Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier.com/locate/ympev

Evolutionary history of the Pasque-flowers (Pulsatilla, Ranunculaceae):Molecular phylogenetics, systematics and rDNA evolution

Gábor Sramkóa,b, Levente Laczkób, Polina A. Volkovac,d, Richard M. Batemane, Jelena Mlinarecf,⁎

aMTA-DE “Lendület” Evolutionary Phylogenomics Research Group, Egyetem tér 1., H-4032 Debrecen, HungarybDepartment of Botany, University of Debrecen, Egyetem tér 1., H-4032 Debrecen, Hungaryc Papanin Institute for Biology of Inland Waters RAS, 152742, Borok, Nekouz District, Yaroslavl Region, RussiadMoscow Grammar School in the South-West No. 1543, 26 Bakinskikh Komissarov Street, 3–5, 119571 Moscow, Russiae Jodrell Laboratory, Royal Botanic Gardens Kew, Richmond, Surrey TW9 3NJ, UKfDivision of Molecular Biology, Department of Biology, Faculty of Science, Horvatovac 102a, HR-10000 Zagreb, Croatia

A R T I C L E I N F O

Keywords:AridificationCentral Asian originClassificationDated phylogenyrDNASteppe development

A B S T R A C T

Pulsatilla (Anemoneae, Ranunculaceae) is sister to Anemone s.s. and contains ca 40 perennial species of con-siderable horticultural and medical importance. We sequenced 31 of those species, plus nine subspecies, twocultivars and six outgroups, for two nuclear regions (high-copy nrITS and low-copy MLH1) and three plastidregions (rbcL, accD–psaI, trnL intron) in order to generate the first comprehensive species-level phylogeny of thegenus. Phylogenetic trees were constructed using both concatenation-based (maximum likelihood and Bayesianinference) and coalescence methods. The better supported among the internal nodes were subjected to molecularclock dating and ancestral area reconstruction, and karyotypic characters identified by us using Fluorescence InSitu Hybridization were mapped across the tree. The preferred species tree from the coalescence analysis formedthe basis of a new infrageneric classification based on monophyly plus degree of divergence. The earliest di-vergent of the three subgenera, Kostyczewianae, is represented by only a single species that is morphologicallyintermediate between Anemone s.s. and ‘core’ Pulsatilla. Subgenus Pulsatilla is considerably richer in species thanits sister subgenus Preonanthus and contains three monophyletic sections. Species possessing nodding flowersand pectinately dissected leaves are phylogenetically derived compared with groups possessing erect flowers andpalmately lobed leaves. Pulsatilla separated from Anemone s.s. at ca 25Ma. Our results indicate a central Asianmountain origin of the genus and an initial diversification correlated with late Tertiary global cooling plusregional mountain uplift, aridification and consequent expansion of grasslands. The more rapid and extensivediversification within subgenus Pulsatilla began at ca 3Ma and continued throughout the Quaternary, driven notonly by major perturbations in global climate but also by well-documented polyploidy.

1. Introduction

We have reconstructed the phylogeny of Pulsatilla Mill. (Pasque-flowers), a genus now classified alongside other horticulturally im-portant genera such as Anemone, Hepatica and Clematis in tribeAnemoneae, subfamily Ranunculoideae, family Ranunculaceae(Tamura, 1993). Its ca 40 species of perennial herbs span the north-temperate, subarctic and montane zones from Europe (nine species) toNorth America (two species), but are concentrated in Asia (29 species:Aichele and Schwegleb, 1957). Most of the species prefer grasslandhabitats.

1.1. Phylogenetics and classification of tribe Anemoneae

Monophyly of tribe Anemoneae has been demonstrated by severalmolecular phylogenetic studies that differed considerably in thenumber of species included and the number and nature of genic regionssequenced (Hoot and Palmer, 1994; Hoot, 1995; Ro et al., 1997; Caiet al., 2010; Hoot et al., 2012; Cossard et al., 2016; Lehtonen et al.,2016; Wang et al., 2016; Jiang et al., 2017; Liu et al., 2018). Criticalcomparison of the resulting topologies suggests the existence of threemain clades: (i) Clematis s.l. plus Anemonclema, (ii) Anemone s.s. plusPulsatilla plus three further small monophyletic genera (Barneoudia,Knowltonia, Oreithales) that are better aggregated as a single genusKnowltonia, and (iii) Hepatica plus Anemonastrum (=Anemone subgenus

https://doi.org/10.1016/j.ympev.2019.02.015Received 7 November 2018; Received in revised form 25 January 2019; Accepted 17 February 2019

⁎ Corresponding author.E-mail address: jelena@biol.pmf.hr (J. Mlinarec).

Molecular Phylogenetics and Evolution 135 (2019) 45–61

Available online 01 March 20191055-7903/ © 2019 Elsevier Inc. All rights reserved.

T

Anemonidium). Cytogenetically, groups (i) and (ii) have x =7 whereasgroup (iii) has the apparently derived number of x =8. However, noconsensus has yet been achieved regarding the relative phylogeneticpositions of these three groups.

This phylogenetic ambiguity has major implications for classifica-tion within tribe Anemoneae, as it has become a battle-ground between‘lumpers’ favouring a broad, inclusive genus Anemone (Hoot et al.,2012) and the majority of taxonomic users, including horticulturalistsand medicinal herbalists, who understandably wish to retain suchprominent genera as Clematis, Hepatica and Pulsatilla. In fact, a utili-tarian classification consistent with monophyly can be achieved simplyby treating the x =7 clade formerly attributed to Anemone s.l. as thegenus Anemonastrum, as advocated by Mosyakin (2016). Within group(ii), Pulsatilla has been placed as sister to either Anemone s.s. (Hootet al., 2012) or Knowltonia (Jiang et al., 2017), but in neither case withstrong statistical support. Fortunately, all previous molecular phyloge-netic studies agreed regarding the unequivocal monophyly of genusPulsatilla.

1.2. Phylogenetics and classification of genus Pulsatilla

In spite of its horticultural (Grey-Wilson, 2014) and medical im-portance (Cheng et al., 2008; Xu et al., 2012; Ling et al., 2016), and theconservation interest expressed in many of its species, Pulsatilla hasreceived only limited attention from molecular phylogeneticists.Zetzsche (2004) studied 12 species of the P. alpina complex by se-quencing nrITS, a genic region also used in a study of hybridizationbetween six Far East members of the genus by Sun et al. (2014).Ronikier et al. (2008) assessed plastid sequence variation among tenEuropean Pulsatilla species to explore their phylogeography, and Wanget al. (2016) sequenced 12 Pulsatilla species but for a fundamentallyecological study. More recently, Jiang et al. (2017) included 17 Pulsa-tilla species in a broader study intended to reconstruct phylogeneticrelationships within tribe Anemoneae. Despite using nrITS plus sixplastid markers, they found only limited phylogenetic resolution withinPulsatilla.

Although several infrageneric classifications have been proposed forPulsatilla, even the three most recent (Aichele and Schwegleb, 1957;Tamura, 1995; Grey-Wilson, 2014) were based on traditional her-barium-based taxonomy with some support from studies of anatomyand chromosomes (Table 1). The only common element is the divisionof the genus – at various taxonomic ranks – into three main groups: (i)Kostyczewianae (containing only the central Asian P. kostyczewii, whichresembles Anemone s.s. in more morphological characters than do otherPulsatilla species), (ii) Preonanthus (arctic-alpine species of Eurasia andNorth America), and (iii) “core Pulsatilla” (most of the widely re-cognized species). None of the traditional infrageneric classifications ofthe genus Pulsatilla has previously been evaluated explicitly against amolecular phylogenetic topology.

1.3. Karyotypic background

Karyotypic changes such as polyploidy, dysploidy and chromosomalrearrangements are widely acknowledged as important evolutionaryforces shaping plant genomes and underpinning angiosperm diversity(Cai et al., 2006; Weiss-Schneeweiss et al., 2007; Leitch and Leitch,2008). Polyploidy can play a particularly significant role in plant spe-ciation due to evolutionary advantages attributed to the plasticity ofplant genomes and to increased genetic variability, generating in-dividuals capable of exploiting new niches (Soltis et al., 2015). Poly-ploidy is linked to numerous epigenetic/genomic changes such aschromosome rearrangements, transposable element mobilization, genesilencing and genome downsizing, a set of processes that are collec-tively termed ‘diploidization’ (e.g., Schnable et al., 2011; Ainoucheet al., 2012). Mapping of tandemly-arrayed repetitive sequences such asthe 5S and 18S-5.8S-26S (=35S) ribosomal RNA genes (rDNA) has

proven an excellent tool for tracing karyotypic changes in many plantgroups, including Anemone (Mlinarec et al., 2012a). The rDNA pheno-types are often species-specific, thereby providing valuable cytotaxo-nomic markers (Lan and Albert, 2011; Mlinarec et al., 2012b).

Although chromosome numbers have played an important role in

Table 1Overview of previous comprehensive taxonomic treatments of the genusPulsatilla.

Aichele and Schwegleb(1957)

Tamura (1995) Grey-Wilson (2014)

P. sect. Pulsatilla P. subg. Pulsatilla P. subg. PulsatillaP. subsect. Pratenses P. sect. Pulsatilla P. sect. PulsatillaP. pratensis P. subsect.

PulsatillaP. ser. Pratenses

P. montana P. ser. Pratenses P. pratensis s.l.P. rubra P. pratensis P. montana

P. montana P. rubraP. subsect. Vulgares P. rubra P. ser. PulsatillaP. vulgaris P. nigricans P. vulgarisP. grandis P. ser. Pulsatilla P. grandisP. taurica P. vulgaris P. halleriP. slavica P. grandis P. turczaninoviiP. velezensis P. halleri P. ser. BungeanaeP. styriaca P. turczaninovii P. bungeanaP. halleri P. ajanensis P. magadenensis

P. ser. Vernales P. millefoliumP. vernalis P. sukaczewii

P. subsect. Patentes P. ser. Patentes P. tenuilobaP. patens P. patens P. ser. AlbanaeP. flavescens P. flavescens P. albanaP. nuttalliana P. nuttalliana P. ambigua

P. multifida P. andinaP. subsect. Albanae P. ser. Albanae P. armenaP. albana P. albana P. campanellaP. campanella P. campanella P. wallichianaP. wallichiana P. wallichiana P. georgicaP. millefolium P. bungeana P. violaceaP. regeliana P. millefolium P. sect.

SemicampanariaP. turczaninovii P. regeliana P. ser.

SemicampanariaP. bungeana P. ser.

TatewakianaeP. cernua

P. ajanensis P. tatewakii P. chinensisP. sugawarae P. dahurica

P. subsect. Vernales P. ser. Cernuae P. tongkangensisP. vernalis P. cernua P. ser. Vernales

P. dahurica P. vernalisP. sect. Semicampanaria P. ser. Chinenses P. sect. PatentesP. subsect. Cernuae P. chinensis P. ser. PatentesP. cernua P. patens s.l.P. dahurica P. ser.

TatekawianaeP. subsect. Chinenses P. ajanensisP. chinensis P. sugawarai

P. sect. Iostemon P. tatewakiiP. subsect.Kostyczewianae

P. subsect.Kostyczewianae

P. subg.Kostyczewianae

P. kostyczewii P. kostyczewii P. kostyczewiiP. subg.

PreonanthusP. sect. Preonanthus P. sect.

PreonanthusP. sect.

PreonanthusP. subsect. Alpinae P. alpina P. alpinaP. alpina P. alba P. aureaP. alba P. aurea P. occidentalisP. aurea P. occidentalis P. scherfeliiP. occidentalis P. apiifolia

P. sect. Preonanthopsis P. sect.Preonanthopsis

P. sect.Preonanthopsis

P. subsect. Taraoianae P. taraoi P. taraoiP. taraoi P. nipponica P. nipponica

P. subg.Miyakea

P. subg. Miyakea

P. integrifolia P. integrifolia

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

46

the systematics of tribe Anemoneae, the base chromosome numberwithin Pulsatilla is uniformly x =8 (Baumberger, 1971; Tamura, 1995;Mlinarec et al., 2012b). More detailed molecular cytogenetic studies ofPulsatilla are limited to two works (Lee et al., 2005; Mlinarec et al.,2012b) that together report data for only eight species. The typicalkaryotype consists of four pairs of metacentric, one pair of submeta-centric or metacentric, and three pairs of acrocentric chromosomes,where the second pair carries a satellite. Although the majority ofPulsatilla taxa are diploid (2n=2x=16), some clades are reliablytetraploid and multiple ploidies have occasionally been observed withinspecies; for example, P. montana contains diploid, tetraploid and hex-aploid races (Baumberger, 1971; Tamura, 1995).

1.4. Present study

Here, we provide a comprehensive molecular phylogenetic frame-work for 31 of the 38 Pulsatilla species listed in the monograph of Grey-Wilson (2014) plus nine subspecies (four subspecies of P. pratensis andfour subspecies of P. patens plus one of P. halleri), based on two nuclearregions (one low copy) and three plastid regions. We also add con-siderably to the range of karyotypes available for the genus, and mapthose karyotypes across the preferred phylogeny. We use that phylo-genetic hypothesis to: (i) develop a revised supraspecific classificationof the genus using the modern principles of monophyly and relativebranch length (Table 1); (ii) estimate divergence times of the better-supported clades; (iii) infer ancestral areas of those major clades; and(iv) interpret karyotype evolution within the genus.

In order to maximize intelligibility, we employ our revised classi-fication of subgenera, sections and series (Table 1, Fig. 2) throughoutthe paper, noting that it differs from all previous classification at thetaxonomic levels of section and series (Table 1).

2. Material and methods

2.1. Taxon sampling

Efforts were made to sample the largest possible number of speciesfor the analysis (Table 2). Altogether, 51 accessions of 42 Pulsatilla taxa(species, subspecies or cultivars) were sampled as recognized in themonograph of Grey-Wilson (2014). Six species were chosen as out-groups (Anemone multifida Poir., A. sylvestris L., A. narcissiflora L., A.fasciculata L., Hepatica nobilis Mill. and H. transylvanica Fuss), guided bytopologies presented in previous, taxonomically broader molecularstudies (Schuettpelz et al., 2002; Ehrendorfer et al., 2009). Karyotypesof six Pulsatilla taxa were revisited (Lee et al., 2005; Mlinarec et al.,2012b) and a further 11 taxa were examined cytogenetically for thefirst time (Table 2). Vouchers were preserved as herbarium/spirit spe-cimens at Herbarium Croaticum of the University of Zagreb (ZA),Herbarium of the University of Debrecen (DE) and Herbarium Uni-versitatis Mosquensis of the University of Moscow (MW) (Seregin,2018) (see Supplementary Table 1).

2.2. Choice of genic regions

The DNA regions analysed were chosen for compatibility with twocontrasting tree-building approaches: concatenation of DNA informa-tion and subsequent phylogenetic analysis of the data (Wiens andCannatella, 1998) versus species-tree inference based on the multi-species coalescent (Heled and Drummond, 2010; Bryant et al., 2012).We therefore included two nucleus-encoded DNA-regions: (i) the multi-copy nuclear ribosomal internal transcribed spacer (nrITS) region(Álvarez and Wendel, 2003; Baldwin et al., 1995; Nieto-Feliner andRosselló, 2007) and (ii) the low-copy nuclear region MLH1 that encodesa DNA mismatch repair protein and was selected because of its single-copy nature documented in a wide-range of angiosperms (Zhang et al.,2012). In addition, we sequenced three plastid regions known to show

contrasting rates of molecular evolution: (iii) the plastid gene encodingthe ribulose-1,5-bisphosphate carboxylase/oxygenase large subunit(rbcL) (Palmer et al., 1988), (iv) the trnLUAA (type I) intron of the trnT-trnF region (Taberlet et al., 1991), and (v) the intergenic spacer se-parating the genes acetyl-CoA carboxylase beta subunit and the pho-tosystem I subunit VIII (accD–psaI).

2.3. DNA extraction, PCR conditions and sequencing

Genomic DNA was isolated from leaves using a modified CTABprotocol (Doyle and Doyle, 1987). In brief, the volumes were scaleddown to microliters, the CTAB-lysed, grounded tissue was treated withchloroform: isoamyl alcohol twice, and then the DNA was precipitatedwith equal volume of isopropanol and 0.08 V ammonium acetate. TheDNA pellet was washed twice with 70% ethanol, air-dried and re-dis-solved in 10mM Tris (for a detailed protocol see Sramkó et al., 2014).

The nrITS region was amplified using the primers ITS1A (Gulyáset al., 2005) and ITS4 (White et al., 1990), whereas the plastid regionswere amplified using universal plant plastid primers (Taberlet et al.,1991; Sang et al., 1997; Small et al., 1998; Shaw et al., 2005; Shawet al., 2007). All PCRs were run under the conditions as specified fornrITS and plastid regions by Sramkó et al. (2014). For the MLH1 regionwe devised new primers based on the available whole-genome sequenceof Aquilegia coerulea v.3.1 (available from the “Aquilegia coeruleaGenome Sequencing Project, http://phytozome.jgi.doe.gov/”), whichwe compared with transcripts of this gene in Ranunculus muricatus(GenBank accession number KM823030) and Anemone flaccida(GEKW01014092). Our novel primers (MLH1_520fw, 5′-TTA GAG GAGAAG CAT TGG C-3′ and MLH1_900rv, 5′-GTC TTC CTT CGA GCC GTC-3′) successfully amplified a ca 390 bp-long region of the MLH1 genelocated on the chromosome 6 of Aquilegia coerulea v.3.1 between bases673,282 and 673,669. This region includes the second exon of the gene(where the forward primer anchors), the whole third exon, the fourthexon of the gene (where the reverse primer anchors), and two intronsseparating the above exons. PCR amplification used Phusion Green HotStart II High-Fidelity DNA Polymerase (Thermo Fisher Scientific Inc.,USA) as recommended by the manufacturer (including thermal cyclingconditions and PCR mixture) but setting the annealing temperature ofthe thermal cycling at 57 °C. After checking on 1% agarose gel, suc-cessful PCR reactions were purified and sequenced, using the originalprimers as sequencing primers, by a commercially available serviceprovider (Macrogen Inc., South Korea). Our original data were sup-plemented with complete plastid and ribosomal array sequences forthree species investigated by Szczecińska and Sawicki (2015).

2.4. Phylogeny reconstruction

For each region, sequences were aligned using the online version ofMultAlign v.5.4.1 (Corpet, 1988). After concatenation, phylogeneticallyinformative regions were selected using default settings via BMGEv.1.12 (Criscuolo and Gribaldo, 2010); only the selected positions wereused in downstream analyses. We generated three matrices: (i) nuclear(nrITS+MLH1), (ii) plastid (rbcL+ accD-psaI+ trnL intron), and (iii)combined nuclear and plastid matrices. For the concatenation ap-proach, the combinability of both the plastid regions and the nuclearand plastid datasets was checked using a ‘hard incongruence’ approach(Wendel and Doyle, 1998); all matrices were used to build phylogenetictrees using the Maximum Likelihood (ML) method as described below,and the resulting trees were checked for hard incongruences that at-tracted support values exceeding 95%. Since no such incongruence wasdetected, we combined the datasets – first the plastid regions, then theplastid plus the nuclear – to increase phylogenetic resolution (e.g.,Wiens and Cannatella, 1998).

We employed two contrasting tree-building approaches. First, weused a Maximum Likelihood (ML) search based on an unpartitioneddataset as implemented in the web-based version of PhyML v.3.0

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

47

(Guindon et al., 2010) with automatic model selection (Vincent Lefort,Jean-Emmanuel Longueville and Olivier Gascuel; available at http://www.atgc-montpellier.fr/phyml-sms/). Branch robustness was testedusing the Approximate Likelihood-Ratio Test (aLRT) approach(Anisimova and Gascuel, 2006), where branch support was categorized

arbitrarily as ‘supported’ (aLRT≥ 0.75) or ‘unsupported’ (aLRT <0.75).

Our second approach used a Bayesian approach to phylogenetic treebuilding based on a partitioned dataset as implemented in MrBayesv.3.2.6 (Ronquist et al., 2012). The sequences were partitioned by their

Table 2Details of the taxa and samples analyzed. Accessions marked by an asterisk were those only used in the cytological part of the work and omitted from the phylo-genetic analyses. More details of sampling and voucher information can be found in Supplementary Table 1.

Taxon Geographic provenance Chromosome number (2n)

Pulsatilla ajanensis Regel & Tiling Russia, Magadan, Khasyn, Yablonevyi PassP. albana Bercht. & J. Presl Azerbaijan, XinaliqP. albana Bercht. & J. Presl* Germany, Göttingen (garden) 16P. alpina Schrank Austria, Kalkmagerrasen, Ötscher 16P. ambigua Turcz. ex Pritz China, Xinjiang, Yuming, Baerkelu Mts.P. andina (Rupr.) Grossh. Russia, Daghestan, GunibP. aurea (N.Busch) Juz. Russia, Krasnodar region, Sochi distr.P. bungeana C.A.Mey. ex Ledeb. Germany, Bayreuth (garden) 32P. campanella Fisch. ex Regel Kyrgyzstan, Thalaskiy AlatauP. cernua (Thunb.) Bercht. & J.Presl China, Liaoning, FushunP. chinensis (Bunge) Regel China, BeijingP. dahurica (Fisch. ex DC.) Spreng. Russia, Yakutia, KyubiuteP. dahurica (Fisch. ex DC.) Spreng. Croatia, Zagreb (garden); originally collected in Russia, "Jak" 16P. dahurica (Fisch. ex DC.) Spreng. China, Heilongjiang, TaheP. georgica Rupr. Germany, Tübingen (garden) 16P. grandis L.* Austria, Eichkogel 32P. grandis L. Slovakia, ZadielP. halleri Willd.* Germany, Potsdam (garden)P. halleri Willd. Italy, Fenestrelle, Forte Serre MarieP. halleri Willd. subsp. slavica (G.Reuss) Zämelis Slovakia, TerchováP. halleri Willd. subsp. styriaca (Pritz.) Zämelis* Germany, Potsdam (garden) 32P. subslavica Futák ex Goliašová* Germany, Potsdam (garden)P. integrifolia (Miyabe & Tatew.) Vorosch. Russia, Sakhalin Island, Mt. DikayaP. kostyczewii (Korsh.) Juz. Kyrgyzstan, Alay valleyP. magadanensis Khokhryakov & Vorosch. Russia, Magadan, Ola left shore of River OksaP. montana Rchb. Romania, Cluj-NapocaP. montana Rchb. Italy, Trieste, Monte SpaccatoP. occidentalis (S. Watson) Freyn. U.S.A., Washington, North Cascades National ParkP. patens (L.) Mill. subsp. patens Romania, RimeteaP. patens (L.) Mill. s.l. Kazakhstan, ZhaksyP. patens (L.) Mill. subsp. patens* Germany, Tübingen (garden) 16P. patens (L.) Mill. subsp. flavescens (Zucc.) Zämelis Russia, Buryatia, Altachejsky Nature ReserveP. patens (L.) Mill. subsp. multifida (Pritz.) Zämelis Kazakhstan, Pavlodar, SherbaktyP. patens (L.) Mill. subsp. nuttalliana (DC.) Grey-Wilson Canada, Yukon, Vuntut National Park, Black Fox Creek drainageP. pratensis Mill. subsp. bohemica Skalický Czech Rep., Brno 16P. pratensis Mill. subsp. hungarica Soó Hungary, Bagamér 16P. pratensis Mill. subsp. nigricans (Störck) Zämelis* Hungary, BársonyosP. pratensis Mill. subsp. nigricans (Störck) Zämelis Hungary, KozárdP. pratensis Mill. subsp. pratensis Poland, BocheniecP. rubra (Lam.) Delarbre Germany, München-Nymphenburg (garden); originally collected in France, Bresse

(garden)32

P. rubra (Lam.) Delarbre Germany, München-Nymphenburg (garden)P. sachalinensis Hara Russia, Sakhalin Island, Smirnykhovsky District, the mouth of the Belkina RiverP. sugawarai Miyabe & Tatew. Russia, Sakhalin Island, Makasovskiy districtP. sukaczewii Juz. Germany, Jena (garden) 16P. taraoi (Makino) Takeda ex Zämelis & Paegle Russia, Kuril islands, vulcano BurevestnikP. tatewakii (Kudo) Kudo Russia, Tymovsky District, near the village ZonalnoeP. tenuiloba (Hayek) Juz. Mongolia, ArbulagP. turczaninovii Krylov & Serg.* Russia, Saian Mountains (origin); seed from Göteborg 2012 16P. turczaninovii Krylov & Serg. Russia, Buryatia, Altachejsky Nature ReserveP. turczaninovii Krylov & Serg. Kazakhstan, Sherbakty (Pavlodar region)P. turczaninovii Krylov & Serg. Mongolia, TsetserlegP. vernalis Mill. Austria, OberdrauburgP. vernalis Mill. PolandP. violacea Rupr. Norway, Oslo (garden)P. vulgaris Mill. Germany, Göttingen (garden)P. vulgaris Mill. ‘coccinea’ Germany, Göttingen (garden)P. vulgaris Mill. ‘coccinea’ Germany, Göttingen (garden)P. vulgaris Mill. ‘rubra’ U.K., Kew DNA Bank (ID, 30125)P. wallichiana Ulbr. Kyrgyzstan, Chuy, Panfilov districtAnemone multifida Poir. Germany, Chemnitz (garden)Anemone narcissifolia L. s.d.Anemone fasciculata L. Russia, Northern Ossetia, TsivtaAnemone sylvestris L. Hungary, Budapest, Sváb HillHepatica nobilis Mill. Romania, Bihor, NucetHepatica transylvanica Fuss Romania, Ciuc, Frumoasa

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

48

function and genomic location: (i) nuclear gene exons (5.8S of nrITSplus exons of MLH1), (ii) nuclear spacers (ITS1 and ITS2 of nrITS) andintrons (of MLH1), (iii) plastid genes (entire rbcL, partial accD and psaI),(iv) plastid intron (trnL) and spacer (accD-psaI). We performed a prioridata partitioning and molecular evolution model selection using Par-titionFinder v.2.1.1 (Lanfear et al., 2017), and the recommended par-titions with the appropriate model settings (GTR+G for the nuclear,GTR+G+I for the plastid partition) were used in the Bayesian phylo-genetic analysis. Four independent Metropolis-Coupled Markov ChainMonte Carlo (MCMCMC) analyses were run with four (three heated)chains for 25 million generations, sampling every 1000th generationand using a heating parameter of 0.2. Convergence of the runs and theeffective sampling of the posterior space were checked using Tracerv.1.6 (available from http://beast.bio.ed.ac.uk/Tracer). A maximumclade credibility tree was drawn from the four combined runs afterdiscarding 30% of the trees as ‘burn-in’ in TreeAnnotator v.2.3.2(Drummond and Rambaut, 2007). Posterior probability (PP) values ofeach branch were categorized as ‘supported’ (PP≥ 0.85) or 'un-supported' (PP < 0.85). All phylogenetic trees were visualized in Fig-Tree v.1.4.2 (available at http://tree.bio.ed.ac.uk/software/figtree/).

2.5. Species-tree inference with molecular dating

Whereas concatenation of sequences will produce a single tree thatbest fits the combined sequence alignment according to the phyloge-netic criterion applied, a full Bayesian multispecies coalescent methodcan in theory infer comparatively accurate species tree from evenweakly informative loci (Szöllősi et al., 2015; Xu and Yang, 2016). Ourdated species-tree of Pulsatilla species was estimated using a multi-species coalescent approach as implemented in StarBEAST2 v.0.15(Ogilvie et al., 2017), an add-on package of BEAST2 v.2.5.0 (Bouckaertet al., 2014). Two independent MCMC analyses were performed byrunning the analyses for 2 billion generations and sampling every 20kth

generation (see below). Our sample set included several accessions ofsome species (e.g., P. turczaninovii, P. dahurica: Table 2). Accessionsregarded a priori as different subspecies of conspecific accessions wereassigned to different species to test for their relationships.

All five DNA regions were treated as unlinked for the site and clockmodel but were linked for the tree model of the plastid regions. Themodel of molecular evolution most appropriate for the DNA regions wasidentified during the MCMC analysis using the Bayesian Model Testv.1.1.0 (Bouckaert and Drummond, 2017), another add-on package ofthe BEAST2 programme. Tajima's (1993) relative rate tests indicatedthat a strict clock should be applied for all DNA regions. Finally, a Birth-Death model (Feller, 1939) was applied as prior when exploring spe-ciation process. To simplify computation, three monophyletic groupsthat received especially strong support during the concatenation ana-lyses (Fig. 1) were topologically constrained during the coalescenceanalysis: (i) the monophyly of the root (i.e., point of separation ofAnemonidium from Anemone s.s.), (ii) the monophyly of genus Pulsatilla(clade I on Fig. 1), and (iii) the monophyly of section Pulsatilla (clade IIIon Fig. 1). The root and the separation of the genus Pulsatilla were time-calibrated using a secondary dating approach, taking the estimates ofWang et al. (2016) as a normally distributed prior with means de-termined at 27.63Ma (sigma 1.4) and 11.64Ma (sigma 1.2), respec-tively. The two independent StarBEAST2 runs were checked for con-vergence in Tracer v.1.6 before being combined using a 50% burn-in.This combined matrix was again checked for satisfactory effectivesample size values in Tracer, and a maximum clade credibility tree withmean heights was drawn using TreeAnnotator v.2.5 of the BEAST2package.

2.6. Ancestral area reconstruction

The reconstructed species-tree (Fig. 2) was used as the basis of an-cestral area reconstruction using Bayes-Lagrange analysis following the

Statistical Dispersal-Extinction-Cladogenesis model (S-DEC: Ree andSmith, 2008; Beaulieu et al., 2013), as implemented in RASP v.3.2 (Yuet al., 2015). This approach takes account of the topological uncertaintyin the phylogenetic tree. All Bayesian phylogenetic trees from theStarBEAST2 analysis were imported for the analysis with 50% ‘burn-in’,and the resulting species tree was used as a fully resolved phylogram todefine the consensus tree. We took 1000 random trees and allowed thereconstruction of a maximum of two ancestral states at each node.

Eight biogeographical areas were defined for the Pulsatilla species,based primarily on the distribution maps and information in Aicheleand Schwegleb (1957): A, mountains of central Asia (Himalayas, Pamir-Alay, Tian Shan, Altai); B, the Far East; C, the eastern part of the Eur-asian steppe zone (from the Altai Mountains to Manchuria); D, theCaucasus Mountains; E, the central and western part of the Eurasiansteppe zone (from the Altai Mountains to the Pannonian basin); F,European grasslands (outside the steppe zone); G, North America; H,European mountains.

2.7. Chromosome analyses

All investigated species included in the cytogenetic analysis weregrown in pots in the Botanical Garden of the University of Zagreb. Forkaryological studies, actively growing root-tip meristems were pre-treated with 0.05% colchicine (Sigma-Aldrich Chemie GmbH,Taufkirchen, Germany) in dark conditions for 4 h at room temperature,fixed in a solution of ethanol and acetic acid (3:1) for 24 h at –4 °C, andstored in 70% ethanol at –20 °C until use.

Chromosomes stained with acetocarmine were used for morpho-metric analyses. Acetocarmine staining followed a standard protocol(Mlinarec et al., 2006). Chromosomes were identified according tochromosome type, total length, arm ratio and their heterochromatinbanding pattern and Fluorescence In Situ Hybridization (FISH) signal;they were ranked in order of decreasing chromosome length. Forchromosome classification we followed the nomenclature of Levan et al.(1964). Banding pattern was inferred based on the presence of het-erochromatin blocks visible after FISH. Chromosome measurementswere performed on three to four individuals per species on at least threewell-spread chromosome plates; each preparation was made from threeto four root-tip meristems. Slides chosen for FISH analysis were re-quired to show at least ten well-spread chromosome plates.

Chromosome preparations for FISH were performed according toMlinarec et al. (2012b). The clone pTa794, containing the complete410 bp BamHI fragment of the 5S rRNA gene and the spacer region ofwheat (Gerlach and Dyer, 1980), was used as the 5S rDNA probe. Theprobe was directly labelled with Cy3-dCTP (Amersham, GE Healthcare,Little Chalfont, UK) using a nick-translation kit according to the man-ufacturer’s instructions (Roche Diagnostics GmbH, Mannheim, Ger-many). The 2.4 kb HindIII fragment of the partial 18S rDNA and ITS1from Cucurbita pepo, cloned into UC19 (Torres-Ruiz and Hemleben,1994), was used as the 35S rDNA probe. The probe was directly labelledwith Cy3-dCTP by nick-translation, as above. The hybridization mixture(40 µl) containing 50% formamide, 10% dextran sulphate, 0.6% sodiumdodecyl sulphate, 2× SSC and 2 ng µl−1 of labelled probe was dena-tured at 76 °C for 15min. Chromosome preparations were denatured at73 °C for 5min after applying the hybridization mixture. Stringentwashes were performed at 42 °C in the following solutions: 2× SSC,0.1× SSC, 2× SSC, 4× SSC/Tween (5min each).

The preparations were mounted in antifade buffer Vectashield(Vector Laboratories, UK) containing DAPI counterstain (2 µgml−1)and were stored at 4 °C. Signals were visualized and photographs cap-tured on an Olympus BX51 microscope, equipped with a highly sensi-tive Olympus DP70 digital camera. Images were merged and contrastedusing Adobe Photoshop 6.0. An average of 15 well-spread metaphaseswas analyzed for each plant.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

49

Fig. 1. Phylogenetic tree of the genus Pulsatilla based on Maximum Likelihood analysis of concatenated sequences of two nuclear regions and three plastid regions.(A) Tree presented as a phylogram with branch lengths proportional to mutation changes and drawn to scale. (B) Tree presented as a cladogram, showing branchsupport values resulting from both ML aLRT and Bayesian posterior probability analysis (hyphens denote branches gaining less than 95% posterior probability) withunsupported branches collapsed into polytomies. Identical taxon names are distinguished by indicating their geographic provenance using ISO country codes. Taxonnames follow those given in Table 2.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

50

3. Results

3.1. Phylogenetic tree reconstruction

The aligned length of raw sequence reads was 3336 bp, thoughBMGE subsequently removed 49 sites as phylogenetically less reliable.This matrix was submitted to PhyML-SMS as unpartitioned dataset, andSmart Model Selection identified the GTR+G model of sequenceevolution as best fitting this dataset under both the AIC and BIC in-formation criteria. For the Bayesian analysis using MrBayes,PartitionFinder2 identified as optimal a tripartite partitioning of data:(1) merged coding nuclear and plastid regions (i.e., the nuclear exonsand plastid genes), (2) non-coding nuclear regions (i.e., the ITS regionsand MLH1 introns) and (3) non-coding plastid regions (i.e., the plastidintron and the spacer). For the non-coding nuclear partition it identifiedthe SYM+G model as best fitting, whereas for the non-coding plastidpartition and for the combined coding regions the GTR+G+ I modelwas preferred. The partitions with a unique model of evolution (i.e.,nuclear exons and plastid gene with non-coding plastid regions treatedas one unit, non-coding nuclear regions as a second unit) were used asthe molecular evolutionary model underlying the Bayesian MCMCanalysis.

We have chosen to present the maximum likelihood tree as both aphylogram (Fig. 1A) and a cladogram (Fig. 1B) because 35% of theinternal branches yielded an aLRT of < 0.75, including the short butimportant branches that specify relationships among the six series andthree sections within subgenus Pulsatilla. The phylogram helpfullyshows differential branch lengths, whereas in the cladogram the 'un-supported' branches are collapsed and other short branches are artifi-cially lengthened to allow the addition of support values for the ML andBayesian analyses.

Most of the branches that receive statistical support in the ML treealso occur in the species-coalescence tree (Fig. 2), and most attractbroadly similar levels of support. The topologies differ in the placementof section Semicampanaria, which is nested within section Pulsatilla inFig. 1 but sister to section Pulsatilla in Fig. 2 (in both cases withoutstatistical support). Positions of the two earliest-divergent specieswithin section Albanae are transposed in the two trees with strongstatistical support in both. Within section Pulsatilla, series Pulsatilla andseries Patentes can be tentatively distinguished only in Fig. 2, andprecise relationships among species of series Patentes differ between thetwo trees, again without statistical support. Among those species thatparticipate in polytomies in Fig. 1A, P. chinensis is placed within seriesSemicampanaria in Fig. 2 and P. magadanensis within section

Fig. 2. Species-tree constructed using a multispecies coalescent approach, derived from a StarBEAST2 analysis based on data for two nuclear and three plastidregions. The tree is drawn as a chronogram that relies upon on secondary dating of both the root and the separation of the genus Pulsatilla from Anemone s.s.Estimated dates (Ma) and error bars are shown only for nodes where statistical robustness exceeds 0.95. Below is graphed global climate change from the Oligoceneto Present, as represented by estimated deep-sea mean annual temperatures (red line); also shown are the periods of aridification of central Asia and progressive upliftof central Asian mountains (adapted from Favre et al., 2015). (For interpretation of the references to colour in this figure legend, the reader is referred to the webversion of this article.)

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

51

Preonanthus, though both putative relationships are unsupported.We have chosen to focus our Discussion on the species-coalescence

tree (Fig. 2) for three reasons, none of them wholly objective: theconceptual appeal of coalescence processes in molecular evolution (e.g.,Szöllősi et al., 2015), the resolved nature of the resulting tree that fa-cilitates supraspecific classification, and the degree of congruence ofthe resulting DNA-based classification with previous morphology-basedclassifications (Table 1) – a topic discussed at greater length in section4.1.

Whereas the monophyly of the genus Pulsatilla is strongly supportedin all analyses (albeit constrained in the coalescence analysis: cf. Figs. 1and 2), support for most nodes along the backbone of the highlyasymmetric tree is much weaker. The division of Pulsatilla into threemain clades (subgenera) of radically different size is strongly statisti-cally supported. The earliest-diverging lineage is the monospecificsubgenus Kostyczewianae. The less species-rich of the two remainingsubgenera, Preonanthus, includes P. magadanensis with only limitedsupport, and its division into two series-level sister groups, Pre-onanthopsis and Preonanthus, is equally tentative.

Within the species-rich subgenus Pulsatilla, section Tatewakianeae isearliest-divergent and its circumscription is strongly supported. Bothsection Pulsatilla and section Semicampanaria are resolved as mono-phyletic sisters in Fig. 2, but without statistical support. Of the twoapparently monophyletic series within section Pulsatilla, series Pulsatilla

receives statistical support but series Patentes does not. Three series areresolved within section Semicampanaria, though the relationships be-tween them are not. Series Pratenses and series Albanae receive statis-tical support, but the perceived monophyly of series Semicampanaria isweakened by the uncertain placements of P. cernua and especially P.chinensis.

The often poorly-resolved relationships between species and sub-species within apparently monophyletic series are discussed in section 4only where they are relevant to larger scale questions. Multiple, geo-graphically disparate accessions of five species were included in the MLtree (Fig. 1), in part to test the reliability of our analytical methods.With the exception of P. patens s.l., all conspecific plants yielded near-identical concatenated sequences. The three accessions of P. turczani-novii and three accessions P. dahurica resolved as statistically supportedmonophyletic groups. The one accession of P. vulgaris ‘rubra’ partici-pated in a polytomy with other taxa of series Pulsatilla. The two ac-cessions of P. montana associated with other named taxa that are oftenviewed as subspecies of P. pratensis s.l. However, the last pair of sup-posedly conspecific accessions, attributed to P. patens, were resolved assisters to other subspecies of P. patens s.l., albeit without statisticalsupport and within the same series, Patentes.

Fig. 3. Reconstruction of ancestral areas in the genus Pulsatilla reconstructed using Bayes-Lagrange analysis. Each analyzed species is coded with its contemporarydistribution (inserted key), and distribution likelihoods are shown for the consecutively numbered internal nodes. Derived states inferred by the software packageRASP during the analysis are also included.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

52

3.2. Molecular dating of lineage divergences

Given that estimates of lineage divergence dates incur wide marginsfor error, and that many of the more distal branches in our trees arevery short, we have chosen to focus our interpretation on the bettersupported branches in the chronogram that we derived from our speciescoalescence analysis (Fig. 2).

The root of the chronogram was assessed in our analysis at 28.1 Ma,close to the 27.6Ma input as prior. However, our estimate for thecrown-node of genus Pulsatilla of 9.8Ma was somewhat younger thanthe input prior of 11.6 Ma. Pulsatilla is estimated to have separated fromAnemone s.s. in the late Oligocene (24.6 Ma), whereas the separation ofHepatica from Anemonidium (the second of two distinct clades that wereuntil recently attributed to Anemone) probably took place slightly later,in the early Miocene (21.5 Ma). Both events approximate the lateOligocene warming phase. The three subgenera diverged during thelate Miocene, that of the monotypic subgenus Kostyczewianae (9.8Ma)preceding that of the two remaining subgenera (7.2Ma).

The only statistically valid date obtainable within subgenusPreonanthus is the early Quaternary (1.3Ma) separation of theCaucasian P. aurea from the Alpine P. alpina. The onset of diversifica-tion within subgenus Pulsatilla coincides with the onset of Quaternaryclimatic oscillations (3.0 Ma), marked by the separation of the newly-promoted section Tatewakianae. Sections Pulsatilla and Semicampanariaseparated a little later (2.2 Ma), immediately presaging the main ra-diation within genus Pulsatilla that affected both of these sections.Statistically valid dates were obtained for two further events: diversi-fication within the European series Pulsatilla is estimated to have oc-curred at ca 310 ka and within series Albanae, the Caucasian species (P.albana, P. andina, P. georgica, P. violacea) separated from the centralAsian montane species (P. campanella, P. wallichiana) at ca 510 ka(Fig. 2).

3.3. Ancestral area reconstruction

As is often the case, our ancestral area reconstructions are

P. patens P. georgicaP. alpina

P. pratensis subsp. bohemica P. turczNaninovii

P. grandis

P. vulgaris P. halleri subsp. styriaca P. vulgaris ‘coccinea’

P. bungeana

A C

E F

G H I

J

B P. tenuiloba

D

K Fig. 4. FISH localization of 35S rDNA (red signals) in (A) Pulsatilla alpina DE-Soo-45645, (B) P. patens ZA41504, (C) P. georgica ZA41501, (D) P. tenuiloba ZA41500,(E) P. turczaninovii ZA41498, (F) P. pratensis subsp. bohemica ZA43780, (G) P. halleri subsp. styriaca ZA41497, (H) P. vulgaris ZA41490, (I) P. vulgaris 'coccinea'ZA41491, (J) P. grandis ZA40579 and (K) P. bungeana ZA41499. Bar= 10 µm.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

53

disappointingly ambiguous toward the base of the chronogram (Fig. 3).The S-DEC method of ancestral trait reconstruction identified the cen-tral Asian mountains and the Far East (AB, 17%) or the central Asianmountains alone (A, 14%) as the most plausible among several poten-tial ancestral areas for the root at the base of genus Pulsatilla (node 83).The most likely ancestral area of the common ancestor of the two re-maining subgenera is the Far East (B, 37%: node 82). Indeed, most ofthe deeper nodes (78–82) resolve the Far East as the most probablereconstructed ancestral area, indicating the importance of this region inthe evolution of the genus.

A Far Eastern origin appears most likely for subgenus Preonanthus,but the only statistically supported node within the subgenus (77)marks the split between the Caucasian and European alpine species ofthe section Preonanthus and suggests that their common ancestorprobably occurred in both areas.

The common ancestor of section Pulsatilla is estimated to have oc-cupied the Far-Eastern and eastern Eurasian steppe zone or the Far East(node 76). Section Tatewakianae (node 44) is confined to the Far East.The common ancestor of all remaining species of subgenus Pulsatilla(node 75) most likely lived in the eastern Eurasian steppe zone.

The ancestor of section Pulsatilla (node 74) most likely occupiedEurasian steppes, though series Pulsatilla (node 67) probably originatedin European grasslands. Branch support is weak in sectionSemicampanaria (node 62) but suggests an origin in the Europeanmountains before spreading through the eastern Eurasian steppes be-fore diverging into the montane regions of the Caucasus and centralAsian mountains (node 51).

3.4. Phylogenetic mapping of ribosomal DNA patterns

The genus Pulsatilla exhibits considerable cytogenetic diversity thatcorrelates well with our tree topologies. The genus contains a mixtureof diploids (2n=2x=16) and tetraploids (2n=4x=32).

No members of subgenus Kostyczewianae or of subgenus Preonanthussection Preonanthopsis were examined. Section Preonanthus (Table 1)was represented only by P. alpina, which has two terminal 35S rDNAloci (Fig. 4A) and two centromeric 5S rDNA loci (Fig. 5A). It can bedistinguished from other Pulsatilla species by the absence of 5S rDNAsignals on acrocentric chromosomes, and is the only investigated di-ploid Pulsatilla having 5S rDNA signals on a submetacentric chromo-some pair – a position otherwise observed only in polyploids of seriesPulsatilla.

Series Patentes was represented by P. patens subsp. patens, whichshowed four 35S rDNA signals located terminally on the short arm ofacrocentric chromosomes (Fig. 4B) and two 5S rDNA signals borneterminally on longer arm of acrocentric chromosome pair (Fig. 5B).

Four of the eight species comprising series Albanae were in-vestigated cytogenetically: P. georgica, P. tenuiloba, P. violacea and P.albana (see also Mlinarec et al., 2012a). All investigated species arediploids (Figs. 4C, 5C) and resemble species of P. patens s.l. in thenumber and position of rDNA loci. Four terminal 35S rDNA signals aretypical (Fig. 4C, D), though P. albana has only two (Mlinarec et al.,2012a); also, heteromorphism is evident in P. georgica (Fig. 4C). 5SrDNA patterns consistently show two telomeric signals on acrocentricchromosomes (Fig. 5C, D).

Series Semicampanaria (represented by P. chinensis, P. dahurica andP. cernua in Fig. 6) consists of diploid species (Lee et al., 2005) with fourterminal 35S rDNA signals positioned on short arm of acrocentricchromosomes. Three or four 5S rDNA signals are located close to thecentromere within the long arms of metacentric chromosome pair 4 andin the terminal regions within long arms of acrocentric chromosomepair 7 (Lee et al., 2005).

Series Pratenses consists of diploids characterized by the presence oftwo terminal 35S rDNA loci (Fig. 4E, F) and three 5S rDNA loci, bothterminal and centromeric (Fig. 5E, F). Pulsatilla turczaninovii can bedistinguished from other diploid species by the presence of telomeric 5S

rDNA signals on metacentric chromosome pair 4 (Fig. 5E), and P. pra-tensis subsp. bohemica is unique in possessing four 5S rDNA-bearingacrocentric chromosomes (Fig. 5F).

All species of series Pulsatilla are tetraploid (Tamura, 1995). Mostmembers of the group show eight major and from three to seven minor35S rDNA signals (Fig. 4G–J), though P. halleri subsp. styriaca has onlyfive major 35S rDNA signals, one being hemizygous (Fig. 4G). Themajor signals are telomeric and located on acrocentric chromosomes,whereas the minor signals are centromeric and positioned on meta-centric chromosomes (Fig. 4G–J). 5S rDNA signal number ranges fromeight to ten, the last two being hemizygous (Fig. 5G–J). In P. coccinea,5S rDNA signals are located close to the centromere on the longer armof two of four metacentric chromosomes 1, close to the chromosometermini on the longer arm of one of four metacentric chromosomes 2and 4, on the shorter arm of two of four submetacentric chromosomes5, and close to the chromosome termini on the longer arm of all fouracrocentric chromosomes 7 (Fig. 5I).

Series Albanae was represented only by P. bungeana, which iscomposed of two chromosome sets distinguishable by size, morphologyand rDNA pattern. Eight telomeric 35S rDNA signals loci were locatedon the short arm of acrocentric chromosomes 6 and 7, the signals onchromosome 7 being highly heteromorphic (Fig. 4K). The seven 5SrDNA signals (one hemizygous) are located in the proximity of thecentromere on the longer arm of one of four metacentric chromosomes1, close to the chromosome termini on the longer arm of two of fourmetacentric chromosomes 4, and close to the chromosome termini onthe long arm of all four acrocentric chromosomes 7 (Fig. 5K).

Most Pulsatilla taxa investigated in this study showed large hetero-chromatic DAPI-positive AT-rich blocks positioned on the longer arm ofan acrocentric chromosome (Fig. 6). The exceptions are P. turczaninovii,which has additional DAPI bend located interstitially on longer arm ofmetacentric chromosome pair, and P. georgica, which lacked DAPIblocks (Fig. 6).

4. Discussion

4.1. Comparison of our classification with previous classifications

Table 1 compares our DNA-based classification following mono-phyly with those of three previous classifications. Here, we will focuson the two most recent classifications, presented in the morphology-based monographs of Tamura (1995) and Grey-Wilson (2014).

The most radical contrast is that both Tamura (1995) and Grey-Wilson recognized P. integrifolia as separate monotypic subgenus,Miyakea, whereas our tree firmly placed this species within series Pa-tentes of subgenus Pulsatilla, where it is the closely similar sister speciesto P. vernalis – a species that was awarded its own series within sectionPulsatilla by Tamura (1995) and was misplaced within section Semi-campanaria by Grey-Wilson (2014). Within section Pulsatilla, Grey-Wilson's ‘series Bungeanae’ proved to be polyphyletic, the majority of itsspecies actually being distributed throughout our expanded circum-scription of series Albanae (Fig. 2). The statistically significant separa-tion of P. bungeana and P. tenuiloba from the rest of series Albanae(Fig. 1) may indicate a more profound separation (as suggested by theearlier classifications that treat these species as a separate series Bun-geanae; Table 1), but the currently applied molecular technique andtaxon sampling does not support the series-level separation of Bun-geanae. In a more radical move, the comparatively poorly known P.magadanensis is here transferred from Grey-Wilson's series Bungeanae ofsection Pulsatilla into section Preonanthus (a decision previously ad-vocated by Luferov, 2004). Whereas the exact placement of P. maga-danensis within the subgenus Preonanthus is currently equivocal, itsbasal position is corroborated by some important morphological char-acters: this species shares the lack of nectariferous staminodes withother species in the subgenera Preonanthus and Kostyczewianae; addi-tional plesiomorphic characters are the greenish-yellow anthers with

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

54

grey stripes (all other species have yellow anthers) and five-sepalledflowers (all others have six-sepalled corolla), making this species mor-phologically exceptional similar to the basal species, P. kostyczewii.Pulsatilla rubra was placed by both Tamura (1995) and Grey-Wilson(2014) in section Pulsatilla series Pratenses but here is placed in seriesPulsatilla, whereas P. turczaninovii makes the converse taxonomicjourney from series Pulsatilla to series Pratenses. We elevate series Ta-tewakianae of Tamura (1995) and Grey-Wilson (2014) to the level ofsection, and confirm that it includes P. ajanensis, a species previouslyplaced in series Pulsatilla by Tamura (1995). Our results also tentativelysupport Grey-Wilson's (2014) decision to place P. chinensis within sec-tion Semicampanaria series Semicampanaria, though we acknowledgethat a credible case remains for Tamura's (1995) recognition of a se-parate monotypic series Chinenses.

Considered in total, and given a genus of ca 40 species, a com-paratively small number of taxonomic changes have proven necessaryto render the past supraspecific classifications congruent with our ownmonophyly-based hierarchy, indicating (though not directly testing) a

considerable degree of conformity between the molecular and mor-phological signals.

4.2. Phylogenetic relationships and evolutionary history of the genusPulsatilla

Monophyly of genus Pulsatilla, as circumscribed to include theAnemone-like central Asian species P. kostyczewii, is strongly supportedby all aspects of our detailed study (Figs. 1 and 2) (see also Zetzsche,2004). Possession by P. kostyczewii of a combination of morphologicalcharacter states that are plesiomorphic (absence of nectariferous sta-minodes; presence of purple stamens resembling those of Anemone s.s.)and apomorphic (leaves divided into long-linear segments; achene awnshairy) for the genus accords with its status as an early-divergent, spe-cies-poor lineage (e.g., Crisp and Cook, 2005); along with the com-paratively long branch separating P. kostyczewii from the remainingPulsatilla species, this amply justifies its status as one of three subgenerain Pulsatilla – subgenus Kostyczewianae.

P. patens P. georgicaP. alpina

P. pratensis subsp. bohemica P. turczNaninovii

P. grandis

P. vulgaris P. halleri subsp. styriaca P. vulgaris ‘coccinea’

P. bungeana

A B C

D E F

G H I

J

P. tenuiloba

D

K Fig. 5. FISH localization of 5S rDNA (red signals) in (A) Pulsatilla alpina DE-Soo-45645, (B) P. patens ZA41504, (C) P. georgica ZA41501, (D) P. tenuiloba ZA41500, (E)P. turczaninovii ZA41498, (F) P. pratensis subsp. bohemica ZA43780, (G) P. halleri subsp. styriaca ZA41497, (H) P. vulgaris ZA41490, (I) P. vulgaris 'coccinea' ZA41491,(J) P. grandis ZA40579 and (K) P. bungeana ZA41499. Bar= 10 µm.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

55

The remaining Pulsatilla species are divided, with high statisticalconfidence, into two main clades that are also best treated as subgenera(Fig. 2). The smaller clade encompasses approximately seven speciesthat constitute subgenus Preonanthus (a taxon viewed as a genus onmorphological grounds by Szentpéteri et al., 2008) whereas the largerclade encompasses the numerous species and subspecies that constitutethe third subgenus, Pulsatilla. This subgeneric distinction evident in ourmolecular data is supported by our cytogenetic data; P. alpina, re-presenting subgenus Preonanthus, exhibited a unique rDNA pattern thatdistinguished it from all other investigated diploid Pulsatilla species.Moreover, P. alpina resembles Anemone s.s. in the chromosomal posi-tions of 5S rDNA loci (Mlinarec et al., 2012b). We will discuss firstsubgenus Preonanthus and then subgenus Pulsatilla, paying particularattention to the hierarchical levels of section and series (Fig. 2, Table 1).

The ML topology would allow recognition of three sections withinsubgenus Preonanthus, but we follow the species coalescence topology

(Fig. 2) by favouring two sections. Within section Preonanthopsis, theenigmatic P. sachalinensis (currently known from only a single locality,on Sakhalin Island, and not accepted by Grey-Wilson, 2014) is tenta-tively placed as sister to the pairing of P. taraoi (also a localized en-demic, on the Southern Kuril Islands) and P. occidentalis, a morewidespread species of North America. These relationships suggest ascenario of eastward migration of the lineage across the Far Easternislands and the Bering Strait into North America, thus allowing allo-patric speciation. This hypothesis could be tested in the future by se-quencing the putatively closely related Japanese endemic, P. nipponica.The Far Eastern P. magadanensis is placed only tentatively in sectionPreonanthus, as sister to the cold-adapted pairing of the Alpine P. alpinaand Caucasian montane P. aurea (see also Zetzsche, 2004). The commonancestor of these montane species may have lived somewhere in thelowlands between the Caucasian and European mountain ranges andbecome divided as the population migrated from the lowlands to

Fig. 6. Karyotype diversification in Pulsatilla as mapped across the coalescence species-tree presented in Fig. 2. Position of 35S and 5S rDNA genes are shown as greenand red dots, respectively; heterochromatin regions are shown as blue blocks. Additional data sources: *Lee et al. (2005), **Mlinarec et al. (2012b).

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

56

mountains during the onset of interglacial periods.Within subgenus Pulsatilla, only the earliest-diverging section,

Tatewakianae, shows substantial internal molecular differentiation. Allthree species in the group possess distinctive ternate basal leaves. Incontrast, a wide range of leaf morphologies is exhibited by species ofsection Pulsatilla, including ternate and ternate-derived palmate leaves(Grey-Wilson, 2014). Most notably, P. integrifolia – an endemic speciesof Sakhalin Island with entire leaves unique in the genus that wasformerly awarded the status of monospecific subgenus (Grey-Wilson,2014) – unexpectedly nested well within series Patentes, nicely illus-trating the basic principle of cladistics that monophyly has priority overphenetic measures based on the amount of character-state divergence(e.g., Bateman, 2009; Stuessy, 2009).

Series Patentes is especially morphologically polymorphic(Zimmermann and Miehlich-Vogel, 1962; Zimmermann, 1963). Pulsa-tilla patens s.l. has palmately divided leaves resembling those of Ane-mone s.s., rather than the apomorphic pinnately divided leaf structurethat characterizes most Pulsatilla species. However, P. patens s.l. appearspolyphyletic in our trees, two species (P. vernalis and P. integrifolia)being nested within P. patens s.l., albeit without strong statistical sup-port. However, the significant genetic structure within P. patens s.l.(Fig. 1) may validate the taxonomic view of previous monographers(Zimmermann and Miehlich-Vogel, 1962; Zimmermann, 1963) whoregarded the subspecies of P. patens s.l. as separate, allopatric species.This remarkable result demands a more detailed phylogeographic studyof the P. patens ‘aggregate’.

The inclusion of P. integrifolia and P. vernalis within series Patentes isan unexpected result as no specialist of this genus has ever consideredthe species included in this clade to be related. Nonetheless, the erectflowers and less-dissected leaf-blades connect the species of seriesPatentes.

Species of the European series Pulsatilla are also known to readilyhybridize and yield a fertile F1, particularly the well-known Europeanhorticultural species P. vulgaris, P. halleri and P. grandis (Grey-Wilson,2014). Morphological synapomorphies of the group include large, erector semi-erect flowers and achenes possessing long awns, and all speciesare tetraploid. Although P. rubra has traditionally been classified insection Semicampanaria series Pratenses, our study revealed P. rubra topossess sequences identical with those of P. vulgaris ‘coccinea’ (also P.vulgaris ‘rubra’: Fig. 1) and it was shown by Mlinarec et al. (2012b) to betetraploid, in contrast with series Pratensis which is uniformly diploid.Although our sample of P. rubra was a garden plant, we are confidentthat we did not confuse it with cultivar ‘coccinea’ or ‘rubra’; the plantshowed the required nodding flowers and finely dissected leaves, andits morphological identification was confirmed by expert A.N. Luferov(I.M. Sechenov First Moscow State Medical University). Unfortunately,the near-uniformity of DNA sequences in series Pulsatilla means that wecan neither confirm nor reject the hypothesis that P. halleri subsp. sla-vica is a stabilized hybrid between P. vulgaris and P. grandis (Mereďaand Hodálová, 2011).

Although the monophyly of section Semicampanaria is only weaklysupported by the species-tree (Fig. 2), this clade credibly unites allspecies that produce a nodding or half-nodding flower (Grey-Wilson,2014). Its three taxonomic series are also only tentatively circum-scribed.

The most surprising inclusion in series Pratenses is the central andeastern Asian species P. turczaninovii, shown by our trees to be earliestdivergent within the series. Its semi-erect or half-nodding, brightlycoloured flowers with widely spreading sepals resemble those of seriesPulsatilla. However, the diploid nature and three 5S rDNA loci of P.turczaninovii are consistent with the remainder of series Pratenses butnot with the uniformly tetraploid series Pulsatilla (Fig. 6). The poly-morphic species P. pratensis s.l. has attracted much taxonomic attentionfrom European taxonomists (Aichele and Schwegleb, 1957;Zimmermann, 1958; Szentpéteri et al., 2007). Our analysis failed todistinguish among the four subspecies of P. pratensis, nor could it

separate P. pratensis s.l. from P. montana (one sample of which wastaken from the locus classicus in Italy), suggesting that these taxa may beconspecific with geographic separation at the subspecies level.

Series Semicampanaria contains P. cernua, P. dahurica and, moretentatively, P. chinensis. These species share simple pinnate leaves withP. vernalis (Grey-Wilson, 2014), but our results suggest that this mor-phological similarity is a parallelism rather than a synapomorphy thatindicates shared ancestry. Jiang et al. (2017) sampled several plants ofeach of these three species for their multigene phylogeny but wereunable to demonstrate the monophyly of any of the three.

The final taxonomic series of Pulsatilla to be considered is seriesAlbanae of section Semicampanaria. This clade of Caucasian and centralAsian high mountain species has as its earliest-divergent species first P.tenuiloba and then P. bungeana, both from the Eastern Eurasian steppes(see also Aichele and Schwegleb, 1957). Pulsatilla sukaczewii and P.tenuiloba, typical of the northeast region of Eurasia (Siberia, Mongoliaand China), were previously regarded as closer to P. bungeana, which isthe only small-bodied, semi-erect flowered species in the series. Pulsa-tilla sukaczewii from the Baikal region of Russia was found to be nearlyidentical to P. georgica in our analysis (Fig. 1). Grey-Wilson (2014) ar-gued that P. sukaczewii is commonly confused with the Caucasian P.georgica in European gardens. Given that our plant of ‘P. sukaczewii’ isone of the few samples we were obliged to obtain from living collec-tions (Table 2), we suspect it was mislabelled. The strongly supportedspecies-pair of the west Himalayan P. wallichiana and the Pamir–Tian-Shan–east Hindu Kush endemic P. campanella (both central Asianmembers of the Albanae group, along with P. ambigua) clearly showsthat a vicariance event induced speciation.

4.3. Karyotype differentiation and evolution

In every investigated Pulsatilla species, whether diploid or poly-ploid, all of the 35S rDNA loci were separated from the 5S rDNA geneclusters and located terminally on the short arm of acrocentric chro-mosomes, whereas there was no preferential chromosomal position for5S rDNA loci; they occupy terminal, subterminal or pericentromericpositions, and can be located on acrocentric, submetacentric or meta-centric chromosomes (Fig. 6). Clearly, 5S rDNA is far more mobile than35S rDNA in Pulsatilla, reflecting a more general pattern that has re-cently been reported across land plants (Garcia et al., 2017).

Mapping of rDNA characters across a topology congruent with thecoalescence species-tree for Pulsatilla permitted evolutionary inter-pretation (Fig. 6). The majority of the species examined, including theearliest-divergent species analyzed (P. alpina of subgenus Preonanthus),possessed two 35S and two 5S rDNA loci, which therefore representsthe most likely plesiomorphic condition. Deletion of 35S rDNA sites wasobserved only in the diploid P. albana and tetraploid P. halleri subsp.styriaca, which had two and five 35S rDNA signals, respectively. Fre-quent duplication or deletion of 5S rDNA loci, leading to large-scalepolymorphism of numbers, sizes and physical positions of signals,characterized the investigated members of series Pulsatilla. Deletion of5S rDNA signals was observed in P. patens and the clade consisting of P.albana, P. georgica and P. violacea, all of which have only one 5S rDNAlocus, as well as in P. cernua, which has three 5S rDNA loci. The absenceof signal from one of the homologues in P. cernua may be consequenceof sequence elimination due to unequal crossing-over, often related to aprocess of concerted evolution of tandemly repeated genes or to theactivities of repetitive DNA such as transposable elements. By contrast,diversification of P. turczaninovii and P. pratensis subsp. bohemica wasprobably accompanied by the duplication of one 5S rDNA locus, whichwas subsequently translocated to a metacentric chromosome in P.turczaninovii and to an acrocentric chromosome in P. pratensis subsp.bohemica.

Polyploidization can precipitate numerous epigenetic and genomicchanges such as chromosome rearrangements, gene silencing andgenome downsizing, together with transposable element mobilization

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

57

(Schnable et al., 2011; Ainouche et al., 2012). There is also significantevidence in the literature of rDNA loci reduction following poly-ploidization (Clarkson et al., 2005; Dvořák, 1990; Vaio et al., 2005;Weiss-Schneeweiss et al., 2007). Hemizygous 5S rDNA loci were ob-served in the majority of the investigated polyploid species of Pulsatilla,most likely occurring as a consequence of reduction of 5S rDNA duringdiploidization, though this process has occurred to differing extentsamong the polyploid taxa. Furthermore, Pulsatilla polyploids exhibitadditional 35S rDNA signals of smaller size and/or lower intensity thatare located in the centromeric/pericentromeric regions of the chro-mosomes, which are enriched in transposable elements (Mlinarec et al.,2016). Interestingly, dispersed 35S rDNA signals were not found inPulsatilla diploids. We postulate that transposable element mobilizationwas induced by polyploidization and that the seeding of rDNA repeatsto ectopic locations in the genome was transposon-mediated. A sig-nificant fraction of the rDNA units in animals are interrupted bytransposable elements that are highly specialized for insertion intoconserved sites within the rDNA genes (Zhang et al., 2008); recentstudies suggested that they may encourage rDNA mobility (Costa et al.,2013).

The origin of polyploid Pulsatilla species has not previously beeninvestigated. The existence in P. grandis, P. vulgaris, P. rubra and P.halleri subsp. styriaca of two diploid chromosome sets distinguishableby size, morphology and rDNA FISH pattern indicates that seriesPulsatilla had a hybrid, allopolyploid origin. Although chromosomemarkers such as rDNA cannot solely reveal the precise origin of poly-ploids, they can nonetheless reveal some interesting patterns (Mlinarecet al., 2012a; Jang et al., 2018). The occurrence and localization of the5S rDNA on a particular chromosome pair can enable the identificationof markers of the two parental species involved in the formation ofhybrid taxa. Pulsatilla polyploids have 5S rDNA signals positionedsubterminally on the longer arm of metacentric chromosomes – a po-sition was found only in P. turczaninovii among the investigated di-ploids. Furthermore, members of series Pulsatilla have 5S rDNA on theshorter arm of submetacentric chromosomes, and a 5S rDNA-bearingsubmetacentric chromosome pair occurred only in P. alpina among theinvestigated diploids. Therefore, based on the phylogenetic distributionof the rDNA FISH pattern, we hypothesize that the members of seriesPulsatilla have a single allopolyploid origin from P. turczaninovii and P.alpina, presumably on the eastern Eurasian steppes (Fig. 3). Confirma-tion of this hypothesis would require application of more informativemethods such as GISH (Mlinarec et al., 2012a).

4.4. Evolutionary history

Polyploidization has frequently been implicated in both diversifi-cation and the colonization of new habitats (Fawcett and Van de Peer,2010). Current mainstream theory states that polyploids are oftenabundant in harsh and unstable environments, and they often occupynew niches in which their diploid progenitors would be less likely topersist (Fawcett and Van de Peer, 2010). Also, recent studies proposedthat the comparative success of certain ancient polyploids might belinked to periods of climatic change. Thus, the colonization of Europeby the polyploid members of series Pulsatilla may have been facilitatedby the periods of drastic climatic change that characterized the Pleis-tocene.

According to our dated phylogeny (Fig. 2) and ancestral-state re-construction (Fig. 3), the ancestral area of the genus Pulsatilla mostlikely lies in the central Asian mountains, a region previously identifiedas an important area of origin for xeric grassland species in severalother families of flowering plants (Zhang et al., 2014; Zhang et al.,2015; Zhang et al., 2016; Zhang et al., 2017). Diversification of thegenus clearly coincided with the prolonged uplift of the northern cen-tral Asian mountain ranges (Pamir-Alay, Tian Shan). Uplift becamemore pronounced in the late Miocene (Favre et al., 2015; Spicer, 2017),and is suspected of having contributed to the significant aridification of

central Asia (Miao et al., 2012).We hypothesize that the ancestor of the genus Pulsatilla – pre-

sumably a member of Anemone s.s. – lived in the central Asian region, inthe generally more humid and warm Miocene evergreen or semi-ever-green forests (Spicer, 2017), but with the progressive uplift of thePamir-Alay and Tian Shan ranges it adapted to the ensuing colder andforest-free mountainous conditions. The subsequent global coolingperiod (Fig. 2) is likely to have facilitated diversification of a high-mountain lineage represented today by subgenus Preonanthus thatsubsequently spread to other high mountain ranges of the NorthernHemisphere, including the Rocky Mountains of North America. The FarEast then became a secondary centre of speciation; the gradual globalcooling and aridification of central Asia allowed them to colonize thegrasslands of lowland Eurasia, which expanded toward the close of theMiocene (Tang et al., 2011).

The formation of extensive grasslands during the Pliocene and –more importantly – during the Quaternary (Guthrie, 2001) had a dra-matic effect on the putative evolutionary radiation of the species; theydiversified into several lineages that migrated through grasslands fromthe Far East via central Asia to Europe, eventually reaching NorthAmerica on two occasions. The major climatic oscillations that char-acterized the subsequent Quaternary period might also have heavilyinfluenced speciation in this genus of predominantly grassland species,inducing repeated contraction and expansion of grasslands during theglacial-interglacial cycles (Frenzel et al., 1992).

5. Conclusions

(1) Our multi-gene phylogeny allowed reclassification within genusPulsatilla; the resulting classification into three subgenera, five sectionsand eight series broadly resembles the most recent classifications ori-ginating from traditional morphological methods but differs in severalimportant respects. Although morphological differentiation is limitedand an objective morphological analysis is still lacking, pinnately dis-sected leaves and nodding flowers are identified as apomorphic ten-dencies within the genus.

(2) Mapping of karyotypes across the poorly resolved sectionsPulsatilla and Semicampanaria reveals phylogenetically correlated var-iation in the number and distribution of rDNA loci that is of value forboth circumscribing and determining the relationships of species.Particular polyploid patterns constitute apomorphies for some sectionsand series, though at least one putative species reputedly includes bothdiploid and tetraploid plants.

(3) Reconstructing divergence dates and phylogeographic patternsof internal nodes of our coalescent species-tree suggest that speciesdiversity remains greatest in the region of origin of the genus in centralAsia, potentially linked to the progressive uplift of the Pamir-Alay andTian-Shan Mountains.

(4) Subsequent diversification of this predominantly grasslandgenus is hypothesized to have been driven by the well-documentedglobal cooling of temperature and growing aridification during lateMiocene, which is likely to have encouraged radical expansion ofgrassland habitats in Eurasia.

(5) More extensive diversification, arguably constituting an evolu-tionary radiation, occurred during the Quaternary, presumably trig-gered by the environmental perturbations and consequent migrationsthat characterize glacial-interglacial cycles and further encouraged byrepeated polyploidization.

(6) Although we have presented the most data-rich and tax-onomically comprehensive phylogeny of the genus to date, our resultsmake clear that next-generation sequencing techniques (NGS: pre-ferably focusing on the nuclear genome) will be necessary in order tobetter resolve relationships at taxonomic levels up to and includingseries, ideally combined with a morphological cladistic analysis rich inmicromorphological characters. NGS will need to be combined withmuch more intensive sampling and morphometric surveys to achieve

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

58

robust circumscription of species and subspecies, especially in the fourseries that contain multiple taxa that have proven molecularly near-identical in our study (Fig. 2).

6. Taxonomic treatment

In light of our results, we provide here a revised classification forthe genus. Asterisked taxa were not included in our study so theirclassification below should be regarded as provisional; one putativespecies suffixed with a hashtag was not recognized by Grey-Wilson(2014):

Genus Pulsatilla Mill.

Subgenus Kostyczewianae (Aichele & Schweg.) Grey-Wilsonspecies: P. kostyczewii (Korosh.) Juz.

Subgenus Preonanthus (DC.) Juz.Section Preonanthus (DC.) Spach

species: P. alpina Schrank, P. alba Rchb.*, P. aurea (N.Busch) Juz., P.magadanensis Khokhryakov & Vorosch, P. scherfelii (Ullep.) Skalický

Section Preonanthopsis Zämelis & Paeglespecies: P. taraoi (Makino) Takeda ex Zämelis & Paegle, P. nipponica(Takeda) Ohwi*, P. occidentalis Freyn, P. sachalinensis Hara#

Subgenus PulsatillaSection Tatewakianae Tamura

species: P. tatewakii Kudô, P. sugawarai Miyabe & Tatew., P. ajanensisRegel & Tiling

Section PulsatillaSeries Patentes (Aichele & Schweg.) Juz. ex. Tamura

species: P. integrifolia (Miyabe & Tatew.) Vorosch., P. patens (L.) Mill. [incl.subsp. flavescens (Zucc.) Zämelis, subsp. multifida (E.Pritz.) Zämelis, subsp.nuttalliana (DC.) Grey-Wilson],P. vernalis Mill.

Series Pulsatillaspecies: P. vulgaris Mill. [incl. cultivar. ‘coccinea’, cultivar. ‘rubra’], P. grandis

L., P. halleri Willd. [incl. subsp. slavica (Reuss) Zämelis, subsp. styriaca (Pritz.)Simonk.], P. rubra (Lam.) DelarbreSection Semicampanaria Zämelis & PaegleSeries Pratenses (Aichele & Schweg.) Juz. ex Tamura

species: P. pratensis (L.) Mill. [incl. subsp. bohemica Skalicky, subsp. hunga-rica

Soó, subsp. nigricans (Störck) Zämelis], P. montana Rchb., P. turczaninoviiKrylov & Serg.

Series Semicampanaria Grey-Wilsonspecies: P. cernua (Thunb.) Bercht. & J.Presl, P. chinensis (Bunge) Regel,P. dahurica (Fisch. ex DC.) Spreng., P. tongkangensis Y.N. Lee & T.C.Lee*

Series Albanae (Aichele & Schweg.) Juz. ex Tamuraspecies: P. albana Bercht. & J. Presl, P. andina (Rupr.) Grossh., P. georgicaRupr., P. violacea Rupr., P. ambigua Turcz. ex Pritz, P. campanella Fisch. ExRegel, P. wallichiana Ulbr., P. tenuiloba (Hayek) Juz., P. sukaczewii Juz.*, P.bungeana C.A.Mey. ex Ledeb., P. armena Rupr.*, P. millefolium Ulbr.*

Acknowledgements

We are grateful to colleagues who contributed samples: BruceBennett, Yanping Guo, Papp László, Aaron Liston, Darko Mihelj, OlgaMochalova, Richard Olmstead, Evgeniy Nikolin, Chu-Ze Shen, AlexeyShipunov, Attila Takács, Alexander Taran and Olga Vokhmina. Wethank Alexander Luferov and Susann Nilsson for identifying some of thesamples analysed by us. We also thank Volker Debus, Andriy Korolyuk,Olga Mochalova, Susann Nilsson, Zsolt Raksányi and Valentin Yakubovfor providing images of particular species. Part of the material wascollected on the territory We thank the staff of Altachejskiy NatureReserve and mountain botanical garden of the Dagestan scientificcentre RAS, and the Sakhalin Botanical Garden of RAS for permittingplant collection on their respective territories. This study was supportedby the International Association for Plant Taxonomy, Slovakia and USA;the Hungarian Scientific Research Fund [OTKA PD109686], Hungary;the Bolyai János Scholarship of the Hungarian Academy of Sciences[BO/00001/15], Hungary; the Russian Fund for Basic Research [RFBR

# 15-29-02486], Russia and an institutional project financed by theUniversity of Zagreb, Croatia.

Author contribution

GS designed and carried out sampling, performed laboratory work,designed and performed phylogenetic analyses, wrote the manuscript;LL performed laboratory work, designed and performed phylogeneticanalyses; PAV carried out sampling and wrote the manuscript; RMBperformed full revision and restructuring of the manuscript; JM ob-tained samples, designed and performed cytogenetic analyses, per-formed laboratory work, wrote the manuscript.

Conflict of interest

The authors declare no conflicts of interest.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ympev.2019.02.015.

References

Aichele, D., Schwegleb, H.W., 1957. Die Taxonomie der Gattung Pulsatilla. Repertoriumnovarum specierum regni vegetabilis 60, 1–230. https://doi.org/10.1002/fedr.19570600103.

Ainouche, M., Chelaifa, H., Ferreira, J., Bellot, S., Ainouche, A., Salmon, A., 2012.Polyploid evolution in Spartina: Dealing with highly redundant hybrid genomes. In:Soltis, P.S., Soltis, D.E. (Eds.), Polyploidy and Genome Evolution. Springer BerlinHeidelberg, Berlin, Heidelberg, pp. 225–243. https://doi.org/10.1007/978-3-642-31442-1_12.

Álvarez, I., Wendel, J.F., 2003. Ribosomal ITS sequences and plant phylogenetic in-ference. Mol. Phylogenet. Evol. 29, 417–434. https://doi.org/10.1016/S1055-7903(03)00208-2.

Anisimova, M., Gascuel, O., 2006. Approximate Likelihood-Ratio test for branches: a fast,accurate, and powerful alternative. Syst. Biol. 55, 539–552. https://doi.org/10.1080/10635150600755453.

Baldwin, B.G., Sanderson, M.J., Porter, J.M., Wojciechowski, M.F., Campbell, C.S.,Donoghue, M.J., 1995. The ITS region of nuclear ribosomal DNA: a valuable source ofevidence on Angiosperm phylogeny. Ann. Missouri Bot. Gard. 82, 247–277. https://doi.org/10.2307/2399880.

Bateman, R.M., 2009. Evolutionary classification of European orchids: the crucial im-portance of maximising explicit evidence and minimising authoritarian speculation.J. Eur. Orch. 41, 243–318.

Baumberger, H., 1971. Chromosomenzahlbestimmungen und Karyotypanalysen bei denGattungen Anemone, Hepatica und Pulsatilla. Berichte der SchweizerischenBotanischen Gesellschaft 80, 17–95.

Beaulieu, J.M., Tank, D.C., Donoghue, M.J., 2013. A Southern Hemisphere origin forcampanulid angiosperms, with traces of the break-up of Gondwana. BMC Evol. Biol.13, 80. https://doi.org/10.1186/1471-2148-13-80.

Bouckaert, R., Heled, J., Kühnert, D., Vaughan, T., Wu, C.-H., Xie, D., Suchard, M.A.,Rambaut, A., Drummond, A.J., 2014. BEAST 2: a software platform for Bayesianevolutionary analysis. PLoS Comput. Biol. 10, e1003537. https://doi.org/10.1371/journal.pcbi.1003537.

Bouckaert, R.R., Drummond, A.J., 2017. bModelTest: Bayesian phylogenetic site modelaveraging and model comparison. BMC Evol. Biol. 17, 42. https://doi.org/10.1186/s12862-017-0890-6.

Bryant, D., Bouckaert, R., Felsenstein, J., Rosenberg, N.A., RoyChoudhury, A., 2012.Inferring species trees directly from biallelic genetic markers: bypassing gene trees ina full coalescent analysis. Molec. Biol. Evol. 29, 1917–1932. https://doi.org/10.1093/molbev/mss086.

Cai, Q., Zhang, D., Liu, Z.-L., Wang, X.-R., 2006. Chromosomal localization of 5S and 18SrDNA in five species of subgenus Strobus and their implications for genome evolutionof Pinus. Ann. Bot. 97, 715–722. https://doi.org/10.1093/aob/mcl030.

Cai, Y.-F., Li, S.-W., Chen, M., Jiang, M.-F., Liu, Y., Xie, Y.-F., Sun, Q., Jiang, H.-Z., Yin, N.-W., Wang, L., Zhang, R., Huang, C.-L., Lei, K., 2010. Molecular phylogeny ofRanunculaceae based on rbcL sequences. Biologia 65, 997–1003. https://doi.org/10.2478/s11756-010-0105-8.

Cheng, L., Zhang, M., Zhang, P., Song, Z., Ma, Z., Qu, H., 2008. Silver complexation andtandem mass spectrometry for differentiation of triterpenoid saponins from the rootsof Pulsatilla chinensis (Bunge) Regel. Rapid Commun. Mass Spectrom. 22, 3783–3790.https://doi.org/10.1002/rcm.3801.

Clarkson, J.J., Lim, K.Y., Kovarik, A., Chase, M.W., Knapp, S., Leitch, A.R., 2005. Long-term genome diploidization in allopolyploid Nicotiana section Repandae (Solanaceae).New Phytol. 168, 241–252. https://doi.org/10.1111/j.1469-8137.2005.01480.x.

Corpet, F., 1988. Multiple sequence alignment with hierarchical clustering. Nucl. AcidsRes. 16, 10881–10890.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

59

Cossard, G., Sannier, J., Sauquet, H., Damerval, C., de Craene, L.R., Jabbour, F., Nadot, S.,2016. Subfamilial and tribal relationships of Ranunculaceae: evidence from eightmolecular markers. Plant Syst. Evol. 302, 419–431. https://doi.org/10.1007/s00606-015-1270-6.

Costa, G.W.W.F., Cioffi, M.B., Bertollo, L.A.C., Molina, W.F., 2013. Transposable elementsin fish chromosomes: a study in the marine Cobia species. Cytogenet. Genome Res.141, 126–132. https://doi.org/10.1159/000354309.

Criscuolo, A., Gribaldo, S., 2010. BMGE (Block Mapping and Gathering with Entropy): anew software for selection of phylogenetic informative regions from multiple se-quence alignments. BMC Evol. Biol. 10, 210. https://doi.org/10.1186/1471-2148-10-210.

Crisp, M.D., Cook, L.G., 2005. Do early branching lineages signify ancestral traits? TrendsEcol. Evol. 20, 122–128. https://doi.org/10.1016/j.tree.2004.11.010.

Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities offresh leaf tissue. Phytochem. Bull. 19, 11–15.

Drummond, A., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by samplingtrees. BMC Evol. Biol. 7, 214. https://doi.org/10.1186/1471-2148-7-214.

Dvořák, J., 1990. Evolution of multigene families: the ribosomal RNA loci of wheat andrelated species. In: Brown, A.H.D., Clegg, M.T., Kahler, A.L., Weir, B.S. (Eds.), Plantpopulation genetics, breeding, and genetic resources. Sinauer Associates Inc.,Sunderland, Massachusetts, pp. 83–97.

Ehrendorfer, F., Ziman, S.N., König, C., Keener, C.S., Dutton, B.E., Tsarenko, O.N., Bulakh,E.V., Boşcaiu, M., Médail, F., Kästner, A., 2009. Taxonomic revision, phylogeneticsand transcontinental distribution of Anemone section Anemone (Ranunculaceae). Bot.J. Linn. Soc. 160, 312–354. https://doi.org/10.1111/j.1095-8339.2009.00861.x.

Favre, A., Päckert, M., Pauls, S.U., Jähnig, S.C., Uhl, D., Michalak, I., Muellner-Riehl,A.N., 2015. The role of the uplift of the Qinghai-Tibetan Plateau for the evolution ofTibetan biotas. Biol. Rev. 90, 236–253. https://doi.org/10.1111/brv.12107.

Fawcett, J.A., Van de Peer, Y., 2010. Angiosperm polyploids and their road to evolu-tionary success. Trends Evol. Biol. 2, e3. https://doi.org/10.4081/eb.2010.e3.

Feller, W., 1939. Die Grundlagen der Volterraschen Theorie des Kampfes ums Dasein inwahrscheinlichkeitstheoretischer Behandlung. Acta Biotheor. 5, 11–40. https://doi.org/10.1007/bf01602932.

Frenzel, B., Pécsi, M., Velicho, A.A. (Eds.), 1992. Atlas of paleoclimates and paleoenvir-onments of the Northern Hemisphere. Geographical Research Institute, HungarianAcademy of Sciences - Gustav Fischer Verlag, Budapest - Stuttgart.

Garcia, S., Kovařík, A., Leitch, A.R., Garnatje, T., 2017. Cytogenetic features of rRNAgenes across land plants: analysis of the Plant rDNA database. Plant J. 89,1020–1030. https://doi.org/10.1111/tpj.13442.

Gerlach, W.L., Dyer, T.A., 1980. Sequence organization of the repeating units in thenucleus of wheat which contain 5S rRNA genes. Nucl. Acids Res. 8, 4851–4865.https://doi.org/10.1093/nar/8.21.4851.

Grey-Wilson, C., 2014. Pasque-flowers. The genus Pulsatilla. The Charlotte-Louise Press,Kenninghall.

Guindon, S., Dufayard, J.-F., Lefort, V., Anisimova, M., Hordijk, W., Gascuel, O., 2010.New algorithms and methods to estimate Maximum-Likelihood phylogenies: asses-sing the performance of PhyML 3.0. Syst. Biol. 59, 307–321. https://doi.org/10.1093/sysbio/syq010.

Gulyás, G., Sramkó, G., Molnár, V.A., Rudnóy, S., Illyés, Z., Balázs, T., Bratek, Z., 2005.Nuclear ribosomal DNA ITS paralogs as evidence of recent interspecific hybridizationin the genus Ophrys (Orchidaceae). Acta Biol. Cracov. Ser. Bot. 47, 61–67.

Guthrie, R.D., 2001. Origin and causes of the mammoth steppe: a story of cloud cover,woolly mammal tooth pits, buckles, and inside-out Beringia. Quat. Sci. Rev. 20,549–574. https://doi.org/10.1016/S0277-3791(00)00099-8.

Heled, J., Drummond, A.J., 2010. Bayesian inference of species trees from multilocusdata. Molec. Biol. Evol. 27, 570–580. https://doi.org/10.1093/molbev/msp274.

Hoot, S.B., 1995. Phylogenetic relationships in Anemone (Ranunculaceae) based on DNArestriction site variation and morphology. In: Jensen, U., Kadereit, J. (Eds.),Systematics and Evolution of the Ranunculiflorae. Springer, Vienna, pp. 295–300.https://doi.org/10.1007/978-3-7091-6612-3_30.

Hoot, S.B., Palmer, J.D., 1994. Structural rearrangements, including parallel inversions,within the chloroplast genome of Anemone and related genera. J. Molec. Evol. 38,274–281. https://doi.org/10.1007/bf00176089.

Hoot, S.B., Meyer, K.M., Manning, J.C., 2012. Phylogeny and reclassification of Anemone(Ranunculaceae), with an emphasis on austral species. Syst. Bot. 37, 139–152.https://doi.org/10.1600/036364412x616729.

Jang, T.-S., Parker, J.S., Emadzade, K., Temsch, E.M., Leitch, A.R., Weiss-Schneeweiss, H.,2018. Multiple origins and nested cycles of hybridization result in high tetraploiddiversity in the monocot Prospero. Front. Plant Sci. 9. https://doi.org/10.3389/fpls.2018.00433.

Jiang, N., Zhou, Z., Yang, J.-B., Zhang, S.-D., Guan, K.-Y., Tan, Y.-H., Yu, W.-B., 2017.Phylogenetic reassessment of tribe Anemoneae (Ranunculaceae): Non-monophyly ofAnemone s.l. revealed by plastid datasets. PLoS ONE 12, e0174792. https://doi.org/10.1371/journal.pone.0174792.

Lan, T., Albert, V.A., 2011. Dynamic distribution patterns of ribosomal DNA and chro-mosomal evolution in Paphiopedilum, a lady's slipper orchid. BMC Plant Biol. 11, 126.https://doi.org/10.1186/1471-2229-11-126.

Lanfear, R., Frandsen, P.B., Wright, A.M., Senfeld, T., Calcott, B., 2017. PartitionFinder 2:new methods for selecting partitioned models of evolution for molecular and mor-phological phylogenetic analyses. Molec. Biol. Evol. 34, 772–773. https://doi.org/10.1093/molbev/msw260.

Lee, W.K., Choi, H.W., Koo, D.H., Young, S.Y., Bang, J.W., 2005. Molecular cytogeneticsof five Pulsatilla species to the 5S, 35S rDNA genes by fluorescence in situ hy-bridization. Korean J. Genet. 27, 179–185.

Lehtonen, S., Christenhusz, M.J.M., Falck, D., 2016. Sensitive phylogenetics of Clematisand its position in Ranunculaceae. Bot. J. Linn. Soc. 182, 825–867. https://doi.org/

10.1111/boj.12477.Leitch, A.R., Leitch, I.J., 2008. Genomic plasticity and the diversity of polyploid plants.

Science 320, 481–483. https://doi.org/10.1126/science.1153585.Levan, A., Fredga, K., Sandberg, A.A., 1964. Nomenclature for centromeric position on

chromosomes. Hereditas 52, 201–220. https://doi.org/10.1111/j.1601-5223.1964.tb01953.x.

Ling, Y., Lin, Z., Zha, W., Lian, T., You, S., 2016. Rapid detection and characterisation oftriterpene saponins from the root of Pulsatilla chinensis (Bunge) Regel by HPLC-ESI-QTOF-MS/MS. Phytochem. Anal. 27, 174–183. https://doi.org/10.1002/pca.2613.

Liu, H., He, J., Ding, C., Lyu, R., Pei, L., Cheng, J., Xie, L., 2018. Comparative analysis ofcomplete chloroplast genomes of Anemoclema, Anemone, Pulsatilla, and Hepatica re-vealing structural variations among genera in tribe Anemoneae (Ranunculaceae).Front. Plant Sci. 9. https://doi.org/10.3389/fpls.2018.01097.

Luferov, A.N., 2004. A taxonomic synopsis of Ranunculaceae of the Far-East of Russia [inRussian]. Turczaninowia 7, 5–84.

Mereďa, P., Hodálová, I., 2011. Cievnaté rastliny. – Vascular plants. In: Ambrós, L., Čejka,T., Černý, J., Darolová, A., Hodálová, I., Krištofík, J., Kubinská, A., Mišíková, K.,Mereďa, P.j., Šoltés, R., Šubová, D., Vidlička, Ľ. (Eds.), Atlas druhov európskehovýznamu pre územia Natura 2000 na Slovensku. – The atlas of species of europeaninterest for Natura 2000 sites in Slovakia. Slovart, Bratislava, pp. 38–119.

Miao, Y., Herrmann, M., Wu, F., Yan, X., Yang, S., 2012. What controlled Mid-LateMiocene long-term aridification in Central Asia? — global cooling or Tibetan Plateauuplift: a review. Earth Sci. Rev. 112, 155–172. https://doi.org/https://doi.org/10.1016/j.earscirev.2012.02.003.

Mlinarec, J., Franjević, D., Harapin, J., Besendorfer, V., 2016. The impact of the Tekaychromoviral elements on genome organisation and evolution of Anemone s.l.(Ranunculaceae). Plant Biol. 18, 332–347. https://doi.org/10.1111/plb.12393.

Mlinarec, J., Papeš, D.A., Besendorfer, V., 2006. Ribosomal, telomeric and hetero-chromatin sequences localization in the karyotype of Anemone hortensis. Bot. J. Linn.Soc. 150, 177–186. https://doi.org/10.1111/j.1095-8339.2006.00467.x.

Mlinarec, J., Šatović, Z., Malenica, N., Ivančić-Baće, I., Besendorfer, V., 2012a. Evolutionof the tetraploid Anemone multifida (2n = 32) and hexaploid A. baldensis (2n = 48)(Ranunculaceae) was accompanied by rDNA loci loss and intergenomic translocation:evidence for their common genome origin. Ann. Bot. 110, 703–712. https://doi.org/10.1093/aob/mcs128.

Mlinarec, J., Šatović, Z., Mihelj, D., Malenica, N., Besendorfer, V., 2012b. Cytogenetic andphylogenetic studies of diploid and polyploid members of Tribe Anemoninae(Ranunculaceae). Plant Biol. 14, 525–536. https://doi.org/10.1111/j.1438-8677.2011.00519.x.

Mosyakin, S.L., 2016. Nomenclatural notes on North American taxa of Anemonastrum andPulsatilla (Ranunculaceae), with comments on the circumscription of Anemone andrelated genera. Phytoneuron 2016–79, 1–12.

Nieto-Feliner, G., Rosselló, J.A., 2007. Better the devil you know? Guidelines for in-sightful utilization of nrDNA ITS in species-level evolutionary studies in plants. Mol.Phylogenet. Evol. 44, 911–919. https://doi.org/10.1016/j.ympev.2007.01.013.

Ogilvie, H.A., Bouckaert, R.R., Drummond, A.J., 2017. StarBEAST2 brings faster speciestree inference and accurate estimates of substitution rates. Molec. Biol. Evol. 34,2101–2114. https://doi.org/10.1093/molbev/msx126.

Palmer, J.D., Jansen, R.K., Michaels, H.J., Chase, M.W., Manhart, J.R., 1988. ChloroplastDNA variation and plant phylogeny. Ann. Missouri Bot. Gard. 75, 1180–1206.https://doi.org/10.2307/2399279.

Ree, R.H., Smith, S.A., 2008. Maximum likelihood inference of geographic range evolu-tion by dispersal, local extinction, and cladogenesis. Syst. Biol. 57, 4–14. https://doi.org/10.1080/10635150701883881.

Ro, K.-E., Keener, C.S., McPheron, B.A., 1997. Molecular phylogenetic study of theRanunculaceae: utility of the nuclear 26S ribosomal DNA in inferring intrafamilialrelationships. Mol. Phylogenet. Evol. 8, 117–127. https://doi.org/10.1006/mpev.1997.0413.

Ronikier, M., Costa, A., Aguilar, J.F., Feliner, G.N., Küpfer, P., Mirek, Z., 2008.Phylogeography of Pulsatilla vernalis (L.) Mill. (Ranunculaceae): chloroplast DNAreveals two evolutionary lineages across central Europe and Scandinavia. J. Biogeogr.35, 1650–1664. https://doi.org/10.1111/j.1365-2699.2008.01907.x.

Ronquist, F., Teslenko, M., van der Mark, P., Ayres, D.L., Darling, A., Höhna, S., Larget, B.,Liu, L., Suchard, M.A., Huelsenbeck, J.P., 2012. MrBayes 3.2: efficient Bayesianphylogenetic inference and model choice across a large model space. Syst. Biol. 61,539–542. https://doi.org/10.1093/sysbio/sys029.

Sang, T., Crawford, D., Stuessy, T., 1997. Chloroplast DNA phylogeny, reticulate evolu-tion, and biogeography of Paeonia (Paeoniaceae). Am. J. Bot. 84, 1120–1136. https://doi.org/10.2307/2446155.

Schnable, J.C., Springer, N.M., Freeling, M., 2011. Differentiation of the maize sub-genomes by genome dominance and both ancient and ongoing gene loss. PNAS 108,4069–4074. https://doi.org/10.1073/pnas.1101368108.

Schuettpelz, E., Hoot, S.B., Samuel, R., Ehrendorfer, F., 2002. Multiple origins of SouthernHemisphere Anemone (Ranunculaceae) based on plastid and nuclear sequence data.Plant Syst. Evol. 231, 143–151. https://doi.org/10.1007/s006060200016.

Seregin, A., 2018. Moscow University Herbarium (MW). Version 1.47. LomonosovMoscow State University. 10.15468/cpnhcc.

Shaw, J., Lickey, E.B., Beck, J.T., Farmer, S.B., Liu, W.S., Miller, J., Siripun, K.C., Winder,C.T., Schilling, E.E., Small, R.L., 2005. The tortoise and the hare II: Relative utility of21 noncoding chloroplast DNA sequences for phylogenetic analysis. Am. J. Bot. 92,142–166. https://doi.org/10.3732/ajb.92.1.142.

Shaw, J., Lickey, E.B., Schilling, E.E., Small, R.L., 2007. Comparison of whole chloroplastgenome sequences to choose noncoding regions for phylogenetic studies in angios-perms: The tortoise and the hare III. Am. J. Bot. 94, 275–288. https://doi.org/10.3732/ajb.94.3.275.

Small, R.L., Ryburn, J.A., Cronn, R.C., Seelanan, T., Wendel, J.F., 1998. The tortoise and

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

60

the hare: choosing between noncoding plastome and nuclear ADH sequences forphylogeny reconstruction in a recently diverged plant group. Am. J. Bot. 85,1301–1315. https://doi.org/10.2307/2446640.

Soltis, P.S., Marchant, D.B., Van de Peer, Y., Soltis, D.E., 2015. Polyploidy and genomeevolution in plants. Curr. Opin. Genet. Dev. 35, 119–125. https://doi.org/10.1016/j.gde.2015.11.003.

Spicer, R.A., 2017. Tibet, the Himalaya, Asian monsoons and biodiversity – In what waysare they related? Plant Divers. 39, 233–244. https://doi.org/10.1016/j.pld.2017.09.001.

Sramkó, G., Molnár, V.A., Hawkins, J.A., Bateman, R.M., 2014. Molecular phylogeny andevolutionary history of the Eurasiatic orchid genus Himantoglossum s.l. (Orchidaceae).Ann. Bot. 114, 1609–1626. https://doi.org/10.1093/aob/mcu179.

Stuessy, T.F., 2009. Plant Taxonomy: the Systematics and Evaluation of ComparativeData. Columbia University Press, New York.

Sun, Y.-L., Park, W.-G., Oh, H.-K., Hong, S.-K., 2014. Genetic diversity and hybridizationof Pulsatilla tongkangensis based on the nrDNA ITS region sequence. Biologia 69,24–31. https://doi.org/10.2478/s11756-013-0284-1.

Szczecińska, M., Sawicki, J., 2015. Genomic resources of three Pulsatilla species revealevolutionary hotspots, species-specific sites and variable plastid structure in the fa-mily Ranunculaceae. Int. J. Mol. Sci. 16, 22258–22279. https://doi.org/10.3390/ijms160922258.

Szentpéteri, J., Szegö, B., Szente, S., 2007. Biogeographical studies and taxonomic reviewof Pulsatilla pratensis var. flavescens. Acta Bot. Hung. 49, 163–178. https://doi.org/10.1556/ABot. 49.2007.1-2.16.

Szentpéteri, L.J., Szente, S., Szegő, B., Stranczinger, S., 2008. Morphological studies andtaxonomic review of Preonanthus (Ranunculaceae). Acta Bot. Hung. 50, 407–415.https://doi.org/10.1556/ABot. 50.2008.3-4.17.

Szöllősi, G.J., Tannier, E., Daubin, V., Boussau, B., 2015. The inference of gene trees withspecies trees. Syst. Biol. 64, e42–e62. https://doi.org/10.1093/sysbio/syu048.

Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for amplification of3 noncoding regions of chloroplast DNA. Plant Mol. Biol. 17, 1105–1109. https://doi.org/10.1007/Bf00037152.

Tajima, F., 1993. Simple methods for testing the molecular evolutionary clock hypothesis.Genetics 135, 599–607.

Tamura, M., 1993. Ranunculaceae. In: Kubitzki, K., Rohwer, J., Bittrich, V. (Eds.),Flowering Plants · Dicotyledons. Springer, Berlin Heidelberg, pp. 563–583. https://doi.org/10.1007/978-3-662-02899-5_67.

Tamura, M., 1995. II. Systematic Part. In: Hiepko, P. (Ed.), Die NatürlichenPflanzenfamilien nebst ihren Gattungen und wichtigsten Arten, inbesondere denNutzpflanzen. Band 17 a IV. Angiospermae: Ordnung Ranunculales, Fam.Ranunculaceae. Duncker & Humbolt, Berlin, pp. 223–519.

Tang, Z., Ding, Z., White, P.D., Dong, X., Ji, J., Jiang, H., Luo, P., Wang, X., 2011. LateCenozoic central Asian drying inferred from a palynological record from the northernTian Shan. Earth Planet. Sci. Lett. 302, 439–447. https://doi.org/https://doi.org/10.1016/j.epsl.2010.12.042.

Torres-Ruiz, R.A., Hemleben, V., 1994. Pattern and degree of methylation in ribosomalRNA genes of Cucurbita pepo L. Plant Mol. Biol. 26, 1167–1179. https://doi.org/10.1007/bf00040697.

Vaio, M., Speranza, P., Valls, J.F., Guerra, M., Mazzella, C., 2005. Localization of the 5Sand 45S rDNA sites and cpDNA sequence analysis in species of the Quadrifaria groupof Paspalum (Poaceae, Paniceae). Ann. Bot. 96, 191–200. https://doi.org/10.1093/aob/mci168.

Wang, W., Lin, L., Xiang, X.-G., Ortiz, R.d.C., Liu, Y., Xiang, K.-L., Yu, S.-X., Xing, Y.-W.,Chen, Z.-D., 2016. The rise of angiosperm-dominated herbaceous floras: insights fromRanunculaceae. Sci. Rep. 6, 27259. https://doi.org/10.1038/srep27259.

Weiss-Schneeweiss, H., Schneeweiss, G.M., Stuessy, T.F., Mabuchi, T., Park, J.-M., Jang,C.-G., Sun, B.-Y., 2007. Chromosomal stasis in diploids contrasts with genome re-structuring in auto- and allopolyploid taxa of Hepatica (Ranunculaceae). New Phytol.174, 669–682. https://doi.org/10.1111/j.1469-8137.2007.02019.x.

Wendel, Jonathan F., Doyle, Jeff J., 1998. Phylogenetic Incongruence: Window intoGenome History and Molecular Evolution. In: Soltis, Douglas E., Soltis, Pamela S.,Doyle, Jeff J. (Eds.), Molecular Systematics of Plants II. Springer US, Boston, MA, pp.265–296. https://doi.org/10.1007/978-1-4615-5419-6_10.

White, T.J., Bruns, T.D., Lee, S., Taylor, J.W., 1990. Amplification and direct sequencingof fungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H.,Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Applications.Academic Press, San-Diego, pp. 315–322.

Wiens, J.J., Cannatella, D., 1998. Combining data sets with different phylogenetic his-tories. Syst. Biol. 47, 568–581. https://doi.org/10.1080/106351598260581.

Xu, B., Yang, Z., 2016. Challenges in species tree estimation under the multispeciescoalescent model. Genetics 204, 1353–1368. https://doi.org/10.1534/genetics.116.190173.

Xu, Q.-M., Shu, Z., He, W.-J., Chen, L.-Y., Yang, S.-L., Yang, G., Liu, Y.-L., Li, X.-R., 2012.Antitumor activity of Pulsatilla chinensis (Bunge) Regel saponins in human liver tumor7402 cells in vitro and in vivo. Phytomedicine 19, 293–300. https://doi.org/10.1016/j.phymed.2011.08.066.

Yu, Y., Harris, A.J., Blair, C., He, X., 2015. RASP (Reconstruct Ancestral State inPhylogenies): A tool for historical biogeography. Mol. Phylogenet. Evol. 87, 46–49.https://doi.org/10.1016/j.ympev.2015.03.008.

Zetzsche, H., 2004. Die Phylogeographie des Artkomplexes Pulsatilla alpina(Ranunculaceae). Mathematisch-Naturwissenschaftlich-Technischen Fakultät derMartin-Luther-Universität Halle-Wittenberg. Martin-Luther-Universität Halle-Wittenberg, Halle (Saale), p. 161.

Zhang, M.-L., Sanderson, S.C., Sun, Y.-X., Byalt, V.V., Hao, X.-L., 2014. Tertiary montaneorigin of the Central Asian flora, evidence inferred from cpDNA sequences ofAtraphaxis (Polygonaceae). J. Integr. Plant Biol. 56, 1125–1135. https://doi.org/doi:10.1111/jipb.12226.

Zhang, M.-L., Wen, Z.-B., Fritsch, P.W., Sanderson, S.C., 2015. Spatiotemporal evolutionof Calophaca (Fabaceae) reveals multiple dispersals in central Asian mountains. PLoSONE 10, e0123228. https://doi.org/10.1371/journal.pone.0123228.

Zhang, M.-L., Xiang, X.-G., Xue, J.-J., Sanderson, S.C., Fritsch, P.W., 2016. Himalayanuplift shaped biomes in Miocene temperate Asia: evidence from leguminousCaragana. Sci. Rep. 6, 36528. https://doi.org/10.1038/srep36528.

Zhang, M.-L., Zeng, X.-Q., Sanderson, S.C., Byalt, V.V., Sukhorukov, A.P., 2017. Insightinto Central Asian flora from the Cenozoic Tianshan montane origin and radiation ofLagochilus (Lamiaceae). PLoS ONE 12, e0178389. https://doi.org/10.1371/journal.pone.0178389.

Zhang, N., Zeng, L., Shan, H., Ma, H., 2012. Highly conserved low-copy nuclear genes aseffective markers for phylogenetic analyses in angiosperms. New Phytol. 195,923–937. https://doi.org/10.1111/j.1469-8137.2012.04212.x.

Zhang, X., Eickbush, M.T., Eickbush, T.H., 2008. Role of recombination in the long-termretention of transposable elements in rRNA gene loci. Genetics 180, 1617–1626.https://doi.org/10.1534/genetics.108.093716.

Zimmermann, W., 1958. Zur Taxonomie der Gattung Pulsatilla Miller I. Repertoriumnovarum specierum regni vegetabilis 61, 94–100. https://doi.org/10.1002/fedr.19580610106.

Zimmermann, W., 1963. Pulsatilla, Werden und Wandeln einer Gattung. Schr.Naturwissen. Verbr. Naturwissen. Kennt. Wien 103, 99–122.

Zimmermann, W., Miehlich-Vogel, G., 1962. Zur Taxonomie der Gattung Pulsatilla MillerIII. Die Subsektion Patentes. Die Kulturpflantze Beiheft 3, 93–133.

G. Sramkó, et al. Molecular Phylogenetics and Evolution 135 (2019) 45–61

61