistance in the field; M. Sonnleitner for plotting the GIS maps; H. Weiss-Schneeweiss, E. M. Temsch and J. Greilhuber as well as M. Barfuss, E. Grasserbauer, V. Klejna and G. Kohl for providing advice for cytogenetic and molecular lab work, respectively; H. Halbritter for SEM-analyses; J. Suda and his team (Charles University in Prague) for flow cytometric analyses. M. A. Fischer, G. Fischer, H. Niklfeld, E. Magyari and T. Stuessy provided literature and translations and F. Tod took care of the living plants of Doronicum in the Botanical Garden of the University of Vienna (HBV). A. Laciny is acknowledged for providing useful comments on a previous version of the manuscript. We thank L. Peruzzi, one anonymous reviewer and the academic editor E. Hörandl for their insightful comments that helped improving our manuscript

Author Contributions

Conceived and designed the experiments: CP PS. Performed the experiments: CP PEG. Analyzed the data: CP MW GMS PS. Wrote the paper: CP GMS PS.

References

  1. 1. Comes HP, Kadereit JW (1998) The effect of Quaternary climatic changes on plant distribution and evolution. Trends Plant Sci 3: 432–438.
  2. 2. Hewitt GM (2004) Genetic consequences of climatic oscillations in the Quaternary. Philos Trans R Soc Lond B Biol Sci 359: 183–195. pmid:15101575
  3. 3. Schmitt T (2007) Molecular biogeography of Europe: Pleistocene cycles and postglacial trends. Frontiers Zool 4: 11. pmid:17439649
  4. 4. Tribsch A, Schönswetter P (2003) Patterns of endemism and comparative phylogeography confirm palaeoenvironmental evidence for Pleistocene refugia in the Eastern Alps. Taxon 52: 477–497. pmid:12857639
  5. 5. Schönswetter P, Stehlik I, Holderegger R, Tribsch A (2005) Molecular evidence for glacial refugia of mountain plants in the European Alps. Mol Ecol 14: 3547–3555. pmid:16156822
  6. 6. Holderegger R, Thiel-Egenter C (2009) A discussion of different types of glacial refugia used in mountain biogeography and phylogeography. J Biogeogr 36: 476–480.
  7. 7. Schneeweiss GM, Schönswetter P (2011) A re-appraisal of nunatak survival in arctic-alpine phylogeography. Mol Ecol 20: 190–192. pmid:21265053
  8. 8. Escobar García P, Winkler M, Flatscher R, Sonnleitner M, Krejčíková J, et al. (2012) Extensive range persistence in peripheral and interior refugia characterizes Pleistocene range dynamics in a widespread Alpine plant species (Senecio carniolicus, Asteraceae). Mol Ecol 21: 1255–1270. pmid:22276934
  9. 9. Birks HH (2008) The Late-Quaternary history of arctic and alpine plants. Plant Ecol Divers 1: 135–146.
  10. 10. Birks HJB, Willis KJ (2008) Alpines, trees, and refugia in Europe. Plant Ecol Divers 1: 147–160.
  11. 11. Svenning JC, Skov F (2007) Could the tree diversity pattern in Europe be generated by postglacial dispersal limitation? Ecol Lett 10: 453–460. pmid:17498144
  12. 12. Dullinger S, Willner W, Plutzar C, Englisch T, Schratt-Ehrendorfer L, et al. (2012) Post-glacial migration lag restricts range filling of plants in the European Alps. Glob Ecol Biogeogr 21: 829–840.
  13. 13. Merxmüller H (1952) Untersuchungen zur Sippengliederung und Arealbildung in den Alpen. I. Jahrb Vereins Schutze Alpenpfl Alpentiere 17: 96–133.
  14. 14. Merxmüller H (1953) Untersuchungen zur Sippengliederung und Arealbildung in den Alpen. II. Jahrb Vereins Schutze Alpenpfl Alpentiere 18: 138–158.
  15. 15. Merxmüller H (1954) Untersuchungen zur Sippengliederung und Arealbildung in den Alpen. III. Jahrb Vereins Schutze Alpenpfl Alpentiere 19: 97–139.
  16. 16. Schneeweiss GM P. Schönswetter P (2010) The wide but disjunct range of the European mountain plant Androsace lactea L. (Primulaceae) reflects Late Pleistocene range fragmentation and post-glacial distributional stasis. J Biogeogr 37: 2016–2025.
  17. 17. Sanz M, Schönswetter P, Vallès J, Schneeweiss GM, Vilatersana R (2014) Southern isolation and northern long-distance dispersal shaped the phylogeography of the widespread, but highly disjunct, European high mountain plant Artemisia eriantha (Asteraceae). Bot J Linn Soc 174: 214–226.
  18. 18. Schönswetter P, Tribsch A, Barfuss M, Niklfeld H (2002) Several Pleistocene refugia detected in the high alpine plant Phyteuma globulariifolium Sternb. & Hoppe (Campanulaceae) in the European Alps. Mol Ecol 11: 2637–2647. pmid:12453246
  19. 19. Schönswetter P, Tribsch A (2005) Vicariance and dispersal in the alpine perennial Bupleurum stellatum L. (Apiaceae). Taxon 54: 725–732.
  20. 20. Mráz P, Gaudeul M, Rioux D, Gielly L, Choler P, et al. (2007) Genetic structure of Hypochaeris uniflora (Asteraceae) suggests vicariance in the Carpathians and rapid post-glacial colonization of the Alps from an eastern Alpine refugium. J Biogeogr 34: 2100–2114.
  21. 21. Hewitt GM (2011) Quaternary phylogeography: the roots of hybrid zones. Genetica 139: 617–638. pmid:21234647
  22. 22. Thiel-Egenter C, Alvarez N, Holderegger R, Tribsch A, Englisch T, et al. (2011) Break zones in the distributions of alleles and species in alpine plants. J Biogeogr 38: 772–782.
  23. 23. Ronikier M, Schneeweiss GM, Schönswetter P (2012) The extreme disjunction between Beringia and Europe in the arctic-alpine Ranunculus glacialis s. l. does not coincide with the deepest genetic split—a story of the importance of temperate mountain ranges in arctic-alpine phylogeography. Mol Ecol 21: 5561–5578. pmid:23057869
  24. 24. Kadereit JW, Griebeler EM, Comes HP (2004) Quaternary diversification in European alpine plants: Pattern and process. Philos Trans R Soc Lond B Biol Sci 359: 265–274. pmid:15101582
  25. 25. Willis KJ, Niklas KJ (2004) The role of Quaternary environmental change in plant macroevolution: the exception or the rule? Philos Trans R Soc Lond B Biol Sci 359: 159–172. pmid:15101573
  26. 26. Otto SP, Whitton J (2000) Polyploid incidence and evolution. Annu Rev Genet 34: 401–437. pmid:11092833
  27. 27. Soltis DE, Soltis PS, Tate JA (2003) Advances in the study of polyploidy since Plant speciation. New Phytol 161: 173–191.
  28. 28. Sobel JM, Chen GF, Watt LR, Schemske DW (2010) The biology of speciation. Evolution 64: 295–315. pmid:19891628
  29. 29. Soltis PS, Soltis DE (2009) The role of hybridization in plant speciation. Annu Rev Plant Biol 60: 561–588. pmid:19575590
  30. 30. Choler P, Erschbamer B, Tribsch A, Gielly L, Taberlet P (2004) Genetic introgression as a potential to widen a species’ niche: Insights from alpine Carex curvula. Proc Natl Acad Sci USA 101: 171–176. pmid:14691249
  31. 31. Burnier J, Buerki S, Arrigo N, Küpfer P, Alvarez N (2009) Genetic structure and evolution of alpine polyploid complexes: Ranunculus kuepferi (Ranunculaceae) as a case study. Mol Ecol 18: 3730–3744. pmid:19674303
  32. 32. Guggisberg A, Mansion G, Conti E (2009) Disentangling reticulate evolution in an arctic-alpine polyploid complex. Syst Biol 58: 55–73. pmid:20525568
  33. 33. Brysting AK, Mathiesen C, Marcussen T (2011) Challenges in polyploidy phylogenetic reconstruction: A case story from the arctic-alpine Cerastium alpinum complex. Taxon 60: 333–347.
  34. 34. Casazza G, Granato L, Minuto L, Conti E (2012) Polyploid evolution and Pleistocene glacial cycles: a case study from the alpine primrose Primula marginata (Primulaceae). BMC Evol Biol 12: 56. pmid:22530870
  35. 35. Schmickl R, Paule J, Klein J, Marhold K, Koch MA (2012) The evolutionary history of the Arabidopsis arenosa complex: diverse tetraploids mask the Western Carpathian center of species and genetic diversity. PLoS ONE 7: e42691. pmid:22880083
  36. 36. Dillenberger MS, Kadereit JW (2013) The phylogeny of the European high mountain genus Adenostyles (Asteraceae-Senecioneae) reveals that edaphic shifts coincide with dispersal events. Amer J Bot 100: 1171–1183.
  37. 37. Magauer M, Schönswetter P, Jang T-S, Frajman B (2014) Disentangling relationships within the disjunctly distributed Alyssum ovirense/A. wulfenianum group (Brassicaceae), including description of a novel species from the north-eastern Alps. Bot J Linn Soc 176: 486–505.
  38. 38. Tribsch A (2004) Areas of endemism of vascular plants in the Eastern Alps in relation to Pleistocene glaciations. J Biogeogr 31: 747–760.
  39. 39. Tribsch A, Englisch T, Gugerli F, Holderegger R, Niklfeld H, et al. (2010) Integrating data across biodiversity levels: the project IntraBioDiv. In: Spehn EM, Körner C, editors. Data mining for global trends in mountain biodiversity. Boca Raton: CRC Press, Taylor & Francis. pp. 89–105.
  40. 40. Essl F, Dullinger S, Plutzar C, Willner W, Rabitsch W (2011) Imprints of glacial history and current environment on correlations between endemic plant and invertebrate species richness. J Biogeogr 38: 604–614.
  41. 41. Hurdu BI, Pușcaș M, Turtureanu PD, Niketić M, Coldea G, et al. (2012) Patterns of plant endemism in the Romanian Carpathians (South-Eastern Carpathians). Contr Bot 47: 25–38.
  42. 42. Taberlet P, Zimmermann NE, Englisch T, Tribsch A, Holderegger R, et al. (2012) Genetic diversity in widespread species is not congruent with species richness in alpine plant communities. Ecol Lett 15: 1439–1448. pmid:23006492
  43. 43. Schönswetter P, Suda J, Popp M, Weiss-Schneeweiss H, Brochmann C (2007) Circumpolar phylogeography of Juncus biglumis (Juncaceae) inferred from AFLP fingerprints, cpDNA sequences, nuclear DNA content and chromosome numbers. Mol Phylogen Evol 42: 92–103. pmid:16905337
  44. 44. Sonnleitner M, Flatscher R, Escobar García P, Rauchová J, Suda J, et al. (2010) Distribution and habitat segregation on different spatial scales among diploid, tetraploid and hexaploid cytotypes of Senecio carniolicus (Asteraceae) in the Eastern Alps. Ann Bot 106: 967–978. pmid:20880930
  45. 45. Winkler M, Tribsch A, Schneeweiss GM, Brodbeck S, Gugerli F, et al. (2012) Tales of the unexpected: phylogeography of the arctic-alpine model plant Saxifraga oppositifolia (Saxifragaceae) revisited. Mol Ecol 21: 4618–4630. pmid:22809067
  46. 46. Winkler M, Tribsch A, Schneeweiss GM, Brodbeck S, Gugerli F, et al. (2013) Strong nuclear differentiation contrasts with widespread sharing of plastid DNA haplotypes across taxa in European purple saxifrages (Saxifraga section Porphyrion subsection Oppositifoliae). Bot J Linn Soc 173: 622–636.
  47. 47. Cavillier F (1907) Étude sur les Doronicum à fruits homomorphes. Annu Conserv Jard Bot Genève 10: 177–251.
  48. 48. Cavillier F (1911) Nouvelles études sur le genre Doronicum. Annu Conserv Jard Bot Genève 13–14: 195–368.
  49. 49. Álvarez Fernández I, Fuertes Aguilar J, Panero J, Nieto Feliner G (2001) A phylogenetic analysis of Doronicum (Asteraceae, Senecioneae) based on morphological, nuclear ribosomal (ITS), and chloroplast (trnL-F) evidence. Mol Phylogen Evol 20: 41–64.
  50. 50. Álvarez Fernández I (2003) Systematics of the Eurasian and North-African genus Doronicum (Asteraceae, Senecioneae). Ann Miss Bot Gard 90: 319–389.
  51. 51. Widder FJ (1934) Die Bastarde der Doronicum-Arten. Mitt Naturwiss Vereins Steiermark 71: 132–146.
  52. 52. Peruzzi L, Bedini G, Andreucci A (2012) Homoploid hybrid speciation in Doronicum L. (Asteraceae)? Morphological, karyological and molecular evidences. Plant Biosyst 146: 867–877.
  53. 53. Álvarez Fernández I, Nieto Feliner G (2001) A multivariate approach to assess the taxonomic utility of morphometric characters in Doronicum (Asteraceae, Senecioneae). Folia Geobot 36: 423–444.
  54. 54. Pachschwöll C (2013) Lectotypification of Doronicum carpaticum (Griseb. & Schenk) Nyman (Asteraceae)), with some remarks on its distribution and taxonomy. Contr Bot 48: 7–14.
  55. 55. Wcisło H (1952) Cytological and embryological studies in Doronicum L. Bull Int Acad Pol Sci Lett, Cl Sci Math Nat Sér B Sci Nat 1951: 147–166.
  56. 56. Fabbro T, Körner C (2004) Altitudinal differences in flower traits and reproductive allocation. Flora 199: 70–81.
  57. 57. Tackenberg O, Stöcklin J (2008) Wind dispersal of alpine plant species: A comparison with lowland species. J Veg Sci 19: 109–118.
  58. 58. Vierhapper F (1900) „Arnica Doronicum Jacquin”und ihre nächsten Verwandten. sterr Bot Z 50: 109–115, 173–178, 202–208, 257–264, 501.
  59. 59. Pachschwöll C, Pușcaș M, Schönswetter P (2011) Distribution of Doronicum clusii and D. stiriacum (Asteraceae) in the Alps and Carpathians. Biologia (Bratislava) 66: 977–987.
  60. 60. Müller H (1881) Alpenblumen, ihre Befruchtung durch Insekten und ihre Anpassungen an dieselben. Leipzig: Engelmann. 616 p.
  61. 61. Günthart A (1926) Die Blütenbiologie der Alpenflora. In: Schröter C, editor. Das Pflanzenleben der Alpen. Eine Schilderung der Hochgebirgsflora. 2nd ed. Zürich: Raustein. pp. 1028–1097.
  62. 62. Després L, Gielly L, Redoutet B, Taberlet P (2003) Using AFLP to resolve phylogenetic relationships in a morphologically diversified plant species complex when nuclear and chloroplast sequences fail to reveal variability. Mol Phylogen Evol 27: 185–196. pmid:12695084
  63. 63. Paun O, Schönswetter P, Winkler M, Intrabiodiv Consortium, Tribsch A (2008) Historical divergence vs. contemporary gene flow: Evolutionary history of the calcicole Ranunculus alpestris group (Ranunculaceae) in the European Alps and the Carpathians. Mol Ecol 17: 4263–4275. pmid:19378404
  64. 64. Álvarez I, Wendel JF (2003) Ribosomal ITS sequences and plant phylogenetic inference. Mol Phylogen Evol 29: 417–434.
  65. 65. Nieto Feliner G, Rosselló JA (2007) Better the devil you know? Guidelines for insightful utilization of nrDNA ITS in species level evolutionary studies in plants. Mol Phylogen Evol 44: 911–919. pmid:17383902
  66. 66. Corriveau JL, Coleman AW (1988) Rapid screening method to detect potential biparental inheritance of plastid DNA and results for over 200 angiosperm species. Amer J Bot 75: 1443–1458.
  67. 67. Shaw J, Lickey EB, Schilling EE, Small RL (2007) Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. Amer J Bot 94: 275–288. pmid:21636401
  68. 68. Borsch T, Quandt D (2009) Mutational dynamics and phylogenetic utility of noncoding chloroplast DNA. Plant Syst Evol 282: 169–199.
  69. 69. Doležel J, Greilhuber J, Suda J (2007) Estimation of nuclear DNA content in plants using flow cytometry. Nat Protoc 2: 2233–2244. pmid:17853881
  70. 70. Kron P, Suda J, Husband BC (2007) Applications of flow cytometry to evolutionary and population biology. Annu Rev Ecol Evol Syst 38: 847–876.
  71. 71. Alvarez N, Thiel-Egenter C, Tribsch A, Holderegger R, Manel S, et al. (2009) History or ecology? Substrate type as a major driver of spatial genetic structure in Alpine plants. Ecol Lett 12: 632–640. pmid:19392716
  72. 72. Frajman B, Pachschwöll C, Schönswetter P (2014) Contributions to the knowledge of the flora of the Dinarides (Balkan Peninsula). Phyton (Horn) 54: 27–46.
  73. 73. Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochem Bull 19: 11–15.
  74. 74. Tel-Zur N, Abbo S, Myslabodski D, Mizrahi Y (1999) Modified CTAB procedure for DNA isolation from epiphytic cacti of genera Hylocereus and Selenicereus (Cactaceae). Plant Mol Biol Rep 17: 249–254.
  75. 75. Suda J, Krahulcová A, Trávníček P, Krahulec F (2006) Ploidy level versus DNA ploidy level: an appeal for consistent terminology. Taxon 55: 447–450.
  76. 76. Suda J, Weiss-Schneeweiss H, Tribsch A, Schneeweiss GM, Trávníček P, et al. (2007) Complex distribution patterns of di-, tetra-, and hexaploid cytotypes in the European high mountain plant Senecio carniolicus (Asteraceae). Amer J Bot 94: 1391–1401. pmid:21636507
  77. 77. Trávníček P, Kubátová B, Čurn V, Rauchová J, Krajníková E, et al. (2011) Remarkable coexistence of multiple cytotypes of the Gymnadenia conopsea aggregate (the fragrant orchid): evidence from flow cytometry. Ann Bot 107: 77–87. pmid:21059612
  78. 78. Kolář F, Lučanová M, Vít P, Urfus T, Chrtek J, et al. (2013) Diversity and endemism in deglaciated areas: ploidy, relative genome size and niche differentiation in the Galium pusillum complex (Rubiaceae) in Northern and Central Europe. Ann Bot 111: 1095–1108. pmid:23589633
  79. 79. Doležel J, Greilhuber J, Lucretti S, Meister A, Lysák MA, et al. (1998) Plant genome size estimation by flow cytometry: inter-laboratory comparison. Ann Bot 82 (Suppl A): 17–26.
  80. 80. Doležel J, Sgorbati S, Lucretti S (1992) Comparison of three DNA fluorochromes for flow cytometric estimation of nuclear DNA content in plants. Physiol Plant 85: 625–631.
  81. 81. Vos P, Hogers R, Bleeker M, Reijans M, Van De Lee T, et al. (1995) AFLP—a new technique for DNA fingerprinting. Nucl Acids Res 23: 4407–4414. pmid:7501463
  82. 82. Schönswetter P, Solstad H, Escobar García P, Elven R (2009) A combined molecular and morphological approach to the taxonomically intricate European mountain plant Papaver alpinum s.l. (Papaveraceae). Taxa or informal phylogeographical groups? Taxon 17: 1326–1343.
  83. 83. Bonin A, Bellemain E, Bronken Eidesen P, Pompanon F, Brochmann C, et al. (2004) How to track and assess genotyping errors in population genetic studies. Mol Ecol 13: 3261–3273. pmid:15487987
  84. 84. Oxelman B, Lidén M, Berglund D (1997) Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst Evol 206: 393–410.
  85. 85. Taberlet P, Gielly L, Pautou G, Bouvet J (1991) Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol Biol 17: 1105–1109. pmid:1932684
  86. 86. Sun Y, Skinner DZ, Liang GH, Hulbert SH (1994) Phylogenetic analysis of Sorghum and related taxa using Internal Transcribed Spacers of nuclear ribosomal DNA. Theor Appl Genet 89: 26–32. pmid:24177765
  87. 87. Benham J, Jeung JU, Jasieniuk M, Kanazin V, Blake T (1999) Genographer: a graphical tool for automated fluorescent AFLP and microsatellite analysis. J Agric Genom 4: 399.
  88. 88. Vekemans X, Beauwens T, Lemaire M, Roldan-Ruiz I (2002) Data from amplified fragment length polymorphism (AFLP) markers show indication of size homoplasy and a relationship between degree of homoplasy and fragment size. Mol Ecol 11: 139–151. pmid:11903911
  89. 89. Nei M (1987) Molecular evolutionary genetics. New York: Columbia Univ Press. 512 p.
  90. 90. Winkler M, Tribsch A, Paun O, Englisch T, IntraBioDiv-Consortium, et al. (2010) Pleistocene distribution range shifts were accompanied by breeding system divergence within Hornungia alpina (Brassicaceae) in the Alps. Mol Phylogen Evol 54: 571–582. pmid:19695334
  91. 91. Ehrich D (2006) AFLPdat: a collection of R functions for convenient handling of AFLP data. Mol Ecol Notes 6: 603–604.
  92. 92. Ortiz MÁ, Tremetsberger K, Terrab A, Stuessy TF, García-Castaño JL, et al. (2008) Phylogeography of the invasive weed Hypochaeris radicata (Asteraceae): from Moroccan origin to worldwide introduced populations. Mol Ecol 17: 3654–3667. pmid:18662226
  93. 93. Slovák M, Kučera J, Turis P, Zozomová-Lihová J (2012) Multiple glacial refugia and postglacial colonization routes inferred for a woodland geophyte, Cyclamen purpurascens: patterns concordant with the Pleistocene history of broadleaved and coniferous tree species. Biol J Linn Soc 105: 741–760.
  94. 94. Bryant D, Moulton V (2004) Neighbor-Net: An agglomerative method for the construction of phylogenetic networks. Mol Biol Evol 21: 255–265. pmid:14660700
  95. 95. Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23: 254–267. pmid:16221896
  96. 96. Hennig C, Hausdorf B (2012) The prabclus package version 2.2–4, Department of Statistical Science, University College London, London. Available: http://cran.r-project.org/web/packages/prabclus/. Accessed 4 August 2014.
  97. 97. Hausdorf B, Hennig C (2010) Species delimitation using dominant and co-dominant multi-locus markers. Syst Biol 59: 491–503. pmid:20693311
  98. 98. Kruskal JB (1964) Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika 29: 1–27.
  99. 99. Felsenstein J (1989) PHYLIP-phylogeny inference package (version 3.2). Cladistics 5: 164–166.
  100. 100. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959. pmid:10835412
  101. 101. Falush D, Stephens M, Pritchard JK (2007) Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol Ecol Notes 7: 574–578. pmid:18784791
  102. 102. Falush D, Stephens M, Pritchard JK (2003) Inference of population structure using multilocus genotype data: linked loci and correlated allele frequencies. Genetics 164: 1567–1587. pmid:12930761
  103. 103. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14: 2611–2620. pmid:15969739
  104. 104. Nordborg M, Hu TT, Ishino Y, Jhaveri J, Toomajian C, et al. (2005) The pattern of polymorphism in Arabidopsis thaliana. PLoS Biol 3: e196. pmid:15907155
  105. 105. Rosenberg NA (2004) DISTRUCT: a program for the graphical display of population structure. Mol Ecol Notes 4: 137–138.
  106. 106. Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl Acids Symp Ser 41: 95–98.
  107. 107. Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate gene genealogies. Mol Ecol 9: 1657–1660. pmid:11050560
  108. 108. Ingvarsson PK, Ribstein S, Taylor DR (2003) Molecular evolution of insertions and deletions in the chloroplast genome of Silene. Mol Biol Evol 20: 1737–1740. pmid:12832644
  109. 109. Joly S, Stevens MI, van Vuuren BJ (2007) Haplotype networks can be misleading in the presence of missing data. Syst Biol 56: 857–862. pmid:17957580
  110. 110. Stamatakis A (2006) RAxML-VI-HPC: Maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22: 2688–2690 pmid:16928733
  111. 111. Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap algorithm for the RAxML web servers. Syst Biol 57: 758–771. pmid:18853362
  112. 112. Suda J, Trávníček P (2006) Estimation of relative nuclear DNA content in dehydrated plant tissues by flow cytometry. Curr Protoc Cytom 38: 7.30.1–7.30.14.
  113. 113. Fischer MA, Adler W, Oswald K (2008) Exkursionsflora von Österreich, Liechtenstein und Südtirol. 3rd ed. Linz: Land Oberösterreich, Biologiezentrum der OÖ Landesmuseen. 1391 p.
  114. 114. Badanina VA (1997) Pro vydovu samostiinist’ Doronicum stiriacum Vill. D.T. u flori Ukrayiny [About species’ independence of Doronicum stiriacum (Vill.) D.T. in the Ukrainian Flora]. Ukr Bot Zh 54: 372–375.
  115. 115. Stuessy TF (2009) Plant taxonomy. The systematic evaluation of comparative data. New York: Columbia Univ Press. 514 p.
  116. 116. Győrffy I (1909) Megjegyzések a tátrai Doronicum Clusii (All.) Tausch ismeretéhez—Bemerkungen zur Kenntnis von Doronicum Clusii (All.) Tausch aus der Hohen Tátra. Magyar Bot Lapok 8: 47–50.
  117. 117. Dixon CJ, Schönswetter P, Suda J, Wiedermann MM, Schneeweiss GM (2009) Reciprocal Pleistocene origin and postglacial range formation of an allopolyploid and its sympatric ancestors (Androsace adfinis group, Primulaceae). Mol Phylogen Evol 50: 74–83.
  118. 118. Stewart JR, Lister AM, Barnes I, Dalén L (2010) Refugia revisited: individualistic responses of species in space and time. Proc R Soc Lond B Biol Sci 277: 661–671
  119. 119. Petit RJ, Aguinagalde I, de Beaulieu JL, Bittkau C, Brewer S, et al. (2003) Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 1563–1565. pmid:12791991
  120. 120. Hewitt GM (1996) Some genetic consequences of ice ages, and their role in divergence and speciation. Biol J Linn Soc 58: 247–276.
  121. 121. Hampe A, Petit RJ (2005) Conserving biodiversity under climate change: the rear edge matters. Ecol Lett 8: 461–467. pmid:21352449
  122. 122. Środoń A (1987) Flora peryglacjalna z Sowlin koło Limanowej (Vistulian, Karpaty Zachodnie) [Periglacial flora of the Vistulian age from Sowliny near Limanowa (W. Carpathians)]. Acta Palaeobot 27: 53–70.
  123. 123. Lindner L, Dzierżek J, Marciniak B, Nitychoruk J (2003) Outline of Quaternary glaciations in the Tatra Mts.: their development, age and limits. Geol Quart 47: 269–280.
  124. 124. Makos M, Nitychoruk J (2011) Last Glacial Maximum climatic conditions in the Polish part of the High Tatra Mountains (Western Carpathians). Geol Quart 55: 253–268.
  125. 125. Ronikier M (2011) Biogeography of high-mountain plants in the Carpathians: An emerging phylogeographical perspective. Taxon 60: 373–389.
  126. 126. Alsos IG, Alm T, Normand S, Brochmann C (2009) Past and future range shifts and loss of diversity in dwarf willow (Salix herbacea L.) inferred from genetics, fossils and modelling. Glob Ecol Biogeogr 18: 223–239.
  127. 127. Lihová J, Carlsen T, Brochmann C, Marhold K (2009) Contrasting phylogeographies inferred for the two alpine sister species Cardamine resedifolia and C. alpina (Brassicaceae). J Biogeogr 36: 104–120.
  128. 128. Tribsch A, Schönswetter P, Stuessy TF (2002) Saponaria pumila (Caryophyllaceae) and the ice-age in the Eastern Alps. Amer J Bot 89: 2024–2033. pmid:21665631
  129. 129. Schönswetter P, Tribsch A, Schneeweiss GM, Niklfeld H (2003) Disjunctions in relict alpine plants: phylogeography of Androsace brevis and A. wulfeniana (Primulaceae). Bot J Linn Soc 141: 437–446.
  130. 130. López P, Tremetsberger K, Kohl G, Stuessy T (2012) Progenitor-derivative speciation in Pozoa (Apiaceae, Azorelloideae) of the southern Andes. Ann Bot 109: 351–363. pmid:22112441
  131. 131. Polatschek A (1966) Cytotaxonomische Beiträge zur Flora der Ostalpenländer II. sterr Bot Z 113: 101–147.
  132. 132. Favarger C (1991) Liens génétiques entre la flore orophile des Tatras et celle des Alpes à la lumière de quelques complexes polyploïdes. Pol Bot Stud 2: 23–38.
  133. 133. Favarger C (1950) Polyploïdie et vicariance dans la flore alpine. Arch Julius Klaus-Stift Vererbungsf 25: 472–477.
  134. 134. Huber W, Baltisberger M (1992) IOPB chromosome data 4. Int Organ Plant Biosyst Newslett 18/19: 6–8.
  135. 135. Lovka M, Sušnik F, Löve Á, Löve D (1972) Doronicum glaciale (Wulf.) Nyman. In: Löve Á, editor. IOPB chromosome number reports XXXVI. Taxon 21: 333–346.
  136. 136. Májovský J (1978) Index of chromosome numbers of Slovakian flora (Part 6). Acta Fac Rerum Nat Univ Comen Bot 26: 1–42.
  137. 137. Tasenkevich L, Vysotskaja EI, Vorobetz NK (1989) Chisla khromosom redkikh i endemichnykh vidov sosudistykh rastenii Ukrainskikh Karpat [Chromosome numbers in rare and endemic species of vascular plants from the Ukrainian Carpathians]. Bot Zh (Moscow & Leningrad) 74: 1669–1670.
  138. 138. Skalińska M (1950) Studies in chromosome numbers of Polish angiosperms. Acta Soc Bot Pol 20: 45–64.
  139. 139. Favarger C (1949) Notes de caryologie alpine. Bull Soc Neuchâtel Sci Nat 72: 15–42.
  140. 140. Favarger C, Küpfer P (1968) Contribution à la cytotaxonomie de la flore alpine des Pyrénées. Collect Bot (Barcelona) 6: 325–352.