molecular evidence for multiple diversification patterns of alpine … · 2008. 6. 17. · ventoux,...

INTRODUCTIONSouthern Europe encompasses elevated mountains

that harbour alpine plants in the Mediterranean regionThe highest European mountains which are still in anuplifting process are the result of the onset of the Alpineorogeny in Mesozoic-Cenozoic times (c 65 Myr) Platetectonic activity of microplates generated main move-ments during the latest uplifting periods inMediterranean Europe Betic range (Jurassic-Miocene)Cantabrian range and the Pyrenees (Paleocene-Eocene)Jura (Miocene-Pliocene) the Alps (Oligocene-Miocene)Apennines (Paleocene-Miocene) and the BalkanPeninsula (Paleocene-Pleistocene) (Ager 1975)Important is that the present configuration of thesemountains was established far earlier than differentiationof alpine species in Europe

Distribution and genetic structure of alpine plantshave been ultimately molded by the geologic complexityof the European continent coupled with Tertiary andQuaternary climatic episodes The occurrence of theMediterranean climate in the Pliocene (c 32 Myr)caused seasonal rhythm and summer drought thatbecame stable approximately 28 Myr ago (Suc 1984)

Accordingly floras of mountain ranges may have alsoevolved during the Pliocene and Pleistocene (Stebbins1984) including surviving floras in high altitudes encir-cled by Mediterranean conditions Glacial-interglacialcycles in the Pleistocene promoted isolation of alpineplants in high-altitude mountains of the MediterraneanBasin during warm (interglacial) periods Converselyplant populations from isolated mountain ranges wereconnected during long glacial stages Most alpine plantsthat we find in southern European mountains are theresult of migrations from lower altitudes after glacialperiods or descendants of colonists from northern Europe(Stebbins 1984)

Molecular phylogenetics help infer historical bio-geography and evolutionary patterns Relationshipsamong living populations of alpine plants in a phylogeo-graphic context seems to be the most appropriateapproach to address diversification patterns of alpineplants in Europe (Avise amp al 1987 Comes amp Kadereit1998 Avise 2000) Intraspecific phylogeography is con-cerned with the principles and processes governing thegeographic distributions of genealogical lineages espe-cially those within and among closely related species(Avise 2000) This approach has been successfully

463

Vargas Alpine plants in Mediterranean Europe 52 August 2003 463ndash476

Molecular evidence for multiple diversification patterns of alpine plants inMediterranean Europe

Pablo Vargas

Real Jardiacuten Botaacutenico CSIC Plaza de Murillo 2 E-28014 Madrid Spain vargasma-rjbcsices

A preliminary synthesis of diversification patterns of alpine plants in the Mediterranean region of Europe ispresented based on seven plant groups displaying morphological differentiation and infraspecific taxa Bothprevious and new phylogenetic results from ITS sequences and fingerprinting data suggest different coloniza-tion routes and modes of speciation in Androsace vitaliana (recent differentiation in the Iberian Peninsula)Anthyllis montana (west-to-east colonization and differentiation in Europe) Arenaria tetraquetra (colonizationand differentiation from SE Iberian mountains to the Pyrenees increasing number of chromosome comple-ments) Saxifraga oppositifolia (colonization from the arctic to the Iberian Peninsula) Saxifraga pentadactylis(differentiation in Mediterranean and Eurosiberian mountains by geographic isolation) and Soldanella alpina(differentiation and colonization from northern Iberia to the Alps and then to the Pyrenees and the BalkanPeninsula) Relative static diversification of Juniperus communis var saxatilis in Europe based on identity ofchloroplast trnL-F sequences is also described Most morphological variation expressed by number of sub-species recognized in previous taxonomic treatments of the seven plant groups appears to have occurred dur-ing the Pleistocene (lt 175 Myr) Recurrent change of Quaternary climatic conditions in the MediterraneanBasin coupled with geographic characteristics life cycle dispersal mechanisms and pre-Holocene geneticstructure are not convincing factors to account for all the observed diversification Additionally stochasticprocesses are also considered for evaluating present-day distributions and processes of speciation

KEYWORDS Androsace Anthyllis Arenaria colonization patterns Juniperus molecular diversificationSaxifraga Soldanella subspecific differentiation

implemented to describe postglacial colonization routesof trees in central Europe (Demesure amp al 1996Dumolin-Lapegue amp al 1997 King amp Ferris 1998Raspeacute amp al 2000 Petit amp al 2002) and to test thehypothesis of nunatak areas for in situ glacial survival inarctic and alpine plants (Brochmann amp al 1996Gabrielsen amp al 1997 Stehlik amp al 2002 Schoumlnswet-ter amp al 2002) Quaternary glaciations may not havedramatically affected the survival of plants in southernEurope in contrast with total extinction of angiospermsin northern Europe due to long-term establishment of icesheets In fact three main refugia in the Iberian Italianand Balkan peninsulas have been proposed from whichre-colonization occurred (Taberlet amp al 1998 Hewitt2000) Genetic structure and speciation patterns of alpinefloras in southern Europe are more complex due to a con-tinuous presence of plants in Mediterranean mountainsalong fluctuating vegetation belts

This paper is a preliminary synthesis of results fromseven plant groups using molecular markers Androsacevitaliana Anthyllis montana Arenaria tetraquetraJuniperus communis var saxatilis Saxifraga oppositifo-lia Saxifraga pentadactylis Soldanella alpina (Vargas2002) Our working hypothesis is that geographic char-acteristics recent climatic events (glaciations) lifecycles dispersal mechanisms and pre-Holocene geneticstructures are mostly responsible for present distributionand differentiation within alpine plants Subspecific taxarepresent morphological differentiation of each speciesin an evolutionary context The main objective is to testthis phylogeographic hypothesis based on variation ofmolecular markers phylogenetic reconstructions andgeographic arrangement of mountains that have led topresent distributions and modes of speciation in Europe

MATERIALS AND METHODSSpecies selection mdash Alpine plants are typically

defined as those obtaining optimal habitat conditionsover 2300 m in the Mediterranean region (Koumlrner 1999)Unfortunately we have found few examples in the liter-ature of alpine vascular plants in southern Europe con-taining subspecific taxa which also have been analyzedusing phylogeographic data Seven plant groups havebeen considered in this paper thanks to previous publica-tions from several authors on Anthyllis montana (Kropfamp al 2002a) Saxifraga oppositifolia (Abbott amp al2000 Holderegger amp al 2002) and Soldanella alpina(Zhang amp al 2001) Additionally we have obtained andanalyzed new data from Androsace vitaliana Arenariatetraquetra Juniperus communis var saxatilis andSaxifraga oppositifolia and extended the sample ofSaxifraga pentadactylis (Vargas 2001) Some other stud-

ies have not been included due to insufficient number ofpopulations low morphological variation or uninforma-tive molecular results

Molecular markers mdash Molecular variation with-in the same species is the basic requirement to infer rela-tionships among populations Fingerprinting techniquesscreening the nuclear genome such as RAPDs ISSRsAFLPs and microsatellites ensure a large number ofdata to be analysed by phylogenetic methodologies(Hillis amp al 1996a) In some cases sequences of nuclearribosomal DNA (ITS ETS) display a minimal number ofphylogenetically informative characters among popula-tions Recombination of biparental nuclear genomes mayobscure historical relationships based on hierarchical ge-nealogies particularly in cases of reticulation Unipa-rental haplotypes either from the mitochondria or chloro-plast alleviate this problem because they are haploid andnon-recombinant macromolecules (Avise amp al 1987Comes amp Kadereit 1998) However low levels of genet-ic variation in organelle genomes together with phylo-genetic reconstructions based on genes inherited by onlyone parental organism (typically maternal inheritance inangiosperms Mogensen 1996) are disadvantages forthe use of these markers in phylogeography of plants andinference of morphological character evolution

Phylogeographic reconstructions mdash Tradi-tional biogeography considers two major possibilities toaccount for the origin of present distributions of popula-tions and species of alpine plants dispersal and vicari-ance (Ronquist 1997) The vicariance hypothesis arguesfor isolation into island-like areas of populations on sun-dered mountains from a continuous area separated duringinterglacial periods In contrast under dispersal distribu-tion of populations on elevated mountains is the result ofcolonization from one or few individuals We abandonany inference based on area cladograms because ourphylogenetic reconstructions are analysed at the popula-tional level and distributions of alpine plants are pro-foundly disparate Genealogies of neutral genes are usedto evaluate colonization patterns and microevolutionaryprocesses (Schaal amp Olsen 2000) The use of nuclearmarkers link morphological differentiation and phylo-geography because most genes responsible for characterevolution are contained in the nuclear genomeNeighbor-joining and parsimony-based analyses ofnuclear fingerprinting and ribosomal ITS sequences areimplemented

The internal transcribed spacer (ITS) region ofnuclear 18S-26S ribosomal DNA has proven to beinformative for both biogeographic studies and infer-ences of character evolution at inter- and intraspecificlevels (Baldwin amp al 1995) Availability of ITSsequences in GenBank is the result of ease of sequencingthe ITS region coupled with an acceptable number of


464

phylogenetically informative characters in some infra-specific studies (Baldwin 1993 Vargas amp al 1999a)Despite intrinsic problems in reconstruction of molecularphylogenies by using multi-copy genes such as the ITSregion affected by concerted evolution (Fuertes amp al1999) reliable phylogenetic signal at the infraspecificlevel has been corroborated using different molecularmarkers (Kropf amp al 2002a Zhang amp al 2001)Additionally the body of ITS sequence analyses allowcomparison across angiosperms and the use of this mol-ecule to estimate maximum age of diversificationthrough molecular clocks (Richardson amp al 2001) Toobtain ITS and trnL-trnF sequences we used a sequenc-ing strategy as described in Vargas amp al (1999a) andTaberlet amp al (1991)

Molecular clocks mdash To include a time frame formost molecular and morphological differentiation ofeach species we searched for maximum ages of diversi-fication Diversification times were calculated from bothprevious estimates (Zhang amp al 2001 Kropf amp al2002a) and new ones based on the logic of Baldwin ampSanderson (1998) and Sanderson (2002) Calibrationpoints depend on the geological events and the plantgroup (eg origin of Madeira island and split of theStrait of Gibraltar in Saxifraga Vargas amp al 1999a) andthe fossil record (Dorofeev 1963 for Primulaceae) Acalibrated rate of Saxifraga in Richardson amp al (2001) isalso considered even though it is clearly divergent incomparison to other calibration points Levels of diversi-fication are also compared using average molecular vari-ation of ITS sequences in a group of angiosperms(Richardson amp al 2001)

RESULTSAndrosace vitaliana (Primulaceae) mdash The for-

mer genus Vitaliana has been lately considered as a sec-tion (Vitaliana) of Androsace (Kress 1997) a circum-scription in agreement with an ITS phylogeny whereaccessions of Androsace vitaliana form part of the line-age of Androsace sect Aretia (L) WDJ Koch (PSchoumlnswetter pers comm) This European endemicoccurs in the Abruzzi Alps Pyrenees and elevatedmountains of the Iberian Peninsula at an altitudinalrange between 1500 and 3300 m It is characterized byfruits in capsules with 2ndash3 seeds each and two ploidallevels (2n = 20 40) The taxonomic treatment ofFerguson (1972) includes five infraspecific taxa underVitaliana primuliflora Bertol subspp assoana M Laiacutenz(S and E Iberia) canescens O Schwarz (SW Alps andPyrenees) cinerea (Suumlnd) I K Ferguson (C and SWAlps and E Pyrenees) praetutiana (Buser ex Suumlnd) I KFerguson (Apennines) and primuliflora (SE Alps) A

most recent taxonomic treatment also includes subspflosjugorum Kress (Cantabrian range) described as theother subspecies under Androsace (Kress 1997) (Table1) The ITS region is 604ndash606 bp in length 222ndash223 bpin ITS-1 164 bp in 58S 219ndash220 bp in ITS-2 (AacutelvarezLucentildeo amp Vargas unpubl) Eight variable and five par-simony-informative characters were obtained andanalysed When using Douglasia alaskana (Coville ampStandl ex Hulteacuten) S Kelso Douglasia beringensis SKelso Androsace ciliata DC A sempervivoidesJaquem ex Duby and A spinilifera as the outgroup theaccessions of A vitalia form a monophyletic group Apolytomy of three clades of A vitaliana was retrievedcontaining three samples from the Alps and MountVentoux eight accessions from the Iberian Peninsulaand one from the Apennines (Fig 1) A large polytomy isthe result of identity of ITS sequences from the entireIberian Peninsula except for two samples from theCantabrian range (subsp flosjugorum) Populations ofsubsp flosjugorum have certain independence as well asthose of subspp cinerea and vitaliana from the Alps andMount Ventoux A maximum age of 07 Myr for the splitof populations of Androsace vitaliana from Iberia is esti-mated from the highest K2P (Kimura-2-Parameter) pair-wise distance between the populations from MountVentoux and the Cantabrian range (116) The popula-tions from the Iberian Peninsula may have diverged inthe last 02 Myr as suggested by the highest K2P pair-wise distance (033) between accessions from theCantabrian range and the rest of the Iberian Peninsula

Anthyllis montana (Leguminosae) mdash Morphol-ogical differentiation is reflected by four subspecies con-sidered in the last taxonomic treatments (Cullen 1968Greuter amp al 1989) They have two ploidal levels (2n =14 28) fruits typically with one seed and occur between200 and 2700 m The ITS region is 605 bp in length 231bp in ITS-1 162 bp in 58S 212 bp in ITS-2 (Kropf ampal 2002a) The alignment of the nine accessions (Table1) rendered six variable sites of which three were parsi-mony-informative When using Hymenocarpos circinna-tus (L) Savi Anthyllis vulneria L and four species of Asect Oreanthyllis Grisebach as the outgroup the ITS treewas resolved into four lineages of A montana in a pecti-nate topology southern Spain Algeria-French Alps cen-tral Italy and Austria-NE Italy-Slovenia-Croatia (Fig 2)Average K2P pairwise distance of sequences among thefour ITS lineages was 036 plusmn 016 which implies amaximum diversification time of 025ndash066 Myr (Kropfamp al 2002a) The AFLP tree is congruent with that ofITS and provides further evidence of population discon-tinuity into western and eastern groups Interestinglypopulations from central and southern Italy belong totwo different groups however weakly supported

Arenaria tetraquetra (Caryophyllaceae) mdash


465

Six ploidal levels have been found in this plant which isendemic to elevated mountains (between 1400 and 3400m) of the Iberian Peninsula subsp amabilis (2x SierraNevada) subsp murcica (3x 4x 5x Betic range) andsubsp tetraquetra (6x Pyrenees 7x Sierra de la Pela)(Loacutepez 1990) A synthetic treatment including only twosubspecies tetraquetra and amabilis is found in Chateramp Halliday (1993) A characteristic cushion-like habitand capsules with seeds adapted to dispersal by rain

drops (Goyder 1987) seem to be mechanisms not favor-able for long-distance dispersal The ITS region is 633 bpin length 258 bp in ITS-1 159 bp in 58S 216 bp in ITS-2 (Valcaacutercel Nieto amp Vargas unpubl) We found a highnumber of variable sites (18) and parsimony-informativecharacters (9) in nine populations sampled across therange of the species Phylogenetic analyses were per-formed using three sequences of A alfacarensis (endem-ic to SE Iberian Peninsula) as the outgroup which is the


466

Table 1 European subspecific taxa and accessions for the seven species investigated including locality of wild popu-lations vouchers molecular markers literature references and GenBank accession numbers

Molecular GenBank Taxon Locality marker accession no Reference

Androsace vitaliana (L) Lapeyr Aacutelvarez amp al unpublssp assoana (M Laiacutenz) Kress Spain Sistema Central (2 populations) ITSssp cinerea (Suumlnderm) Kress France Mount Ventoux ITSssp flosjugorum Kress Spain Cantabrian range Montes Aquilianos ITSssp nevadensis (Chiarugi) Kress Spain Sierra Nevada ITSssp praetutiana (Buser) Kress Italy Apennines ITSssp vitaliana Spain Pyrenees (3 pops) Switzerland Alps ITS

Anthyllis montana L Kropf amp al 2002assp atropurpurea (Vuk) Pignatti Italy Friuli ITS AJ315504

ssp hispanica (Degen amp Hervier) Cullen Spain Sierra de Baza ITS AJ315508

ssp jacquini (AKerner) Hayek Austria Geissberg Mt Croatia Velebit ITS AJ315503 AJ315501Mts Slovenia Nanos Mt AJ315502

ssp montana Algeria Massif du Djurdjura France Haute- ITS AJ315509 AJ315507Savoie Hautes-Alpes Italy Abruzzi AJ315506 AJ315505

Arenaria tetraquetra L Valcaacutercel amp al un-publ

ssp amabilis (Bory) H Lindb fil Spain Sierra Nevada (3 populations) ITSssp murcica (Font Quer) Favarger amp Spain Betic range (3 populations) ITS

Nieto Felssp tetraquetra Spain Sierra de la Pela Pyrenees (2 pops) ITS

Juniperus communis Lssp communis var communis Spain Burgos trnL-trnF AY354286ssp communis var saxatilis Pall France Hautes-Pyrenees Iceland Stoumlgn trnL-trnF AY354288 AY354289

Italy Scanno Spain Aacutevila Granada Jaeacuten AY354290 AY354291Madrid Switzerland Alps Turkey Caucasus Mts AY354287 AY354292

AY354293 AY254294AY354295

Saxifraga oppositifoliassp oppositifolia Austria Salzburg Iceland Reykjavik Italy ITS AY354303

Dolomites Norway Oppland Spain Cantabrian AY354304 + AY354306range (Cantabria Palencia) Sierra Nevada AY354396 AY354302(Veleta Peak) Pyrenees (Huesca) Sierra de AY354298 AY354301Urbioacuten (La Rioja) AY354297 AY298906

ssp paradoxa D A Webb Spain S Pyrenees (Sierra del Cadiacute) N Pyrenees ITS AY354299 AY354300(Aran Valley)

Saxifraga pentadactylis Lapyr AY354307 AJ133031 Vargas 2001AY354308

ssp almanzorii P Vargas Spain Sistema Central (Sierra de Gredos) ITS AJ133032 + AJ133026ssp pentadactylis Andorra Pyrenees (El Serrat) ITS AJ233862ssp willkommiana (Boiss Spain Cantabrian range (Pentildea Prieta) Sistema ITSex Willk) M Laiacutenz Central (Sierra de Guadarrama) Sistema

Ibeacuterico (Sierra de Urbioacuten)Soldanella alpina L Zhang amp al 2001

ssp alpina Austria Carinthia Spain Basque Country ITS AJ306323 AJ306322 Switzerland Waadt Yugoslavia Montenegro AJ306321 AJ306324

ssp cantabrica Kress Spain Cantabria ITS AJ306325

most closely related species in the phylogeny of A sectPlinthine (Reichenb) Pau As a result a pectinate topol-ogy of the ITS tree was obtained (Fig 3) where popula-tions from SE Iberia (Sierra Nevada and Betic range) arebasal and northern accessions (Sierra de la Pela andPyrenees) are at terminal nodes Populations of subspmurcica are polyphyletic accessions of tetraquetra are

not fully resolved and all accessions of subsp amabilisform a monophyletic group Considering molecular-clock evolution of ITS sequences and using variationaverage within angiosperms (Richardson amp al 2001)ITS sequence variation (0ndash16 K2P pairwise distance)is high enough to hypothesize a maximum age lt 192Myr for present distribution and increment of chromo-


467

Fig 1 Distribution (A) and phylogeographic hypothesis (B) for Androsace vitaliana (C subsp vitaliana SpainPyrenees Bonaigua pass) based on ITS sequences (Aacutelvarez Lucentildeo amp Vargas unpubl) Hypothetical barrier repre-sented by a hatched bar

Fig 2 Distribution (A) and phylogeographic hypothesis (B) for Anthyllis montana (C subsp montana Spain PyreneesSallent de Gallego) based on ITS sequences and AFLP variation Hypothetical migration routes (arrows) inferred fromKropf amp al (2002a)

some complements to 7xJuniperus communis (Cupressaceae) mdash In

Europe two taxa have recently been recognized withinthis species var communis and var saxatilis (Farjon2001) However a different taxonomic treatment consid-ered three subspecies (Franco 1993) subspp communishemisphaerica (J amp C Presl) Nyman and alpina (Suter)Celak This juniper occurs primarily between 1900 and2600 m in southern Europe has a uniform chromosomenumber (2n = 22) and displays blue fleshy cones Sixchlorotypes based on nucleotide substitutions and largeindels in trnL-trnF sequences have been found within aMediterranean juniper (J oxycedrus L Martiacutenez ampVargas 2002) This species also belongs to sectJuniperus and contains four subspecies In contrast trnL-F sequences of nine European populations of J commu-nis var saxatilis (Table 1) display no molecular varia-tion Identical trnL-trnF sequences have also beenobtained when sampling former infraspecific taxa con-sidered within J communis (subsp communis var com-munis subsp hemisphaerica) Distribution of Juniperuscommunis var saxatilis in central and southern Europeand the sample of nine populations (numbered from 1 to9) are shown in Figure 4

Saxifraga oppositifolia (Saxifragaceae) mdashSynthetic taxonomic treatments of this species can befound in Webb amp Gornall (1989) and Webb (1993)Seven subspecies are considered of which five occur inEurope (subspp blepharophylla oppositifolia para-doxa rudolphiana speciosa) The species occursbetween 1700 and 3000 m in southern Europe The most

common chromosome number is 2n = 26 but 2n = 52 hasalso been reported from arctic areas In Saxifraga fruitsare typically capsules with numerous seeds (Webb ampGornall 1989) We have sequenced nine populations ofsubsp oppositifolia and two of subsp paradoxa (Table1) The ITS region is 659 bp in length 277 bp in ITS-1163 bp in 58S 219 bp in ITS-2 Eleven nucleotide sub-stitutions and five parsimony-informative characterswere found Chromatograms with nucleotide peaks over-lapping at the same sites were observed in six popula-tions after upstream and downstream sequencing reac-tions IUPAC symbols were used for the analyses Thesemistrict consensus ITS tree of 140 minimum-lengthFitch parsimony trees yielded four major clades (Fig 5)when using S spathularis and S aizoides as the out-group The first one supports monophyly of S oppositi-folia (95 bootstrap) a support value similar to thatobtained when including four more species from sectPorphyrion (Conti amp al 1999 results not shown) Abasal polytomy includes four accessions of S oppositifo-lia three of samples from southern Norway Iceland andAustrian Alps and one of the rest of the accessions (68bootstrap Fig 5) This second clade is resolved in partby a sequential branching pattern where one accessionfrom the Central Pyrenees comes first followed by agroup (third clade 72 bootstrap) of four accessions ata basal position [Dolomites Cantabrian range (Palencia)Sierra de Urbioacuten (La Rioja) and Sierra Nevada] and afourth clade (45 bootstrap) containing one sample ofsubsp oppositifolia from the Cantabrian range(Cantabria) and the two samples of subsp paradoxa


468

Fig 3 Distribution (A) andphylogeographic hypothesis(B) for Arenaria tetraquetra (Csubsp amabilis Spain SierraNevada Veleta Peak) basedon ITS sequences (ValcaacutercelNieto amp Vargas unpubl)Colonization and increment ofchromosome complements in-dicated by an arrow

from the Pyrenees (Sierra del Cadiacute and Aran Valley)Four separated populations from the Alps (Dolomites)Sierra de Urbioacuten (La Rioja) Cantabrian range (Palencia)and SE Iberia (Sierra Nevada) have the same ITSsequence (Fig 5) The highest K2P pairwise distancewithin Iberian accessions (031) is found between sam-ples from Sierra de Urbioacuten and Aran Valley Maximumage of diversification of S oppositifolia in Iberia is esti-mated between 088 and 030 Myr using different cali-bration events unspecific calibration in Saxifraga (088Myr Richardson amp al 2001) origin of Madeira (060Myr Vargas amp al 1999a) split of the Strait of Gibraltar(043 Myr Vargas amp al 1999a) estimation average inangiosperms (03 Myr Richardson amp al 2001)

Saxifraga pentadactylis (Saxifragaceae) mdashThis saxifrage is endemic to the northern half of theIberian Peninsula and occurs between 1500 and 3000 m(Vargas 1997) Neighborhood diffusion (Shigesada ampal 1995) seems to be a predominant dispersal mechan-ism of plants with capsules containing over 100 seeds asare populations of the three endemic taxa of S pen-tadactylis (Vargas amp Nieto 1996) subsp pentadactylis(Pyrenees) subsp willkommiana (central part of north-ern half of Iberia) subsp almanzorii (part of Sierra deGredos) (Fig 6) A single chromosome number has beenfound (2n = 32) An extended sample (Table 1) with twomore accessions from Sistema Central (Sierra de Urbioacuten)and Cantabrian range (Pentildea Prieta) provided figures ofITS sequence variation within the range described in aprevious publication (Vargas 2001) 682ndash683 bp inlength (281 bp in ITS-1 168 bp in 58S 233ndash234 bp inITS-2) six variable sites three parsimony-informative

characters Similar phylogenetic conditions were used asin Vargas (2001) The ITS consensus tree displays abiphyletic topology (Fig 6) with one clade weakly sup-ported (52 bootstrap) of samples from northern Iberiaand a second one of two samples from central Iberia(70 bootstrap) Although a fully resolved tree was notobtained this topology indicates that subsp willkommi-ana is not monophyletic The average K2P pairwise dis-tance between ITS accessions was 058 being thehighest distance of 121 (between Sierra de Urbioacutenand Sierra de Guadarrama) Separation times are sug-gested at a maximum between 355 and 121 Myr by theuse of different calibration events unspecific calibrationin Saxifraga (355 Myr Richardson amp al 2001) originof Madeira (242 Myr Vargas amp al 1999) split of theStrait of Gibraltar (172 Myr Vargas amp al 1999a) esti-mation average in angiosperms (121 Myr Richardson ampal 2001)

Soldanella alpina (Primulaceae) mdash Two sub-species are recognised (subspp alpina and cantabrica)based on floral shapes and sizes (Kress 1984 Zhang ampal 2001) Both of them are 2n = 40 The ITS region is644ndash645 bp in length 249 bp in ITS-1 165 bp in 58S230ndash231 bp in ITS-2 Number of nucleotide substitutionswithin Soldanella was 30 (15 parsimony-informativecharacters) and only one within S alpina A polytomyincluded all accessions in the ITS tree The AFLP recon-struction using Primula latifolia and P bulleyana as theoutgroup depicted a pectinate tree in which subspcantabrica is the basal-most lineage followed by popula-tions from the Alps Populations from the Pyrenees andthe Balkan Peninsula are intermingled among those from


469

Fig 4 Distribution (A) ofJuniperus communis varsaxatilis (B Iceland Skaf-tafel National Park) in cen-tral and southern Europe(A) The sequencing ofnine populations (num-bered from 1 to 9) renderedan identical sequence forthe trnL-trnF spacer

the Alps A clock-like evolution of ITS sequences ofSoldanella has been tested (Zhang amp al 2001) A singlenucleotide substitution between subsp alpina andcantabrica (~ 020 K2P pairwise distance) indicates arecent divergence time (lt 013 Myr) using a conserva-tive rate

DISCUSSIONRecent migration and differentiation of

Androsace vitaliana in Iberia mdash The branching pat-tern obtained from analyses of ITS sequences indicatesisolation between populations from the IberianPeninsula Alps and Apennines Two polytomies preventinferring patterns of migration within Europe and withinthe Iberian Peninsula (Fig 1) A fragmentation processseparating populations from Iberia and the Alps is inter-preted from relatively significant DNA sequence diver-gence between them (116 K2P pairwise distance) butsimilar levels occur within the two groups (033 and016 respectively) Similarly Alpine populations areisolated from that from the Apennines (066ndash083)Separation times based on the same logic as in Zhang ampal (2001) suggest split of the three lineages in the latePleistocene (lt 07 Myr) and then differentiation in theIberian Peninsula in the last 024 Myr Lack of variationin ITS sequences indicates recent isolation in four moun-tain ranges of Iberia Pyrenees Sistema Iberico Sistema

Central and Sierra Nevada (Fig 1) None of the two pre-vious taxonomic treatments are in agreement with theITS tree where populations of subspp vitaliana from theAlps and Pyrenees are not monophyletic (Fig 1)

West-to-east differentiation of Anthyllis mon-tana mdash A pectinate branching pattern of the ITS andAFLP phylogenies is described (Kropf amp al 2002a) inwhich the Iberian populations are basal followed bythose from the western southern central and easternAlps and the Balkan Peninsula These phylogeneticreconstructions suggest an eastward migration (Fig 2)Tree resolution of European populations also revealedfragmentation into western and eastern areas correspon-ding approximately with subspp hispanicamontana andjacquiniiatropurpurea respectively Character evolutionof floral traits based on morphometric and AFLP analy-ses leads to the conclusion that A montana contains onlytwo infraspecific taxa (subspp montana and jacquinii) asa result of vicariance in W-E Europe (Kropf amp al2002b) Evidence of a secondary contact was observed inthe Alps MaritimesLiguria region Estimation of a max-imum age of 700000 years indicates infraspecific diver-gence in Late Quaternary times (Kropf amp al 2002a)

South-to-north colonization and incrementof polyploidy in Arenaria tetraquetra mdash As part ofa wider investigation in Arenaria sect Plinthine(Valcaacutercel Nieto amp Vargas unpubl) preliminary data ofITS sequence evolution reveal that populations of subsptetraquetra may have colonized the Pyrenees from SE


470

Fig 5 Distribution (A) of Saxifraga oppositifolia in southern Europe Colored circles indicate the five sequences ofnuclear ribosomal ITS sequences found Semistrict consensus tree of 140 minimum-length Fitch parsimony trees (B)from analysis of ITS sequences (90 steps CI 099 RI 096 including cladistically uninformative characters) Saxifragaspathularis and S aizoides were used as the outgroup (not shown) Bootstrap values are above branches Photographof subsp oppositifolia (C) from Spain (Pyrenees Envalira pass)

Iberia (Fig 3) Differentiation northwards from SEIberian populations of subsp amabilis (2x) and subspmurcica (3x 4x 5x) is in agreement with a previoushypothesis of increment of chromosome complements(Nieto amp Favarger 1988) Assuming a clock-like evolu-tion of the ITS region this mode of speciation in aploidal series is occurring since the onset of thePleistocene (c 175 Myr) The possibility of increment ofchromosome complements via hybridisation (allopoly-ploiditation) should be cautiously contemplated (Nieto ampFavarger 1988) Phylogenetic reconstructions based onchloroplast markers (already in progress) may shed fur-ther light on modes of speciation and colonization of Atetraquetra

Static diversification in Juniperus communismdash Lack of molecular variation not only within var sax-atilis from Iceland to the Caucasus but also with respectto var communis suggests a single origin of this speciesand a relative static diversification pattern in Europe(Fig 4) Higher levels of variation have been obtained byusing AFLP data in J communis (van der Merwe amp al2000) suggesting multiple colonization patterns at alocal scale (the British Islands) Identity of trnL-Fsequences among populations separated over 3000 kmcoupled with active dispersal of fleshy cones of J com-

munis (Jordano 1993 Martiacutenez amp Vargas 2002) lead usto interpret that relatively recent colonization may haveoccurred across Europe A more variable molecularmarker is necessary to test this interpretation

Colonization of Saxifraga oppositifolia in theIberian Peninsula from the arctic mdash Survival ofan angiosperm (Saxifraga oppositifolia) in the arctic andfurther circumpolar migration from Beringia has beenrecently documented (Abbott amp al 2000) It has alsobeen hypothesized that lack of geographical genetic pat-terns in the Alps may be the result of both immigrationsafter glaciations and in situ survival (Holderegger amp al2002) Phylogeographic studies including populationsfrom Mediterranean Europe have not been addressed forthis species Our results based on ITS sequences (Fig 5)are in agreement with those based on chloroplast varia-tion in the Eurosiberian region (Abbott amp al 2000Holderegger amp al 2002) Terminal placement of ITSaccessions in a sequential branching pattern indicatesthat colonization in the Iberian Peninsula occurred fromthe north Ancient arctic lineages may have been presentacross Europe pre-dating Quaternary episodes (Abbott ampal 2000) Maximum divergence times between 088 and030 Myr is herein estimated for the Iberian populationsInterestingly identity of ITS sequences in four separated


471

Fig 6 Distribution (A) of Saxifraga pentadactylis Strict consensus tree (B) of two minimum-length Fitch parsimonytrees from analysis of ITS sequences (122 steps CI 087 RI 089) Saxifraga spathularis was used as the outgroup (notshown) as well as sequences and phylogenetic parameters from Vargas (2001) Bootstrap values over 50 are shownabove tree branches Photograph of subsp willkommiana (C) from Spain (Sierra de Guadarrama Madrid)

mountain ranges of Europe (Dolomites PyreneesCantabrian range and Sierra Nevada) indicates geneticaffinities (Fig 5) Presence both of ancient and recentlineages depicts a complex genetic structure not only inthe Eurosiberian region of Europe (Abbott amp al 2000)but also in the Mediterranean region The finding ofadditive polymorphism in ITS sequences supports thelikelihood of hybridization processes Some genotypessurviving in Iberia after Quaternary climatic changesmay have had secondary contacts as in northern Europe(Gabrielsen amp al 1997) and the Alps (Holderegger ampal 2002) Further investigations using markers unaffect-ed by concerted evolution (dissimilar to ITS sequences)are needed to evaluate the impact of secondary contactson the genetic structure of southern European popula-tions Acquisition of alternate leaves in subsp paradoxa(Pyrenees) seems to be derived from opposite leaves acharacter described for all the other subspecies (Fig 5)A larger sample of the entire subsection Oppositifoliae(already in progress by R Holderegger pers comm)may shed light on character evolution within S oppositi-folia

Differentiation of Eurosiberian andMediterranean populations of Saxifraga pen-tadactylis mdash Our phylogenetic hypothesis (Fig 6) is inagreement with previous results of genetic drift in moun-tains of the Iberian Peninsula (Vargas 2001)Populations of subsp wilkommiana do not form a natu-ral group because morphological differentiation asdescribed in two taxonomic treatments (Webb amp Gornall1989 Vargas 1997) do not reflect isolation betweenEurosiberian (Cantabrian range Pyrenees Sierra deUrbioacuten) and Mediterranean (Sierra de GuadarramaSierra de Gredos) mountains In the same mountainrange (Sistema Central) genetic isolation through pre-dominant autogamy in subsp almanzorii (Sierra de

Gredos) from populations of subsp willkommiana hasalready been described (Vargas 2001) Differentiation ofleaf and flower morphologies is estimated to occur atmost in the last 355 Myr even though it is most plausi-ble at the onset of the Pleistocene (c 175 Myr)

Differentiation from relict Cantabrian pop-ulations of Soldanella alpina mdash Low levels of ITSsequence variation (one nucleotide substitution betweensubspp alpina and cantabrica) were found (Zhang amp al2001) The AFLP tree supports a basalmost placement ofsubsp cantabrica followed by populations of subspalpina in a sequential fashion where populations fromthe Pyrenees and the Balkan Peninsula are included ingroups of populations from the Alps (Fig 7) As thePyrenees are situated between the Alps and theCantabrian range two interpretations are considered forthe occurrence of S alpina in the Pyrenees (1) coloniza-tion to the Pyrenees from the Alps after long-distancedispersal from the Cantabrian range to the Alps and (2)early colonization from the Cantabrian range extinctionand re-colonization from the Alps Seed dispersalinferred by using chloroplast markers would help to testthese hypotheses particularly in this case because evi-dence of hybridisation and introgression based on mor-phological nuclear DNA and distributional patterns hasbeen documented (Zhang amp al 2001) Using a conser-vative estimated rate divergence between subspp alpinaand cantabrica may have taken place in the last 150000years coinciding with the Late Penultimate Glacial (vanAndel amp Tzedakis 1996)

To explore patterns of colonization and evolution ofalpine plants cladistic analyses based on DNA sequenceshelp infer hierarchical relationships of populationsBecause of limitations in cytotaxonomic interpretationsfrom chromosome counts Stebbins (1984) recommend-ed use of cladistics and sequences of mitochondrial


472

Fig 7 Distribution (A) andphylogeographic hypothe-sis (B) for Soldanella alpi-na (C subsp alpina SpainPyrenees Bonaigua Pass)based on ITS sequencesand AFLP variationHypothetical migrationroutes (arrows) inferredfrom Zhang amp al (2001)

chloroplast and nuclear ribosomal DNA for comparisonsbetween closely related species and races of the samespecies This author warned that no absolute rules orgeneralizations could be made on migration patterns ofarctic-alpine flora The above-discussed plant groupsoccurring in southern Europe support this prediction alsofor alpine species in Mediterranean mountains Somepopulations of the seven species studied in the presentpaper cluster together into monophyletic groups of sub-species A search for naturalness in the evolution ofalpine plants reveals that populations of Androsace vital-iana subsp flosjugorum (Fig 1) Anthyllis montanasubsp hispanica (Fig 2) Arenaria tetraquetra subspamabilis (Fig 3) and Soldanella alpina subsp alpina(Fig 7) form natural groups in the phylogenetic recon-structions Lack of resolution in some reconstructions isthe result of identity of sequences (Juniperus communis)low number of characters (Androsace vitaliana) or char-acter additivity (Saxifraga oppositifolia) In contrastsome subspecies recognized in recent classifications arenot supported by molecular evidence for monophyleticpopulations Anthyllis montana subspp montana andjacquinii (Fig 2) Arenaria tetraquetra subsp tetraque-tra and murcica (Fig 3) and Saxifraga pentadactylissubsp willkommiana (Fig 6) Island-like distribution ofalpine plants in southern Europe suggests that allopatryappears to be the main force for divergence and isolationof subspecies Infraspecific taxa are considered to beintermediate stages in the speciation process in a geo-graphic context (Stuessy 1990) and it is expected thatthe phylogenetic status will vary over time to achievemonophyly via sorting and extinction of lineages in asequence of polyphyly reg paraphyly reg monophyly(Rieseberg amp Brouillet 1994)

Time of diversification and colonization was out-lined by applying a conservative rate of ITS sequenceevolution at the highest pairwise distances in all casesAs a result we found that maximum age of morphologi-cal differentiation in southern Europe for six of the sevenstudy cases expressed by subspecific taxa is significant-ly different for each plant group Androsace vitaliana (lt066 Myr) Anthyllis montana (lt 088 Myr) Arenariatetraquetra (lt 192 Myr) Saxifraga oppositifolia (lt 137Myr) Saxifraga pentadactylis (lt 175 Myr) andSoldanella alpina (lt 015 Myr) In any case subspecifictaxa appear to have differentiated in the Pleistocene (lt175 Myr wwwiugsorgiugspubsintstratcharthtm) inmost of the estimates using different calibration pointsThe above figures are higher than time of speciation cal-culated in oceanic islands (lt 073 for silverswordsalliance Baldwin amp Sanderson 1998 lt 079 forDendroseris Sang amp al 1994) but similar to diversifi-cation rates within herbaceous species from Medi-terranean floristic regions of America (lt 15 Myr for

Sanicula Vargas amp al 1999b) In contrast speciation ofa group of seven gentians (Gentiana sect Ciminalis) hasbeen hypothesized to have occurred in the Pleistocene(Hungerer amp Kadereit 1998) Interestingly the estimatesof infraspecific variation herein presented based onmolecular data are mostly in consonance with previousestimates of speciation rates for herbs (115 Myr) basedon chromosomal evolution (Levin amp Wilson 1976)Generalization of diversification times for angiospermsherbaceous plants members of the same family or evenspecies within the same genus should be cautiously con-templated because every plant group undergoes uniquedifferentiation histories (Hillis amp al 1996b)

Diversification rates across taxa are the result ofidiosyncratic characteristics of each species extinctionof their populations and taxonomic artefacts (Barra-clough amp Nee 2001) Morphological differentiation ofthe seven species has resulted in fluctuating taxonomictreatments at the infraspecific level Infraspecific taxahave been long discussed and subjected to historicalinterpretations (Stuessy 1990) Irrespective of taxonom-ic subjectivity present populations of alpine Europeantaxa display morphological differentiation as a result ofsorting and extinction primarily in the PleistoceneFurther studies using taxonomic treatments of additionalgenera morphological differentiation and moleculardivergence rates will serve to test the hypothesis thatinfraspecific variation does not predate the Quaternaryclimatic changes in a significant number of cases

Three basic hypotheses on glacial survival are pro-posed (1) total extinction through glaciations (tabularasa) as considered for thermophilous plants in the arc-tic (Brochmann amp al 1996) and trees in central andnorthern Europe (Demesure amp al 1996 Dumolin-Lapegue amp al 1997 King amp Ferris 1998 Raspeacute amp al2000 Petit amp al 2002) (2) in situ persistence of popu-lations within glaciated regions in ice-free mountains(nunataks) (Stehlik 2000 Gugerli amp Holderegger2001) (3) glacial-induced migration to lowland and peri-pheral refugia followed by recolonization of alpine habi-tats after ice-sheet retreat (altitudinal migration)(Schoumlnswetter amp al 2002 Gutieacuterrez amp al 2002) Un-like extinction of most trees and arctic plants in centraland northern Europe under the tabula rasa hypothesispopulations of alpine species in Mediterranean moun-tains may have primarily survived by altitudinal migra-tions due to absence of an ice sheet across lowlands insouthern Europe (Van Andel amp Tzedakis 1996 Hewitt2000) The importance of re-colonization from nunatakareas should be analysed in future investigations There-fore recurrent altitudinal re-colonizations of neighbour-ing populations as found in Armeria from Sierra Nevada(Gutieacuterrez amp al 2002) and migrations from northernregions as inferred in Saxifraga oppositifolia from the


473

arctic (Holderegger amp al 2002 this paper) mayaccount in part for present distribution and geneticstructure of alpine plants in Mediterranean EuropeAdditionally colonization patterns in the four geograph-ic directions (north-to-south south-to-north east-to-west and west-to-east) are herein also documented

Multiple diversification patterns are indeed promot-ed by disparate biological characteristics (populationstructure phylogenetic relationships breeding systemdispersal syndromes and ecological requirements)together with geologic complexity of the European con-tinent recent climatic events (glaciations) and pre-Holocene genetic structure (Taberlet amp al 1998 Comesamp Kadereit 1998 Hewitt 2000) Lack of parallelism indiversification patterns among our seven plant groupswhich share similar phylogenetic relationships(Primulaceae Saxifraga spp) dispersal syndromes(neighborhood diffusion through seeds in capsules forfive cases) habitat requirements (rocky slopes) breedingsystems (predominant allogamy) and existence in thesame continent during Quaternary episodes indicatesadditional causes Most alpine species are distributedindependently of one another along mountain ranges insouthern Europe which implies per se different colo-nization and extinction histories Recent studies byTaberlet (2002) and collaborators on phylogeography ofsix alpine species sharing similar distributions and habi-tats also illustrate that phylogeographic patterns are notconcordant (see also Gugerli amp Holderegger 2001)Additional biotic and abiotic causes may be responsiblefor disparate evolutionary patterns of alpine plantsincluding random processes Stochastic dispersal inSaxifraga with a non-specific long-distance syndromeappears to have been involved in colonization of Madeirafrom the continent (c 1000 km) but isolation of Medi-terranean populations of S globulifera by the formationof the Strait of Gibraltar (separated by only c 14 km)appears to have been successful (Vargas amp al 1999a)This surprising result is in agreement with a considerablehigh number (c 29 ) of colonizers of Macaronesia dis-playing non-specific long-distance syndromes (Vargasin press) Ecological facilitation (Callaway 1995) inwhich positive interactions among plants facilitate estab-lishment and survival of species has also played animportant role in stochastic processes for colonization Inthis context it has been recently demonstrated that inter-actions among alpine plants of physically harsh environ-ments in mountains of America and Eurasia are predom-inantly positive (Callaway amp al 2002) includingJuniperus communis Saxifraga oppositifolia and Solda-nella alpina in the study Comparative phylogeographyof alpine plants lead us to conclude that we are not onlyfar from compiling data from a significant number ofspecies to infer common patterns but also to compre-

hend all factors responsible for present distributions andpredominant modes of speciation in MediterraneanEurope

ACKNOWLEDGEMENTSI thank the symposium group at Patras (Greece) our seminar

participants Rafael Aacutelvarez Rolf Holderegger Modesto LucentildeoGonzalo Nieto Virginia Valcaacutercel and an anonymous reviewer forhelpful comments on the manuscript J Muntildeoz for his help withmap arrangements and Tove Gabrielsen for leaf material ofSaxifraga oppositifolia from Norway This research has been sup-ported by the Spanish Direccioacuten General de InvestigacioacutenCientiacutefica y Teacutecnica (DGICYT) and Comunidad Autoacutenoma deMadrid through the projects PB98-0535 and 07M01242000

LITERATURE CITEDAbbott R J Smith L C Milne R I Crawford R M

M Wolff K amp Balfour J 2000 Molecular analysis ofplant migration and refugia in the arctic Science 2891343ndash1346

Ager D V 1975 The geological evolution of Europe ProcGeol Ass 86 127ndash154

Avise J C 2000 Phylogeography The History andFormation of Species Harvard Univ Press CambridgeMassachusetts

Avise J C Arnold J Ball R M Berminghan E LambT Neigal J E Reeb C A amp Saunders N C 1987Intraspecific phylogeography the mitochondrial DNAbridge between population genetics and systematics AnnReview Ecol Syst 18 489ndash522

Baldwin B G 1993 Molecular phylogenetics of Calycadenia(Compositae) based on ITS sequences of nuclear riboso-mal DNA chromosomal and morphological evolutionreexamined Amer J Bot 80 222ndash238

Baldwin B G Sanderson M J Porter J MWojciechowski M F Campbell C S amp DonoghueM J 1995 The ITS region of nuclear ribosomal DNA avaluable source of evidence on angiosperm phylogenyAnn Missouri Bot Gard 82 247ndash277

Baldwin B G amp Sanderson M J 1998 Age and rate ofdiversification of the Hawaiian silversword alliance(Compositae) Proc Natl Acad Sci USA 959402ndash9406

Barraclough T G amp Nee S 2001 Phylogenetics and speci-ation Trends Ecol Evol 16 391ndash 399

Brochmann C Gabrielsen T M Hagen A amp TollefsrudM M 1996 Seed dispersal and molecular phylogeogra-phy glacial survival tabula rasa or does it really matterNorsk Vidensk-Akad Mat-Naturvidensk Kl Avh 1854ndash68

Callaway R M 1995 Positive interactions among plantsBot Rev 61 306ndash349

Callaway R M Brooker R W Choler P Kikvidze ZLortie C J Michalet R Paolini L Puignaire F INewingham B Aschehoug E T Armas C Kikodze


474

D amp Cook B J 2002 Positive interactions among alpineplants increase with stress Nature 417 844ndash848

Chater A O amp Halladay G 1993 Arenaria L Pp 140ndash148in Tutin T G Burges N A Chater A O EdmondsonJ R Heywood V H Moore D M Valentine D HWalters S M amp Webb D A (eds) Flora Europaea ed2 vol 1 Cambridge Univ Press Cambridge

Comes P H amp Kadereit J W 1998 The effect ofQuaternary climatic changes on plant distribution and evo-lution Trends Plant Sci 3 432ndash438

Conti E Soltis D E Hardig T M amp Schneider J 1999Phylogenetic relationships of the silver saxifrages(Saxifraga Sect Ligulatae Haworth) implications for theevolution of substrate specificity life histories and bio-geography Molec Phyl Evol 13 536ndash555

Cullen J 1968 Anthyllis L Pp 177ndash182 in Tutin T GHeywood V H Burges N A Moore D M ValentineD H Walters S M amp Webb D A (eds) FloraEuropaea vol 2 Cambridge University Press London

Demesure B Comps B amp Petit R J 1996 ChloroplastDNA phylogeography of commom beech (Fagus sylvaticaL) in Europe Evolution 50 2515ndash2520

Dorofeev P I 1963 Primulaceae Pp 517ndash518 in TakhtajanA L (ed) Oznovij Paleontologii Akademia NaukaMoscow

Dumolin-Lapegravegue S Demesure B Fineschi S Le CorreV amp Petit R J 1997 Phylogeographic structure of whiteoaks throughout the European continent Genetics 1461475ndash1487

Farjon A 2001 World Checklist and Bibliography ofConifers ed 2 Royal Botanic Gardens Kew

Ferguson I K 1972 Vitaliana Sesler P 20 in Tutin T GHeywood V H Burges N A Moore D M ValentineD H Walters S M amp Webb D A (eds) FloraEuropaea vol 3 Cambridge Univ Press Cambridge

Franco J A 1993 Juniperus L Pp 46ndash48 in Tutin T GBurges N A Chater A O Edmondson J R HeywoodV H Moore D M Valentine D H Walters S M ampWebb D A (eds) Flora Europaea ed 2 vol 1Cambridge Univ Press Cambridge

Fuertes J Rosselloacute J A amp Nieto G 1999 Nuclear riboso-mal DNA (nrDNA) concerted evolution in natural and arti-ficial hybrids of Armeria (Plumbaginaceae) Molec Ecol8 1341ndash1346

Gabrielsen T M Bachmann K Jacobson K S ampBrochmann C 1997 Glacial survival does not matterRAPD phylogeography of Nordic Saxifraga oppositifoliaMolec Ecol 6 831ndash842

Goyder D J 1987 Observations on the geographical distri-bution reproductive biology and ecology of Arenariaalfacarensis Pamp Anales Jard Bot Madrid 44 285ndash297

Greuter W Burdet H M amp Long G 1989 Med-Checklistvol 4 Dicotyledones (Lauraceae-Rhamnaceae) GenevaSwitzerland

Gugerli F amp Holderegger R 2001 Nunatak survival tabu-la rasa and the influence of the Pleistocene ice-areas onplant evolution in mountain areas Trends Plant Sci 6397ndash398

Gutieacuterrez B Fuertes J amp Nieto G 2002 Glacial-inducedaltitudinal migrations in Armeria (Plumbaginaceae)inferred from patterns of chloroplast DNA haplotype shar-ing Molec Ecol 11 1965ndash1974

Hewitt G M 2000 The genetic legacy of the Quaternary iceages Nature 405 907ndash913

Hillis D M Mable B K amp Moritz C 1996b Applicationsof molecular systematics the state of the field and look tothe future Pp 515ndash543 in Hillis D M Moritz C ampMable B K (eds) Molecular Systematics SinauerAssociates Sunderland Massachusetts

Hillis D M Moritz C amp Mable B K 1996a MolecularSystematics Sinauer Associates Sunderland Massa-chusetts

Holderegger R Stehlik I amp Abbott R J 2002 Molecularanalysis of the Pleistocene history of Saxifraga oppositifo-lia Molec Ecol 11 1409ndash1418

Hungerer K B amp Kadereit J W 1998 The phylogeny andbiogeography of Gentiana L sect Ciminalis (Adans)Dumort a historical interpretation of distribution rangesin the European high mountains Pers Pl Ecol Evol Syst1 121ndash135

Jordano P 1993 Geographical ecology and variation ofplant-seed disperser interactions southern Spanishjunipers and frugivorous thrushes Vegetatio 10710885ndash104

King R amp Ferris C 1998 Chloroplast DNA phylogeogra-phy of Alnus glutinosa L Molec Ecol 7 1157ndash1161

Koumlrner C 1999 Alpine Plant Life Functional Plant Ecologyof High Mountain Ecosystems Springer-Verlag Heidel-berg

Kress A 1984 Primulaceen-Studien 7 Chromosomenzaumlh-lungen an verschiedenen Primulaceen Teil B SoldanellaBotanical Garden Munich

Kress A 1997 Androsace Pp 22ndash40 in Castroviejo SAedo C Laiacutenz M Morales R Muntildeoz Garmendia FNieto Feliner G amp Paiva J (eds) Flora Iberica vol 5CSIC Madrid

Kropf M Kadereit J W amp Comes P 2002a LateQuaternary distribution stasis in the submediterraneanmountain plant Anthyllis montana L (Fabaceae) inferredfrom ITS sequences and amplified fragment length poly-morphism markers Molec Ecol 11 447ndash463

Kropf M Kadereit J W amp Comes P 2002b LatePleistocene range fragment within the mountain plantAnthyllis montana L (Fabaceae) revealed by congruentpatterns of morphometric and genetic variation P 301 inSixth International Congress of Systematic andEvolutionary Biology September 9ndash16 Patras Greece[Abstr]

Levin D A amp Wilson A C 1976 Rates of evolution in seedplants net increase in diversity of chromosome numbersand species numbers through time Proc Natl Acad SciUSA 73 2086ndash2090

Loacutepez G 1990 Arenaria L Pp 172ndash224 in Castroviejo SAedo C Laiacutenz M Loacutepez G Morales R MuntildeozGarmendia F Nieto Feliner G amp Paiva J (eds) FloraIberica vol 2 CSIC Madrid

Martiacutenez J amp Vargas P 2002 Chloroplast differentiationand active dispersal of Juniperus oxycedrus in theMediterranean basin based on trnL-trnF sequences Pp 26in Phylogeography in Southern European RefugiaEvolutionary Perspectives on the Origins andConservation of European Biodiversity InternationalSymposium Vairatildeo (Portugal) March 11ndash15 Univ PortoVairatildeo [Abstr]


475

Mogensen H L 1996 The hows and whys of cytoplasmicinheritance in seed plants Amer J Bot 83 383ndash404

Nieto G amp Favarger C 1988 On the races of Arenaria tetra-quetra L (Caryophyllaceae) Bot J Linn Soc 97 1ndash8

Petit R J Csaikl U M Bordaacutecs S Burg K Coart ECotrell J Van Dam B Deans J D Dumolin-Lapeacutegue S Fineschi S Finkeldey R Gillies AGlaz I Giocoechea P G Jensen J S Koumlnig A OLowe A J Madsen S F Maacutetyaacutes G Munro R COlalde M Pemonge M-H Popescu F Slade DTabbener H Taurchini D de Vries S G MZiegenhagen B amp Kremer A 2002 Chloroplast DNAvariation in European white oaks phylogeography andpatterns of diversity based on data from over 2600 popu-lations Forest Ecol Management 156 5ndash26

Raspeacute O Saumitou-Laprade P Cuguen J amp Jacque-mart A-L 2000 Chloroplast DNA haplotype variationand population differentiation in Sorbus aucuparia L(Rosaceae Maloideae) Molec Ecol 9 1113ndash1122

Richardson J E Pennington R T Pennington T D ampHollingsworth P M 2001 Rapid diversification of aspecies-rich genus of neotropical rain forest trees Science293 2242ndash2245

Rieseberg L R amp Brouillet L 1994 Are many plant speciesparaphyletic Taxon 43 21ndash32

Ronquist F 1997 Dispersal-vicariance analysis a newapproach to the quantification of historical biogeographySyst Biol 46 195ndash203

Sanderson M J 2002 Estimating absolute rates of molecularevolution and divergence times a penalized likelihoodapproach Molec Biol Evol 19 101ndash109

Sang T Crawford D J Kim S-C amp Stuessy T F 1994Radiation of the endemic genus Dendroseris (Asteraceae)on the Juan Fernaacutendez Islands evidence from sequencesof the ITS regions of nuclear ribosomal DNA Amer JBot 81 1494ndash1501

Schaal B A amp Olsen K M 2000 Gene genealogies andpopulation variation in plants Proc Natl Acad SciUSA 96 302ndash306

Schoumlnswetter P Tribsch A Barfuss M amp Niklfeld H2002 Several Pleistocene refugia detected in the highalpine plant Phyteuma globulariifolium Sternb amp Hoppe(Campanulaceae) in the European Alps Molec Ecol 112637ndash2647

Shigesada N Kawasaki K amp Takeda Y 1995 Modelingstratified diffusion in biological invasions Amer Nat146 229 251

Stebbins G L 1984 Polyploidy and the distribution of thearctic-alpine flora new evidence and a new approach BotHelv 94 1ndash13

Stehlik I 2000 Nunatak and peripheral refugia for alpineplants during Quaternary glaciation in the middle part ofthe Alps Bot Helv 110 25ndash30

Stehlik I Blattner F R Holderegger R amp Bachmann K2002 Nunatak survival of the high alpine plantEritrichium nanum (L) Gaudin in the central Alps duringthe ice ages Molec Ecol 11 2027ndash2036

Stuessy T F 1990 Plant Taxonomy The SystematicEvaluation of Comparative Data Columbia Univ PressNew York

Suc J P 1984 Origin and evolution of the Mediterranean veg-etation and climate in Europe Nature 307 429ndash432

Taberlet P 2002 Comparative phylogeography of alpineplants P 16 in Phylogeography in Southern EuropeanRefugia evolutionary perspectives on the origins andconservation of European biodiversity InternationalSymposium Vairatildeo (Portugal) March 11ndash15 Univ PortoVairatildeo [Abstr]

Taberlet P Fumagalli L Wust-Saucy AG amp CossonsJ-F 1998 Comparative phylogeography and postglacialcolonization routes in Europe Molec Ecol 7 453ndash464

Taberlet P Gielly L Pautou G amp Bouvet J 1991Universal primers for amplification of three non-codingregions of chloroplast DNA Pl Molec Biol 171105ndash1109

Templeton A R 1980 The theory of speciation via founderprinciple Genetics 94 1011ndash1038

Van Andel T H amp Tzedakis P C 1996 Paleolithic land-scapes of Europe and environs 150000ndash25000 years agoan overview Quat Sci Rev 15 481ndash500

Van der Merwe M Winfield M O Arnold G M ampParker J S 2000 Spatial and temporal aspects of thegenetic structure of Juniperus communis populationsMolec Ecol 9 379ndash386

Vargas P 1997 Saxifraga L Pp 162ndash242 in Castroviejo SAedo C Laiacutenz M Morales R Muntildeoz Garmendia FNieto Feliner G amp Paiva J (eds) Flora Iberica vol 5CSIC Madrid

Vargas P 2001 Phylogenetic and evolutionary insights in theSaxifraga pentadactylis complex (Saxifragaceae) varia-tion of nrITS sequences Nordic J Bot 21 75ndash82

Vargas P 2002 Multiple diversification patterns of alpineplants in the Mediterranean Europe P 229 SixthInternational Congress of Systematic and EvolutionaryBiology September 9ndash16 Patras Greece

Vargas P In press Are Macaronesian islands refugia of relictplant lineages a molecular survey In Weiss S J ampFerrand N (eds) Phylogeography in Southern EuropeanRefugia Evolutionary Perspectives on the Origins andConservation of European Biodiversity Kluwer AcademicPublishers Dordrecht The Netherlands

Vargas P Baldwin B G amp Constance L 1999b A phylo-genetic study of Sanicula sect Sanicoria and S sectSandwicenses (Apiaceae) based on nuclear rDNA andmorphological data Syst Bot 24 228-248

Vargas P Morton C amp Jury S L 1999a Biogeographicpatterns in Mediterranean and Macaronesian species ofSaxifraga (Saxifragaceae) inferred from phylogeneticanalyses of ITS sequences Amer J Bot 86 724ndash734

Vargas P amp Nieto Feliner G 1996 Artificial hybridizationwithin Saxifraga pentadactylis (Saxifragaceae) Nord JBot 16 257ndash266

Webb D A 1993 Saxifraga L Pp 437ndash458 in Tutin T GBurges N A Chater A O Edmondson J R HeywoodV H Moore D M Valentine D H Walters S M ampWebb D A (eds) Flora Europaea vol 1 ed 2Cambridge Univ Press Cambridge

Webb D A amp Gornall R J 1989 Saxifrages of EuropeChristopher Helm London

Zhang L Comes H P amp Kadereit J 2001 Phylogeny andQuaternary history of the European montanealpineendemic Soldanella (Primulaceae) based on ITS andAFLP variation Amer J Bot 88 2331ndash2345


476

implemented to describe postglacial colonization routesof trees in central Europe (Demesure amp al 1996Dumolin-Lapegue amp al 1997 King amp Ferris 1998Raspeacute amp al 2000 Petit amp al 2002) and to test thehypothesis of nunatak areas for in situ glacial survival inarctic and alpine plants (Brochmann amp al 1996Gabrielsen amp al 1997 Stehlik amp al 2002 Schoumlnswet-ter amp al 2002) Quaternary glaciations may not havedramatically affected the survival of plants in southernEurope in contrast with total extinction of angiospermsin northern Europe due to long-term establishment of icesheets In fact three main refugia in the Iberian Italianand Balkan peninsulas have been proposed from whichre-colonization occurred (Taberlet amp al 1998 Hewitt2000) Genetic structure and speciation patterns of alpinefloras in southern Europe are more complex due to a con-tinuous presence of plants in Mediterranean mountainsalong fluctuating vegetation belts

This paper is a preliminary synthesis of results fromseven plant groups using molecular markers Androsacevitaliana Anthyllis montana Arenaria tetraquetraJuniperus communis var saxatilis Saxifraga oppositifo-lia Saxifraga pentadactylis Soldanella alpina (Vargas2002) Our working hypothesis is that geographic char-acteristics recent climatic events (glaciations) lifecycles dispersal mechanisms and pre-Holocene geneticstructures are mostly responsible for present distributionand differentiation within alpine plants Subspecific taxarepresent morphological differentiation of each speciesin an evolutionary context The main objective is to testthis phylogeographic hypothesis based on variation ofmolecular markers phylogenetic reconstructions andgeographic arrangement of mountains that have led topresent distributions and modes of speciation in Europe

MATERIALS AND METHODSSpecies selection mdash Alpine plants are typically

defined as those obtaining optimal habitat conditionsover 2300 m in the Mediterranean region (Koumlrner 1999)Unfortunately we have found few examples in the liter-ature of alpine vascular plants in southern Europe con-taining subspecific taxa which also have been analyzedusing phylogeographic data Seven plant groups havebeen considered in this paper thanks to previous publica-tions from several authors on Anthyllis montana (Kropfamp al 2002a) Saxifraga oppositifolia (Abbott amp al2000 Holderegger amp al 2002) and Soldanella alpina(Zhang amp al 2001) Additionally we have obtained andanalyzed new data from Androsace vitaliana Arenariatetraquetra Juniperus communis var saxatilis andSaxifraga oppositifolia and extended the sample ofSaxifraga pentadactylis (Vargas 2001) Some other stud-

ies have not been included due to insufficient number ofpopulations low morphological variation or uninforma-tive molecular results

Molecular markers mdash Molecular variation with-in the same species is the basic requirement to infer rela-tionships among populations Fingerprinting techniquesscreening the nuclear genome such as RAPDs ISSRsAFLPs and microsatellites ensure a large number ofdata to be analysed by phylogenetic methodologies(Hillis amp al 1996a) In some cases sequences of nuclearribosomal DNA (ITS ETS) display a minimal number ofphylogenetically informative characters among popula-tions Recombination of biparental nuclear genomes mayobscure historical relationships based on hierarchical ge-nealogies particularly in cases of reticulation Unipa-rental haplotypes either from the mitochondria or chloro-plast alleviate this problem because they are haploid andnon-recombinant macromolecules (Avise amp al 1987Comes amp Kadereit 1998) However low levels of genet-ic variation in organelle genomes together with phylo-genetic reconstructions based on genes inherited by onlyone parental organism (typically maternal inheritance inangiosperms Mogensen 1996) are disadvantages forthe use of these markers in phylogeography of plants andinference of morphological character evolution

Phylogeographic reconstructions mdash Tradi-tional biogeography considers two major possibilities toaccount for the origin of present distributions of popula-tions and species of alpine plants dispersal and vicari-ance (Ronquist 1997) The vicariance hypothesis arguesfor isolation into island-like areas of populations on sun-dered mountains from a continuous area separated duringinterglacial periods In contrast under dispersal distribu-tion of populations on elevated mountains is the result ofcolonization from one or few individuals We abandonany inference based on area cladograms because ourphylogenetic reconstructions are analysed at the popula-tional level and distributions of alpine plants are pro-foundly disparate Genealogies of neutral genes are usedto evaluate colonization patterns and microevolutionaryprocesses (Schaal amp Olsen 2000) The use of nuclearmarkers link morphological differentiation and phylo-geography because most genes responsible for characterevolution are contained in the nuclear genomeNeighbor-joining and parsimony-based analyses ofnuclear fingerprinting and ribosomal ITS sequences areimplemented

The internal transcribed spacer (ITS) region ofnuclear 18S-26S ribosomal DNA has proven to beinformative for both biogeographic studies and infer-ences of character evolution at inter- and intraspecificlevels (Baldwin amp al 1995) Availability of ITSsequences in GenBank is the result of ease of sequencingthe ITS region coupled with an acceptable number of


464









465




466













AY354293 AY254294AY354295












467








468







469









470







471








472








473



















474


































475



































476









465




466













AY354293 AY254294AY354295












467








468







469









470







471








472








473



















474


































475



































476




466













AY354293 AY254294AY354295












467








468







469









470







471








472








473



















474


































475



































476




467








468







469









470







471








472








473



















474


































475



































476






468







469









470







471








472








473



















474


































475



































476






469









470







471








472








473



















474


































475



































476








470







471








472








473



















474


































475



































476






471








472








473



















474


































475



































476







472








473



















474


































475



































476







473



















474


































475



































476



















474


































475



































476


































475



































476



































476

molecular evidence for multiple diversification patterns of alpine … · 2008. 6. 17. · ventoux,...

Documents