Phylogeny and historical biogeography of Isodon (Lamiaceae): Rapid radiation in south-west China and...

Molecular Phylogenetics and Evolution 77 (2014) 183ndash194

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage wwwelsevier com locate ympev

Phylogeny and historical biogeography of Isodon (Lamiaceae) Rapidradiation in south-west China and Miocene overland dispersal into Africa

httpdxdoiorg101016jympev2014040171055-7903 2014 Elsevier Inc All rights reserved

uArr Corresponding author at Laboratory of Plant Phylogenetics and ConservationXishuangbanna Tropical Botanical Garden Chinese Academy of Sciences KunmingYunnan 650223 PR China Fax +86 871 65160916

E-mail address jielixtbgaccn (J Li)

Xiang-Qin Yu abi Masayuki Maki c Bryan T Drew d Alan J Paton e Hsi-Wen Li f Jian-Li Zhao gJohn G Conran h Jie Li abuArra Laboratory of Plant Phylogenetics and Conservation Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Kunming Yunnan 650223 PR Chinab Center for Integrative Conservation Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Menglun Yunnan 666303 PR Chinac Division of Ecology and Evolutionary Biology Graduate School of Life Sciences Tohoku University Aoba Sendai 980-8578 Japand Museum of Natural History Department of Biology University of Florida USAe The Herbarium Royal Botanic Gardens Kew Richmond TW9 3AB UKf Herbarium (KUN) Kunming Institute of Botany Chinese Academy of Sciences Kunming Yunnan 650204 PR Chinag Key Laboratory of Tropical Forest Ecology Xishuangbanna Tropical Botanical Garden Chinese Academy of Sciences Menglun Yunnan 666303 PR Chinah Centre for Evolutionary Biology and Biodiversity amp Sprigg Geobiology Centre School of Earth and Environmental Sciences Benham Bldg DX 650 312 The University of AdelaideSA 5005 Australiai University of Chinese Academy of Sciences Beijing 100049 PR China

a r t i c l e i n f o

Article historyReceived 27 December 2013Revised 10 April 2014Accepted 16 April 2014Available online 30 April 2014

KeywordsIntercontinental disjunctionIsodonOverland dispersalHengduan Mountains regionQinghaindashTibetan PlateauRapid radiation

a b s t r a c t

Rapid organismal radiations occurring on the QinghaindashTibetan Plateau (QTP) and the mechanismsunderlying AsiandashAfrica intercontinental disjunctions have both attracted much attention from evolution-ary biologists Here we use the genus Isodon (Lamiaceae) a primarily East Asian lineage with disjunctspecies in central and southern Africa as a case study to shed light upon these processes The molecularphylogeny and biogeographic history of Isodon were reconstructed using sequences of three plastidmarkers the nuclear ribosomal internal transcribed spacer (nrITS) and a low-copy nuclear gene (LEAFYintron II) The evolution of chromosome numbers in this genus was also investigated using probabilisticmodels Our results support a monophyletic Isodon that includes the two disjunct African species both ofwhich likely formed through allopolyploidy An overland migration from Asia to Africa through Arabiaduring the early Miocene is proposed as the most likely explanation for the present disjunct distributionof Isodon The opening of the Red Sea in the middle Miocene may appear to have had a major role indisrupting floristic exchange between Asia and Africa In addition a rapid radiation of Isodon wassuggested to occur in the late Miocene It corresponds with one of the major uplifts of the QTP andsubsequent aridification events Our results support the hypothesis that geological and climatic eventsplay important roles in driving biological diversification of organisms distributed in the QTP area

2014 Elsevier Inc All rights reserved

1 Introduction

Orogenic and environmental processes play a major role inorganismal evolution (Richardson et al 2001 Mao et al 2012)The geological configuration of the Asian plate was suggested tobe greatly shaped by the IndiandashEurasia collision and the ArabiandashEurasia collision (Yin 2010) The former triggered the formationof the QinghaindashTibetan Plateau (QTP) uplift of which werereported to promote the rapid radiation of numerous organisms

(Liu et al 2006 Sun et al 2012) while the latter also served asa bridge for floral and faunal exchange between Africa and Asia(Stewart and Disotell 1998 Zhou et al 2011) Although both tec-tonic events probably altered species distribution patterns dramat-ically few studies have addressed the questions directly or testedhypotheses involving the two processes using appropriate taxa

The QinghaindashTibetan Plateau (QTP) complex sometimes calledthe lsquoroof of the worldrsquo harbors extremely rich species diversityand endemism (Wu 1988) Elucidating the mechanisms drivingplant speciation in this region is of great interest Organismaldiversification in this region has possibly been elevated at leastin part by the severe alteration of topography and accompanyingclimate change that ensued since the uplift of the QTP during theMiocene and Quaternary (Liu et al 2006 Sun et al 2012) Rapid

184 X-Q Yu et al Molecular Phylogenetics and Evolution 77 (2014) 183ndash194

radiation as inferred by short internal branch lengths (Richardsonet al 2001) and large polytomies (low DNA sequence divergenceand poor resolution) has been reported for many plant groupssuch as Saxifragales (Jian et al 2008) and Malpighiales (Xi et al2012) Research investigating rapid radiations of species-rich plantgroups in the QTP area may thus enrich our understanding of theevolutionary origins of biodiversity (Erkens et al 2012)

Intercontinental disjunct distributions between closely relatedspecies are a remarkable feature of angiosperm biogeography(Raven and Axelrod 1974) Disjunct distribution patterns fromthe Northern Hemisphere especially the Eastern AsiandashEasternNorth America disjunctions have been investigated extensively(Wen 1999 Nie et al 2006) In contrast mechanisms responsiblefor Southern Hemisphere disjunctions (eg AsiandashAfrica disjunc-tion) have received far less attention (eg Yuan et al 2005 Zhouet al 2011) The (Africandash) ArabiandashEurasia collision was consideredequally important to the IndiandashEurasia collision (Yin 2010) andmight have played a significant role in floral and faunal exchangebetween Africa and Asia

Overland migration between Africa and Eurasia across the Ara-bian Peninsula has been proposed for Hoplobatrachus Peters 1863(Amphibia Ranidae) through a connection between the ArabianPeninsula and Asia in the Miocene (Kosuch et al 2001) LychnisL (Caryophyllaceae) from Eurasia to the Ethiopian highlands viathe Arabian Peninsula (Popp et al 2008) and Uvaria L (Annona-ceae) which has been associated with the lsquoout-of-Africarsquo dispersaland radiation of primates during the Miocene (Zhou et al 2011)Other studies (eg Kulju et al 2007) have also discussed the pos-sibility of overland dispersal between Africa and Asia Howevercompeting hypotheses also exist the lsquoIndian raftingrsquo hypothesis(during the Cretaceous) (Conti et al 2002) dispersal of high-lati-tude boreotropical floras (from the late Paleocene to middleEocene) (Davis et al 2002) and trans-oceanic long-distance dis-persal (Yuan et al 2005) are three other mechanisms commonlyinvoked to explain AsiandashAfrica disjunctions

Isodon (Schrad ex Benth) Spach (tribe Ocimeae Dumort sub-family Nepetoideae) is a genus of ca 100 species in Lamiaceae(Harley et al 2004) The genus is distributed primarily in tropicaland subtropical Asia with a center of species diversity (ca 70) inthe Hengduan Mountains region (HMR southeastern corner of theQTP) of south-west China The genus also ranges west to Afghani-stan and Pakistan with a further two disjunct species I ramosissi-mus (Wall ex Benth) Murata and I schimperi (Vatke) JK Morton incentral and southern Africa Isodon includes shrubs subshrubs andperennial herbs with paniculate inflorescences composed of many-flowered cymes The corollas of Isodon species can differ in sizetube length color and the color of spots on the lips possibly asan adaptive response to pollinators The majority of Isodon speciesoccurring in the HMR region occupy cold and arid highland habi-tats but a few species extend to humid and warm tropical regionsIn contrast the two African Isodon species grow in more humidhabitats such as the margins of evergreen forests or in riparianzones (pers obs Li 1988)

Li (1988) monographed the genus in China dividing it intofour sections based on morphological characters such as inflores-cence density and the shape of the fruiting calyx Later the twoAfrican species were added (Ryding 1993 Morton 1998) buttheir molecular phylogenetic position within Isodon has not beentested In addition there was apparently at least one chromo-some number shift in Isodon the African species I ramosissimushas a chromosome number of 2n = 42 and a base chromosomenumber proposed to be x = 7 (Morton 1962 1993) or x = 14(Yamashiro et al 2005) In contrast the Asian species havenumbers of 2n = 24 or 2n = 36 with a suggested base chromo-some number of x = 12 (Jin and Sha 2004 Yamashiro et al2005 Huang 2011)

Our previous research focused mainly on the relationshipbetween Isodon and related genera and the two African specieswere not involved in (Zhong et al 2010) Accordingly in this studywe aim to obtain a more comprehensive phylogenetic reconstruc-tion of Isodon with the two African species included to illuminatethe evolution of this species-rich genus We are also using Isodonas a case study to understand better apparently rapid species radi-ations in the Hengduan Mountains region of south-west China andthe possible mechanisms triggering AsiandashAfrica disjunctionsMolecular phylogenetic approaches are used to reconstruct thepossible historical biogeography of Isodon and investigate thebroad topics mentioned above as well as to answer the followingspecific questions

1 Does the radiation of Isodon correlate with the suggestedgeological hypotheses concerning the QTP region and associ-ated climate change

2 What is the phylogenetic position of the two endemic Africanspecies with ostensibly aberrant chromosome number

3 If the two endemic African species are confirmed to bemembers of Isodon what mechanisms might be responsiblefor this AsiandashAfrica disjunction

2 Materials and methods

21 Taxon sampling

The classification of Isodon largely follows Li (1988) but eightspecies [Isodon anisochilus CY Wu I brachythyrsus CY Wu etHW Li I forrestii var intermedius CY Wu et HW Li I kunmingen-sis CY Wu et HW Li I pluriflorus CY Wu et HW Li I polystachys(Sun ex C H Hu) CY Wu et HW Li I setschwanensis var yungsh-engensis CY Wu et HW Li and I taliensis (CYWu) Hara] weretreated as distinct taxa following Wu and Li (1977) in order to rep-resent the species diversity of this genus better Isodon umbrosus(Maxim) H Hara and its varieties which were not included in Li(1988) were added following Flora of Japan (Murata andYamazaki 1993) and the inclusion of I oreophilus (Diels) AJ Patonet Ryding follows Zhong et al (2010) Isodon sp aff flavidus (Hand-Mazz) H Hara was treated as a separate taxonomic entity due toits allopatric distribution with respect to I flavidus (Hand -Mazz)H Hara and the samples labelled I eriocalyx (Dunn) Kudocirc 214 and Ihenryi (Hemsl) Kudocirc 232 were included as putative hybrids Afurther six specimens that could not be identified accurately toknown taxa were also treated as separate Isodon OTUs (operationaltaxonomic units) in the analyses

Samples of DNA for the two African species were obtained fromthe DNA Bank of the Royal Botanic Gardens Kew In total 73 acces-sions representing 71 species of Isodon were included in this studyThe complete list of voucher specimen numbers collection locali-ties distribution codes available chromosome numbers and Gen-Bank accession numbers of samples used are given inSupplementary Table S1 nrITS sequences (Supplementary mate-rial Table S2) for 26 Isodon species (based on different vouchers)sequenced in our previous study were also added to the presentdata set Voucher specimens of samples from Africa and Japanare deposited at the herbarium of the Royal Botanic GardensKew (K) and the herbarium of Botanical Gardens in Tohoku Univer-sity of Japan (TUS) respectively Others are deposited at Laboratoryof Plant Phylogenetics and Conservation Xishuangbanna TropicalBotanical Garden Chinese Academy of Sciences (XTBG) To aug-ment our sampling we also downloaded 21 nrITS sequences fromother Ocimeae genera from GenBank (Supplementary materialTable S2) Similarly we also downloaded 39 rpl32ndashtrnL sequencesfrom various genera in Lamiaceae (Supplementary materialTable S3)

X-Q Yu et al Molecular Phylogenetics and Evolution 77 (2014) 183ndash194 185

Based on the work of Zhong et al (2010) our outgroup taxaincluded seven species representing six genera in Ocimeae forthe cpDNA data set and 22 species representing ten genera in Oci-meae for the nrITS data set Since the Yule model used in ourmolecular dating analysis assumes more or less even samplingthroughout the tree (Velasco 2008) we sequenced rpl32ndashtrnL fromthree additional species of Elsholtzia Willd for this study to makethe Elsholtzieae (with only two published accessions available)better represented

22 DNA extraction PCR and DNA sequencing

Total genomic DNA was extracted from leaf tissue using a mod-ified cetyltrimethyl ammonium bromide (CTAB) method (Doyleand Doyle 1987) We sequenced three chloroplast (cpDNA) inter-genic spacer regions (trnDndashtrnT psbAndashtrnH and rpl32ndashtrnL) Twonuclear DNA fragments ribosomal internal transcribed spacers(nrITS) and the second intron of the LEAFY gene (part sequencesfrom the second and third exons) were also used The primersand corresponding annealing temperatures used in this study arelisted in Supplementary Table S4 To amplify and sequence the sec-ond intron of LEAFY Isodon-specific primers were designed usingthe degenerate primers LFsxl-2 and LFtxr (Frohlich andMeyerowitz 1997) as a starting point

Polymerase chain reaction conditions product purificationcloning and cycle sequencing followed protocols described else-where (Li et al 2011) Multiple bands obtained from LEAFY ampli-fication were separated by excision and elution from the gels usingan OMEGA quick Gel Extraction Kit (OMEGA) and then used forcloning Direct sequencing of nrITS PCR products from universalprimers produced lsquolsquocleanrsquorsquo sequences except for Isodon eriocalyx214 I medilungensis (CY Wu et HW Li) H Hara I sp 2 and Idawoensis (Hand -Mazz) H Hara Those sequences were obtainedthrough cloning with at least five positive clones sequenced foreach individual We could not obtain the nrITS sequence data fromI henryi 232

For the LEAFY intron II fragment two distinct copies with differ-ent sizes (A and B) were found in four species of Isodon consistentwith the findings of Aagaard et al (2005) There might be selectiveamplification of the two copies as copy A was obtained in all of theingroup while copy B was only obtained from four species of Iso-don and with an almost identical sequence It is possible that copyA was easier to be detected by our specific primers Copy B fromthe four species of Isodon was used as an outgroup because ofthe great difficulties in alignment between ingroup and outgrouptaxa

23 Sequence alignment and phylogenetic analyses

DNA sequences were aligned with the MAFFT algorithm (Katohet al 2005) in SATeacute v225 (Liu et al 2012) using default settingsand edited manually in MEGA v50 (Tamura et al 2011) Align-ment gaps were treated as missing data For multiple clones fromeach individual a single representative sequence was chosen ran-domly for phylogenetic analysis if all the clones formed a well-sup-ported clade in the preliminary analysis But multiple sequenceswere retained if they grouped with different accessions Maximumparsimony (MP) analyses were conducted using the heuristicsearch option in PAUPfrasl v40b10 (Swofford 2003) The parametersare 1000 random addition sequence replicates tree-bisection-reconnection (TBR) swapping MulTrees on with 1000 trees savedfrom each random addition sequence replicate Bootstrap supportvalues (BS) for the internal nodes were obtained with 100 boot-strap replicates using the same settings presented above We usedjModelTest v011 (Posada 2008) to select the best-fitting nucleo-tide substitution models for analyses of Bayesian inference (BI)

and BEAST according to the Akaike information criterion (AICAkaike 1974) Nucleotide models were as follows cpDNA GTR+InrITS GTR+G LEAFY exons K80+I LEAFY intron GTR+G BI analy-ses were performed using the program MrBayes v312 (Ronquistand Huelsenbeck 2003) Two runs were conducted in parallel withfour Markov chains with each running for 2000000 generationsand sampled every 100th generation Examination of the log-likelihood values suggested that stationarity was reached in about50000 generations Thus the first 200000 generations (10) werediscarded to make sure the burn-in period was sufficiently longFor purposes of the results and discussion in this paper cladeswith posterior probability values of 050ndash079 were consideredweakly supported 080ndash094 moderately supported and 095ndash100strongly (well-) supported

In this study incongruence between different gene trees wasidentified in two ways (1) we compared individual gene topolo-gies and considered topologies incongruent when conflicting nodeswere supported by at least 70 bootstrap support (BS) and 09Bayesian posterior probability in one of the two gene trees or(2) we conducted the incongruence length difference (ILD) test(Farris et al 1995) When incongruence is observed between dif-ferent gene trees the trees can then be used to calculate a networkwhich represents potential hybridization events (Russell et al2010) For the incongruent position of the African lineage foundbetween the cpDNA and nrITS trees (see results) the phylogeneticnetwork method was employed using both the 70 and 90 (forcomparison) Bayesian consensus trees in Dendroscope v30(Huson and Scornavacca 2012)

24 Molecular dating of Isodon

As there are no known fossils for Isodon molecular dating in thisstudy relies on fossils from related clades in Lamiaceae such as theearly Eocene fossil hexacolpate putative pollen of Ocimum L iden-tified by Kar (1996) and the earlyndashmiddle Oligocene fruit fossil ofMelissa L (Reid and Chandler 1926 Martiacutenez-Millaacuten 2010) Thesetwo fossils have also been employed in recent phylogenetic studiesof Lamiaceae (Drew and Sytsma 2012 2013) Those authors placedthe hexacolpate pollen described by Kar (1996) as a putative Oci-mum at the crown of Nepetoideae (as opposed to the crown ofOciminae) and the Melissa fossil at the stem of this genus We agreewith this conservative placement

Divergence times were estimated using a Bayesian methodimplemented in the program BEAST v172 (Drummond et al2006 Drummond and Rambaut 2007) For all BEAST analyseswe implemented a Yule process speciation prior and both theuncorrelated exponential (UCED) and lognormal (UCLN) relaxedclock model of rate change were conducted Bayes factors calcu-lated with Tracer v15 (Rambaut and Drummond 2007) were usedto compare the results of UCED and UCLN Two separate runs wereconducted each with 20000000 generations sampled every 1000generations and the resulting log files and trees from each runwere combined with the program Log Combiner v172 Tree Anno-tator v172 was used to summarize the set of post burn-in (10)trees and their parameters The combined log file was checkedusing Tracer v15 to ensure the effective sample sizes (ESS) wereall above 200 Finally the maximum clade credibility (MCC) chro-nogram was visualized using the program FigTree v131 (Rambautand Drummond 2010)

Estimation of Isodon divergence times was done in two stepsFirstly stem and crown ages of Isodon was estimated using therpl32ndashtrnL data set which included 61 species representing allthree tribes (Elsholtzieae Mentheae and Ocimeae) in Nepetoideae(Harley et al 2004) and four outgroups (Supplementary Table S3)In this initial analysis only 12 species representing the majorclades of Isodon were included as the Yule process assumes that


the sampling was even throughout the tree (Velasco 2008) TheGTR+G model of substitution was suggested for this data set Thelognormal prior distribution was used for the fossil calibrationsas it is perhaps the most appropriate distribution for summarizingpaleontological information (Ho and Phillips 2009) Prior parame-ters for the two fossil calibrations in this study followed Drew andSytsma (2012) but we did not employ temporal constraints on theroot of the tree

Secondly we conducted three detailed analyses within Isodonusing a combined cpDNA nrITS and LEAFY intron II data setrespectively It has been suggested that relaxed clock dating meth-ods are unable to reconstruct the pattern of molecular rate changeaccurately with only one internal calibration (Sauquet et al 2012)Three crown ages (see Section 3) within Isodon obtained from ourfirst analysis were used as lsquosecondary calibrationrsquo points using alognormal prior distribution Through adjusting the value of themean and standard deviation the 95 highest posterior density(HPD) of main nodes yielded from the detailed analyses was madeto be consistent with that obtained from our first analysis

25 Diversification rates of Isodon

The diversification rate change over time within Isodon wasexplored using several programs based on R applications All ofour three data sets (cpDNA nrITS and LEAFY intron II) were usedfor analyses Firstly we used a maximum likelihood method asimplemented in LASER v 24-1 (Rabosky 2006) to determinewhether diversification rates of Isodon have changed over timeThe test statistic for diversification rate-constancy was calculatedas DAICRC = AICRC AICRV where AICRC was the AIC score for thebest fitting rate-constant diversification model and AICRV wasthe AIC for the best fitting rate-variable diversification model Apositive value for DAICRC indicates that the data are best explainedby a rate-variable model of diversification Two rate-constant (purebirth and birthndashdeath model) and three rate-flexible diversificationmodels (a logistic density-dependent speciation rate an exponen-tial density-dependent and a yule2rate model) were evaluated inour study Secondly the Relative Cladogenesis (RC) test was per-formed in GEIGER v 201 (Harmon et al 2008) to detect shifts ofdiversification rates within Isodon using the MCC chronogram gen-erated from BEAST analyses Bonferroni correction was used toadjust the P-values Thirdly we calculated the semi-logarithmiclineage through time (LTT) plots by APE v 31-1 (Paradis et al2004) We used 3000 trees chosen randomly from the BEASTtree-chains to generate the confidence intervals

26 Ancestral area reconstructions

Five biogeographic regions were delimited based on speciesdistributions (Morton 1962 Wu and Li 1977 Li 1988 Friis andVollesen 2005) (A) QTP and adjacent regions including the QTPnortheastern India Nepal Bhutan Afghanistan Pakistan and Yun-nan Plateau (B) tropical Asia including southern India Sri LankaBangladesh and other tropical areas in China and southeasternAsia (C) eastern Asia including areas of south central north eastand northeast China Korea and parts of Russia (D) Japan (E) trop-ical Africa including Equatorial Guinea Cameroon Burundi Tanza-nia Sierra Leone southern Zimbabwe southern Sudan Ugandaand Ethiopia

Ancestral area reconstructions (AAR) were conducted usingboth the Statistical DispersalndashVicariance (S-DIVA) approach (Yuet al 2010) and likelihood analysis under the dispersalndashextinc-tionndashcladogenesis (DEC) model (Ree et al 2005 Ree and Smith2008) Analyses were implemented in RASP v21 (Yu et al 2011)and Lagrange respectively For the latter analysis we consideredthe connection between Asia and Africa to have been present from

ca 20 Ma based on previous studies (Roumlgl 1998 1999) and migra-tion between eastern Asia and Japan was allowed since ca 5 Ma(Kizaki and Oshiro 1977) For both S-DIVA and Lagrange approachwe conducted two analyses with the maximum area number ateach node set as 2 and 3 respectively since no species includedin our study occurred in more than three of our defined biogeo-graphic regions

AAR analyses were conducted using both cpDNA and nrITS dataset However the analysis using the LEAFY intron II data set wasnot attempted since two homoeologous copies were obtained fromsome putative hybrids including the two African species but mul-tiple copies were not present in all species Hanceola sinensis(Hemsl) Kudocirc and Orthosiphon wulfenioides (Diels) Hand-Mazzwere chosen as outgroups based on the results from our molecularphylogenetic reconstruction Since relationships among species ofIsodon were not fully resolved in this study AARs could not beestimated unambiguously within all the major clades recoveredThus only biogeographic events of nodes or branches withstrongly-supported posterior probability (above 095) wereinferred

27 Evolution of chromosome numbers

We used the program chromEvol v13 (Mayrose et al 2010) toinfer ancestral chromosome numbers in Isodon It estimateschanges in chromosome number along a phylogenetic tree byimplementing eight likelihood models Besides gain loss and poly-ploidy lsquolsquodemi-polyploidizationrsquorsquo which represents an increase inchromosome number by the union of a reduced and an unreducedgamete (Mayrose et al 2010) was also taken into account ABayesian consensus tree (constructed from the combined cpDNAmatrix) pruned to include only the species with reported chromo-some numbers was used as an input tree Species with chromo-some numbers and the corresponding references are shown inSupplementary Table S1 All clades except the monospecific cladeII had species represented for this analysis (see Section 3) Becausedifferent chromosome numbers have been reported in Ocimumbasilicum L (Paton and Putievsky 1996 Mukherjee et al 2005)we only used the most common count for analysis The programwas run under the default parameters using the best-fitting modelselected according to the likelihood ratio tests using the Akaikeinformation criterion (AIC)

3 Results

31 Phylogenetic analysis

The three cpDNA gene regions were combined since no conflictswere observed among partitions However we did not combine thetwo nuclear data sets as the ILD test suggested significant incon-gruence between the nrITS and LEAFY intron II data sets(p lt 001) Moreover the cpDNA data set was not combined witheither the nrITS or LEAFY intron II data set based on both theresults from the ILD test and the incongruent phylogenetic positionof the two African species The aligned lengths of the psbAndashtrnHtrnDndashtrnT rpl32ndashtrnL nrITS and LEAFY intron II data sets were465 1019 1051 650 and 2364 bp respectively The combinedplastid data set was 2535 bp in length

The monophyly of Isodon including the two African specieswas supported strongly by the cpDNA (BS = 98 PP = 10 Fig 1A)nrITS (BS = 77 PP = 10 Fig 1B) and LEAFY intron II (BS = 100PP = 10 Fig 2) trees Isodon was found to contain four main cladesThree of these clades were well-supported Asian lineages consis-tent with previous findings (Zhong et al 2010) an early-branchingclade I a monospecific clade II and clade III (two well-supported

(A) (B) Isodon shikokianus var occidentalisI shikokianus var intermediusI trichocarpusI longitubusI serraI umbrosusI rubescensI shikokianus I coetsaI adenanthusI inflexus 2I effususI umbrosus var excisinflexusI japonicusI japonicus var glaucocalyxI lihsienensisI amethystoidesI rosthorni iI pharicusI sp 4I adenolomusI sp 2I wikstroemioidesI grandifolius var atuntzeensis I bul leyanusI calcicolusI anisochi lusI sp 1I sp 3I parvifoliusI inflexus 1I coetsa var cavalerieiI brachythyrsusI forrestii var intermedius 1I scopariusI pluriflorusI polystachysI eriocalyxI eriocalyx 214 aI kunmingensisI flexicaul isI phyl lostachysI eriocalyx 214 b I oresbiusI forrestiiI smithianusI rugosiformisI hi rtellusI forrestii var intermedius 2I sp 6I setschwanensis var yungshengensisI irroratusI sp 5I setschwanensisI leucophyllus I nervosusI sculponeatusI dawoensisI medilungensisI flabelliformisI glutinosusI taliensisI ternifoliusI yuennanensisI henryiI phyllopodusI sp aff flavidusI oreophilusI lophanthoides I lophanthoides var gerardianusI lophanthoides var micranthusI ramosissimusI schimperiHyptis brevipesHyptis alataEriope latifoliaHyptis emoryiAsterohyptis mocinianaOrthosiphon aristatusOrthosiphon stamineusOrthosiphon wulfenioidesOcimum gratissimum var suave Ocimum gratissimumOcimum basilicum Coleus xanthanthusPlectranthus calycinus Plectranthus barbatusPlectranthus caninusHanceola sinensisHanceola exsertaLavandula multi fidaLavandula spLavandula angustifoliaSiphocranion nudipesSiphocranion macranthum

1

1

1

1

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

095

09

098

099

097

092095

091

100

100100

100

100

95

100

73

56

80

9997

78

93

77

9199

100

78

8178

91

97

97

59

88

8159

82

58

97

63

64

9261

88

Outgroup

IV

II

I

III

Isodon adenanthusI lihsienensisI inflexus 1

I effususI umbrosus var excisinflexus

I shikokianusI shikokianus var occidentalis

I umbrosusI trichocarpusI calcicolusI phyllostachysI flexicaulis

I parvifoliusI rugosiformis

I sp 4I medilungensisI forrestii var intermedius 2I grandifolius var atuntzeensis

I leucophyllusI pharicus

I coetsa var cavalerieiI irroratusI taliensis

I amethystoidesI serra

I bulleyanusI longitubusI inflexus 2I shikokianus var intermedius

I japonicusI rubescens

I kunmingensisI sp 1

I sp 3I rosthorniiI dawoensisI forrestii var intermedius 1I adenolomusI brachythyrsusI smithianusI oresbiusI sp 2I wikstroemioidesI flabelliformisI eriocalyx 214I hirtellus

I sculponeatusI forrestii

I coetsaI eriocalyxI scopariusI setschwanensisI glutinosusI polystachysI anisochilus

I nervosusI sp 6

I setchwanensis var yungshengensisI pluriflorus

I japonicus var glaucocalysI sp 5

I ternifoliusI ramosissimus

I schimperiI lophanthoides var gerardianusI lophanthoidesI yuennanensisI sp aff flavidus

I lophanthoides var micranthusI henryiI henryi 232I phyllopodus

I oreophilusPlectranthus calycinus

Ocimum basilicumOcimum gratissimum var suave

Orthosiphon wulfenioidesHanceola sinensis

Lavandula spSiphocranion macranthum

5 changes

1

1

11

1

091

1

1

1

1

1

1

1

1

100

92

100100

100

99

100

98

75

92

98

100

84

98

Outgroup

I

IIIV

III

Fig 1 Fifty percent majority rule consensus trees from Bayesian analysis of combined cpDNA sequence data set (with branch lengths) (A) and nrITS sequence data set (B)Support values of nodes P 09 PP (posterior probability) and 50 BS (bootstrap support) are shown above and below the branches respectively and only those for major nodesare shown in the cpDNA tree Scale bars indicate expected substitutions lsquolsquoarsquorsquo and lsquolsquobrsquorsquo in the nrITS tree indicate distinct types of sequences isolated from a single individualbranches in bold represent the African clade indicates conflict between the parsimony and Bayesian analyses


subclade III-1 and III-2 in LEAFY intron II tree) that includes theremaining Asian species (Figs 1 and 2) The two species endemicto Africa formed a strongly supported clade (clade IV) in both thecpDNA (BS = 98 PP = 10) and nrITS (BS = 100 PP = 10) treesHowever the position of the African lineage differed between theanalyses using the above two data sets

The African clade was embedded in Asian clades and shared acommon ancestor with clades II and III in the cpDNA tree(BS = 100 PP = 10) (Fig 1A) In the nrITS tree the African cladewas grouped with clade I with weak Bayesian support (PP = 071)(Fig 1B) but was placed as the sister to all other ingroup taxa inthe MP analysis (BS lt 50) Nevertheless it was distant from theclade IIndashIII a different pattern from the cpDNA tree The topologiesrecovered from the Parsimony and Bayesian analyses of the threeDNA data sets were generally identical except for the Africanclade Thus for each data set only the Bayesian consensus treewas shown but with both the parsimony bootstrap support andBayesian posterior probability values indicated

For the LEAFY intron II data set all the sequences of copy Aformed a well-supported clade (BS = 100 PP = 10) It provided evi-dence that these sequences are orthologues of LEAFY intron IINotably two distinct types of sequences of LEAFY intron II (IV-aBS = 100 PP = 10 IV-b BS = 100 PP = 10) were obtained for thetwo African species Clade IV-a grouped with clade I and IV-bappeared as sister to the clade IIndashIII (Fig 2) However clade IV-bwas placed as the sister to all other ingroup in the MP analysiswhich was also found in the analysis of nrITS data set Within cladeIII polytomies were found in all three phylogenetic trees and onlythe cpDNA tree with branch lengths from the BI analysis wasshown (Fig 1A) The extremely short internal branch lengths were

typical of the whole genus For the two putative hybrids two dis-tinct types of sequences were obtained from both nrITS and LEAFYintron II in I eriocalyx 214 One grouped with I eriocalyx (nrITSBS = 64 PP = 10 LEAFY intron II BS = 54 PP = 094) and the othergrouped with I phyllostachys (nrITS BS = 92 PP = 10 LEAFY intronII BS = 82 PP = 10) However I eriocalyx 214 appeared as sister toI hirtellus with strong support (BS = 78 PP = 10) in the cpDNA treeAlthough we could not obtain the nrITS sequence for I henryi 232two distinct sequences obtained from LEAFY intron II for this sam-ple grouped with I sp aff flavidus and I henryi with strong(BS = 84 PP = 10) and moderate (BS lt 50 PP = 082) supportrespectively

Results of the phylogenetic network analysis using both 70and 90 Bayesian consensus trees between cpDNA and nrITS con-sistently showed reticulation in clade III (Supplementary Fig S1)Although the analysis using the 70 Bayesian consensus treesmight be more likely to suffer from stochastic error it suggestedthat the African clade was of hybrid origin between members ofclade I and clade IIndashIII This result was further supported by thegene tree generated from LEAFY intron II data set (see above)

32 Divergence time estimates

The comparison of Bayes factors that were calculated fromUCED and UCLN relaxed clock models indicated that UCED wasbetter than UCLN in all of our BEAST analyses Thus we only pro-vided the results under the UCED relaxed clock model Althoughthe root of the tree was not constrained divergence times of majornodes of Nepetoideae inferred from the BEAST analysis of the

Fig 2 Fifty percent majority rule consensus tree from Bayesian analysis of LEAFY intron II sequence data set Support values of nodes P 09 PP (posterior probability) and 50BS (bootstrap support) are shown above and below the branches respectively lsquolsquoArsquorsquo and lsquolsquoBrsquorsquo indicate different copies of LEAFY lsquolsquoarsquorsquo and lsquolsquobrsquorsquo indicate distinct types of sequencesisolated from a single individual branches in bold represent the African clade indicates conflict between the parsimony and Bayesian analyses

Table 1Divergence time estimates of BEAST analyses for major nodes of Nepetoideae and Isodon

Node Age estimated in this study Age estimated in Drew and Sytsma (2012)

cpDNA ITS LEAFY intron II cpDNA

Mean 95 HPD Mean 95 HPD Mean 95 HPD Mean 95 HPD

C1 Nepetoideae crown 6136 5264ndash7478 ndash ndash ndash ndash 5729 524ndash637C2 Melissa stem 3248 2956ndash3716 ndash ndash ndash ndash 3201 2968ndash3563a Mentheae crown 4553 349ndash586 ndash ndash ndash ndash 4572 4001ndash5208b Ocimeae crown 4304 2855ndash5789 ndash ndash ndash ndash 437 3432ndash5284c Isodon stem 2732 1703ndash3984 ndash ndash ndash ndash 2490 1719ndash3235d Isodon crown 1961 1466ndash2644 1978 1415ndash2757 2142 1510ndash3081 ndash ndashe Split between IIndashIII and IV 1361 981ndash1853 ndash ndash ndash ndash ndash ndashf Split between II and III 1076 750ndash1493 1320 831ndash1955 1026 643ndash1504 ndash ndashg I lineage crown 812 389ndash1358 814 373ndash1377 575 233ndash1043 ndash ndashh III lineage crown 928 626ndash1309 987 591ndash1491 762 470ndash1163 ndash ndashi Crown of III lineage exclude I sp 5 784 514ndash1117 ndash ndash ndash ndash ndash ndashj IV lineage crown 536 172ndash1025 510 171ndash1001 ndash ndash ndash ndashj1 IV-a lineage crown ndash ndash ndash ndash 608 255ndash1140 ndash ndashj2 IV-b lineage crown ndash ndash ndash ndash 551 208ndash1029 ndash ndashk Split between I and IV ndash ndash 1534 919ndash2253 ndash ndash ndash ndashl Split between I and IV-a ndash ndash ndash ndash 1454 784ndash2269 ndash ndashm Split between IIndashIII and IV-b ndash ndash ndash ndash 1575 1014ndash2362 ndash ndashn III-2 lineage crown ndash ndash ndash ndash 631 380ndash970 ndash ndash

HPD highest posterior density lsquolsquondashrsquorsquo indicates no data available all estimated time with a unit of Ma


rpl32ndashtrnL data set were largely consistent with earlier findings(Drew and Sytsma 2012) and were summarized in Table 1

The results of the rpl32ndashtrnL BEAST analysis suggested thatIsodon originated ca 2732 million years ago (Ma) in the late

Oligocene (95 HPD = 3984ndash1703 Ma node c in Table 1 andSupplementary Fig S2) This was highly congruent with the agesestimated by Drew and Sytsma (2012) Results obtained from thecombined cpDNA BEAST analysis indicated that Isodon began to

(A)

(B)

Fig 3 Chronogram of Isodon derived from BEAST analysis of combined cpDNA data set (A) with the optimizations of ancestral distributions represented by the pie charts(relative frequencies of the areas) at each node distribution of the extant species of Isodon and letters coding for areas are indicated on the map the estimated dispersalevents are shown beside the branches Blue bars represent 95 credibility intervals for each node with posterior probability above 09 time scale is shown at the bottom Thehypothesized tectonic events occurring on the Arabian Plate are also shown (B) The star at node i indicates the shift of diversification rate based on the Relative Cladogenesis(RC) test


diversify ca 1961 Ma in the early Miocene (95 HPD = 2644ndash1466 Ma node d in Fig 3A and Supplementary Fig S2) The splitbetween the African lineage and the Asian clade might occur ca1361 Ma (95 HPD = 1853ndash981 Ma node e in Fig 3A) The diver-sification of most species of Isodon probably occurred ca 928 Ma(95 HPD = 1309ndash626 Ma node h in Table 1 and Fig 3A) in thelate Miocene Results of the nrITS and LEAFY intron II BEAST anal-yses were highly consistent with that of cpDNA (Table 1) with onlyminor discrepancies between some nodes Moreover the splitbetween clade IV and I inferred from the nrITS BEAST analysiswas ca 1534 Ma (95 HPD = 2253ndash919 Ma node k in Table 1)It agreed with our cpDNA BEAST analysis Based on the LEAFYintron II BEAST analysis the split between clades I and IV-a (nodel in Table 1) and the split between clades IIndashIII and IV-b (node m inTable 1) occurred virtually simultaneously


The LASER results rejected the null hypothesis of temporallyhomogeneous diversification rates and suggested variable

diversification rates within Isodon All of the three data sets betterfitted a rate-variable model of diversification with the diversifica-tion rate-constancy statistic DAICRC being 774 for cpDNA 678 fornrITS and 1191 for LEAFY intron II data set respectively The yule2-rate diversification model was selected as the best fitting model forall of our three data sets The Relative Cladogenesis (RC) analysesdetected significant shifts of diversification rates within Isodon atca 784 Ma based on cpDNA data set (node i in Fig 3 p = 0002)at ca 987 Ma based on nrITS data set (node h in SupplementaryFig S3 p = 0035) and at ca 631 Ma based on LEAFY intron II dataset (node n in Supplementary Fig S4 p = 0004) These eventsmight have resulted in the unsolved phylogenetic relationshipswithin clade III Our analyses of the LTT plots also supported thefindings obtained from the LASER and GERGER analyses Thesemi-logarithmic LTT plots derived from our three data sets clearlyindicated an increase of the diversification rate (acceleratedlineage accumulation) within Isodon after ca 10 Ma in the lateMiocene and towards the current time (Fig 4) in contrastrelatively stable rates of diversification were illustrated before thattime

15 10 5 0

12

510

2050

Time (million years ago)

Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

sTime (million years ago)

nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II

Fig 4 Lineage through time plots (with 95 confidence intervals) of Isodon derived from cpDNA nrITS and LEAFY intron II data sets The bold line corresponds to themaximum credibility tree from the BEAST analyses


34 Ancestral area reconstruction

The ancestral distributions and the dispersal events inferredfrom the S-DIVA and the Lagrange analysis produced similarresults Setting the lsquoMaxarearsquo limit to 2 and 3 provided almostidentical results In addition the results obtained from cpDNAand nrITS data sets were highly congruent (Supplementary mate-rial Table S5) Thus only the results with lsquoMaxarearsquo set as 2 of S-DIVA analysis based on cpDNA data set were shown in Fig 3A

The QTP and adjacent regions were inferred as the most likelyancestral area of Isodon species occurring in Africa were estimatedto have dispersed from the QTP and adjacent regions in the early tomiddle Miocene (ca 20ndash14 Ma nodes d and e in Table 1 andFig 3A) The common ancestor of clades II III and IV might haveoccurred in the QTP and adjacent regions as well as tropical AfricaThe stem lineage of clades II III and IV subsequently underwent avicariant split of the combined ancestral area (node e in Fig 3A)The ancestral area for the early-branching clade I (node g inFig 3A) clades II and III (node f in Fig 3A) and clade III (node hin Fig 3A) was also inferred as the QTP and adjacent regions

35 Inference of chromosome number change

The lsquolsquoConst_rate_demirsquorsquo model (Loglikelihood = 2722AIC = 6044) with the chromosome duplication rate equal to thedemi-duplication rate was selected as the best-fitting model forthe data set The inferred change events of chromosome numberwithin Isodon were shown in Fig 5 The ancestral haploid chromo-some number of Isodon was estimated to be n = 12 (probabil-ity = 099) Gain and demi-polyploidization (the union of areduced and an unreduced gamete) may have been involved inthe speciation of I lophanthoides and the African I ramosissimus

4 Discussion

41 Molecular phylogeny of Isodon

Phylogenetic reconstructions using three different data setsconsistently show that Isodon is monophyletic with four majorwell-supported clades (I II III and IV Figs 1 and 2) Three of theclades (I II and III) recovered are consistent with previous findings(Zhong et al 2010) In addition the two African-endemic speciesone of which has an aberrant chromosome number formed astrongly-supported clade IV within Isodon As in other putativerapid radiations (eg Hughes and Eastwood 2006 Liu et al2006) large polytomies were found in all three gene trees in our

study A similar situation was encountered with Solms-laubachiaMuschl (Brassicaceae) which also has a center of diversity in theHengduan Mountains region (Yue et al 2009) This pattern reflectsthe challenge of recovering recent rapid radiations that exhibitextremely short internal branch lengths (Richardson et al 2001Shavit et al 2007) Resolving the species phylogeny for a groupthat radiated explosively is particularly difficult because reticula-tion may occur during species divergence due to incomplete repro-ductive isolation (see below Rokas et al 2003 Belfiore et al2008) For example it has been suggested that P25000 bp areneeded in Saxifragales to resolve a rapid radiation with high sup-port values (Jian et al 2008) Future research applying large-scalegenomic sequence data (eg Fior et al 2013) therefore may beneeded to resolve species-level relationships within Isodon

Rapid diversification may trigger extensive hybridization andintrogression because intrinsic barriers to reproductive isolationamong recently evolved lineages that prevent gene flow betweenmany species may have had too little time to develop fully(Mallet 2007) Indeed within Isodon seven natural interspecifichybrids have been recorded in Japan alone (Murata andYamazaki 1993) Our comparison between the cpDNA and nuclearDNA trees indicated that either I eriocalyx or I phyllostachys wasthe possible paternal donor of I eriocalyx 214 with I hirtellus asthe possible maternal donor This hybrid was more similarmorphologically to I eriocalyx based on our field observations Inaddition I henryi 232 was also inferred to be a hybrid between Isp aff flavidus (paternal) and I henryi (maternal) Our phylogeneticnetwork analysis shows evidence for additional hybridizationevents (Supplementary Fig S1B) which indicates that ongoinghybridization may play an important role in the evolutionary his-tory of Isodon and thus greatly inhibit proper classification

In this study the cpDNA and nrITS gene trees showed incongru-ent phylogenetic positions for the African clade (Fig 1A and B) Thephylogenetic network between cpDNA and nrITS also indicatedthat the African clade was of hybrid origin between distantly-related parent species one from clade I and the other from cladeIIndashIII (Supplementary Fig S1A) This result was further corrobo-rated by the gene tree of the low-copy nuclear gene LEAFY intronII which showed that two divergent homoeologous copies of theAfrican clade grouped with putative maternal (clade IIndashIII) andpaternal (clade I) parents respectively (clades IV-a and IV-b inFig 2) Based on current data there are different plausible scenar-ios for the paternal (maternal) parent of the African clade It couldbe the common ancestor of clade I (clade IIndashIII) or alternatively theprogenitor could have been an extinct lineage that was closelyrelated to clade I (clade IIndashIII) The fact that only one copy of nrITSwas found indicates that homogenization to the paternal type

Fig 5 Chromosome number evolution of Isodon inferred using chromEvol over a pruned Bayesian tree from the combined cpDNA data set Only taxa with reportedchromosome numbers are included Inferred ancestral haploid numbers (n) are displayed in a pie chart at each node a whole pie chart represents the haploid number withprobability P 099 Different portions of the pie charts represent two or more haploid numbers with different probability values


might have occurred in the two African species as reported byother hybridization studies (eg Kovarik et al 2005)

Analysis of changes in chromosome number based on our lim-ited species sampling (Fig 5) suggested that the ancestral haploidchromosome number of Isodon was n = 12 with high probability(099) A demi-polyploidization event was inferred to be involvedin the origin of the African species I ramosissimus although the ori-gin of the other African species can not be inferred because of miss-ing chromosome number data A likely formation of the inferreddemi-polyploidization of I ramosissimus is the fusion of two unre-duced gametes to yield n = 24 followed by chromosome fusions(Lysak et al 2006) or loss to yield n = 21 Taken together ourresults from DNA sequence data and chromosome number evolu-tion suggest that the African clade was likely formed through allo-polyploidy However it must be noted that owing to limitedinformation on chromosome numbers in Isodon further compre-hensive cytological studies are needed to confirm our findings

42 Rapid radiation of Isodon in the Hengduan Mountains region

Polytomies found in all three gene trees together with extre-mely short internal branch lengths indicate that a rapid radiationmight have occurred in the evolutionary history of Isodon(Richardson et al 2001 Shavit et al 2007) Additionally our bio-geographic reconstruction inferred that the QTP and adjacentregions are the most likely ancestral area (Fig 3A) with the HMR(southeastern edge of the QTP) currently containing the greatestdiversity of Isodon species (ca 70) The question raised here isthis what triggered the extensive diversification of this species-rich genus in this region

Several studies have attempted to illustrate the relationshipbetween geological events and the origin and diversification oforganisms within the QTP area (eg Liu et al 2006) Howeverdue to the lack of convincing fossil records divergence timeestimates of previous studies have been hampered because ofusing only one calibration point (Liu et al 2006) or secondary cal-ibrations based on results obtained from previous studies (Zhangand Fritsch 2010) Some studies have even attributed the low res-olution of species-level phylogenies solely to a rapid radiationwithout divergence time estimation (eg Lu et al 2010) Othersonly use the DNA sequence substitution rate instead of fossil

calibrations in their estimates of divergence times (eg Liu et al2002)

In this study the divergence time estimation relies on two fos-sils of Lamiaceae that are well accepted (Harley et al 2004Martiacutenez-Millaacuten 2010) and have been used in recent studieswithin the Nepetoideae (Drew and Sytsma 2012 2013) Theresults of our molecular dating suggest that the diversification ofthe most diverse lineage III within Isodon may begin ca 928 Ma(95 HPD = 1309ndash626 Ma) in the late Miocene (node h in Table 1and Fig 3A) based on cpDNA data set ca 987 Ma (95HPD = 1491ndash591 Ma) on nrITS data set and ca 762 Ma (95HPD = 1163ndash470 Ma) on LEAFY intron II data set To confirm ifthe diversification rate in Isodon has changed over time we con-ducted several analyses as implemented in packages based on Rapplication Rate-constancy statistic DAICRC calculated in LASERbased on cpDNA nrITS and LEAFY intron II data sets all suggestedvariable diversification rates within Isodon The Relative Cladogen-esis (RC) analyses revealed shifts of diversification rates withinIsodon at ca 784 Ma (cpDNA Fig 3) ca 987 Ma (nrITS Supple-mentary Fig S3) and ca 631 Ma (LEAFY intron II SupplementaryFig S4) The semi-logarithmic LTT plots derived from our threedata sets also indicated an accelerated lineage accumulation afterca 10 Ma (Fig 4) These estimates correspond well with the thirdof four major QTP uplifts believed to have occurred ca 22ndash2015ndash13 10ndash8 and 36ndash0 Ma (Harrison et al 1992 Coleman andHodges 1995 Shi et al 1999 Spicer et al 2003) although theexact timings of these uplifts are still being debated The extensiveuplift of the QTP and the aridification in this region due to theonset of the East Asian and Indian monsoon climate occurringaround 8 Ma was also corroborated by researchers applyingpaleobotanical data (eg Molnar et al 1993 An et al 2001)The large number of arid-adapted Isodon species (Li 1988) appear-ing at that time may also be a reflection of aridification in thisregion

Combining previous findings with our results we propose thatbursts of speciation triggered by geological andor climatic changeson the QTP may be a common phenomenon and of significantimportance in generating the current biodiversity within thisregion Similar scenarios were reported within Andean regions(Hughes and Eastwood 2006 Drew and Sytsma 2013 Hugheset al 2013)


43 Mechanisms of the AsiandashAfrica disjunction

Our data provide evidence that the two Isodon species (I ramo-sissimus and I schimperi) endemic to tropical Africa (Li 1988Ryding 1993 Morton 1998) belong in Isodon (Figs 1 and 2) Thisgrouping illustrates a considerable disjunction within the genusbetween Asia and Africa as well as illustrating uneven species rich-ness between the two areas As mentioned previously there arefour explanations usually espoused to explain AsiandashAfrica disjunc-tions the Indian rafting hypothesis (Gondwana vicariance) (Contiet al 2002) dispersal of a high-latitude boreotropical paleoflora(Davis et al 2002) transoceanic long-distance dispersal (Yuanet al 2005) and Miocene overland migration via the Arabian Pen-insula (Kosuch et al 2001 Zhou et al 2011)

The former two hypotheses are rejected as our divergence timeestimates suggested that Isodon possibly originated in the late Oli-gocene ca 27 Ma (95 HPD = 3984ndash1703 Ma node c in Table 1)This time is far too young to be associated with the breakup ofGondwana rafting of India or the once extensive boreotropicalpaleoflora Besides direct trans-oceanic dispersal between Asiaand Africa (Nie et al 2013) dispersal via EocenendashOligocenelsquolsquoLemurian stepping-stonesrsquorsquo (Schatz 1996) which connected IndiaSri Lanka the Seychelles and Madagascar has also been suggestedto be a possible route for organisms with AsiandashAfrica disjunctions(Yuan et al 2005) However the above two pathways seem lesslikely for Isodon because although little is known about fruit dis-persal in Isodon some species in Lamiaceae with similar fruit sizeand characters were reported to be dispersed primarily by passiveballistics across only limited distances with secondary dispersionby water flow (Zhou et al 1999) Similarly unlike the wingeddiaspores of Paederia L (Rubiaceae) that were suggested to beadapted to dispersal by ocean currents (Nie et al 2013) the smalldry inconspicuous nutlets of Isodon lack any apparent dispersaladaptations The absence of any obvious epizoic adaptions alsomeans that its seeds are not likely to be dispersed by birds or con-sumed by primates Hence long-distance transoceanic dispersalseems unlikely

Similarly our ancestral area optimization suggests that dis-persal from Asia to Africa might have occurred ca 20ndash14 Ma dur-ing the early Miocene (nodes d and e in Table 1 and Fig 3A) Bythat time the Lemurian stepping-stones may have no longerexisted Furthermore there is no evidence that Isodon has everoccurred on the Seychelles or Madagascar Moreover very fewextant species of tribe Ocimeae naturally co-occur in Africa andAsia and most of those that do are very widespread and likely tohave been transported anthropogenically for medicinal andorculinary uses (eg Ocimum tenuiflorum L) The ability to disperseover relatively short distances favors overland migration possiblythrough Arabia as the most plausible explanation for the currentdistribution of Isodon Some species of Isodon currently reach asfar west as Afghanistan and Pakistan and the possibility of plantdispersal across northern Africa and Arabia is also supported byfossil evidence of Meliaceae in Europe Africa and Asia (Muellneret al 2006)

The species of Isodon occurring in Africa were estimated to havedispersed from those in the QTP and adjacent regions ca 20ndash14 Maduring the early Miocene Subsequent vicariance between the QTPand adjacent regions and Africa possibly occurred ca 14 Ma (nodesd and e in Table 1 and Fig 3A) Despite the uncertainty of thetiming of the allopolyploidy event our results suggest the commonancestor of the African clade most likely formed in Asia and subse-quently dispersed to Africa where it gave rise to two daughter spe-cies perhaps using Arabia as a stepping stone Although theestimated age of collision between the Arabian and Eurasian platesvaries widely ranging from the Eocene to the late Miocene(reviewed in Yin 2010) our findings suggest that floristic exchange

between the Asian and African continents might have been feasiblesince at least 20 Ma (Fig 3B) This is consistent with previous sug-gestions that the collision between the Arabian plate and Eurasiaand the closure of the Tethys Sea occurred ca 20 Ma connectingAfrica and western Asia (Roumlgl 1998 1999) By integrating molecu-lar and fossil evidence hominoid species were suggested to havedispersed out of Africa and into Eurasia through Arabia at this time(Stewart and Disotell 1998) A recent study also support the viewof overland dispersal from Asia to Africa via southwest Asia and theArabian peninsula in Macaranga Thouars (Euphorbiaceae vanWelzen et al 2014)

Importantly two stages were suggested to be involved in theopening of the Red Sea The estimated age for the first movementvaried from 41ndash34 Ma to 20 Ma onwards and the second move-ment was dated to 5ndash4 Ma onwards (Bartov et al 1980) Moreovera drastic rifting in the Gulf of Suez (northwest of the Red Sea) mayhave occurred during the middle Miocene (Garfunkel and Bartov1977) This raises the possibility that the continuous opening ofthe Red Sea may have separated the Arabian Peninsula from conti-nental Africa ca 14 Ma to an extent (Fig 3B) that disrupted floristicexchange between Asia and Africa An alternative scenario is thatthis vicariance was caused by climate changes occurring on theArabian Peninsula (eg Menzies et al 1992) However no Isodonspecies and very few other species of tribe Ocimeae are currentlyfound in Northern Africa or Arabia This could be possibly due tochanges in climate and habitats in these regions such as thedesertification of Northern Africa in the late Miocene (Schusteret al 2006) Other members of Ocimeae only persist in areaswhere there is some seasonal rainfall such as western Yemen orOman

Allopolyploids are known to benefit potentially from lsquolsquohybridvigorrsquorsquo and lsquolsquointrinsic fitness advantagersquorsquo which allow them toexploit habitats previously unavailable to their diploid progenitors(Stebbins 1985) The reason why only the allopolyploid clade ofIsodon successfully survives in Africa might be attributed to theadvantages of allopolyploids This is analogous to the diversifica-tion of allopolyploid Nicotiana L section Suaveolentes Goodsp(Solanaceae) species in Australia following their origin dispersaland then extinction in South America ca 10 Ma also despite thelack of any obvious long-distance diaspore characteristics (Chaseet al 2003 Ladiges et al 2011 Marks and Ladiges 2011) Possiblecauses for the uneven number of species of Isodon between Asiaand Africa could be i) the allopolyploid African lineage did notdiversify until the beginning of the Pliocene and there might havebeen too little time for them to speciate further andor ii) therecould be native African habitat competitors such as PlectranthusL which occupies a similar range of habitats in Africa to Isodonin Asia There are no species of Plectranthus in the QTP regionwhere Isodon is most diverse and only two Isodon species occurin tropical Africa where there are ca 180 Plectranthus species

5 Conclusions

The reconstructed biogeographic history of Isodon providesinsights into the mechanisms triggering AsiandashAfrica disjunctionsIsodon was suggested to originate in the QTP and adjacent regionsin the late Oligocene Subsequent migration into Africa in early tomiddle Miocene via overland migration through Arabia was con-sidered as the most plausible dispersal scenario The opening ofthe Red Sea in the middle Miocene (ca 14 Ma) may also haveplayed a critical role in disrupting floristic exchange between Asiaand Africa Furthermore allopolyploidy of the African species mayhave facilitated the dispersal of Isodon from Asia to Africa butother organisms showing similar distribution pattern should bestudied to find further evidence for the overland migration


hypothesis In addition a rapid radiation of Isodon was dated to thelate Miocene which corresponds well with one of the major upliftsof the QTP and subsequent aridification events Combined withprevious studies our results indicate that bursts of speciationtriggered by geologicalclimatic changes on the QTP may be a com-mon phenomenon and of significant importance in generating thecurrent high levels of biodiversity seen within this region

Acknowledgments

We would like to thank J-S Zhong L Li Y-F Chang X-Q Ci HLi and S Suddee for their assistance in sample collection andespecially E Kapinos and Z Davis for their help for providingDNA samples from the Royal Botanic Gardens Kew We also thankP Soltis and D Soltis for their constructive suggestions and helpwith revising the manuscript We are grateful to the curators andstaff at KUN for allowing us to examine and study their specimensThis work was supported by the National Natural Science Founda-tion of China (Grant No 31370245) and the Science Foundation ofthe Chinese Academy of Sciences 135 program XTBG-T01

Appendix A Supplementary material

Supplementary data associated with this article can be found inthe online version at httpdxdoiorg101016jympev201404017

References

Aagaard JE Olmstead RG Willis JH Phillips PC 2005 Duplication of floralregulatory genes in the Lamiales Am J Bot 92 1284ndash1293

Akaike H 1974 A new look at the statistical model identification IEEE TransAutomat Contr 19 716ndash723

An ZS Kutzbach JE Prell WL Porter SC 2001 Evolution of Asian monsoonsand phased uplift of the Himalaya-Tibetan plateau since Late Miocene timesNature 411 62ndash66

Bartov Y Steinitz G Eyal M Eyal Y 1980 Sinistral movement along the Gulf ofAqabamdashits age and relation to the opening of the Red Sea Nature 285 220ndash222

Belfiore NM Liu L Moritz C 2008 Multilocus phylogenetics of a rapid radiationin the genus Thomomys (Rodentia Geomyidae) Syst Biol 57 294ndash310

Chase MW Knapp S Cox AV Clarkson JJ Butsko Y Joseph J Savolainen VParokonny AS 2003 Molecular systematics GISH and the origin of hybrid taxain Nicotiana (Solanaceae) Ann Bot 92 107ndash127

Coleman M Hodges K 1995 Evidence for Tibetan plateau uplift before 14 Myrago from a new minimum age for east-west extension Nature 374 49ndash52

Conti E Eriksson T Schonenberger J Sytsma KJ Baum DA 2002 Early tertiaryout-of-India dispersal of Crypteroniaceae evidence from phylogeny andmolecular dating Evolution 56 1931ndash1942

Davis CC Bell CD Fritsch PW Mathews S 2002 Phylogeny of Acridocarpus-Brachylophon (Malpighiaceae) implications for Tertiary tropical floras andAfroasian biogeography Evolution 56 2395ndash2405

Doyle JJ Doyle JL 1987 A rapid DNA isolation procedure for small quantities offresh leaf tissue Phytochem Bull 19 11ndash15

Drew BT Sytsma KJ 2012 Phylogenetics biogeography and staminal evolutionin the tribe Mentheae (Lamiaceae) Am J Bot 99 933ndash953

Drew BT Sytsma KJ 2013 The South American radiation of Lepechinia(Lamiaceae) phylogenetics divergence times and evolution of dioecy Bot JLinn Soc 171 171ndash190

Drummond A Rambaut A 2007 BEAST Bayesian evolutionary analysis bysampling trees BMC Evol Biol 7 214

Drummond AJ Ho SYW Phillips MJ Rambaut A 2006 Relaxed phylogeneticsand dating with confidence PLoS Biol 4 e88

Erkens RHJ Chatrou LW Couvreur TLP 2012 Radiations and key innovationsin an early branching angiosperm lineage (Annonaceae Magnoliales) Bot JLinn Soc 169 117ndash134

Farris JS Kaumlllersjouml M Kluge AG Bult C 1995 Testing significance ofincongruence Cladistics 10 315ndash319

Fior S Li M Oxelman B Viola R Hodges SA Ometto L Varotto C 2013Spatiotemporal reconstruction of the Aquilegia rapid radiation through next-generation sequencing of rapidly evolving cpDNA regions New Phytol 198579ndash592

Friis I Vollesen K 2005 Flora of the Sudan-Uganda border area east of the NileBiol Skrif 51 465ndash466

Frohlich MW Meyerowitz EM 1997 The search for flower homeotic genehomologs in basal angiosperms and Gnetales a potential new source of data onthe evolutionary origin of flowers Int J Plant Sci 158 131ndash142

Garfunkel Z Bartov Y 1977 The Tectonics of the Suez rift Geological survey ofIsrael Jerusalem

Harley RM Atkins S Budantsev AL Cantino PD Conn BJ Grayer R HarleyMM De Kok R Krestovskaja T Morales R Paton AJ Ryding O Upson T2004 Labiatae In Kadereit JW (Ed) The families and Genera of VascularPlants VI (Lamiales) Springer Berlin pp 167ndash275

Harmon LJ Weir JT Brock CD Glor RE Challenger W 2008 GEIGERinvestigating evolutionary radiations Bioinformatics 24 129ndash131

Harrison TM Copeland P Kidd W Yin A 1992 Raising Tibet Science 2551663ndash1670

Ho SYW Phillips MJ 2009 Accounting for calibration uncertainty inphylogenetic estimation of evolutionary divergence times Syst Biol 58367ndash380

Huang SS 2011 Studies on chromosome number of four original species ofChinese medicine Xihuangcao J Trop Subtrop Bot 19 374ndash376

Hughes C Eastwood R 2006 Island radiation on a continental scale exceptionalrates of plant diversification after uplift of the Andes Proc Natl Acad Sci USA103 10334ndash10339

Hughes CE Pennington RT Antonelli A 2013 Neotropical plant evolutionassembling the big picture Bot J Linn Soc 171 1ndash18

Huson DH Scornavacca C 2012 Dendroscope 3 an interactive tool for rootedphylogenetic trees and networks Syst Biol 61 1061ndash1067

Jian S Soltis PS Gitzendanner MA Moore MJ Li R Hendry TA Qiu YLDhingra A Bell CD Soltis DE 2008 Resolving an ancient rapid radiation inSaxifragales Syst Biol 57 38ndash57

Jin ZM Sha W 2004 The karyotype study on Isodon japonica var glaucocalyx andLeonurus japonicus Guangxi Sci 11 78ndash80

Kar R 1996 On the Indian origin of Ocimum (Lamiaceae) a palynologicalapproach Palaeobotanist 43 43ndash50

Katoh K Kuma K Toh H Miyata T 2005 MAFFT version 5 improvement inaccuracy of multiple sequence alignment Nucleic Acids Res 33 511ndash518

Kizaki K Oshiro I 1977 Paleogeography of the Ryukyu Islands Mar Sci Monthly9 542ndash549

Kosuch J Vences M Dubois A Ohler A Bohme W 2001 Out of Asiamitochondrial DNA evidence for an oriental origin of tiger frogs genusHoplobatrachus Mol Phylogenet Evol 21 398ndash407

Kovarik A Pires JC Leitch AR Lim KY Sherwood A Matyasek R Rocca JSoltis D Soltis P 2005 Rapid concerted evolution of nuclear ribosomal DNA intwo Tragopogon allopolyploids of recent and recurrent origin Genetics 169931ndash944

Kulju KKM Sierra SEC Draisma SGA Samuel R van Welzen PC 2007Molecular phylogeny of Macaranga Mallotus and related genera(Euphorbiaceae ss) insights from plastid and nuclear DNA sequence dataAm J Bot 94 1726ndash1743

Ladiges PY Marks CE Nelson G 2011 Biogeography of Nicotiana sectionSuaveolentes (Solanaceae) reveals geographical tracks in arid Australia JBiogeogr 38 2066ndash2077

Li HW 1988 Taxonomic review of Isodon (Labiatae) J Arnold Arbor 69 289ndash400Li L Li J Rohwer JG van der Werff H Wang ZH Li HW 2011 Molecular

phylogenetic analysis of the Persea group (Lauraceae) and its biogeographicimplications on the evolution of tropical and subtropical Amphi-Pacificdisjunctions Am J Bot 98 1520ndash1536

Liu JQ Gao TG Chen ZD Lu AM 2002 Molecular phylogeny and biogeographyof the Qinghai-Tibet Plateau endemic Nannoglottis (Asteraceae) MolPhylogenet Evol 23 307ndash325

Liu JQ Wang YJ Wang AL Hideaki O Abbott RJ 2006 Radiation anddiversification within the Ligularia-Cremanthodium-Parasenecio complex(Asteraceae) triggered by uplift of the Qinghai-Tibetan Plateau MolPhylogenet Evol 38 31ndash49

Liu K Warnow TJ Holder MT Nelesen SM Yu J Stamatakis AP Linder CR2012 SATeacute-II very fast and accurate simultaneous estimation of multiplesequence alignments and phylogenetic trees Syst Biol 61 90ndash106

Lu L Fritsch PW Cruz BC Wang H Li DZ 2010 Reticulate evolution crypticspecies and character convergence in the core East Asian clade of Gaultheria(Ericaceae) Mol Phylogenet Evol 57 364ndash379

Lysak MA Berr A Pecinka A Schmidt R McBreen K Schubert I 2006Mechanisms of chromosome number reduction in Arabidopsis thaliana andrelated Brassicaceae species Proc Natl Acad Sci USA 103 5224ndash5229

Mallet J 2007 Hybrid speciation Nature 446 279ndash283Mao KS Milne RI Zhang LB Peng YL Liu JQ Thomas P Mill RR Renner

SS 2012 Distribution of living Cupressaceae reflects the breakup of PangeaProc Natl Acad Sci USA 109 7793ndash7798

Marks CE Ladiges PY 2011 Comparative morphology and phylogeny ofNicotiana section Suaveolentes (Solanaceae) in Australia and the South PacificAust Syst Bot 24 61ndash86

Martiacutenez-Millaacuten M 2010 Fossil record and age of the Asteridae Bot Rev 7683ndash135

Mayrose I Barker MS Otto SP 2010 Probabilistic models of chromosomenumber evolution and the inference of polyploidy Syst Biol 59 132ndash144

Menzies MA Baker J Bosence D Dart C Davison I Hurford A AlrsquoKadasi MMcClay K Nichols G AlrsquoSubbary A 1992 The timing of magmatism upliftand crustal extension preliminary observations from Yemen In Storey B(Ed) Magmatism and Continental Break Up Geological Society London pp293ndash304

Molnar P England P Martinod J 1993 Mantle dynamics uplift of the TibetanPlateau and the Indian monsoon Rev Geophys 31 357ndash396


Morton JK 1962 Cytotaxonomic studies on the west African Labiatae J Linn SocLond Bot 58 231ndash283

Morton JK 1993 Chromosome numbers and polyploidy in the flora of CameroonsMountain Opera Bot 159ndash172

Morton JK 1998 New names in Plectranthus (Lamiaceae) and allied genera fromthe Ethiopian Region Novon 8 265ndash266

Muellner AN Savolainen V Samuel R Chase MW 2006 The mahogany familylsquolsquoout-of-Africarsquorsquo divergence time estimation global biogeographic patternsinferred from plastid rbcL DNA sequences extant and fossil distribution ofdiversity Mol Phylogenet Evol 40 236ndash250

Mukherjee M Datta AK Maiti GG 2005 Chromosome number variation inOcimum basilicum L Cytologia 70 455ndash458

Murata G Yamazaki T 1993 Isodon In Iwatsuki K Yamazaki T Boufford DEOhba H (Eds) Flora of Japan Iiia Kodansha Tokyo pp 309ndash314

Nie ZL Deng T Meng Y Sun H Wen J 2013 Post-Boreotropical dispersalsexplain the pantropical disjunction in Paederia (Rubiaceae) Ann Bot 111873ndash886

Nie ZL Sun H Beardsley PM Olmstead RG Wen J 2006 Evolution ofbiogeographic disjunction between eastern Asia and eastern North America inPhryma (Phrymaceae) Am J Bot 93 1343ndash1356

Paradis E Claude J Strimmer K 2004 APE analyses of phylogenetics andevolution in R language Bioinformatics 20 289ndash290

Paton AJ Putievsky E 1996 Taxonomic problems and cytotaxonomicrelationships between and within varieties of Ocimum basilicum and relatedspecies (Labiatae) Kew Bull 51 509ndash524

Popp M Gizaw A Nemomissa S Suda J Brochmann C 2008 Colonization anddiversification in the African lsquosky islandsrsquo by Eurasian Lychnis L(Caryophyllaceae) J Biogeogr 35 1016ndash1029

Posada D 2008 JModelTest phylogenetic model averaging Mol Biol Evol 251253ndash1256

Roumlgl F 1998 Palaeogeographic considerations for Mediterranean and Paratethysseaways (Oligocene to Miocene) Ann Naturhist Mus Wien 99 279ndash310

Roumlgl F 1999 Circum-Mediterranean Miocene paleogeography In Roumlssner GHeissig K (Eds) The Miocene Land Mammals of Europe Dr Fritz Pfeil VerlagMunich pp 39ndash48

Rabosky DL 2006 LASER a maximum likelihood toolkit for detecting temporalshifts in diversification rates from molecular phylogenies Evol BioinformOnline 2 247

Rambaut A Drummond A 2007 Tracer 15 lthttpbeastbioedacukTracergtRambaut A Drummond A 2010 FigTree 13 1 lthttptreebioedacuksoftware

figtreegtRaven PH Axelrod DI 1974 Angiosperm biogeography and past continental

movements Ann Mo Bot Gard 61 539ndash673Ree RH Moore BR Webb CO Donoghue MJ 2005 A likelihood framework for

inferring the evolution of geographic range on phylogenetic trees Evolution 592299ndash2311

Ree RH Smith SA 2008 Maximum likelihood inference of geographic rangeevolution by dispersal local extinction and cladogenesis Syst Biol 57 4ndash14

Reid EM Chandler MEJ 1926 Catalogue of Cainzoic plants in the Department ofGeology vol 1 The Brembridge flora British Museum (Natural History)London

Richardson JE Pennington RT Pennington TD Hollingsworth PM 2001 Rapiddiversification of a species-rich genus of neotropical rain forest trees Science293 2242ndash2245

Rokas A Melika G Abe Y Nieves-Aldrey JL Cook JM Stone GN 2003Lifecycle closure lineage sorting and hybridization revealed in a phylogeneticanalysis of European oak gallwasps (Hymenoptera Cynipidae Cynipini) usingmitochondrial sequence data Mol Phylogenet Evol 26 36ndash45

Ronquist F Huelsenbeck JP 2003 MrBayes 3 Bayesian phylogenetic inferenceunder mixed models Bioinformatics 19 1572ndash1574

Russell A Samuel R Klejna V Barfuss MHJ Rupp B Chase MW 2010Reticulate evolution in diploid and tetraploid species of Polystachya(Orchidaceae) as shown by plastid DNA sequences and low-copy nucleargenes Ann Bot 106 37ndash56

Ryding O 1993 A reconsideration of the genus Rabdosiella (LamiaceaeNepetoideae Ocimeae) Plant Syst Evol 185 91ndash97

Sauquet H Ho SYW Gandolfo MA Jordan GJ Wilf P Cantrill DJ Bayly MJBromham L Brown GK Carpenter RJ 2012 Testing the impact of calibrationon molecular divergence times using a fossil-rich group the case of Nothofagus(Fagales) Syst Biol 61 289ndash313

Schatz G 1996 MalagasyIndo-australo-malesian phytogeographic connectionsIn Lourenco WR (Ed) Biogeography of Madagascar Orstom Paris pp 73ndash83

Schuster M Duringer P Ghienne JF Vignaud P Mackaye HT Likius A BrunetM 2006 The age of the Sahara desert Science 311 821

Shavit L Penny D Hendy MD Holland BR 2007 The problem of rooting rapidradiations Mol Biol Evol 24 2400ndash2411

Shi YF Li JJ Li BY Yao TD Wang SM Li SJ Cui ZJ Wang FB Pan BTFang XM Zhang QS 1999 Uplift of the Qinghai-Xizang (Tibetan) plateau andeast Asia environmental change during late Cenozoic Acta Geograph Sin 5410ndash21

Spicer RA Harris NBW Widdowson M Herman AB Guo S Valdes PJ WolfeJA Kelley SP 2003 Constant elevation of southern Tibet over the past 15million years Nature 421 622ndash624

Stebbins GL 1985 Polyploidy hybridization and the invasion of new habitatsAnn Mo Bot Gard 824ndash832

Stewart CB Disotell TR 1998 Primate evolutionndashin and out of Africa Curr Biol8 R582ndashR588

Sun YS Wang AL Wan DS Wang Q Liu JQ 2012 Rapid radiation of Rheum(Polygonaceae) and parallel evolution of morphological traits Mol PhylogenetEvol 63 150ndash158

Swofford DL 2003 Phylogenetic analysis using parsimony (frasland other methods)version 40b10 Sinauer Associates Sunderland Massachusetts

Tamura K Peterson D Peterson N Stecher G Nei M Kumar S 2011 MEGA5molecular evolutionary genetics analysis using maximum likelihoodevolutionary distance and maximum parsimony methods Mol Biol Evol 282731ndash2739

van Welzen PC Strijk JS van Konijnenburg-van Cittert JH Nucete M MerckxVS 2014 Dated phylogenies of the sister genera Macaranga and Mallotus(Euphorbiaceae) congruence in historical biogeographic patterns PLoS ONE 9e85713

Velasco JD 2008 The prior probabilities of phylogenetic trees Biol Philos 23455ndash473

Wen J 1999 Evolution of eastern Asian and eastern North American disjunctdistributions in flowering plants Annu Rev Ecol Syst 30 421ndash455

Wu ZY 1988 Hengduan mountain flora and her significance J Jpn Bot 63297ndash311

Wu CY Li HW (Eds) 1977 Labiatae Flora Reipublicae Popularis Sinicae 65 (2)66 Science Press Beijing

Xi Z Ruhfel BR Schaefer H Amorim AM Sugumaran M Wurdack KJEndress PK Matthews ML Stevens PF Mathews S 2012 Phylogenomicsand a posteriori data partitioning resolve the Cretaceous angiosperm radiationMalpighiales Proc Natl Acad Sci USA 109 17519ndash17524

Yamashiro T Suzuki K Masayuki M 2005 Chromosome Numbers of Isodon(Lamiaceae) in Japan Acta Phytotax Geobot 56 241ndash246

Yin A 2010 Cenozoic tectonic evolution of Asia a preliminary synthesisTectonophysics 488 293ndash325

Yu Y Harris AJ He XJ 2010 S-DIVA (Statistical Dispersal-Vicariance Analysis) atool for inferring biogeographic histories Mol Phylogenet Evol 56 848ndash850

Yu Y Harris AJ He XJ 2011 RASP (Reconstruct Ancestral State in Phylogenies)21 beta lthttpmnhscueducnsoftblogRASPgt

Yuan YM Wohlhauser S Moumlller M Klackenberg J Callmander MW Kuumlpfer P2005 Phylogeny and biogeography of Exacum (Gentianaceae) a disjunctivedistribution in the Indian Ocean Basin resulted from long distance dispersal andextensive radiation Syst Biol 54 21ndash34

Yue JP Sun H Baum DA Li JH Al-Shehbaz IA Ree RH 2009 Molecularphylogeny of Solms-laubachia (Brassicaceae) sl based on multiple nuclear andplastid DNA sequences and its biogeographic implications J Syst Evol 47402ndash415

Zhang ML Fritsch PW 2010 Evolutionary response of Caragana (Fabaceae) toQinghai-Tibetan Plateau uplift and Asian interior aridification Plant Syst Evol288 191ndash199

Zhong JS Li J Li L Conran JG Li HW 2010 Phylogeny of Isodon (Schrad exBenth) Spach (Lamiaceae) and Related Genera Inferred from Nuclear RibosomalITS trnL-trnF Region and rps16 Intron Sequences and Morphology Syst Bot 35207ndash219

Zhou LL Su YCF Thomas DC Saunders RMK 2011 lsquoOut-of-Africarsquo dispersalof tropical floras during the Miocene climatic optimum evidence from Uvaria(Annonaceae) J Biogeogr 39 322ndash335

Zhou SL Pan KY Hong DY 1999 Nutlet dispersal of Mosla hangchouensisMatsuda (Labiatae) Guihaia 19 176ndash179


radiation as inferred by short internal branch lengths (Richardsonet al 2001) and large polytomies (low DNA sequence divergenceand poor resolution) has been reported for many plant groupssuch as Saxifragales (Jian et al 2008) and Malpighiales (Xi et al2012) Research investigating rapid radiations of species-rich plantgroups in the QTP area may thus enrich our understanding of theevolutionary origins of biodiversity (Erkens et al 2012)

Intercontinental disjunct distributions between closely relatedspecies are a remarkable feature of angiosperm biogeography(Raven and Axelrod 1974) Disjunct distribution patterns fromthe Northern Hemisphere especially the Eastern AsiandashEasternNorth America disjunctions have been investigated extensively(Wen 1999 Nie et al 2006) In contrast mechanisms responsiblefor Southern Hemisphere disjunctions (eg AsiandashAfrica disjunc-tion) have received far less attention (eg Yuan et al 2005 Zhouet al 2011) The (Africandash) ArabiandashEurasia collision was consideredequally important to the IndiandashEurasia collision (Yin 2010) andmight have played a significant role in floral and faunal exchangebetween Africa and Asia

Overland migration between Africa and Eurasia across the Ara-bian Peninsula has been proposed for Hoplobatrachus Peters 1863(Amphibia Ranidae) through a connection between the ArabianPeninsula and Asia in the Miocene (Kosuch et al 2001) LychnisL (Caryophyllaceae) from Eurasia to the Ethiopian highlands viathe Arabian Peninsula (Popp et al 2008) and Uvaria L (Annona-ceae) which has been associated with the lsquoout-of-Africarsquo dispersaland radiation of primates during the Miocene (Zhou et al 2011)Other studies (eg Kulju et al 2007) have also discussed the pos-sibility of overland dispersal between Africa and Asia Howevercompeting hypotheses also exist the lsquoIndian raftingrsquo hypothesis(during the Cretaceous) (Conti et al 2002) dispersal of high-lati-tude boreotropical floras (from the late Paleocene to middleEocene) (Davis et al 2002) and trans-oceanic long-distance dis-persal (Yuan et al 2005) are three other mechanisms commonlyinvoked to explain AsiandashAfrica disjunctions

Isodon (Schrad ex Benth) Spach (tribe Ocimeae Dumort sub-family Nepetoideae) is a genus of ca 100 species in Lamiaceae(Harley et al 2004) The genus is distributed primarily in tropicaland subtropical Asia with a center of species diversity (ca 70) inthe Hengduan Mountains region (HMR southeastern corner of theQTP) of south-west China The genus also ranges west to Afghani-stan and Pakistan with a further two disjunct species I ramosissi-mus (Wall ex Benth) Murata and I schimperi (Vatke) JK Morton incentral and southern Africa Isodon includes shrubs subshrubs andperennial herbs with paniculate inflorescences composed of many-flowered cymes The corollas of Isodon species can differ in sizetube length color and the color of spots on the lips possibly asan adaptive response to pollinators The majority of Isodon speciesoccurring in the HMR region occupy cold and arid highland habi-tats but a few species extend to humid and warm tropical regionsIn contrast the two African Isodon species grow in more humidhabitats such as the margins of evergreen forests or in riparianzones (pers obs Li 1988)

Li (1988) monographed the genus in China dividing it intofour sections based on morphological characters such as inflores-cence density and the shape of the fruiting calyx Later the twoAfrican species were added (Ryding 1993 Morton 1998) buttheir molecular phylogenetic position within Isodon has not beentested In addition there was apparently at least one chromo-some number shift in Isodon the African species I ramosissimushas a chromosome number of 2n = 42 and a base chromosomenumber proposed to be x = 7 (Morton 1962 1993) or x = 14(Yamashiro et al 2005) In contrast the Asian species havenumbers of 2n = 24 or 2n = 36 with a suggested base chromo-some number of x = 12 (Jin and Sha 2004 Yamashiro et al2005 Huang 2011)

Our previous research focused mainly on the relationshipbetween Isodon and related genera and the two African specieswere not involved in (Zhong et al 2010) Accordingly in this studywe aim to obtain a more comprehensive phylogenetic reconstruc-tion of Isodon with the two African species included to illuminatethe evolution of this species-rich genus We are also using Isodonas a case study to understand better apparently rapid species radi-ations in the Hengduan Mountains region of south-west China andthe possible mechanisms triggering AsiandashAfrica disjunctionsMolecular phylogenetic approaches are used to reconstruct thepossible historical biogeography of Isodon and investigate thebroad topics mentioned above as well as to answer the followingspecific questions

1 Does the radiation of Isodon correlate with the suggestedgeological hypotheses concerning the QTP region and associ-ated climate change

2 What is the phylogenetic position of the two endemic Africanspecies with ostensibly aberrant chromosome number

3 If the two endemic African species are confirmed to bemembers of Isodon what mechanisms might be responsiblefor this AsiandashAfrica disjunction

2 Materials and methods

21 Taxon sampling

The classification of Isodon largely follows Li (1988) but eightspecies [Isodon anisochilus CY Wu I brachythyrsus CY Wu etHW Li I forrestii var intermedius CY Wu et HW Li I kunmingen-sis CY Wu et HW Li I pluriflorus CY Wu et HW Li I polystachys(Sun ex C H Hu) CY Wu et HW Li I setschwanensis var yungsh-engensis CY Wu et HW Li and I taliensis (CYWu) Hara] weretreated as distinct taxa following Wu and Li (1977) in order to rep-resent the species diversity of this genus better Isodon umbrosus(Maxim) H Hara and its varieties which were not included in Li(1988) were added following Flora of Japan (Murata andYamazaki 1993) and the inclusion of I oreophilus (Diels) AJ Patonet Ryding follows Zhong et al (2010) Isodon sp aff flavidus (Hand-Mazz) H Hara was treated as a separate taxonomic entity due toits allopatric distribution with respect to I flavidus (Hand -Mazz)H Hara and the samples labelled I eriocalyx (Dunn) Kudocirc 214 and Ihenryi (Hemsl) Kudocirc 232 were included as putative hybrids Afurther six specimens that could not be identified accurately toknown taxa were also treated as separate Isodon OTUs (operationaltaxonomic units) in the analyses

Samples of DNA for the two African species were obtained fromthe DNA Bank of the Royal Botanic Gardens Kew In total 73 acces-sions representing 71 species of Isodon were included in this studyThe complete list of voucher specimen numbers collection locali-ties distribution codes available chromosome numbers and Gen-Bank accession numbers of samples used are given inSupplementary Table S1 nrITS sequences (Supplementary mate-rial Table S2) for 26 Isodon species (based on different vouchers)sequenced in our previous study were also added to the presentdata set Voucher specimens of samples from Africa and Japanare deposited at the herbarium of the Royal Botanic GardensKew (K) and the herbarium of Botanical Gardens in Tohoku Univer-sity of Japan (TUS) respectively Others are deposited at Laboratoryof Plant Phylogenetics and Conservation Xishuangbanna TropicalBotanical Garden Chinese Academy of Sciences (XTBG) To aug-ment our sampling we also downloaded 21 nrITS sequences fromother Ocimeae genera from GenBank (Supplementary materialTable S2) Similarly we also downloaded 39 rpl32ndashtrnL sequencesfrom various genera in Lamiaceae (Supplementary materialTable S3)



























3 Results





1

1

1

1

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

095

09

098

099

097

092095

091

100

100100

100

100

95

100

73

56

80

9997

78

93

77

9199

100

78

8178

91

97

97

59

88

8159

82

58

97

63

64

9261

88

Outgroup

IV

II

I

III
















I nervosusI sp 6










5 changes

1

1

11

1

091

1

1

1

1

1

1

1

1

100

92

100100

100

99

100

98

75

92

98

100

84

98

Outgroup

I

IIIV

III





















(A)

(B)







15 10 5 0

12

510

2050


Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage


nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II








4 Discussion

























5 Conclusions




Acknowledgments




References




































































































































3 Results





1

1

1

1

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

095

09

098

099

097

092095

091

100

100100

100

100

95

100

73

56

80

9997

78

93

77

9199

100

78

8178

91

97

97

59

88

8159

82

58

97

63

64

9261

88

Outgroup

IV

II

I

III
















I nervosusI sp 6










5 changes

1

1

11

1

091

1

1

1

1

1

1

1

1

100

92

100100

100

99

100

98

75

92

98

100

84

98

Outgroup

I

IIIV

III





















(A)

(B)







15 10 5 0

12

510

2050


Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage


nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II








4 Discussion

























5 Conclusions




Acknowledgments




References











































































































1

1

1

1

1

11

1

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

095

09

098

099

097

092095

091

100

100100

100

100

95

100

73

56

80

9997

78

93

77

9199

100

78

8178

91

97

97

59

88

8159

82

58

97

63

64

9261

88

Outgroup

IV

II

I

III
















I nervosusI sp 6










5 changes

1

1

11

1

091

1

1

1

1

1

1

1

1

100

92

100100

100

99

100

98

75

92

98

100

84

98

Outgroup

I

IIIV

III





















(A)

(B)







15 10 5 0

12

510

2050


Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage


nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II








4 Discussion

























5 Conclusions




Acknowledgments




References





















































































































(A)

(B)







15 10 5 0

12

510

2050


Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage


nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II








4 Discussion

























5 Conclusions




Acknowledgments




References










































































































15 10 5 0

12

510

2050


Num

ber o

f lin

eage

s

cpDNA

15 10 5 0

12

510

2050

Num

ber o

f lin

eage


nrITS

15 10 5 0

12

510

2050

Num

ber o

f lin

eage

s


LEAFY intron II








4 Discussion

























5 Conclusions




Acknowledgments




References





























































































































5 Conclusions




Acknowledgments




References












































































































Acknowledgments




References










































































































Phylogeny and historical biogeography of Isodon (Lamiaceae): Rapid radiation in south-west China and...

Documents

Transcript of Phylogeny and historical biogeography of Isodon (Lamiaceae): Rapid radiation in south-west China and...