JPlantRes(2015)128:73–90 DOI10.1007/s10265-014-0666-7 REGULAR PAPER Progressive migration and anagenesis in Drimys confertifolia of the Juan Ferna´ndez Archipelago, Chile Patricio Lo´pez-Sepu´lveda • Koji Takayama • Josef Greimler • Daniel J. Crawford • Patricio Pen˜ailillo • Marcelo Baeza • Eduardo Ruiz • Gudrun Kohl • Karin Tremetsberger • Alejandro Gatica • Luis Letelier • Patricio Novoa • Johannes Novak • Tod F. Stuessy Received:19December2013/Accepted:12June2014/Publishedonline:8October2014 (cid:2)TheAuthor(s)2014.ThisarticleispublishedwithopenaccessatSpringerlink.com Abstract A common mode of speciation in oceanic populations of D. confertifolia and the continental species islandsisbyanagenesis,whereinanimmigrantarrivesand D. winteri and D. andina, and to test probable migration through time transforms by mutation, recombination, and routes between the major islands. Population genetic drift into a morphologically and genetically distinct spe- analyses were conducted using AFLPs and nuclear cies, with the new species accumulating a high level of microsatellites of 421 individuals from 42 populations genetic diversity. We investigate speciation in Drimys fromtheJuanFerna´ndezislandsandthecontinent.Drimys confertifolia, endemicto the two major islands ofthe Juan confertifolia shows a wide genetic variation within popu- Ferna´ndez Archipelago, Chile, to determine genetic con- lations on both islands, and values of genetic diversity sequences of anagenesis, to examine relationships among within populations are similar to those found within pop- ulations of the continental progenitor. The genetic results arecompatiblewiththehypothesisofhighlevelsofgenetic Electronicsupplementarymaterial Theonlineversionofthis variation accumulating within anagenetically derived spe- article(doi:10.1007/s10265-014-0666-7)containssupplementary cies inoceanic islands, and with the conceptoflittle or no material,whichisavailabletoauthorizedusers. P.Lo´pez-Sepu´lveda(cid:2)M.Baeza(cid:2)E.Ruiz A.Gatica DepartamentodeBota´nica,UniversidaddeConcepcio´n, LaboratoriodeEcofisiolog´ıaVegetal,DepartamentodeBiolog´ıa, Casilla160-C,Concepcio´n,Chile FacultaddeCiencias,UniversidaddeLaSerena, Casilla599,LaSerena,Chile K.Takayama TheUniversityMuseum,TheUniversityofTokyo, L.Letelier Hongo7-3-1,Bunkyo-ku,Tokyo113-0033,Japan CentrodeInvestigacionesenEcosistemas,UniversidadNacional Auto´nomadeMe´xico,C.P.58190Morelia,Michoaca´n,Mexico J.Greimler(cid:2)G.Kohl DepartmentofSystematicandEvolutionaryBotany, P.Novoa BiodiversityCenter,UniversityofVienna,Rennweg14, Jard´ınBota´nicodeVin˜adelMar,Corporacio´nNacionalForestal, 1030Vienna,Austria CaminoElOlivar305,Vin˜adelMar,Chile D.J.Crawford J.Novak DepartmentofEcologyandEvolutionaryBiologyandthe InstituteforAppliedBotanyandPharmacognosy,Universityof BiodiversityInstitute,UniversityofKansas,Lawrence, VeterinaryMedicine,Veterina¨rplatz1,1210Vienna,Austria KS60045,USA T.F.Stuessy(&) P.Pen˜ailillo Herbarium,DepartmentofEvolution,Ecology,andOrganismal InstitutodeBiolog´ıaVegetalyBiotecnolog´ıa,Universidadde Biology,TheOhioStateUniversity,1315KinnearRoad, Talca,2Norte685,Talca,Chile Columbus,OH43212,USA e-mail:[email protected] K.Tremetsberger DepartmentofIntegrativeBiologyandBiodiversityResearch, InstituteofBotany,UniversityofNaturalResourcesandLife Sciences,GregorMendelStraße33,1180Vienna,Austria 123 74 JPlantRes(2015)128:73–90 geographical partitioning of this variation over the land- and mutation. The final result is a new species that differs scape. Analysis of the probability of migration within the genetically and morphologically from its ancestor, with archipelago confirms colonization from the older island, levels of genetic variation approximating those of the Robinson Crusoe, tothe younger island Alejandro Selkirk. progenitor species (Stuessy et al. 2006). Examples of this type of speciation are less frequent, but it has been docu- Keywords AFLPs (cid:2) Anagenesis (cid:2) Genetic variation (cid:2) mented in Dystaenia (Apiaceae; Pfosser et al. 2006) and Microsatellites (cid:2) Oceanic islands (cid:2) Migration Acer (Sapindaceae; Takayama et al. 2012a, b) of Ullung Island,Korea,andinWeigela(Caprifoliaceae;Yamadaand Maki 2012) of the Izu Islands, Japan. The genetic conse- Introduction quences of anagenetic speciation show relatively high levels of genetic differentiation within and among popu- Patternsandprocessesofspeciationinoceanicislandshave lations of the island endemic relative to continental source long captured the attention of evolutionary biologists populations. Obviously, many factors impact levels of (Drake et al. 2002; Rosindell and Phillimore 2011; genetic variation within island populations, such as Schaefer et al. 2011; Stuessy and Ono 1998). Important breeding systems, island age, human impact, etc. (Stuessy attributes of oceanic islands, such as geographical isola- et al. 2013), but mode of speciation is particularly tion, clearly delimited area, and restricted fauna and flora, significant. have led to islands being regarded as natural laboratories Alsoimportantforunderstandingpatternsandprocesses for the study of evolution. They offer countless opportu- of evolution in oceanic islands is inferring routes of nities for investigating evolutionary processes in detail, migrationamongislandswithinarchipelagos.Theclassical especially for studying genetic, ecological, biogeographic, hypothesis regarding archipelagos has assumed a single reproductive, and morphological divergence (Moore et al. colonizationeventfromacontinentalareafirsttotheoldest 2002; Mort et al. 2002). island and subsequent colonization of the younger islands Numerous hypotheses and discussions regarding pro- (Juan et al. 2000; Gillespie and Roderick 2002), the so- cesses of speciation in oceanic islands have occurred called ‘‘progression rule’’ (Funk and Wagner 1995). (Carlquist1974;Grantetal.1996;Stuessyetal.2006).The Although this concept relies on parsimony, which is not mostcommonlydescribedspeciationmechanisminislands unreasonable, other possibilities have been demonstrated, is through cladogenesis. In this process, after a single suchasreversecolonization(BallemainandRicklefs2008; introduction, numerous lineages diverge rapidly from the Carineetal.2004),colonizationfollowedbyextinction,or foundingpopulationastheyadapttodifferenthabitatswith migration from younger to older islands (Emerson 2002; appropriate adaptations (Schluter 2000). The genetic con- Juanetal.2000).Obviouslyimportantalsoareavailability sequence of this process is low level of genetic variation oftransportationvectors(GillespieandRoderick2002)and within and among populations of each species (Baldwin adaptationanddispersalofpropagules(CowieandHolland et al. 1998; Crawford and Stuessy 1997; Emerson 2002; 2006). Johnson et al. 2000;Stuessy et al. 2006).Examples of this An appropriate group of islands in which to study ana- mechanismofdivergenceandspeciationinoceanicislands genetic speciation and migration is the Juan Ferna´ndez arenumerous,suchasAeonium(Crassulaceae)andEchium Archipelago, located in the Pacific Ocean 667 km W of (Boraginaceae) in the Canary Islands (Bo¨hle et al. 1996; continentalChile(33(cid:3)S/78–80(cid:3)W,Fig. 1).Thearchipelago Jorgensen and Olesen 2001), Dendroseris and Robinsonia consists of two main islands, Robinson Crusoe (=Masati- (Asteraceae)intheJuanFerna´ndezArchipelago(Crawford erra) and Alejandro Selkirk (=Masafuera), separated by et al. 1998), Bidens (Asteraceae), Schiedea (Caryophylla- 181 kms.Atpresenttheislandsareapproximatelyofequal ceae), Cyanea, Lobelia and Trematolobelia (Lobeliaceae) size (50 km2, Stuessy 1995), but they differ in geological in the Hawaiian Islands (Givnish et al. 2009; Knope et al. age, c. 4millionyearsoldfor RobinsonCrusoe Island and 2012; Price and Wagner 2004), and Scalesia (Asteraceae) 1–2millionyearsoldforAlejandroSelkirkIsland(Stuessy in the Gala´pagos Islands (Eliasson 1974; Schilling et al. et al. 1984). The native vascular flora of the archipelago 1994). includes 75 families, 213 genera, and 361 species, with a The other major type of speciation in oceanic islands is 12 % endemism at the generic level, and 60 % at the anagenesis, also called simple geographic or phyletic spe- specific level (Marticorena et al. 1998). ciation (Simpson 1953). In this case, after colonizers A suitable genus to study the genetic consequences of establish a population on a new island, the processes of anagenetic speciation and migration inthe Juan Ferna´ndez drift,recombination,andmutationmodifythecomposition Archipelago is Drimys (Winteraceae). The genus contains of the original pioneer population and over time genetic seven species distributed in Central and South America variation accumulates by the processes of recombination, (Ehrendorferetal.1979;Marqu´ınezetal.2009;Rodr´ıguez 123 JPlantRes(2015)128:73–90 75 Fig.1 Geographicalpositionof populationssampledofDrimys winteriandD.andinain continentalChile(a)and Drimysconfertifoliain RobinsonCrusoe(b)and AlejandroSelkirk(c)Islands and Quezada 2001; Smith 1943), two of which, D. andina the most probable migration route(s) of the endemic spe- and D. winteri (with two varieties, var. winteri and var. cies between islands; and (3) assess genetic consequences chilensis), grow in continental Chile, with the others dis- of island anagenetic speciation. junct in Brazil and northwestern South America (with extensionsnorthwardintoMexico).Drimysconfertifoliais endemic to the Juan Ferna´ndez Archipelago, common on Materials and methods Robinson Crusoe Island, and with a patchy distribution on Alejandro Selkirk Island. These three species comprise a Species closely related complex, as evidenced by very similar ITS sequences (Ruiz et al. 2008). Drimys confertifolia Phil., ‘‘Canelo’’ (Fig. 2a, b), is a In recent years the number of molecular markers protogynous (and therefore out-crossing) tree (Bernardello available to study genetic diversity and phylogeny in et al. 2001) to 15 m tall, endemic to the Juan Ferna´ndez islands has increased significantly (Emerson 2002). Archipelago. Chromosomally the species is known as Amplified Fragment Length Polymorphisms (AFLPs, Vos n = c. 43 (Sun et al. 1990), which is probably at the do- etal.1995),adominantmarker,hasbeenshowntoprovide decaploid level (based on x = 7, Raven and Kyhos 1965). agoodoverallmeasureofgeneticdiversityandstructureat This is the same level reported for D. winteri (Raven and the population level (Tremetsberger et al. 2003). Nuclear Kyhos1965)andD.granadensis(Ehrendorferetal.1979). microsatellites are co-dominant highly polymorphic In Robinson Crusoe Island D. confertifolia is common, markers that are widely used to study the genetic structure growing together with Myrceugenia fernandeziana (Myrt- ofpopulationsandmigrationroutes(Hardyetal.2006).In aceae), Fagara mayu (Rutaceae), and Juania australis this study, therefore, we selected both methodologies to (Arecaceae) (Greimler et al. 2002). In Alejandro Selkirk analyze the genetic consequences of anagenesis and Islanditoccursinpatchesorisscattered,notforminglarge migration in Drimys confertifolia. purestands,andgrowingtogetherwiththefernsBlechnum Theobjectivesofthispaperareto:(1)determinegenetic cycadifolium (Blechnaceae), Dicksonia externa (Dickson- relationships and variation within and among populations iaceae), Histiopteris incisa (Dennstaedtiaceae), and Lo- ofD.confertifolia,D.andina,andD.winteri;(2)establish phosoria quadripinnata (Dicksoniaceae) (Greimler et al. 123 76 JPlantRes(2015)128:73–90 Fig.2 Habitatsandflowers(insets)ofD.confertifoliaonRobinsonCrusoeIsland(a),D.confertifoliaonAlejandroSelkirkIsland(b),D.andina (c),andDrimyswinterivar.winteri(d) 2013). Drimys confertifolia is hermaphroditic and wind- Collection and DNA isolation pollinated (Bernardello et al. 2001), flowering from November to January (Rodr´ıguez and Quezada 2001). The species were collected during expeditions to the Juan The continental species included in this study are Ferna´ndez Archipelago in 2010 and 2011. Leaves of D. andina (Reiche) R.A.Rodr. et Quez. This species D. confertifolia were collected in silica gel from indi- (‘‘CaneloEnano’’Fig. 2c),growsasasmallshrubto1.5 m viduals of 16 populations on Robinson Crusoe Island tall, and is endemic to the subantarctic forest distributed (Nos. 1–16) and from 15 populations on Alejandro Sel- alongtheAndesandoccasionallyintheCoastalCordillera kirk Island (Nos. 17–31) (Fig. 1). In continental Chile, (37(cid:3)–41(cid:3) S) (Rodr´ıguez and Quezada 2001). The other samples came from one population of D. andina (No. 32), species is D. winteri J.R.Forst. et G.Forst. (Fig. 2d) with two populations of D. winteri var. winteri (Nos. 41 and tworecognizedvarieties,var.winteri(‘‘Canelo,foye’’)and 42), and eight populations of D. winteri var. chilensis var. chilensis (DC.) A.Gray (‘‘Canelo’’). The former vari- (Nos. 33–40) were collected (Fig. 1; Table 1). Voucher etyisatreeto17 m,andisfoundinthesubantarcticforests specimens of each population are deposited in the her- intheextremesouthernportionofChile(45(cid:3)440–55(cid:3)580 S) barium of the University of Vienna (WU). The DNeasy (Rodr´ıguezandQuezada2001).Thelattervarietyisalarge 96 Plant Kit (Qiagen, Hilden, Germany) was used for tree (to 20 m) and is common and endemic to continental extraction of DNA for AFLP and microsatellite analyses. Chile,growingfrom0to1,700 mintheCoastalCordillera Details of numbers of individuals and populations used and the Andes Mountains between 30(cid:3)200–46(cid:3)250 S (Rod- for AFLP and microsatellites, and their distributions, are r´ıguez and Quezada 2001). given in Table 1 and Fig. 1. 123 JPlantRes(2015)128:73–90 77 nteriand NLCA 0.75 0.50 0.50 0.63 0.75 0.88 0.63 0.63 0.63 1.00 0.75 0.63 0.50 0.63 1.50 1.00 0.75(±0.26) 0.50 0.63 0.38 0.25 0.38 0.00 0.13 0.13 0.13 0.00 0.00 0.50 0.50 0.88 0.38 0.50(±0.23) wivar. 0.078) 0.05) eri NPA 0.00 0.00 0.00 0.13 0.00 0.13 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.25 0.13 0.00 0.04(± 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.13 0.00 0.02(± nt wi 9) 5) 4 4 D. 0.0 0.0 dina, HE 0.63 0.64 0.61 0.65 0.64 0.64 0.61 0.57 0.49 0.65 0.52 0.62 0.60 0.66 0.65 0.64 0.62(± 0.48 0.46 0.31 0.27 0.41 0.19 0.42 0.25 0.36 0.38 0.44 0.54 0.50 0.48 0.41 0.47(± an 8) 7) 3 2 D. 0. 0. 3 3 5 9 6 0 5 6 7 9 4 2 4 5 1 4± 5 0 7 5 9 1 9 ± olia, AR 4.7 4.2 4.3 4.7 4.1 4.3 4.2 – 4.0 4.4 3.9 4.6 4.5 4.0 3.4 3.6 4.2( 3.2 3.5 – – – – 3.2 – – – – 3.2 2.8 3.3 3.7 3.3( nfertif 0 8 8 5 0 0 5 3 0 8 5 8 3 0 0 8 9±0.48) 5 3 3 8 5 8 3 0 5 8 5 5 0 3 5 3±0.50) o A 5 8 8 2 0 5 7 1 0 3 2 3 6 5 5 8 8( 2 1 6 8 2 3 1 5 2 3 2 2 0 1 2 7( c N 4. 4. 4. 5. 5. 5. 4. 3. 4. 5. 4. 4. 4. 5. 5. 4. 4. 4. 4. 1. 1. 2. 1. 3. 1. 2. 2. 2. 3. 4. 4. 3. 3. s my 1) 6) Dri 9* 5* 1 9* 9* 4 8* 0 2 6* 5 8* 3 6* 3 4 1±0.1 6 9* 00 14 5 00 4 00 0 1 14 6 9* 9* 7 1±0.0 of FIS 0.2 0.2 0.2 0.2 0.1 0.2 0.2 0.0 0.0 0.2 0.1 0.4 0.1 0.2 0.1 0.0 0.2( 0.1 0.1 –1. –0. 0.3 –1. 0.1 –1. 0.1 0.1 –0. 0.2 0.2 0.1 0.2 0.2( s pulation HE 66 67 64 68 66 67 64 68 52 68 53 66 62 68 67 66 64(±0.051) 51 48 63 33 49 38 44 50 43 46 58 59 52 49 44 50(±0.05) o u 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. p analysisin42 Microsatellites HN(cid:3)O 100.44 120.48 120.50 100.46 150.52 110.51 100.45 30.54 90.46 130.48 150.43 80.31 130.49 140.47 140.54 150.60 0.48(±0.063) 100.40 160.31 10.63 30.29 30.21 10.38 100.38 10.50 30.29 30.33 20.50 70.41 120.30 180.36 70.27 0.35(±0.05) e crosatellit NPB 0 2 0 1 0 0 1 0 0 1 0 0 0 0 0 1 0.40(±0.63) 0 0 0 0 0 1 0 0 1 0 0 0 0 1 0 0.14(±0.38) mi 1) 9) uclear RI 1.64 1.71 1.56 1.79 1.44 1.59 1.78 1.27 1.34 1.41 1.59 1.36 1.35 1.41 1.37 1.38 1.50(±1.6 1.69 2.02 0.00 1.60 1.88 0.00 1.69 0.00 2.06 1.58 1.47 2.14 2.17 1.59 1.05 1.76(±0.3 n LPand GDOL 26 25 25 26 23 27 28 23 23 24 25 23 24 23 23 22 24(±0.17) 21 25 00 25 19 00 20 00 19 21 24 23 23 21 18 22(±0.02) F A 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. A on 8) 99) based SDI 106.84 109.33 105.65 108.55 99.91 114.60 112.58 76.58 96.89 103.15 111.59 90.76 102.99 100.53 102.29 94.05 103.98(±6.8 86.01 108.87 0.00 49.91 54.71 0.00 85.89 0.00 53.25 59.49 47.83 90.75 92.61 96.56 73.05 90.53(±10. e c en 5) 4) g 8. 6. diver TNB 442 456 441 448 431 453 445 319 404 426 451 390 430 435 434 416 433.5(±1 398 474 220 302 328 154 409 258 327 319 272 431 440 426 325 414.7±4( d an 5) 2) Table1ValuesofgeneticdiversityDwinterichilensis.var. AFLPs NTaxonandNPPB(cid:3)vouchers D.confertifoliaRobinsonCrusoeIsland 1913011168.4 1913421270.7 1918531266.6 1917341066.9 1920951467.6 1910861171.7 1919171070.3 192908442.4 192849960.0 19227101267.4 19275111572.9 1919712855.2 19268131467.9 19256141465.2 19260151466.9 19248161463.6 Average66.8(±SD)(±4. D.confertifoliaAlejandroSelkirkIsland 2004117954.5 20014181773.8 200131910.0 1981320224.7 2000521328.8 200232210.0 19803231054.2 196382410.0 1981025328.1 2004926330.9 1961627223.7 1965228857.3 1965629961.2 19657301965.4 1965831845.1 58.8Average(±SD)(±9. 123 78 JPlantRes(2015)128:73–90 HNNEPALCA 0.480.251.00 0.630.000.75 0.660.001.63 0.590.000.75 0.580.000.75 0.560.131.00 0.660.251.63 0.670.001.75 0.570.131.13 0.550.000.50 0.520.001.00 0.600.051.09(±0.05)(±0.09)(±0.44) RINPBdiversityoverloci,rarityindex,Nofheterozygotes,numberofprivatePAFmoreindividuals,*significantvaluesIS Table1continued AFLPsMicrosatellites NHuHFNATaxonandNPPBTNBSDIAGDOLRINPBN(cid:3)(cid:3)OEISARvouchers D.andina 4324321660.438391.920.201.180130.370.490.174.133.92 D.winterichilensisvar. 4331455.634482.730.181.191130.380.650.41*4.633.83 4314341437.724360.360.140.780130.440.690.33*5.133.71 4301351247.030173.140.170.990110.530.620.074.134.24 4300361247.230270.040.161.031120.510.600.114.384.65 2426371145.829971.450.171.010110.440.580.144.254.54 4321381540.326461.320.140.760130.370.690.44*5.634.09 4330391853.032081.650.180.840180.490.690.27*6.003.75 240140858.8355101.190.241.041100.490.600.114.633.84 D.winteriwinterivar. 2402411041.727264.470.160.920100.580.58–0.053.753.53 2420421243.230366.760.150.970120.460.550.094.633.44 Average47.0300.373.310.170.950.30.470.620.194.713.96(±SD)(±6.9)(±34.5)(±12.39)(±0.03)(±0.13)(±0.48)(±0.06)(±0.05)(±0.16)(±0.69)(±0.41) Populationdatainitalics,withfourorfewerindividuals,werenotincludedinthestatisticalanalyses NNPPBTNBSDIAGDOLassignedpopulationnumber,numberofindividualsanalyzed,percentageofpolymorphicbands,totalnumberofbands,Shannondiversityindex,averagegene(cid:3)HFNAHHlobservedheterozygosity,expectedheterozygosity,inbreedingcoefficient,numberofallelesperlocus,allelicrichness,expectedproportionnumberofprivatebands,EISAREONnumberoflocallycommonalleles(freq.C5%)foundin25%orfewerpopulations,–nodatabecausethenumberwascalculatedwithpopulationsthatcontainedsevenoralleles,LCAP\(0.05)afterBonferronicorrection 123 JPlantRes(2015)128:73–90 79 Genetic markers 60 (cid:3)Cfor30 min.Theamplifiedfragmentswererunwitha size standard (GeneScan 600, Applied Biosystems, Foster For AFLPs we followed the protocol of Vos et al. (1995) City,CA,USA) onanautomated sequencer (ABI3130xl). with modifications by Tremetsberger et al. (2003). For GeneMarker ver. 1.85 (SoftGenetics LLC) was used for selection of primers, a trial was done with 85 primer scoring. combinations and four individuals from each of six popu- lationsrepresentativeofalltaxaandislands.Thefollowing Data analysis five primer combinations were selected: MseI–CTG/ EcoRI–ACA (FAM); MseI–CTC/EcoRI–ACA (FAM); For AFLPs the program ARLEQUIN 3.5.1.2 (Excoffier MseI–CAC/EcoRI–ATG (VIC); MseI–CAG/EcoRI–AAG et al. 2005) was used for calculating the total number of (VIC);andMseI–CTC/EcoRI–AGC(NED).Atotalof421 different phenotypes in each population and the average individuals was analyzed, 279 from D. confertifolia, 16 gene diversity over loci (AGDOL; the probability that fromD.andina,104fromD.winterivar.chilensis,and22 two homologous band sites, randomly chosen, are dif- from D. winteri var. winteri (for details see Table 1). ferent). The estimations of other genetic diversity Amplified fragments of DNA were run on an automated parameters by populations, i.e., percentage of poly- sequencer (ABI 3130xl, Applied Biosystems, CA, USA). morphic bands (PPB), total number of AFLP bands Fragments were scored using the program GeneMarker (TNB), and Shannon Diversity Index (SDI) (H = -R Sh ver. 1.85 (SoftGenetics LLC, PA, USA). The range for [p ln(p)] where p is the frequency of the ith band in i i i allele call was 150–510 base pairs. Samples with size the respective population based on all AFLP bands calibration below 90 % were manually adjusted. An auto- recorded) were performed using FAMD ver. 1.25 matic panel editor was generated for each selective primer (Schlu¨ter and Harris 2006). combination (Curtin et al. 2007) and then manually mod- With respect to genetic divergence parameters and ified. For analysis, the matrices generated for each primer population structure, the number of private bands (NPB) combination were combined into one matrix (Wooten and was calculated with FAMD ver. 1.25 (Schlu¨ter and Harris Tolley-Jordan 2009). Ten percent of the total individuals 2006),andtheRarityIndex(RI)(Scho¨nswetterandTribsch were replicated, the error rate being calculated as the ratio 2005)wasestimatedwithR-scriptAFLPdat(Ehrich2006). of number of fragment differences/total number of com- Pairwise F were calculated according to Weir and ST parisons (Bonin et al. 2004). Cockerham (1984), and the probabilities of random Nine nuclear microsatellites markers were selected and departure from zero for obtaining F were calculated ST isolatedfromD.confertifolia(Takayamaetal.2011)based using the exact test with 10,000 permutations in ARLE- on repeatability and scoring suitability. A total of 281 QUIN3.5.1.2.(Excoffieretal.2005).ANeighborNetalgo- individuals of D. confertifolia, 13 of D. andina, 101 of D. rithm (Bryant and Moulton 2004), implemented by the winterivar.chilensis,and22ofD.winterivar.winteriwas software SplitsTree4 ver. 4.10 (Huson and Bryant 2006), analyzed. For fluorescent labeling of PCR amplified frag- was executed using a Nei-Li distance matrix calculated ments, we used the 50-tailed primer method (Boutin- from the original AFLP matrix. An analysis of molecular Ganache et al. 2001) following Takayama et al. (2011). variance (AMOVA) with ARLEQUIN 3.5.1.2 (Excoffier Different dyes (6-FAM, NED, PET, and VIC) for four et al. 2005) was implemented for estimating genetic dif- combinations of multiplex PCR amplification were used, ferentiation among and within populations (hierarchical following a modified protocol of the Qiagen Multiplex structuring). For calculation of probabilities, 1,023 per- PCR Kit (Qiagen, Hilden, Germany). Four multiplex PCR mutations were used, which reveals significance of the reactions were done as follows: AWU5A, AOH4B with variance components. For appraisement of population 6-FAM, A3340, ATEAE with VIC, AWL0 W, A9194, structure(withaBayesianclusteringmethod),theprogram AX48Q with NED, BDYVU, AOES3 with PET. Each STRUCTURE 2.3.3(Falushetal.2007;Hubiszetal.2009; reaction was performed in a final volume of 3 lL con- Pritchardetal.2000)wasemployed.Toassignindividuals taining 0.2 lM of each reverse primer, 0.04 lM of each into K clusters, an admixture model with correlated allele forward primer, and 0.6 lM of fluorescent dye-labeled frequencies (Falush et al. 2003) was used. The number of primer.Touchdownthermalcyclingprogramswereusedas stepswas100,000,with50,000iterations, and10replicate follows: initial denaturation at 95 (cid:3)C for 5 min, followed runs ineach K from 1 to10. The highest level of structure by20cyclesofdenaturationat95 (cid:3)Cfor30 s,annealingat was deduced from a posterior probability of the data for a 63 (cid:3)Cfor 90 s(decreased0.5 (cid:3)C percycle),andextension given K and DK value (Evanno et al. 2005). A Pearson at 72 (cid:3)C for 60 s, plus 25 cycles of denaturation at 95 (cid:3)C correlation was used to test correlation between values of for 30 s, annealing at 55 (cid:3)C for 90 s, and extension at genetic diversity and genetic divergence, and a Mann– 72 (cid:3)C for 60 s; a final extension step was performed at Whitney U test was performed to compare means of 123 80 JPlantRes(2015)128:73–90 independent samples for the previous parameters. In both Results cases the program SPSS ver. 15.0 ((cid:2)SPSS Inc.) was used. For microsatellites the program GENEPOP 4.0 (Ray- AFLPs mond and Rousset 1995) was used to test linkage dis- equilibrium (LD) and significant deviation from Hardy– Among the species of Drimys, we found a total of 583 Weinberg equilibrium (HWE) between loci in each pop- fragments, of which 574 are polymorphic (98.5 %, ulation (Markov chain method 10,000 dememorisation Table 1). In D. confertifolia the total number of fragments steps, 1,000 batches, 500 iterations per batch). The fre- is 576 (100 % polymorphic bands), in D. andina 383 quency of null allele in each marker was calculated fol- fragments (373 bands polymorphic, 97.4 %), and D. win- lowing Brookfield (1996) using Micro-Checker 2.2.3 (van teri 485 fragments (438 polymorphic, 90.3 %, Table 1). Oosterhout et al. 2004). The genetic diversity parameters With each pair of primers, the numbers of fragments in for each species and population, allelic richness (A ), each species (D. confertifolia/D. andina/D. winteri) are: R expected proportion of heterozygotes (H ), number of 153/81/126/for primers MseI–CTG/EcoRI–ACA (FAM); E alleles per locus (N ), and inbreeding coefficient (F ), 120/76/102 for MseI–CAC/EcoRI–ATG (VIC); 112/72/80 A IS were calculated using FSTAT 2.9.3.2 (Goudet 1995). for MseI–CTC/EcoRI–ACA (FAM); 104/58/76 for MseI– GENALEX 6 (Peakall and Smouse 2006) was used to CAG/EcoRI–AAG (VIC); and 77/42/54 for MseI–CTC/ estimate the genetic divergence parameters, number of EcoRI–AGC (NED). All individuals had unique AFLP private alleles (N ), and number of locally common phenotypes. The reproducibility of the bands was 95 %. PA alleles (N ). Allelic richness was calculated for popu- LCA lations that contained seven or more individuals using the Microsatellites rarefaction method (Hurlbert 1971). The genetic rela- tionships among populations were evaluated by con- All microsatellite markers, except for one marker AOES3, structing a neighbor-joining tree based on D genetic were successfully genotyped in 281 individuals of D. A distance (Nei et al. 1983) using the program Populations confertifolia, 123 of D. winteri, and 13 of D. andina. The 1.2.30 (Langella 1999). Pairwise F , AMOVA, marker AOES3 resulted in no amplification in two popu- ST STRUCTURE, and Pearson correlation were estimated in lations of D. confertifolia, and several populations of D. the same way as with AFLP. winteri and of D. andina. Hence, we used the remaining The direction of migration between island populations eightmarkersforfurtherpopulationanalyses.Anexacttest ofD.confertifoliawasinferredbyusingmicrosatellitedata for HWE across populations and loci showed 14 of 248 in based on the log Bayes factor (LBF) as per the model D.confertifolia,sixof80inD.winteri,andoneofeightin ranking method (Beerli and Palczewski 2010) using the D. andina deviating from HWE with a positive F , IS coalescentbasedMCMCmethodimplementedinMigrate- (P\0.05) after Bonferroni correction. The frequency of 3 (Beerli and Felsenstein 1999, 2001). The LBFs based on nullallelesacrosspopulations andlociwasestimatedusing the Be´zier approximation score and harmonic mean were Micro-Checker,resultinginthehighestfrequencyof0.188 estimated using microsatellite data. We compared the null (AWL0W),withanaveragefrequencyof0.080inallofthe modelofnomigration(M )tothemodelsofunidirectional eight markers. Significant LD was not found between any 0 migration from Alejandro Selkirk to Robinson Crusoe pairwisecombinationoflociinallpopulationswithineach Island (M ) and vice versa (M ). In order to standardize species (P\0.05) after Bonferroni correction. 1 2 population sizes, the 16 populations of D. confertifolia in Robinson Crusoe Island and 15 populations in Alejandro Genetic diversity SelkirkIslandwerepooledeachandrandomlysampledfor each replicate (50 individuals from each region). Runs The AFLP measures of genetic diversity in populations of were carried out under a Brownian model multiple times Drimys analyzed are shown in Table 1. The average esti- with varying parameter settings to achieve convergence, mates of genetic diversity in D. confertifolia, from Rob- and for the final MCMC parameters were one long chain inson Crusoe/Alejandro Selkirk Islands, are for percentage with200,000recordedstepsand100-stepincrementwitha ofpolymorphicbands(PPB)66.8/58.8;fortotalnumberof burn-in of 10,000. Uniform priors (minimum, maximum, bands (TNB) 433.5/414.7; for Shannon Diversity Index and delta) were placed for both theta (0, 20, and 2) and (SDI) 103.98/90.53; andfor average genetic diversity over M (0, 20, and 2). Ten replicates of single long Markov loci (AGDOL) 0.24/0.22. All estimates reveal a similar chains were implemented using different random number patternandarehighlycorrelatedwithrrangingfrom0.897 seeds. The MIGRATE analysis was performed at the to 0.956 (n = 22, P\0.001). This pattern holds when University of Oslo Bioportal (https://www.bioportal.uio. those correlations on the single islands were tested. Com- no/). parisons of genetic diversity between populations on the 123 JPlantRes(2015)128:73–90 81 Table2 GeneticdiversitywithAFLPandmicrosatellitesforDrimys When dividing the populations of Robinson Crusoe confertifolia,D.winteri,andD.andina Island into northern (1–7) and southern (9–16) sections, they show no significant difference. The same analysis Species AFLP Microsatellites AGDOL H holdsforAlejandroSelkirkIsland,revealingnosignificant E differencesbetweentheparametersN ,A ,andH ,among D.confertifoliaRobinsonCrusoeIsland 0.259 0.685 A R E northern (17, 18, 23) and southern populations (28–31). D.confertifoliaAlejandroSelkirkIsland 0.234 0.509 The genetic diversity (H ) within species was 0.675 in E D.confertifolia(combined) 0.276 0.675 D.confertifolia,0.733inD.winteri,and0.501inD.andina D.winteri 0.208 0.733 (Table 2). Calculating this value for D. confertifolia on D.andina 0.203 0.501 eachofthetwoislandsrevealed0.685onRobinsonCrusoe AGDOLaveragegenediversityoverloci,H expectedproportionof Island and 0.509 on Alejandro Selkirk Island. E heterozygotes Genetic divergence northern side of Robinson Crusoe Island (1–7) with pop- ulations on the southern side (9–16) show significant dif- For AFLPs, the values of genetic divergence are shown in ferences in the parameters TNB, SDI, and AGDOL. In Table 1. In Drimys confertifolia the number of private Alejandro Selkirk Island, the northern populations (17, 18, bands (NPB) and rarity index (RI) in Robinson Crusoe/ 23) are not significantly different in all genetic diversity Alejandro Selkirk Islands are 0.40/0.14 and 1.50/1.76 parametersfromthesouthernpopulations(28–31).Overall respectively. These values are not correlated, (r = 0.065) the genetic diversity within populations is higher on Rob- [n = 22, P = 0.775]. The same lack of correlation occurs inson Crusoe Island (average SDI 103.98 ± 6.88) than on when analyzing Robinson Crusoe and Alejandro Selkirk Alejandro Selkirk Island (average SDI 90.53 ± 10.99). Islands separately. Comparison between northern popula- For each of the species, the total genetic variation tionswithsouthernpopulationsonRobinsonCrusoeIsland (AGDOL) was 0.276 in D. confertifolia, 0.203 in D. an- reveals significant differences only for the Rarity Index. dina, and 0.208 in Drimys winteri. In D. confertifolia a The same analysis in Alejandro Selkirk Island reveals no highervaluewasfoundonRobinsonCrusoeIsland(0.259) significant differences in both measures of divergence. compared to Alejandro Selkirk Island (0.234) (Table 2). PairwiseF valuesbetweenD.confertifoliaonthetwo In microsatellites, the values of genetic diversity ST islandsandcontinentalD.andinaandD.winterirevealthe parametersforallspeciesofDrimysareshowninTable 1. highest differentiation (F = 0.357) between D. winteri The average values in D. confertifolia from Robinson ST and D. confertifolia of Alejandro Selkirk island, and the Crusoe/AlejandroSelkirk Islands for number ofalleles per lowestbetweenD.confertifoliafromRobinsonCrusoeand locus(N )are4.89/3.73;forallelicrichness(A )are4.24/ A R Alejandro Selkirk (F = 0.186) (Table 3). The mean F 3.30; and the expected proportion of heterozygotes (H ) ST ST E valueamongpopulationswithinD.confertifoliawas0.141 are 0.62/0.47. All estimates are correlated; H was highly E and0.212inD.winteri.AhighproportionofpairwiseF correlated with N (r = 0.868) [n = 22, P = 0.000], and ST A values were significant in D. confertifolia, and all values A (r = 0.634) [n = 22, P = 0.002]. When analyzing this R were significant in D. winteri (Table S1, S2). parameter by island, only the pair N and H is correlated A E For microsatellites, the parameters for estimating (r = 0.810) [n = 15, P = 0.000] in Robinson Crusoe genetic divergence within Drimys species are shown in Island; no correlations were found between these three Table 1. In D. confertifolia the number of private alleles parametersinAlejandroSelkirkisland.PositiveF values IS (N ) and number of locally common alleles (N ) in (P\0.05) after Bonferroni correction were significant in PA LCA Robinson Crusoe/Alejandro Selkirk Islands are 0.04/0.02 11 populations of D. confertifolia and in four populations and 0.75/0.50, respectively. The differences between the of D. winteri. Table3 F valuesamongDrimysconfertifoliainRobinsonCrusoeandAlejandroSelkirkIsland,D.andina,andD.winteri ST D.confertifolia(RC) D.confertifolia(AS) D.andina D.winteri D.confertifolia(RC) 0.186 0.204 0.266 D.confertifolia(AS) 0.150 0.263 0.357 D.andina 0.245 0.283 0.203 D.winteri 0.060 0.101 0.179 AbovediagonalareestimatesfromAFLPdata,belowdiagonalfrommicrosatellites RCRobinsonCrusoeIsland,ASAlejandroSelkirkIsland 123 82 JPlantRes(2015)128:73–90 values of this parameter are not correlated (r = 0.361) visible,withonlyaweakdivergenceofpopulations23,30, [n = 22, P = 0.099]. The same tendency is seen when we 31 on Alejandro Selkirk Island and population 13 on analyzed separately each of the islands. We also compare RobinsonCrusoeIsland.TheclusterofD.winteridoesnot the N between the two island populations of D. conf- show a separation between variety winteri and var. chil- PA ertifolia.TheN ofRobinsonCrusoepopulationsis3.125 ensis. No geographical partitioning is observed in all PA (±0.875), and that of Alejandro Selkirk is 0.75 (±0.366). species. Geographical division between northern (17, 18, 23) and The Neighbour-Joining tree using microsatellite data at southern populations (28–31) in Alejandro Selkirk Island the population level is shown in Fig. 3b. Drimys conferti- showed no significant differences. The same lack of foliaformstwogroups,eachrestrictedtoadifferentisland. divergencebetweennorthernandsouthernpopulationsalso The populations from Alejandro Selkirk Island are linked occurs on Robinson Crusoe Island. morecloselytoD.winterithantotheotherpopulationson The highest values of F between D. confertifolia in RobinsonCrusoeIsland.Nogeographicaldivisionisfound ST Robinson Crusoe and Alejandro Selkirk Island and the in D. winteri, and no division between D. winteri var. other species of Drimys correspond to the pair D. conf- winteriandD.winterivar.chilensisisseen.Drimysandina ertifolia (Alejandro Selkirk island)–D. andina is connected to D. winteri. (F = 0.283), and the lowest to D. confertifolia (Robin- Testing for genetically coherent groups in D. conferti- ST son Crusoe)–D. winteri (F = 0.060) (Table 3). In folia using the Bayesian approach STRUCTURE 2.3.3 ST D.confertifoliathemeanF valueamongpopulationswas (Fig. 4), the group number K = 2 explained best the ST 0.109, and in D. winteri 0.139. The pairwise genetic dif- groupings found by both AFLPs and microsatellites and ferentiation values were significant in all species under revealedaverylowdegreeofadmixtureamongtheislands. study (Table S1, S2). Analyzingeachoftheislandsseparately,K = 2againbest explains the grouping for both molecular markers, how- Genetic structure ever, with a high proportion of individuals showing strong admixture. Resultsoftheanalysisofmolecularvariance(AMOVA)at Mantel tests between genetic differentiation (F /(1 - ST differenthierarchicallevelsinD.confertifoliaareshownin F )) (Rousset 1997) and geographical distance did not ST Table 4. A high percentage of variation was found within reveal any significant correlation for the D. confertifolia populations, with values of 76 % for AFLPs and 50 % for populations.ThetestsforRobinsonCrusoeIslandprovided microsatellites. Among populations within the islands, the r2 = 0.0344 and r2 = 0.0036 for AFLP and microsatel- values for AFLPs and microsatellites were low and 6 and lites, respectively. For Alejandro Selkirk the values were 5 %, respectively, whereas a high percentage of variation r2 = 0.0333forAFLPandr2 = 0.0080formicrosatellites. between Robinson Crusoe and Alejandro Selkirk Islands Results of the MIGRATE analyses (Table 5) showed was found to be 18 and 15 %, respectively. consistent results with smooth histograms of theta and M, The NeighbourNet tree based on AFLP data using all suggesting that the Markov chain had converged on the individualsofD.confertifolia,D.andina,andD.winteriis stationarydistribution.ThebothLBFsbasedontheBe´zier showninFig. 3a.Aseparationbetweenspeciesisobserved approximation score and the harmonic mean constantly with D. confertifolia forming two groups, the first clearly indicate that two models of migration (M and M ) are 1 2 differentiated and corresponding to Alejandro Selkirk more likely than the null hypothesis of no gene flow (M ) 0 Island, and the second to the populations of Robinson betweenthetwoislands (Table 5).TheLBFsarehigherin Crusoe Island. No clear separation between populations is aunidirectionalmigrationmodelfromRobinsonCrusoeto Table4 Summaryofanalysesofmolecularvariance(AMOVA)forAFLPsandmicrosatellitesinDrimysconfertifolia Sourceofvariation AFLPs Microsatellites df SS Variance Totalvariance df SS Variance Totalvariance components (%) components (%) Betweenislands 1 1925.61 16.19 17.97 1 102.7 0.4 14.61 Amongpopulationswithin 20 2,592.67 5.22 5.79 20 119.2 0.2 5.09 island Withinpopulations 237 16,278.05 68.68 76.24 500 1,194.8 2.4 50.30 Thetotalvariancecontributedbyeachcomponent(%),anditsassociatedsignificance(n=1,023permutations)areshown dfdegreesoffreedom,SSsumofsquares 123
Description: