Phylogeny and Genetic Structure of Erophaca (Leguminosae), a East–West Mediterranean Disjunct Genus from the Tertiary

Home , Anthyllis montana, Astragalus, Erophaca, Stauracanthus

Molecular Phylogenetics and Evolution 56 (2010) 441–450

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

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

Phylogeny and genetic structure of Erophaca (Leguminosae), a East–West Mediterranean disjunct genus from the Tertiary

Ramón Casimiro-Soriguer a,*, María Talavera a, Francisco Balao a, Anass Terrab a,b, Javier Herrera a, Salvador Talavera a a Departamento de Biología Vegetal y Ecología, Facultad de Biología, Universidad de Sevilla, Apdo. 1095, E-41080 Sevilla, Spain b Department of Systematic and Evolutionary Botany, University of Vienna, Rennweg 14, A-1030 Vienna, Austria article info abstract

Article history: The genus Erophaca comprises a single herbaceous perennial species with two subspecies distributed at Received 24 November 2009 opposite ends of the Mediterranean region. We used nrDNA ITS to investigate the phylogeny of the genus, Revised 3 February 2010 and AFLP markers (9 primers, 20 populations) to establish the genetic relationship between subspecies, Accepted 19 February 2010 and among populations at each side of the Gibraltar Strait. According to nrDNA ITS, Erophaca is monophy- Available online 26 February 2010 letic, old (Miocene), and sister to the Astragalean clade. Life form attributes and molecular clock estimates suggest that Erophaca is one of the many Tertiary relicts that form part of the present Keywords: Mediterranean ﬂora. Within the occidental subspecies, European plants are clearly derived from AFLP North-African populations (Morocco) which, despite being rare on a regional scale, present the highest Astragalus lusitanicus Erophaca baetica genetic diversity (as estimated by private and rare fragment numbers). In general, genetic diversity Genetic diversity decreased with increasing distance from Morocco. AFLP and nrDNA ITS markers evidenced that the East- Gibraltar Strait ern and the Western subspecies are genetically distinct. Possible causes for their disjunct distribution are Leguminosae discussed. nrDNA ITS Ó 2010 Elsevier Inc. All rights reserved. Mediterranean Phylogeography Phylogeny

1. Introduction a formidable obstacle to plant expansion (e.g., Terrab et al., 2007; Cano-Maqueda et al., 2008), but there are numerous examples of The Mediterranean region is considered one of the World’s biodi- species that have successfully crossed this barrier, both from the versity ‘‘hotspots” (Myers et al., 2000), but actually it comprises two North to the South (e.g., Caujape-Castells and Jansen, 2003; Escudero separate areas of very high diversity, to the East (specially the Bal- et al., 2008) and from South to North (e.g., Carranza et al., 2006; Ortiz kans and Turkey) and the West (Iberian Peninsula and NW Africa) et al., 2009). (Medail and Quezel, 1999; Schmitt, 2007). Many species that now The Legume family is one of the most diverse groups of flower- inhabit these areas experienced intermittent shifts in their altitudi- ing plants with about 19,000 species (Culham, 2007). Many are nal/latitudinal ranges as a result of the climatic oscillations of the major constituents of Mediterranean scrub that covers vast areas Pleistocene (Hewitt, 1996, 2000; Taberlet et al., 1998; Thompson, in Morocco and the Iberian Peninsula, and they are, therefore, of 2005; Rodríguez-Sánchez et al., 2008), and such readjustments have great ecological importance in the region. However, few molecular resulted in a number of interesting biogeographical situations, studies have been undertaken in this area on legume taxa. Excep- including a high incidence of endemism and many disjunct distribu- tions are the use of chloroplast markers in Ulex (Cubas et al., 2005) tions of plants and animals (e.g., Davis and Hedge, 1971; Rosselló and Stauracanthus (Pardo et al., 2008), and allozymes in Calicotome et al., 2007). In the West Mediterranean, for example, lowering of (Arroyo et al., 2008). These studies have shown that the Strait of sea levels during Pliocene and Pleistocene glacial maxima could Gibraltar can be an effective barrier to gene flow in some cases have permitted migrations to occur across the Strait of Gibraltar (e.g., Ulex, Stauracanthus) but not in others (Calicotome). The AFLP (Collina-Girard, 2001; Ortiz et al., 2007), while warmer periods (Amplified Fragment Length Polymorphism) technique is appropri- allowed some species to expand and others to become restricted ate for phylogeographic studies because it allows genotyping of or extinct (Taberlet et al., 1998). Most often the Strait is considered large numbers of loci in many different individuals, and hence the elucidation of genetic population structures and recent phylog-

* Corresponding author. Fax: +34 954557059. enies (Weising et al., 2005). Phylogeographic studies using AFLP E-mail address: [email protected] (R. Casimiro-Soriguer). have been successfully carried out of legume taxa in Europe and

America (e.g., Anthyllis montana; Kropf et al., 2002; Astragalus (Casimiro-Soriguer, unpublished). Ovaries develop into inflated le- ampullarioides; Breinholt et al., 2009; Astragalus cremnophylax; gumes with up to nine, relatively large (7–10 mm) seeds and Travis et al., 1996; Oxytropis campestris; Chung et al., 2004), as well explosive dehiscence. Biotic seed dispersal agents are unknown. as in Mediterranean species of diverse plant groups (Ortiz et al., By mid spring (May) the plants shed their leaves, all above ground 2008; Piñeiro et al., 2007; Terrab et al., 2008). plant parts senesce and become dry, and the buried xylopodia be- The monotypic genus Erophaca Boiss. is endemic to the Mediter- come dormant until next winter. Throughout its distribution range ranean and its sole species, E. baetica (L.) Boiss., was long included populations occur as isolated patches, usually in cork oak (Quercus in Astragalus (as A. lusitanicus Lam.) until transferred to Erophaca by suber) woodlands from sea level up to 1300 m. The foliage is very Podlech (1993), who assigned the genus to the tribe Sophoreae on toxic to sheep, goats and cattle (Bel-Kassaoui et al., 2008) and for the basis of morphology. Also on morphological data, Talavera this reason plants are sometimes cut down by shepherds and farm- (1999) retained Erophaca in the tribe Astragaleae but proposed a ers, presumably to prevent them from spreading (Casimiro-Sori- new subtribe for it, the Erophacinae. Molecular data (Liston, guer, personal observation). Consequently, in parts of its range 1995; Wojciechowski et al., 2000) point to a phylogenetic relation- (particularly Morocco) humans may have decimated populations ship between Erophaca and the Astragalean clade, but as of yet the for centuries. Developing seeds can be heavily preyed upon by phylogeny of the genus is not fully understood. Furthermore, E. Lycaenid butterfly larvae (Jordano et al., 1990). baetica has two subspecies distributed at opposite ends of the Med- We studied 20 populations covering most of the distribution iterranean area, and is thus a good subject for studying evolution- range (Fig. 1). Due to the rarity of E. baetica in North Africa, and de- ary changes below the species level over a relatively large spite considerable sampling effort, only five Moroccan populations geographical range. Lastly, the Western subspecies lives at both could be analyzed. Leaves were picked at random from around 10 sides of the Strait of Gibraltar, and this gives an opportunity to individuals per population, stored in silica gel until processed for check the effects of a major geographic barrier on the genetic dif- DNA isolation, and vouchers deposited in the herbarium of the Uni- ferentiation of populations. As part of a larger study on the ecology, versity of Seville, Spain (SEV). Furthermore, and in order to inves- distribution and evolution of E. baetica (Casimiro-Soriguer, unpub- tigate the relationship among subspecies, 11 individuals of the lished), here we used nuclear ribosomal Internal Transcribed Spac- Eastern subspecies were sampled from a Greek population. Both ers DNA sequencing (nrDNA ITS) and AFLP analyses to: (1) for the nrDNA ITS and AFLP analyses, genomic DNA was extracted contribute to clarify the phylogenetic status of Erophaca within following the CTAB protocol (Doyle and Doyle, 1987) with modifi- the Leguminosae; (2) assess the genetic differences among its sub- cations (Terrab et al., 2008). Extracts were checked for the presence species; and (3) provide insights into the phylogeographic relation- and amount of DNA on 1% TAE agarose gels. ships between North African and S. Iberian populations. 2.2. nrDNA ITS

2. Materials and methods Two randomly chosen plants per population, from nine populations of E. baetica subsp. baetica, and one population of subsp. ori- 2.1. Study plant, populations sampled and DNA isolation entalis, were analyzed (Genbank No. FN58233 through FN58252). Using universal primers ITS4 and ITS5 of White et al. (1990), the Subtribal assignments used in the present study follow Polhill nuclear ribosomal Internal Transcribed Spacers regions ITS1, 5.8S (1981), and Talavera (1999), and the tribe to which Erophaca be- and ITS2 were amplified under the following conditions of the longs will be referred to as Galegeae in accordance with most mod- polymerase chain reaction (PCR): 18 ll of PCR Reddy mix at 1.4 ern systematic works (e.g., Lewis et al., 2005). As for the subspecies (supplied by Abgene, Vienna, Austria), 0.32 ll (20 pM) of each pri- of E. baetica that populate the West and the East Mediterranean, mer (forward or reverse), 0.4 ll DMSO, and 1 ll (2–8 ng/ll) of DNA they are, respectively, the subsp. baetica (S. Iberian Peninsula, Mor- template. Amplified products were purified adding 0.5 ll ExoI and occo and Algeria), and the subsp. orientalis (Chater & Meikle) Pod- 2 ll CIAP (Fermentas, St. Leon-Rot, Germany), then heated for lech (Greece, Turkey, Cyprus and Lebanon). These subspecies can 45 min at 37 °C, and 15 min at 85 °C. The polymerase chain reac- be distinguished morphologically on the basis of leaflet hairiness tion (PCR) sequence profile was one cycle of 1 min at 96 °C, fol- (appressed hairs present in subsp. orientalis vs. glabrous leaflets lowed by 35 amplification cycles. Each cycle consisted of 10 s in subsp. baetica); the teeth of the calyx (longer than half calyx denaturation at 96 °C, 5 s annealing at 50 °C, and 3 min elongation in the Eastern subspecies vs. much shorter in the Western one); at 60 °C. Lastly, the sample was extended at 72 °C for 8 min. The the color of their ‘stipelles’ (i.e., the stipule-like outgrowths on purified fragments were directly sequenced using dye terminator the base of the leaflets) which are either purple (Eastern) or yel- chemistry and the following cycling conditions: 38 cycles of 10 s lowish (Western subspecies); and the fruits (usually smaller than at 96 °C, 25 s at 50 °C, 4 min at 60 °C. The samples were run on a 70 25 cm in subsp. orientalis, and larger than that in subsp. 3130xl Genetic Analyzer capillary sequencer (Applied Biosystems). baetica). Forward and reverse sequences were assembled with AutoAssem- Our study focused primarily on E. baetica subsp. baetica, ende- bler ver. 1.4.0 (Applied Biosystems), and alignment of the complete mic to the Iberian Peninsula, Morocco and Algeria (Fig. 1, inset). ITS1, 5.8S and ITS2 sequences were performed manually using This is a long-lived, perennial plant with a woody tuber-like organ MEGA version 4 (Tamura et al., 2007). To select the nucleotide (xylopodium) and numerous, fast growing herbaceous stems that substitution model that best fitted the data, Modeltest ver. 3.06 start to sprout at the beginning of winter (December), in synchrony (Posada and Crandall, 1998) was used. According to Akaike Infor- with the main rainy season under the Mediterranean climate. mation Criterium the optimal model was GTR+c. Plants up to 2 m in height can be found on shady places, but more The sequences obtained, together with GenBank sequences of a often they are around 1 m tall. The stems converge at ground level number of closely or distantly related taxa, were used to construct on the xylopodium, so that individuals can be easily differentiated a phylogeny based on heuristic parsimony searches, based on 21 in the field. Vegetative reproduction (e.g., by stolons) has never taxa and 779 characters. Computations were performed in PAUP* been observed but, in order to minimize the risk of genotype ver. 4.0 beta 10 (Swofford, 2002), with bootstrapping values calcu- resampling, only plants more than 4 m apart were used in the pres- lated using the full heuristic search option (10,000 replicates). ent study. From December to March the stems bear clusters of GenBank provided sequences for the following taxa (accession showy white, nectariferous flowers that are visited mainly by bees numbers): Anagyris foetida (AF330637); Astragalus boeticus (AF12 R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450 443

Fig. 1. Distribution and genetic structure in 19 populations of E. baetica subsp. baetica, and geographic range of the taxon (gray area). Populations are identiﬁed by their numbers in Table 1, and colors indicate genetic groups inferred from Bayesian clustering. The known distribution of the genus Erophaca is shown as an insert.

1679); A. membranaceus (AF121675); Caragana hololeuca (AB26 chosen to represent the sister-group of Erophaca, while the genus 2524); C. tragacanthoides (AB262536); Chesneya deshungarica (U5 Anagyris was used to root the tree. Other assumptions of the model 0350); E. baetica subsp. orientalis (EU070920); Galega orientalis were: (1) uncorrelated, log-normal prior model of rate change; (2) (AY553709); Glycyrrhiza glabra (AY065623); Hedysarum coronari- Yule model for speciation; and (3) automatic tuning of operators. um (AY772225); Oxytropis multiceps (AF121760); O. splendens Analyses were run for 100 million generations (sampling every (AF121761); Sophora flavescens (AF123452); Swainsona purpurea 1000th), with a burn-in of one million generations. (AF113864); and S. stenodonta (AF113865). Divergence times (either within E. baetica, or from its ancestors) 2.3. AFLP were estimated from phylogeny using the Bayesian relaxed molecular clock method in BEAST 1.4.7 (Drummond and Rambaut, 2007). Overall, 200 plants from N. Africa, the Iberian Peninsula, and the BEAST estimates divergence times (branch lengths) using the topol- E. Mediterranean were studied (Table 1). Following the protocols ogy it generates, and a calibration time (±uncertainty) that is pro- established by Vos et al. (1995), genomic DNA (0.5 lg on average) vided as input. Because the fossil record is poor for the clade to was digested with two restriction endonucleases (EcoRI and MseI), which Erophaca is presumed to belong (i.e., the one that lacks the in- fragments ligated to double-stranded adaptors (EcoRI and MseI) at verted repeat, or IRLC clade of Wojciechowski et al. (1999)) the clock 37 °C for 2 h., then diluted 20-fold with TE0.1 buffer. Fragments was calibrated using the estimate by Lavin et al. (2005) of with the matching nucleotides were amplified (downstream of 14.8 ± 2 mya for the divergence of Colutea arborescens and Astragalus the restriction sites) using preselective primers based on EcoRI americanus (as representative case examples of the subtribe Colutei- and MseI adaptors. This resulted in a 16-fold decrease in the nae and Astragalinae). Furthermore (and as suggested by the topol- number of fragments eventually amplified. Preselective and selec- ogy of the tree) the genera Swainsona, Astragalus and Oxytropis were tive amplifications were performed in a thermal cycler (Gene 444

Table 1 Population-based estimates of genetic diversity in E. baetica. Averages for major geographic areas in bold.

a b c d d Area Pop. Localities Elevation (m) Coordinates Collector No. N AFLP fragments % Polymorphic HD Fragpriv DW .Csmr-oiure l oeua hlgntc n vlto 6(00 441–450 (2010) 56 Evolution and Phylogenetics Molecular / al. et Casimiro-Soriguer R. Sub-area Mean (s.e.m) a b a b Iberian Peninsula Iberian Massif 1 Miranda * 553 40°290/5°570 RCS et al. 315/06 11 246 73.6 0.10 (0.052) 1 3 1.8 4.4 2 Guadalupe 670 39°220/5°200 RCS et al. 321/06 10 252 74.2 0.11 (0.060) 2 2.0 3 Nieﬂa * 806 38°310/4°220 RCS et al. 204/06 9 281 83.3 0.13 (0.070) 4 2.5 4 Cerro Hierro * 658 37°570/5°370 RCS et al. 2/07 10 251 74.9 0.11 (0.059) 2 4 1.8 4.2 5 Almonaster 689 37°520/6°460 RCS et al. 323/06 9 270 77.4 0.11 (0.059) 0 2.0 6 Ronquillo 353 37°430/6°110 RCS 194/06 10 279 78.5 0.13 (0.068) 6 2.9 7 Aljezur 40 37°200/8°500 RCS et al. 233/06 6 244 61.9 0.10 (0.059) 5 2.1 8 Hinojos 90 37°170/6°250 RCS 192/07 11 265 76.9 0.11 (0.058) 0 1.8 9 Loulé * 280 37°120/7°570 RCS et al. 218/06 12 277 78.3 0.11 (0.058) 6 10 2.4 5.3 10 Doñana * 22 37°000/6°300 RCS 193/07 10 255 72.2 0.10 (0.051) 1 1.9 11 Puerto Real * 69 36°310/6°040 FB 18/07 10 280 77.9 0.12 (0.064) 2 2.1 263 75.4 0.11 5.7 4.6 Betic Ranges 12 Cartagena 277 37°380/1°110 RCS & MT 17/07 11 241 76.8 0.11 (0.055) 6 13 2.7 5.2 13 Bédar * 563 37°120/2°000 RCS & MT 16/07 12 235 73.6 0.09 (0.047) 1 1 1.8 3.4 14 Cortes de la Frontera 510 36°400/5°160 RCS 1/07 12 270 77.4 0.12 (0.061) 1 7 2.1 4.5 248 75.9 0.11 7.0 4.4 N. Africa Rif Mountain 15 Gourugou 603 35°140/2°580 RCS et al. 116/07 7 275 76.7 0.13 (0.075) 2 4.2 16 Asifane 900 35°080/4°570 RCS et al. 57/07 9 312 79.2 0.15 (0.078) 8 4.6 17 Targuist * 1300 34°570/4°190 RCS et al. 58/07 10 319 80.6 0.16 (0.082) 13 35 5.0 9.1 Middle Atlas 18 Romanni 111 33°380/7°570 RCS et al. 170/06 9 329 79.9 0.15 (0.081) 11 20 4.7 8.7 19 Ait Atab * 850 32°090/6°460 FB et al. 205/07 11 312 82.4 0.15 (0.077) 9 12 4.2 7.5 309 80.0 0.15 22.3 8.4 Greece 20 Olympia * 203 37°380/21°440E RS 19/07 11 182 47.8 0.05 (0.026) 42 7.9

a Asterisks sequenced for ITS variation. b All N/W except otherwise indicated. c Mean genetic diversity. d a, based on all studied plants; b, sub-area based, with nine plants per population and three (randomly chosen) populations per group. R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450 445

AmpÒ PCR system 9700, PE Applied Biosystems). In order to assess barriers (i.e., Strait of Gibraltar, Guadalquivir River Valley, or both) the reliability of the method a random fraction (N = 23, i.e., 12%) of was also studied with AMOVA. the samples was analyzed twice. It was found that duplicate anal- Genetic relationships among populations were further explored yses were largely indistinguishable, with a repeatability of 98.8%. with a Bayesian assignment technique implemented in STRUC- A screening of selective primers (consisting of a battery of 72) TURE 2.2.3 (Falush et al., 2003; Pritchard et al., 2000) and carried was run on eight individuals from eight populations, and the nine out at the University of Oslo (http://www.bioportal.uio.no/). This primers that generated the most polymorphic fragments were se- approach uses individual multilocus genotypes to construct clus- lected. These were: EcoRI (Fam)-ACT/MseI-CAT; EcoRI (Fam)-ATC/ ters of genetically similar plants, in a way that maximizes MseI-CAG; EcoRI (Vic)-ACG/MseI CAG; EcoRI (Vic)-ACG/MseI Hardy–Weinberg equilibrium and minimizes linkage disequilib- CAA; EcoRI (Ned)-AAC/MseI-CAC; EcoRI (Ned)-AGC/MseI-CTA; Eco- rium within clusters. Following Evanno et al. (2005) and Falush RI (Ned)-ACA/MseI-CAT; EcoRI (Vic)-ATG/MseI-CAG; and EcoRI et al. (2003) we used the ‘admixture’ and the ‘correlated allele fre- (Fam)-ACA/MseI-CAA. Fluorescence-labeled selective ampliﬁcation quency’ models, a combination judged appropriate when the aim is products were separated on a 3130xl Genetic Analyzers (Applied to infer subtle population structures, as expected in our dataset. To Biosystems) capillary sequencer with an internal size standard ensure convergence of the Monte Carlo Markov Chain (MCMC) we (GeneScan-500 ROX, PE Applied Biosystems). Alignment of the used 50,000 burn-ins followed by 500,000 iterations. The true raw data with the size standard was performed with ABI Prism number of genetically distinct sets of populations (i.e., ‘archipela- GeneScan software 2.1 (PE Applied Biosystems). AFLP fragments gos’, sensu Evanno et al., 2005) was determined from analyses that were scored for bands 80–500 bp in length with software devel- had varying values of K, from 1 to a maximum of 19 (i.e., the total oped in the Montana State University (Genographer ver. 1.6.0, number of populations). Ten independent runs were performed available at http://hordeum.oscs.montana.edu/genographer) using and averaged for each value of K. The best estimate of the true the ‘thumbnail’ option. Results were summarized as a matrix of number of archipelagos (Delta-K, as described by Evanno et al. presence/absence. (2005)) was computed with Structure-sum 2.2 (Ehrich, 2006).

2.3.1. AFLP-based genetic diversity and population differentiation 3. Results Total and percent polymorphic fragments were determined for each population using FAMD version 1.08 (Schlüter and Harris, 3.1. nrDNA ITS phylogeny 2006). The numbers of private fragments (Fragpriv), and of private fragments shared by any pair of populations (Fragsh) were also cal- Fig. 2 presents one of two equally-parsimonious trees based on culated. Mean genetic diversity (HD) was computed with Arlequin nrDNA ITS. The tree has 30 steps, a Consistency Index of 0.804, and version 3.01 (Excoffier et al., 2005), and the Rarity 1 Index (equiv- a Retention Index of 0.746. As regards their nrDNA ITS sequences, alent to the frequency of down-weighted marker values; i.e., DW) the studied populations of E. baetica subsp. baetica were undistin- determined with R Statistical Software ver. 2.8.1 (R Development guishable. Nevertheless, they could be differentiated from subsp. Core Team) and AFLPdat (Ehrich, 2006). As both the number of pri- orientalis on the basis of four parsimoniously informative changes vate fragments and the rarity index are highly dependent on the in region ITS1. Furthermore, the genus Erophaca itself appeared spatial distances that separate the populations and the number as monophyletic and separated from the other genera with a boot- of plants analyzed (Tremetsberger et al., 2009), for simplicity we strap value of 100. The tree also showed that Erophaca was only calculated these parameters using three populations chosen at ran- distantly related to Sophora, but should be considered sister to Oxy- dom per geographic group, and with nine randomly chosen indi- tropis, Astragalus, and Swainsona. viduals per population (Schönswetter and Tribsch, 2005). General According to the relaxed molecular clock, the most recent com- Linear Models (module GLM in SPSS) were used to study the effect mon ancestor of the two subspecies of E. baetica lived 11.9 mya of geographic origin on HD, Fragpoly, Fragpriv, and DW. (Fig. 2), although this estimate had considerable uncertainty (95% The extent of genetic differentiation (measured as FST, the fixa- Highest Posterior Density confidence interval, HPD = 3.3–19.1 my; tion index) was determined both for populations and for geograph- i.e., from the Late Miocene to the Early Pliocene). As for the genus ically-based sets of populations (Arlequin 3.01; Excoffier et al., Erophaca itself, it would have split away from its sister clade 2005), and affinities between individuals and populations checked 16.6 mya (HPD = 8.5–20.5 my), during the Mid Miocene. using the random reallocation method implemented in AFLPOP v 1.1 (Duchesne and Bernatchez, 2002). The relationship between pairwise F values and geographical distances was investigated ST 3.2. AFLP-based genetic diversity and population differentiation with Mantel tests based on Spearman rank correlations with 100,000 permutations (package ade4 in R Statistical Software; Dray A total of 561 unambiguously scoreable DNA fragments were and Dufour, 2007). Lastly, the relationship between the two sub- detected in E. baetica subspecies baetica, of which 527 (94%) were species of E. baetica was depicted in a dendrogram built (on the ba- polymorphic. Population-based estimates in Table 1 indicate a sis of among-population pairwise F distances) in SPLITSTREE 4.6 ST maximum of 329 fragments per population, and a minimum of (Huson and Bryant, 2006), with neighbor-joining clustering and 235 (61–82% polymorphic). Mean genetic diversity (H ) was mark- 10,000 replicates. D edly dependent on the geographic location of populations (i.e. Ibe-

rian Massif, Betic Ranges, or North Africa; F2,16 = 19.31, p < 0.001; 2.3.2. Population and plant clustering General Linear Model ANOVA). On average, HD was signiﬁcantly Among-plant genetic distances (Nei and Li, 1979) were calcu- higher for Moroccan populations than for Iberian counterparts lated from the matrix of AFLP scores and represented graphically (p < 0.001; Tukey’s pairwise comparisons). Values of the Rarity In- in a consensus tree built with PAUP* ver. 4.0 beta 10 (Swofford, dex (i.e., the sub-area based DW in Table 1) were also dependent

2002; neighbor-joining clustering, 10,000 replicates). Furthermore, on geographic location (F2,6 = 25.36, p = 0.001), and again signiﬁ- the partitioning (within and among populations) of genetic varia- cantly higher for North-African populations than for the rest (8.4 tion was determined with an Analysis of Molecular Variance on average; p < 0.001, Tukey’s test). Populations from Morocco (AMOVA; Excofﬁer et al., 2005) performed over all the studied were more diverse than those from Iberia also as regards polymor- samples. The effect on genetic variation of evident geographical phic and private fragment numbers. 446 R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450

Fig. 2. The phylogenetic position of Erophaca baetica within the Leguminosae. The tree is based on nrDNA ITS data gathered during the present study (bold) and from GenBank (asterisks). Numerals near branch nodes indicate bootstrapping values (10,000 replicates). Arrows point to the nodes of divergence for E. baetica and its two subspecies, dated, respectively, under the relaxed molecular clock at 16.6 mya (HPD = 8.5–20.5) and 11.9 mya (HPD = 3.3–19.1). The node used to calibrate the clock is marked with a black dot (14.8 ± 2 mya; Lavin et al., 2005).

Table 2 reports genetic distances and numbers of shared private based on AFLP data: two well-supported (98% bootstrap) clusters fragments among populations of Erophaca baetica. The 14 popula- existed on each side of the Strait of Gibraltar, that corresponded tions from the Iberian Peninsula had few (at most two) or no private to Morocco and the Iberian Peninsula. Within Morocco, two sub- fragments in common, while the five Moroccan shared up to 11. clusters could be distinguished, one including Middle Atlas plants According to Mantel tests, the genetic dissimilarity among any (i.e., populations 18 and 19) and another with plants from the Rif two populations (as FST) increased with increasing spatial distance. Mountains (populations 15, 16, and 17). Within the Iberian Pen- This proved significant for the Iberian subset of populations insula, population 12 appeared distinctly separated from the rest (r = 0.561, N = 14, p < 0.001); for the Moroccan subset (r = 0.721, at 100% bootstrap, while the rest encompassed plants from the N =5, p < 0.05); and for the 19 West Mediterranean populations Betic Ranges on the one hand (populations 13 and 14), and plants altogether (r = 0.633, N = 19, p < 0.001). In addition, following ran- from the Iberian Massif on the other (i.e., populations 1 through dom reallocation in AFLPOP, plants were most often correctly as- 11). signed to their populations of origin (with the implication that The population-based graph in Fig. 3 (right) reiterated the ge- populations had genetic profiles definite enough to become predict- netic relationships described above and, additionally, suggested able). Specifically, all Moroccan plants except one (i.e., 98%) were that subsp. orientalis could relate more to Moroccan plants (of allocated to their true source, all but two from the Betic Ranges subsp. baetica) than to Iberian plants. Moreover, Moroccan popula- (94%), and all but 12 from the Iberian Massif (89%). tions also shared a few private fragments with the (single) Greek population (see Table 2). 3.3. Genetic structure and geographic areas Estimates of genetic diversity and differentiation calculated specifically for geographic sub-areas are shown in Table 3. As ex- The Bayesian analysis of individual genotypes showed that the pected, fewer fragments were shared across continents (from 1 most informative representation of overall genetic structure was to 11) than within the Iberian Peninsula (31) or North Africa achieved with K = 3 geographical groups, namely Morocco, the (36). Sub-areal dissimilarity (estimated from fixation indices) Betic Ranges, and the Iberian Massif. The same can be concluded showed an equivalent trend: most dissimilar were plants from from Fig. 3 (left), depicting neighbor-joined individual plants Middle Atlas and the Iberian Massif (FST = 0.28), while those from R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450 447

Table 2

Fixation indices (FST, below diagonal), and numbers of shared private fragments (Fragpriv, above diagonal) in populations of Erophaca baetica subsp. baetica from the Iberian

Peninsula (columns 1–14) and N. Africa (15–19). Data for a single population of the subsp. orientalis from the East Mediterranean are also given (column 20). Extreme FST values are in bold, and empty cells denote that the number of shared fragments equals zero. Numerals identify population as in Table 1.

Populations 1 2 3 4 5 6 7 8 9 1011121314151617181920 1– 2 0.13 – 3 0.19 0.14 – 1 1 4 0.17 0.08 0.17 – 5 0.20 0.15 0.07 0.15 – 2 6 0.19 0.14 0.18 0.14 0.17 – 1 2 7 0.24 0.22 0.26 0.23 0.25 0.20 – 8 0.23 0.18 0.11 0.18 0.10 0.21 0.29 – 9 0.16 0.13 0.19 0.15 0.18 0.14 0.15 0.23 – 10 0.30 0.24 0.19 0.26 0.20 0.26 0.36 0.16 0.26 – 11 0.24 0.18 0.09 0.20 0.13 0.20 0.27 0.13 0.20 0.21 – 12 0.46 0.40 0.34 0.39 0.40 0.39 0.46 0.38 0.41 0.44 0.31 – 1 2 1 13 0.40 0.33 0.36 0.30 0.40 0.33 0.41 0.37 0.35 0.43 0.29 0.35 – 14 0.30 0.23 0.26 0.23 0.30 0.23 0.31 0.30 0.26 0.36 0.20 0.29 0.08 – 15 0.45 0.40 0.36 0.38 0.41 0.37 0.42 0.41 0.39 0.45 0.36 0.37 0.42 0.34 – 10 1 16 0.38 0.36 0.35 0.33 0.38 0.33 0.37 0.37 0.37 0.44 0.33 0.37 0.39 0.32 0.24 – 3 3 1 2 17 0.37 0.34 0.33 0.31 0.35 0.32 0.34 0.35 0.35 0.41 0.31 0.32 0.35 0.30 0.17 0.12 – 2 3 18 0.36 0.32 0.30 0.32 0.35 0.31 0.32 0.35 0.34 0.39 0.31 0.36 0.37 0.30 0.25 0.19 0.18 – 11 2 19 0.38 0.36 0.34 0.35 0.38 0.32 0.36 0.38 0.37 0.42 0.34 0.39 0.39 0.32 0.26 0.19 0.21 0.11 – 3 20 0.77 0.76 0.74 0.75 0.77 0.74 0.80 0.76 0.75 0.79 0.74 0.77 0.78 0.74 0.69 0.72 0.69 0.71 0.71 –

Fig. 3. Left: Neighbor-joining unrooted dendrogram (AFLP-based) of E. baetica subsp. baetica individuals, with bootstrapping values higher than 50% indicated near nodes (10,000 permutations). Colors denote genetic structure inferred from Bayesian clustering. Right: Genetic relationships among populations and subspecies, numbered as in Table 1. Note that the scale is not shared across plots.

North African sub-areas (e.g., Rif Mountain and Middle Atlas) were 4. Discussion comparatively alike (FST = 0.13). Table 4 summarizes results of several AMOVA models analyzing 4.1. Evolution, phylogeny and systematics of the genus Erophaca the effects of major geographic barriers on the partitioning of molecular variance. In general, more than half of the genetic vari- Our relaxed molecular clock estimate for the origin of Erophaca ation occurred among plants within populations, but this was par- (16.6 my) had considerable uncertainty, and yet it matched fairly ticularly evident if no geographic barrier was included in the well previous estimates for the Astragalean clade as a whole models. Including the Strait of Gibraltar explained 18% of variation, (16.1 my; Wojciechowski, 2005). Interestingly, this predates the whereas the Guadalquivir River Valley accounted for around 15%. onset of Mediterranean climate which occurred approximately The model that explained a larger fraction of variation (19%) on 3 mya (Suc, 1984), and is compatible with estimates for the evolu- the basis of geographical barriers was the one considering both tion of coexisting trees (e.g., Quercus suber; Magri et al., 2007). continental (i.e., Strait of Gibraltar) and regional (Guadalquivir Therefore, the genus Erophaca could be reasonably included in River) differentiation. the group of (mostly woody) Tertiary relicts that form part of the 448 R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450

Table 3 identifiable variants of this subspecies. The Strait of Gibraltar Genetic affinities among geographic sub-areas inhabited by E. baetica subsp. baetica. undoubtedly represents an important geographic barrier to gene Pairwise fixation indices (F ) appear below, and numbers of shared private fragments ST flow, and the long-lasting separation between African and Iberian (Fragpriv) above the diagonal. The sub-area-specific numbers of private fragments

(FragPriv) are given in bold. populations (the opening of the Gibraltar Strait is dated around 5.3 mya; Duggen et al., 2003) has allowed genetic differentiation Sub-areas Frag Priv to accumulate, just as in many other herbaceous taxa (Escudero Iberian Betic Rif Mountain Middle Atlas et al., 2008), woody legumes (Pardo et al., 2008), and even tree spe- Massif Ranges cies (Terrab et al., 2008). It must also be noted that African popu- Iberian Massif – 31 5 8 23 lations are much more diverse genetically than European ones, Betic Ranges 0.19 – 11 1 9 despite the plant being (currently) relatively rare in N. Africa. Rif Mountain 0.27 0.22 – 36 41 Middle Atlas 0.28 0.27 0.13 – 31 Our data for Erophaca support the view that Morocco is an important reservoir of genetic diversity in plants (Medail and Diadema, 2009). This diversity can be due to the ancestral nature of populations (in the case of Erophaca, they present the largest numbers of Mediterranean flora at present (e.g., Chamaerops, Arbutus, Olea, Pis- private and rare fragments which tend to accumulate over time; tacia,orPhillyrea; see Palamarev (1989) for a review), and the fact Stehlik et al., 2002; Schönswetter and Tribsch, 2005) or, alterna- that it presents some attributes relatively unusual in typical Med- tively, be a by-product of larger population sizes during Pleistocene iterranean plants (e.g., blooming in winter, and a woody xylopode climatic fluctuations. that acts as the underground storage organ) is hardly surprising. As regards regional genetic differentiation at each side of the This relictic, pre-Mediterranean status has also been claimed for Guadalquivir River Valley, even though this has been demon- other legume genera considerably younger than Erophaca (e.g., strated in other taxa (e.g., Hypochaeris radicata; Ortiz et al., Anagyris; Ortega-Olivencia and Catalán, 2009). 2008), in the case of E. baetica the role of this barrier impeding gene Erophaca was long included in Astragalus (as A. lusitanicus Lam.) flow should be considered less important (compared to the Strait until Podlech (1993) placed it (on the basis of morphology) in a dif- of Gibraltar). According to Mantel tests, southernmost Iberian pop- ferent genus. This is strongly supported by molecular data in the ulations (i.e., on the Betic Sierras) were genetically similar to the present study (nrDNA ITS, Fig. 2) which, in addition, identified African ancestral pool, whereas those from the Iberian Massif ap- Astragalus, Oxytropis and Swainsona as the closest relatives of peared relatively impoverished and poorly structured (suggesting Erophaca. Furthermore, on the basis of nrDNA ITS the genus should they are relatively recent). In general, however, genetic structuring not be considered close to Chesneya (but see Wojciechowski et al. was modest within the Iberian Peninsula and more than half of ge- (2000)). Nevertheless, nrDNA ITS sequences for Gueldenstaedtia netic variance occurred among plants within populations, as could and other genera are still unavailable, so more data are needed be- be expected in a long-lived, obligate outcrosser with low seed dis- fore the phylogenetic relationships among these genera can be persability (Hamrick et al.,1992; Nybom, 2004). clearly established. A phylogenetic tree based on matK sequences (Wojciechowski, personal communication) also distances Erophaca from Chesneya and Gueldenstaedtia. 4.3. Subspecies evolution and extant distribution A few morphological attributes of Erophaca appear also in the Sophoreae (e.g., leaflets with stipelles, and a characteristic mor- According to data in the present study, the most recent ancestor phology of the stigma), so it is understandable that the genus shared by the two disjunct subspecies of E. baetica would be was once included in this tribe (Podlech, 1993). There is general around 11.9 mya old, so they most likely evolved during the agreement at present that Erophaca belongs to the Galegeae (Lewis cool-dry cycles of the Miocene that led to disappearance of many et al., 2005), whereas the existence of a subtribe Erophacinae tropical elements of the flora (Quezel, 1978; Thompson, 2005; (Talavera, 1999) is debatable. Molecular data in the present study Postigo Mijarra et al., 2009). It may seem at first sight that the suggest that Erophaca alone would be the distinct, immediate sister two subspecies have accumulated very few differences after being to the Astragalean clade (including Astragalus, Oxytropis, and separated for such a long period of time (just four nrDNA ITS sub- Swainsona among others), so placing Erophaca in a subtribe of its stitutions in nearly 12 my), but this is not unexpected in long-lived own could be justified. herbaceous perennials (Kay et al., 2006). E–W disjunctions abound in the Mediterranean flora, and are 4.2. Genetic differences across continental and regional barriers in often explained on the basis of contraction of once more continu- E. baetica subsp. baetica ous areas (i.e., vicariance: Davis and Hedge, 1971; Rosselló et al., 2007), mainly brought about by an accentuation of aridity (Hsu AFLP data in the present study identified Morocco, the Betic et al., 1973; Griffin, 2002; Thompson, 2005). This process can pro- Sierras, and the Iberian Massif as sub-areas that shelter genetically vide a likely explanation for the present distribution of our two

Table 4 Genetic variation in E. baetica subsp. baetica as affected by geographic barriers. In each AMOVA model, the ‘Among groups’ source of variation accounts for the effect of the barrier(s), whereas ‘Among plants’ is equivalent to the error term. Fixation indices (FST) measure genetic distances across barriers.

Barrier Sources of variation d.f. Variance components % of variance FST (95% CI) None Among populations 18 17.6 30.0 0.30 (0.28–0.33) Among plants 170 41.2 70.0 Strait of Gibraltar Among groups 1 11.9 18.0 0.37 (0.35–0.40) Among populations 17 13.0 19.7 Among plants 170 41.2 62.3 Guadalquivir River Valley Among groups 1 9.3 14.7 0.35 (0.33–0.37) Among populations 17 12.8 20.3 Among plants 170 41.2 65.0 Guadalquivir River Valley + Strait of Gibraltar Among groups 2 12.3 19.3 0.35 (0.33–0.37) Among populations 16 10.1 15.9 Among plants 170 41.2 64.8 R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450 449 subspecies. On the other hand, molecular data could be interpreted Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by to provide an alternative scenario, namely that the subsp. orientalis sampling trees. BMC Evol. Biol. 7, 214. Duchesne, P., Bernatchez, L., 2002. Aflpop: a computer program for simulated derives from the Western subspecies. If this was the case, and judg- and real population allocation, based on AFLP data. Mol. Ecol. Notes 2, ing from the number of shared private fragments (Table 2), candi- 380–383. dates to represent the ancestral pool would be the N. African Duggen, S., Hoernle, K., van den Bogaard, P., Rupke, L., Morgan, J.P., 2003. Deep roots of the Messinian salinity crisis. Nature 422, 602–606. populations. It is often assumed in the biogeography of the Medi- Ehrich, D., 2006. Aflpdat: a collection of r functions for convenient handling of AFLP terranean flora that ‘Eastern’ is equivalent to ancestral, so the re- data. Mol. Ecol. Notes 6, 603–604. verse scenario just described would be quite exceptional. To our Escudero, M., Vargas, P., Valcarcel, V., Luceño, M., 2008. Strait of Gibraltar: an effective gene-flow barrier for wind-pollinated Carex helodes (Cyperaceae) as knowledge, the only example of a similar (but non-Mediterranean) revealed by DNA sequences, AFLP, and cytogenetic variation. Am. J. Bot. 95, biogeographical pattern is that of the montane subspecies of 745–755. Anthyllis montana, in which the western ones have been demon- Evanno, G., Regnaut, S., Goudet, J., 2005. Detecting the number of clusters of individuals using the software structure: a simulation study. Mol. Ecol. 14, strated to be ancestral, whereas Eastern subspecies are derived 2611–2620. (Kropf et al., 2002). Palaeogeographical reconstructions for Medi- Excoffier, L., Laval, G., Schneider, S., 2005. Arlequin version 3.0: an integrated terranean animal groups have often proposed the North of Africa software package for population genetics data analysis. Evol. Bioinform. Online as an ancestral area (Oosterbroek and Arntzen, 1992; Sanmartin, 1, 47–50. Falush, D., Stephens, M., Pritchard, J.K., 2003. Inference of population structure 2003; Carranza et al., 2006). using multilocus genotype data linked loci and correlated allele frequencies. Genetics 164, 1567–1587. Griffin, D.L., 2002. Aridity and humidity: two aspects of the late Miocene climate of Acknowledgments North Africa and the Mediterranean. Palaeogeogr. Palaeoclimatol. 182, 65–91. Hamrick, J.L., Godt, M.J.W., Sherman-Broyles, S.L., 1992. Factors influencing levels of genetic diversity in woody plant species. New For. 6, 95–124. We are grateful to Prof. P.E. Gibbs for a critical reading of an Hewitt, G.M., 1996. Some genetic consequences of ice ages, and their role, in early version of the manuscript, and to Prof. M. Wojciechowski divergence and speciation. Biol. J. Linn. Soc. 54, 246–276. Hewitt, G., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907– for valuable suggestions on the phylogeny of the Astragalean clade. 913. Prof. T.F. Stuessy allowed access to laboratory facilities at the Hsu, K.J., Ryan, W.B., Cita, M.B., 1973. Late Miocene desiccation of the University of Vienna, and staff of the Biology Laboratory Service Mediterranean. Nature 242, 239–243. (CITIUS, Seville University) assisted during DNA extractions. R. Huson, D.H., Bryant, D., 2006. Application of phylogenetic networks in evolutionary studies. Mol. Biol. Evol. 23, 254–267. Santos kindly supplied leaves of E. baetica subsp. orientalis. This Jordano, D., Haeger, J.F., Rodríguez, J., 1990. The effect of seed predation by Tomares study was financially supported by a predoctoral grant of the Junta ballus (Lepidoptera: Lycaenidae) on Astragalus lusitanicus(Fabaceae): de Andalucía to R. Casimiro-Soriguer; by Proyecto de Excelencia determinants of differences among patches. Oikos 57, 250–256. Kay, K.M., Whittall, J.B., Hodges, S.A., 2006. A survey of nuclear ribosomal internal 2005/RNM848 of the Junta de Andalucía; and by Grant CGL2006- transcribed spacer substitution rates across angiosperms: an approximate 00817 of the Spanish Ministerio de Educación y Ciencia to S. Tala- molecular clock with life history effects. BMC Evol. Biol. 6, 36. vera. The Doñana National Park(Project 33/2004) provided permis- Kropf, M., Kadereit, J.W., Comes, H.P., 2002. Late Quaternary distributional stasis in the submediterranean mountain plant Anthyllis montana L. (Fabaceae) inferred sion to access sampling sites. from ITS sequences and amplified fragment length polymorphism markers. Mol. Ecol. 11, 447–463. Lavin, M., Herendeen, P.S., Wojciechowski, M.F., 2005. Evolutionary rates analysis of References Leguminosae implicates a rapid diversification of lineages during the Tertiary. Syst. Biol. 54, 575–594. Liston, A., 1995. Use of the polymerase chain reaction to survey for the loss of the Arroyo, J., Aparicio, A., Albaladejo, R.G., Muñoz, J., Braza, R., 2008. Genetic structure inverted repeat in the legume chloroplast genome. In: Crisp, M.D., Doyle, J.J. and population differentiation of the Mediterranean pioneer spiny broom (Eds.), Advances in Legumes Systematics, Part 7, Phylogeny. Royal Botanic Calicotome villosa across the Strait of Gibraltar. Biol. J. Linn. Soc. 93, 39–51. Gards, Kew, pp. 31–40. Bel-Kassaoui, H., Lamnaouer, D., Jossang, A., el Abdennebi, H., Charrouf, Z., Bodo, B., Lewis, G., Schrire, B., Mackinder, B., Lock, M. (Eds.), 2005. Legumes of the World. 2008. Role of hypaphorine in the toxicity of Astragalus lusitanicus. Nat. Prod. Res. Royal Botanic Gards, Kew. 22, 461–465. Magri, D., Fineschi, S., Bellarosa, R., Buonamici, A., Sebastiani, F., Schirone, B., Breinholt, J.W., Van Buren, R., Kopp, O.R., Stephen, C.L., 2009. Population genetic Simeone, M.C., Vendramin, G.G., 2007. The distribution of Quercus suber structure of an endangered Utah endemic, Astragalus ampullarioides (Fabaceae). chloroplast haplotypes matches the palaeogeographical history of the western Am. J. Bot. 96, 661–667. Mediterranean. Mol. Ecol. 16, 5259–5266. Cano-Maqueda, J., Talavera, S., Arista, M., Catalan, P., 2008. Speciation and Medail, F., Diadema, K., 2009. Glacial refugia influence plant diversity patterns in biogeographical history of the Campanula lusitanica complex the Mediterranean Basin. J. Biogeogr. 36, 1333–1345. (Campanulaceae) in the Western Mediterranean region. Taxon 57, 1252–1266. Medail, F., Quezel, P., 1999. Biodiversity hotspots in the Mediterranean basin: Carranza, S., Arnold, E.N., Pleguezuelos, J.M., 2006. Phylogeny, biogeography, and setting global conservation priorities. Conserv. Biol. 13, 1510–1513. evolution of two Mediterranean snakes, Malpolon monspessulanus and Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Hemorrhois hippocrepis (Squamata, Colubridae), using mtDNA sequences. Mol. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. Phylogenet. Evol. 40, 532–546. Nei, M., Li, W.H., 1979. Mathematical model for studying genetic variation in terms Caujape-Castells, J., Jansen, R.K., 2003. The influence of the Miocene Mediterranean of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76, 5269–5273. desiccation on the geographical expansion and genetic variation of Nybom, H., 2004. Comparison of different nuclear DNA markers for estimating Androcymbium gramineum (Cav.) McBride (Colchicaceae). Mol. Ecol. 12, intraspecific genetic diversity in plants. Mol. Ecol. 36, 1143–1155. 1515–1525. 333, 233–240. Oosterbroek, P., Arntzen, J.W., 1992. Area-cladograms of Circum-Mediterranean Collina-Girard, J., 2001. L’Atlantide devant le détroit de Gibraltar? Mythe et géologie taxa in relation to Mediterranean palaeogeography. J. Biogeogr. 19, 3–20. Atlantis off the Gibraltar Strait? Myth and geology. Earth Planet. Sci. 333, 233– Ortega-Olivencia, A., Catalán, P., 2009. Systematics and evolutionary history of the 240. circum-Mediterranean genus Anagyris L. (Fabaceae) based on morphological Cubas, P., Pardo, C., Tahiri, H., 2005. Genetic variation and relationships among Ulex and molecular data. Taxon 58, 1290–1306. (Fabaceae) species in southern Spain and northern Morocco assessed by Ortiz, M.A., Tremetsberger, K., Talavera, S., Stuessy, T., Garcia-Castaño, J.L., 2007. chloroplast microsatellite (cpSSR) markers 1. Am. J. Bot. 92, 2031–2043. Population structure of Hypochaeris salzmanniana DC. (Asteraceae), an endemic Chung, M., Gelembiuk, G., Givnish, T.J., 2004. Population genetics and species to the Atlantic coast on both sides of the Strait of Gibraltar, in relation to phylogeography of endangered Oxytropis campestris var. chartacea and Quaternary sea level changes. Mol. Ecol. 16, 541–552. relatives: arctic-alpine disjuncts in eastern North America. Mol. Ecol. 13, Ortiz, M.Á., Tremetsberger, K., Terrab, A., Stuessy, T.F., García-Castaño, J.L., Urtubey, 3657–3673. E., Baeza, C.M., Ruas, C.F., Gibbs, P.E., Talavera, S., 2008. Phylogeography of the Culham, A., 2007. Leguminosae. In: Heywood, V.H., Brummitt, R.K., Culham, A., invasive weed Hypochaeris radicata (Asteraceae): from Moroccan origin to Seberg, O. (Eds.), Flowering Plants of the World, 185–188. Royal Botanic Gards, worldwide introduced populations. Mol. Ecol. 17, 3654–3667. Kew. Ortiz, M.Á., Tremetsberger, K., Stuessy, T.F., Terrab, A., García-Castaño, J.L., Talavera, Davis, P.H., Hedge, I.C., 1971. Floristic links between NW Africa and SW Asia. Ann. S., 2009. Phylogeographic patterns in Hypochaeris section Hypochaeris Naturhist. Mus. Wien 75, 43–57. (Asteraceae, Lactuceae) of the western Mediterranean. J. Biogeogr. 36, 1384– Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of 1397. fresh leaf tissue. Phytochem. Bull. 19, 11–15. Palamarev, E., 1989. Paleobotanical evidences of the Tertiary history and origin of Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram the Mediterranean sclerophyll dendroflora. Plant Syst. Evol. 162, 93–107. for ecologists. J. Stat. Softw. 22 (4), 1–20. 450 R. Casimiro-Soriguer et al. / Molecular Phylogenetics and Evolution 56 (2010) 441–450

Pardo, C., Cubas, P., Tahiri, H., 2008. Genetic variation and phylogeography of Taberlet, P., Fumagalli, L., Wust-Saucy, A.G., Cosson, J.F., 1998. Comparative Stauracanthus (Fabaceae, Genisteae) from the Iberian Peninsula and northern phylogeography and postglacial colonization routes in Europe. Mol. Ecol. 7, Morocco assessed by chloroplast microsatellite (cpSSR) markers. Am. J. Bot. 95, 453–464. 98–109. Talavera, S., 1999. Sobre el tratamiento de la tribu Astragaleae Bercht. et J. Presl Piñeiro, R., Aguilar, J.F., Munt, D.D., Feliner, G.N., 2007. Ecology matters: Atlantic– (Papilionoideae, Leguminosae) en Flora Iberica. Ann. Jard. Bot. Madrid 57, 218–220. Mediterranean disjunction in the sand-dune shrub Armeria pungens Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: molecular evolutionary (Plumbaginaceae). Mol. Ecol. 16, 2155–2171. genetics analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599. Podlech, D., 1993. Beiträge zur Kenntnis der Gattung Astragalus L. (Leguminosae). Terrab, A., Talavera, S., Arista, M., Paun, O., Stuessy, T.F., Tremetsberger, K., 2007. Sendtera 1, 267–272. Genetic diversity at chloroplast microsatellites (cpSSRs) and geographic Polhill, R.M., 1981. Galegeae. 1. In: Polhill, R.M., Raven, P.H. (Eds.), Advances in structure in endangered West Mediterranean firs (Abies spp., Pinaceae). Legume Systematics. Royal Botanic Gards, Kew, pp. 357–363. Taxon 56, 409–416. Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Terrab, A., Schönswetter, P., Talavera, S., Vela, E., Stuessy, T.F., 2008. Range-wide Bioinformatics 14, 817–818. phylogeography of Juniperus thurifera L., a presumptive keystone species of Postigo Mijarra, J.M., Barrón, E., Gómez Manzanaque, F., Morla, C., 2009. Floristic western Mediterranean vegetation during cold stages of the Pleistocene. Mol. changes in the Iberian Peninsula and Balearic Islands (south-west Europe) Phylogenet. Evol. 48, 94–102. during the Cenozoic. J. Biogeogr. 36, 2025–2043. Thompson, J.D., 2005. Plant Evolution in the Mediterranean. Oxford University Pritchard, J.K., Stephens, M., Donnelly, P., 2000. Inference of population structure Press, USA. using multilocus genotype data. Genetics 155, 945–959. Travis, S.E., Maschinski, J., Keim, P., 1996. An analysis of genetic variation in Quezel, P., 1978. Analysis of the flora of Mediterranean and Saharan Africa. Ann. Mo. Astragalus cremnophylax var. cremnophylax, a critically endangered plant, using Bot. Gard., 479–534. AFLP markers. Mol. Ecol. 5, 735–745. Rodríguez-Sánchez, F., Pérez-Barrales, R., Ojeda, F., Vargas, P., Arroyo, J., 2008. The Tremetsberger, K., Urtubey, E., Terrab, A., Baeza, C.M., Ortiz, M.A., Talavera, M., Strait of Gibraltar as a melting pot for plant biodiversity. Q. Sci. Rev. 27, 2100– König, C., Temsch, E., Kohl, G., Talavera, S., Stuessy, T.F., 2009. Pleistocene 2117. refugia and polytopic replacement of diploids by tetraploids in the Patagonian Rosselló, J.A., Lazaro, A., Cosin, R., Molins, A., 2007. A phylogeographic split in Buxus and Subantarctic plant Hypochaeris incana (Asteraceae, Cichorieae). Mol. Ecol. balearica (Buxaceae) as evidenced by nuclear ribosomal markers: when ITS 18, 3668–3682. paralogues are welcome. J. Mol. Evol. 64, 143–157. Vos, P., Hogers, R., Bleeker, M., Reijans, M., van de Lee, T., Hornes, M., Frijters, A., Pot, Sanmartin, I., 2003. Dispersal vs. vicariance in the Mediterranean: historical J., Peleman, J., Kuiper, M., Zabeau, M., 1995. AFLP: a new concept for DNA biogeography of the Palearctic Pachydeminae (Coleoptera, Scarabaeoidea). J. fingerprinting. Nucleic Acids Res. 23, 4407–4414. Biogeogr. 30, 1883–1897. Weising, K., Nybom, H., Wolff, K., Kahl, G., 2005. DNA Fingerprinting in Plants: Schmitt, T., 2007. Molecular biogeography of Europe: Pleistocene cycles and Principles, Methods, and Applications, 2nd ed. CRC Press, Boca Raton. postglacial trends. Front. Zool. 7, 11. White, T.J., Bruns, T., Lee, S., Taylor, J., 1990. Amplification and direct sequencing of Schlüter, P.M., Harris, S.A., 2006. Analysis of multilocus fingerprinting data sets fungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H., containing missing data. Mol. Ecol. Notes 6, 569–572. Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Schönswetter, P., Tribsch, A., 2005. Vicariance and dispersal in the alpine perennial Applications. Academic Press Inc., San Diego, pp. 315–322. Bupleurum stellatum L. (Apiaceae). Taxon 54, 725–732. Wojciechowski, M.F., 2005. Astragalus (Fabaceae): a molecular phylogenetic Stehlik, I., Schneller, J.J., Bachmann, K., 2002. Immigration and in situ glacial perspective. Brittonia 57, 382–396. survival of the low-alpine Erinus alpinus (Scrophulariaceae). Biol. J. Linn. Soc. Wojciechowski, M.F., Sanderson, M.J., Hu, J.M., 1999. Evidence on the monophyly of 77, 87–103. Astragalus Fabaceae) and its major subgroups based on nuclear ribosomal DNA Suc, J.P., 1984. Origin and evolution of the Mediterranean vegetation and climate in ITS and chloroplast DNA trnL intron data. Syst. Bot. 24, 409–437. Europe. Nature 307, 429–432. Wojciechowski, M.F., Sanderson, M.J., Steele, K.P., Liston, A., 2000. Molecular Swofford, D.L., 2002. PAUP* 4.0 (beta version). Computer Program and phylogeny of the ‘‘Temperate Herbaceous Tribes” of Papilionoid legumes: a Documentation Distributed by Sinauer, Sunderland, Mass. supertree approach. Adv. Legume Syst. 9, 277–298.