Origins of Allium ampeloprasum horticultural groups and a molecular phylogeny of the section Allium (Allium: Alliaceae)

Molecular Phylogenetics and Evolution 54 (2010) 488–497

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev

Origins of Allium ampeloprasum horticultural groups and a molecular phylogeny of the section Allium (Allium: Alliaceae) Pablo Hirschegger a, Jernej Jakše a, Peter Trontelj b, Borut Bohanec a,* a b

Biotechnical Faculty, University of Ljubljana, Agronomy Department, Jamnikarjeva 101, 1000 Ljubljana, Slovenia Biotechnical Faculty, University of Ljubljana, Department of Biology, Vecˇna pot 111, 1000 Ljubljana, Slovenia

a r t i c l e

i n f o

Article history: Received 12 May 2009 Revised 25 August 2009 Accepted 28 August 2009 Available online 3 September 2009 Keywords: Allium subgenus Allium Allium ampeloprasum Horticultural groups Nuclear ribosomal internal transcribed spacer (ITS) Intra-individual polymorphism Chloroplast trnD-trnT Chloroplast trnL-trnF Molecular phylogeny

a b s t r a c t The subgenus Allium section Allium includes economically important species, such as garlic and leek, as well as other polyploid minor crops. Phylogenetic studies within this section, with a focus on horticultural groups within A. ampeloprasum, were performed on 31 accessions of 17 species using the nuclear ribosomal DNA internal transcribed spacer (ITS) region and the chloroplast trnL-F and trnD-T regions. The results confirmed the monophyly of section Allium. Four main clades were identified on all ITS analyses but the relationships among those and the remaining species studied within section Allium remained unresolved. Trees based on cpDNA recovered two major clades and a topology only partly congruent with that of the ITS tree. Intra-individual polymorphism of the ITS region proved useful in tracking putative parent species of polyploid taxa. The allopolyploid origin of great headed garlic (GHG), A. iranicum and A. polyanthum was confirmed. No signs of hybridization in leek or kurrat were detected but possible introgression events were identified in pearl onion and bulbous leek. Although GHG is often used as a garlic substitute, molecular analysis revealed only a distant relationship with garlic. We also clarified the previous incorrect classification of cultivated forms within A. ampeloprasum, by showing that leek, kurrat, pearl onion, and bulbous leek should be considered separately from GHG. Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction The genus Allium comprises around 750 species according to Stearn (1992). This is in fair agreement with the current online version of the World Checklist of Selected Plant Families maintained by Royal Botanic Gardens, KEW (UK, http://apps.kew.org/wcsp/reportbuilder.do), which recognizes 860 species. The latest intrageneric classification divides the genus Allium into 15 subgenera and 72 sections (Friesen et al., 2006). The subgenus Allium is the largest, comprising around 280 species (Hanelt et al., 1992), 114 of which compose its largest section, Allium (Mathew, 1996). This section includes economically important species, such as garlic (A. sativum L.) and leek (A. ampeloprasum L.), as well as other minor crops of local importance, such as great headed garlic (GHG), and kurrat. Extensive research has been done in garlic, not only for its culinary and economic value but also for its health related benefits (e.g. Tattelman, 2005). Similar properties have also been described in other species of this section, such as A. ampeloprasum (particularly for leek and GHG) and A. polyanthum Schultes & Schultes fil. (Messiaen et al., 1993).

* Corresponding author. E-mail addresses: [email protected] (P. Hirschegger), jernej.jakse@ bf.uni-lj.si (J. Jakše), [email protected] (P. Trontelj), [email protected] (B. Bohanec). 1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.08.030

Despite the major importance of the section Allium, it has not been subjected to a comprehensive molecular taxonomic evaluation; only partial molecular genetic studies that involved a limited set of species have been published (Kik et al., 1997; Havey and Leite, 1999; Bohanec et al., 2005). Interspecific and infraspecific relationships within this section still remain unresolved. As reviewed by Mathew (1996), polyploidy is a common feature in section Allium. However, to a certain extent, it has been left unexplored, leaving the origin of polyploid species undetermined. A good example of this is A. ampeloprasum, whose boundaries are particularly difficult to establish. This is reflected in the extensive number of subspecies and varieties, as well as synonymous species that have been proposed and revised by different authors over the years (Regel, 1875; Boissier, 1884; Vvedensky, 1935; Feinbrun, 1943; Helm, 1956; Kollmann, 1971; Wendelbo, 1971; Mathew, 1996). Bothmer (1970) studying the section Allium in Greece, introduced the concept of ‘‘A. ampeloprasum complex”, a group of closely related species including A. ampeloprasum, A. bourgeaui Rech. fil., and A. commutatum Guss. Later on, Bothmer (1974) also included A. atroviolaceum Boiss. According to Guern et al. (1991), A. polyanthum is another species that should be included in the ‘‘A. ampeloprasum complex”. It is generally accepted that A. ampeloprasum is composed of four horticultural groups, known as the leek group, kurrat group, pearl onion group and GHG group (Jones and Mann, 1963) as well

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

as of wild representatives. Other cultivated alliums were later recognized as parts of A. ampeloprasum, e.g. tarée irani (Tahbaz, 1971), poireau perpetuél and prei anak (Van der Meer and Hanelt, 1990), Mushuu-ninniku (Ariga et al., 2002), and a form of bulbous leek cultivated in China (Bohanec et al., 2005). This ‘‘species complex” is characterized by different ploidy levels (see Mathew, 1996; Hirschegger et al., 2006) and surrounded by controversies concerning abnormal meiotic behavior (see Khazanehdari and Jones, 1997) and genome constitution (Smilde et al., 1999), features that makes it very interesting to study. Allium ampeloprasum is extremely variable. As a wild plant, it is widely distributed across the Mediterranean basin (sometimes as weedy plants) through the Middle East into western and southern former U.S.S.R. (Jones and Mann, 1963). Minor crops of A. ampeloprasum are locally cultivated from Asia Minor to Iran and the Caucasus, and sporadically in California and other regions of America and Europe (Fritsch and Friesen, 2002). Due to its economic importance, leek has been subjected to intensive cultivation and included in breeding programs in Europe and, to some extent, in other continents. These programs have often encountered problems due to the above mentioned uncommon features, problems related to the tetraploid nature of leek; its tetrasomic inheritance and high inbreeding depression (see Schweisguth, 1970; Smith and Crowther, 1995). Our studies aim to clarify phylogenetic relationships within the subgenus Allium section Allium. Particular emphasis was laid on the re-evaluation of the position of the horticultural groups within A. ampeloprasum since evidence, conflicting with the traditional classification, has been observed in our previous molecular and karyotype studies on bulbous leek-like accessions (Bohanec et al., 2005) and fertile accessions of octoploid GHG (Hirschegger et al., 2006). To address these questions, molecular analyses based on polymorphic nuclear (ITS) and chloroplast DNA regions were performed on cultivated alliums and representative wild species belonging to the section Allium, as well as to the wider subgenus Allium. 2. Material and methods 2.1. Plant materials We included 31 accessions of seventeen Allium species in this study. Several species related to A. ampeloprasum according to the informal classification made by Mathew (1996) were gathered from various institutions: Royal Botanic Gardens Kew (http://data.kew.org/dnabank/homepage.html) (Richmond, UK), Leibniz-Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK) (Gaterleben, Germany), Centre for Genetic Resources (CGN) (Wageningen, The Netherlands), Botanischer Garten und Botanisches Museum Berlin-Dahlem (Berlin, Germany) and EEAA INTA La Consulta (Mendoza, Argentina). Material kindly donated by C. M. Messiaen (Le Teil, France), collected from the wild as well as purchased from seed companies, was also included. Sequences of two more species from the section Allium, as well as twenty-five other species of the subgenus Allium, were obtained from the National Center for Biotechnology Information (NCBI) GenBank databases (http://www.ncbi.nlm.nih.gov/). The source of plant materials and vouchers of the accessions used in this study, as well as sequence accession numbers, are listed and described in Table S1 (Supplemental material). 2.2. Molecular methods Total DNA was extracted from true young leaves cut from bulbs or seedlings following the common CTAB procedure with modifications made by Kump and Javornik (1996).

489

The ITS region was PCR-amplified using primers ITS5 (50 GGAAG TAAAAGTCGTAACAAGG30 ) and ITS4 (50 TCCTCCGCTTATTGATATG C30 ) (White et al., 1990). Both primers anneal to small and large ribosomal subunits, so both ITS regions and the 5.8S gene are complete. PCR was carried out in a total volume of 25 ll containing 40 ng of plant DNA, 1 supplied PCR buffer (Fermentas Life sciences, Lithuania), 2.5 mM MgCl2, 0.2 mM of each dNTP, 0.5 lM of each primer and 1.25 U of Taq DNA polymerase (Fermentas Life sciences, Lithuania). Samples were amplified in either Thermoblock TGRADIENT (Biometra, Germany) or a 2720 Thermal Cycler (Applied Biosystems, USA) without noticeable differences in reaction outcome, with the following amplification protocol: initial denaturation 94 °C for 2 min, followed by 30 cycles of 94 °C 1 min, 61 °C 1 min and 72 °C 2 min, reactions were stopped by incubation at 72 °C for 7 min and cooled to 12 °C until recovery. Two cpDNA regions, trnD-trnT and trnL-trnF, were amplified with universal primers developed by Taberlet et al. (1991): trnL (UAA) F (Tab C) (50 CGAAATCGGTAGACGCTACG30 ) and trnF (GAA) (TabF) (50 ATTTGAACTGGTGACACGAG30 ); and by Demesure et al. (1995): trnD (GUC) (50 ACCAATTGAACTACAATCCC30 ) and trnT (GGU) (50 CTACCACTGAGTTAAAAGGG30 ). PCR chemical components were the same as for ITS amplification, and the amplification protocols were as follows. For the trnD-trnT region, touchdown PCR was employed according to Friesen et al. (2000): initial denaturation 94 °C for 2 min, followed by 10 cycles of 94 °C 20 s, 65 °C (-1 °C/cycle) 1 min and 72 °C 1 min, then 40 cycles of 94 °C 20 s, 55 °C 45 s and 72 °C 1 minute, reactions were stopped by incubation at 72 °C for 7 min and cooled to 12 °C until recovery. For trnL-trnF, the following conditions for amplification were applied: initial denaturation 94 °C for 5 min, followed by 32 cycles of 94 °C 1 minute, 61° 1 min and 72 °C 1 min 30 s, reactions were stopped by incubation at 72 °C for 7 min and cooled to 12 °C until recovery. PCR products were sequenced from both sides using amplification primers by a sequencing service (MACROGEN inc., Seoul, Korea). Heterogeneous PCR amplicons were cloned and sequenced using universal vector primers. This approach was applied to the following accessions: – A. ampeloprasum: cult. ‘‘Joliet” (leek), cult. ‘‘Balady-Saidii” CGN 14710 (kurrat), ‘‘Chinese” (bulbous leek), ”Pearl Onion” (pearl onion), ‘‘Clone FCA” (GHG), ‘‘Clone 1007” (GHG), ‘‘Clone Pazin” (GHG), ‘‘ALL1355” (GHG) and ‘‘Clone Spain 9” (GHG). – A. ampeloprasum var. babingtonii (Borrer) Syme, Chase: 24451 (KEW). – A. polyanthum (provided by Messiaen). – A. truncatum (Feinbr.) Kollmann & D. Zohary, Chase: 24455 (KEW). – A. iranicum (Wend.) Wend., Chase: 24454 (KEW). – A. pyrenaicum Costa et Vayreda, Chase: 24769 (KEW). All sequences were submitted to GenBank and are available under accession numbers EU626238–EU626263, EU626264– EU626289, EU626290–EU626396, FJ664286–FJ664340, FJ628600– FJ628604 and FJ628605–FJ628609. 2.3. Data analyses Forward and reverse sequence trace files were edited and assembled using CodonCode Aligner (version 2.0.6; CodonCode Corp., Dedham, MA, USA). Sequences were aligned with Muscle v3.6 (Edgar, 2004) under default options. Modelgenerator (Keane et al., 2006) was used to determine the best fit model of evolution. Rate heterogeneity across sites was approximated by six gamma categories. Models were selected according to the second order Akaike information criterion (AIC2). Nuclear (ITS) and chloroplast DNA sequences were analyzed separately since they yielded incongruent topologies and differed

490

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

substantially in the amount and strength of phylogenetic signal. Moreover, several allopolyploids among the analyzed taxa implied reticulated relationships. To resolve them, different modes of evolution and inheritance (concerted evolution in ITS, maternal inheritance in cpDNA) were taken into account in separate analyses of the two data sets. The nuclear data set included one hundred and 69 ITS sequences of selected species of the section Allium. Fourteen individuals showing intra-individual sequence heterogeneity were represented by ten sequences obtained by cloning (see list above). The selected model of evolution was Tamura and Nei (1993) with gamma distributed rate heterogeneity and no invariable sites (TrN+G). Twenty-eight ITS sequences from species of other sections of the subgenus Allium as well as A. cepa and A. fistulosum (subgenus Cepa) were chosen as outgroups. The chloroplast data set consisted of 31 assembled sequences from both chloroplast regions. The selected model of evolution was Kimura (1981) three-parameters with unequal base frequencies (K81uf) in the case of the trnD-trnT region, and the Hasegawa et al. (1985) model (HKY) for the trnL-trnF region. A. oleraceum L. (section Codonoprasum) and A. cepa L. and A. fistulosum L. (subgenus Cepa) were chosen as outgroups. Bayesian inference phylogenetic analyses were run in MrBayes v3.1.2. (Ronquist and Huelsenbeck, 2003). All searches were performed in four parallel runs with four chains each for 4  106 generations for ITS analysis and 1  106 for chloroplast DNA, sampled every 100th generation. Searches with combined sequence data were performed under mixed models, all parameters except topology and branch lengths unlinked. After discarding the first 25% of the sampled trees, final topologies were consented following the 50% majority rule. In addition to Bayesian inference, both data sets were analyzed under the parsimony criterion using PAUP 4.0 b10 (Swofford, 2002). All characters were treated as unordered and equally weighted. Gaps were treated as missing, multistate positions interpreted as uncertainty. Starting trees were obtained via stepwise addition. One hundred starting trees were obtained by random addition of sequences and branch-swapped using tree-bisection and reconnection (TBR). Node support was assessed by 1000 non-parametric bootstrap replicates. Finally, ITS sequence clones from each putative polyploid were analyzed together with their potential parent species, using the same methods and models as described above. Potential parent species were chosen based on the topology obtained from the full nuclear data set.

3. Results 3.1. DNA sequences The ITS region ranged from 678 bp in A. ampeloprasum ‘‘Clone Pazin” (EU626396) to 772 bp in A. ampeloprasum leek group cult. ‘‘Joliet” (FJ664304). The final alignment was 793 bp long with 504 variable sites, of which 378 were parsimony-informative. No major length polymorphisms were observed among amplicons except for an amplicon of A. ampeloprasum GHG group ‘‘Clone Pazin” (EU626396), which possessed two deletions, one in the 5.8S gene and one in the ITS2 region. Inclusion or exclusion of this amplicon did not affect the overall topology of the phylogenetic tree, so it was kept for further analyses. The length of the unaligned trnD-trnT region ranged from 898 bp in A. oleraceum (FJ628607) to 945 bp in A. cepa (FJ628608). The final alignment was 949 bp long with 98 variable sites, of which 32 were parsimony-informative. As expected, almost all variable sites occurred within the spacers.

The length of the trnL-trnF region ranged from 657 bp in A. truncatum (EU626244) to 673 bp in A. oleraceum (FJ628602). The final alignment was 684 bp long with 47 variable sites, of which 23 were parsimony-informative. No variable sites occurred within the first intron of the trnL gene and the trnF gene. On the other hand, two substitutions were found in the second intron of the trnL gene. 3.2. Phylogenetic analyses Bayesian inference and parsimony analysis of the nuclear data set yielded similar tree topologies. All major relationships were relatively well supported by both methods). The results are presented as a condensed tree (Fig. 1); the original complete tree is available upon request. All representatives of the section Allium formed a single, well-supported clade. Within this section, two main groups were recovered: Clade A: (see Fig. 1) included Group 2 and Group 3 (see Table 1), amplicons of A. ampeloprasum var. babingtonii, A. polyanthum, A. iranicum; A. atroviolaceum and A. truncatum. Group 2 included amplicons of hexaploid and octoploid forms of A. ampeloprasum and A. polyanthum. Group 3 comprised amplicons of tetraploid and octoploid forms of A. ampeloprasum and A. iranicum. Clade B: included A. sativum, A. leucanthum C. Koch, A. pseudoampeloprasum Miscz. ex Grossh., A. atroviolaceum and Group 1, which included A. pyrenaicum plus hexaploid and octoploid amplicons of A. ampeloprasum. Other minor clades included a monophyletic group formed by different accessions of A. scorodoprasum L. and A. dregeanum Kunth, and one formed by A. commutatum and A. bourgeaui. It is worth mentioning that there was a 99% match between the ITS sequences of A. bourgeaui and each of the 3 accessions of A. commutatum. This close relationship was also recovered in the cpDNA, also with 99% match in the trnD-trnT region and 100% match in the trnL-trnF region. Analysis of the chloroplast data set also showed section Allium to be monophyletic (Fig. 2). However, until more species are sampled, this result should be considered as preliminary. The species of section Allium were placed in two main clades, the first one being divided into two subclades (Fig. 2): Clade I-a: included all tetraploid accessions of A. ampeloprasum (three horticultural groups: leek, kurrat and pearl onion) which appear to be closely related to A. atroviolaceum and A. iranicum. This group was positioned as sister to either A. pyrenaicum or to a clade including A. truncatum, A. polyanthum and A. acutiflorum Loisel. The subdivision of clade I-a was poorly supported by Bayesian inference and bootstrap analyses. Clade I-b: included all hexa- and octoploid accessions of A. ampeloprasum (GHG) and the accessions of A. commutatum and A. bourgeaui. In this case, the groups had good statistical support. Clade II: included A. tuncelianum (Kollmann) N. Özhatay, B. Mathew et Sßiraneci, A. sativum, A. pseudoampeloprasum, A. leucanthum, and A. scorodoprasum. Its subdivision was well supported by Bayesian inference. 3.3. Data sets incongruence Incongruence between the nuclear and chloroplast data set, as shown in Fig. 2, predominantly reflects differences caused by the allopolyploid nature of the investigated species. Although some species are clustered together in a similar way, higher relationships between them are quite different. Considerable incongruence can be detected within section Allium even if putatively allopolyploid

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

species are extracted from both datasets (data not shown). This observation indicates possible past interspecific hybridization events at the diploid level followed by backcrossing. The most obvious difference is the position of hexa- and octoploid samples of A. ampeloprasum and A. ampeloprasum var. babingtonii. According to the results based on the chloroplast DNA

491

dataset (Fig. 2), these accessions formed a single group closely related to A. commutatum and A. bourgeaui, while on the nuclear dataset, amplicons of these accessions were split into two (for hexaploids) or three (for octoploids and A. ampeloprasum var. babingtonii) distinct groups, none of them directly related to A. commutatum or A. bourgeaui (Fig. 2).

Fig. 1. Phylogenetic tree resulting from Bayesian analysis of the ITS region of the subgenus Allium with emphasis on the section Allium. Bayesian posterior probabilities and parsimony bootstrap support values (%) are given in the respective order. Closely related amplicons are presented in condensed form by triangles, their size being proportional to the number of amplicons. A description of groups 1–3 is given in Table 1. The taxonomy of sections follows the classification by Friesen et al. (2006).

492

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

Table 1 List of amplicons included in each group recovered in the ITS phylogenetic analyses. Group 1 A. ampeloprasum ‘‘Clone FCA” A. ampeloprasum ‘‘Clone 1007” A. ampeloprasum ‘‘Clone Pazin” A. ampeloprasum ‘‘ALL 1355” A. ampeloprasum ‘‘Clone Spain 9” A. ampeloprasum A. pyrenaicum Group 2 A. ampeloprasum ‘‘Clone FCA” A. ampeloprasum ‘‘Clone 1007” A. ampeloprasum ‘‘Clone Pazin” A. ampeloprasum ‘‘ALL 1355” A. ampeloprasum ‘‘Clone Spain 9” A. ampeloprasum A. polyanthum

GHG group (8)

EU626370, EU626361, EU626362, EU626335, EU626351, EU626371

GHG group (8)

EU626369, EU626358, EU626336, EU626352

GHG group (6)

EU626368, EU626363, EU626364, EU626344, EU626353

GHG group (6)

EU626367, EU626366, EU626365, EU626355, EU626354, EU626340, EU626349

GHG group (6)

EU626357, EU626356, EU626359, EU626338, EU626339, EU626350

var. babingtonii

EU626345, EU626346, EU626342, EU626341, EU626347, EU626337, EU626348 EU626343

GHG group (8)

EU626307, EU626293, EU626298

GHG group (8)

EU626305, EU626294, EU626290, EU626309

GHG group (6)

EU626303, EU626297, EU626291, EU626302

GHG group (6)

EU626304, EU626299, EU626295

GHG group (6)

EU626301, EU626300, EU626296, EU626292

var. babingtonii

EU626306 FJ664293, FJ664297, FJ664296, FJ664295, FJ664294, FJ664298, FJ664299, FJ664292

Group 3 A. ampeloprasum GHG group (8) ‘‘Clone FCA” A. ampeloprasum GHG group (8) ‘‘Clone 1007” A. ampeloprasum leek group (4) A. ampeloprasum kurrat group (4) A. ampeloprasum pearl onion group (4) A. ampeloprasum bulbous leek group (4) A. iranicum

EU626310 EU626311, EU626317 FJ664307, FJ664308, FJ664306, FJ664317, FJ664316, FJ664313, FJ664312, FJ664303, FJ664304, FJ664332, AY427543, AY427529, DQ176006, AY427553, AY427530 FJ664330, FJ664329, FJ664328, FJ664331, FJ664335, FJ664334, FJ664333, FJ664322, FJ664305, FJ664336, AY427531 FJ664311, FJ664339, FJ664340, FJ664338, FJ664337, FJ664327, FJ664325, FJ664320, FJ664326 FJ664315, FJ664314, FJ664310, FJ664309, FJ664324, FJ664318, FJ664319, FJ664321, FJ664323 EU626312, EU626313, EU626315, EU626318, EU626323

(*1) A. pyrenaicum

EU626376, EU626378, EU626377, EU626379, EU626381, EU626382, EU626383, EU626380, EU626384

(*2) A. truncatum

EU626326, EU626327, EU626329, EU626328, EU626332, EU626331, EU626330, EU626334, EU626333

There were also incongruencies in the positions of other species, such as A. acutiflorum, A. scorodoprasum, and A. pyrenaicum. In contrast, A. pseudoampeloprasum and A. leucanthum were positioned close to A. sativum on both trees. 3.4. Phylogenetic analysis of complex polyploid species Separate analyses of ten cloned ITS amplicons of each polyploid accession together with potential closest relatives revealed the presence of several highly divergent ITS alleles within single individuals in almost all cases. They formed polyphyletic assemblages, with their closest relatives pointing to potential parental species. The distribution and affinities of amplicons of different forms of A. ampeloprasum with their close relatives are presented in Fig. 3. Amplicons of octoploid forms of GHG and A. ampeloprasum var. babingtonii were divided into three groups (Fig. 3a and c) while hexaploid forms of GHG (Fig. 3b) and the tetraploids pearl onion (Fig. 3f) and bulbous leek (Fig. 3g) were divided into two groups. Contrastingly, ITS amplicons of the remaining tetraploid samples, leek (Fig. 3d) and kurrat (Fig. 3e), formed single groups. A. iranicum, A. polyanthum, A. truncatum, and A. pyrenaicum also showed divergent ITS amplicons (Fig. 2).

4. Discussion Our studies, focused predominantly on the elucidation of the genetic constitution of horticultural groups within the section Allium, revealed novel insights, particularly on the origin of some investigated allopolyploid species. The genus Allium is a taxonomically difficult group (for a review see Stearn, 1978; Friesen et al., 2006); sections that include several polyploid species are particularly demanding. Rapid homogenization of multiple copies of ITS sequences — through concerted evolution or other mechanisms — is assumed to lead to ITS uniformity within species (for a review see Álvarez and Wendel, 2003). However, increasing evidence has accumulated suggesting that allopolyploid species maintain both parental ITS types (Liu et al., 2008; Soltis et al., 2008). Soltis et al. (2008) proposed that incomplete homogenization is present in recently formed polyploid species. This phenomenon, also termed intra-individual polymorphism, has already been noted in several Allium species (Friesen et al., 2006; Gurushidze et al., 2007; Gurushidze et al., 2008) but variability of amplicons has not been fully exploited for phylogenetic studies. Similarly, incomplete homogenization of ribosomal external transcribed spacers (ETS) was recently used to reveal phylogenetic relationships of the polyploid series of Centaurea toletana (Garcia-Jacas et al., 2009).

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

493

Fig. 2. Phylogenetic relationships as inferred by Bayesian analysis from the combined trnL-trnF and trnD-trn-T dataset (right-hand side). Bayesian posterior probabilities and bootstrap support values (%) are given at nodes. A simplified summary of relationships from the ITS tree (Fig. 1) region is shown on the left side to depict incongruences between the two data sets. Boxes indicate splits of polyploid species on the ITS cladogram based on segregation of amplicons. Asterisks indicate sequences obtained from GenBank.

Several accessions investigated in this study are predominantly clonally propagated (see Supplementary Table S1). Therefore, it can be assumed that processes causing the phenomenon of concerted evolution (e.g. unequal crossing over, gene conversion) have low impact, and intra-individual polymorphism is to be expected. In fact, intra-individual genetic divergence of the ITS region was detected and proved useful in tracking parental lines in our studies. Although both analyzed chloroplast regions were highly conserved and the total number of variable sites was low, the topology of the cpDNA dendrogram was well-resolved. This confirms the value of cpDNA based phylogenetic analyses in the genus Allium, as previously noted by Havey (1991), Friesen et al. (2000) and Meerow et al. (1999). Our study is the first larger phylogenetic analysis based on cpDNA of the section Allium, since it includes the majority of species (17 out of 25) of the first informal group proposed by Mathew (1996). The origin of putative hybrid species among various groups of different ploidy levels is discussed below in the light of evidence from both DNA regions. 4.1. Leek group (2n = 4x = 32) All studied tetraploid forms of A. ampeloprasum (leek, bulbous leek, kurrat and pearl onion) resolved completely or almost completely in a single clade in both chloroplast and nuclear analyses (Figs. 1 and 2). They can therefore be considered to be variants of the same species. Only one amplicon of pearl onion and one of bulbous leek were positioned elsewhere (Figs. 1 and 3). The resemblance of leek with its bulbous forms was also supported by genome size, interfertility, genomic in situ hybridization and vari-

ation of restriction patterns of nuclear and chloroplast DNA (Bohanec et al., 2005). The origin of leek, which is the most economically important crop within this group, has been extensively studied. A. ampeloprasum was proposed as its wild progenitor by Stearn (1978). However, our results on cpDNA show that A. iranicum and A. atroviolaceum are closer to leek than is A. ampeloprasum (Fig. 2). Mathew (1996) mentioned the possibility that other species than A. ampeloprasum could be the progenitor of leek. Potentially, all remaining species from clade I-a (A. polyanthum, A. truncatum, A. acutiflorum, and A. pyrenaicum) could be considered as possible wild progenitors of leek, but we think that this is unlikely due to their different area of distribution (Fig. 4). It has been assumed that leek is an autotetraploid; supporting evidence was found in studies based on meiotic behavior (Levan, 1940; Kadry and Kamel, 1955; Stack and Roelofs, 1996), marker chromosome shape (Murin, 1964), and chlorophyll deficiencies (Berninger and Buret, 1967). This hypothesis has been challenged, based on cytological observations made by Koul and Gohil (1970) and Khazanehdari et al. (1995), who concluded that leek is represented by the genomic formula AA A0 A00 and that it could be a weak segmental allopolyploid. Even though our studies did not reveal any signs of introgression or hybridization in leek or kurrat, it is possible that increased selection pressure, as well as the almost strict sexual mode of reproduction have lead to a complete homogenization of ITS copies within the two crops. On the other hand, hybridization or introgression events seem very probable in pearl onion and bulbous leek, which still have the ability to reproduce vegetatively and, therefore, could have retained divergent alleles (Fig. 3).

494

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

Fig. 3. Unrooted trees showing the distribution and affinities of cloned ITS amplicons of different forms of A. ampeloprasum in relation to putative related species. (a) ‘‘Clone FCA” (GHG group — 2n = 8x = 64); (b) ‘‘Clone Pazin” (GHG group — 2n = 6x = 48); (c) var. babingtonii (2n = unknown); (d) cult. ‘‘Jolant” (leek group — 2n = 4x = 32); (e) cult. ‘‘Balady-Saidii” CGN 14710 (kurrat group — 2n = 4x = 32); (f) ‘‘Pearl Onion” (pearl onion group — 2n = 4x = 32); and (g) ‘‘Chinese” (bulbous leek group — 2n = 4x = 32). The size of triangles is proportional to the number of amplicons; circles represent single amplicons positioned outside main groups.

4.2. Great headed garlic group (2n = 6x = 48; 2n = 8x = 64) All hexa- and octoploid accessions of GHG, as well as A. ampeloprasum var. babingtonii, formed a single clade closely related to A. commutatum and A. bourgeaui in our cpDNA analyses (Fig. 2). However, ITS data revealed large intra-individual divergence and confirmed the allopolyploid nature of these accessions, as previously

proposed by Khoshoo et al. (1960), Messiaen et al. (1993) and Hirschegger et al. (2006) (Fig. 1). Results on A. ampeloprasum var. babingtonii confirmed that this isoclonal plant is a form of GHG, as previously suggested (Stearn, 1978; Treu et al., 2001). Based on cpDNA tree topology, GHG should be considered to be separate from the tetraploid forms of A. ampeloprasum, as these forms do not share the same maternal ancestors. Previous reports

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

495

Fig. 4. Interpretation of geographic distribution of Allium species included in the present phylogenetic analyses based on species descriptions made by Stearn (1978), Mathew (1996), Pospichal (1897), Khassanov (1996), Bothmer (1974), Maab and Klaas (1995), and from information gathered from http://epic.kew.org/ and http://database.prota.org/.

provide further evidence of this. In mtDNA studies of Kik et al. (1997), accessions of GHG and A. ampeloprasum var. babingtonii were placed together in a separate clade distinct from that formed by leek cultivars, kurrat and prei anak. Ariga et al. (2002) found that even though some morphological characteristics were shared by both odorless garlic ‘‘Mushuu-ninniku” (2n = 32, a form of A. ampeloprasum) and GHG, their RFLP pattern, alliinase sequence and alliinase activity were quite different. Bohanec et al. (2005) found that the genome of GHG was 17% larger than that of leek and bulbous leek accessions; the chloroplast DNA restriction patterns suggested that leek possesses derived states while GHG shares the same progenitor states as A. polyanthum and A. commutatum. Our results on cpDNA also recovered A. polyanthum as a close relative of all studied tetraploid forms of A. ampeloprasum. Moreover, in ITS analyses, an even closer relationship between A. polyanthum and GHG was found as part of their amplicons clustered together in Group 2. On the other hand, Group 1 amplicons were more closely related to A. atroviolaceum, A. pseudoampeloprasum, and A. leucanthum than to A. sativum, supporting the early findings of Khoshoo et al. (1960) who found that GHG accessions gathered in India were chemically different from A. sativum. GHG presented lower volatile oil content and had saponins, a feature that garlic lacked. Our cpDNA results also support this hypothesis, since no direct relationship between GHG and A. sativum was found (Fig. 2). 4.3. Non-cultivated polyploid species of the section Allium Our results on ITS clearly support the allopolyploid nature of A. iranicum (Figs. 1 and 3) already proposed by Ghaffari (2006) based on cytogenetic analyses of tetraploid accessions. Chloroplast DNA analysis placed A. iranicum next to A. atroviolaceum and all studied tetraploid forms of A. ampeloprasum.

Ipek et al. (2008) discussed and rejected the hypothesis that A. tuncelianum is the immediate progenitor species of A. sativum, based on AFLP and ITS phylogenetic analyses. Our results on ITS did not reveal any specific relationship between these two species. However, cpDNA results showed that A. tuncelianum and A. sativum (as well as A. pseudoampeloprasum, A. leucanthum and A. scorodoprasum) share a most recent common ancestor. Our molecular data does not support the suggestion of Ipek et al. (2008) that A. tuncelianum could be another form of leek-like species (Figs. 1 and 2). Two different accessions of A. atroviolaceum (one of them from GenBank) failed to form a single clade in the ITS phylogeny. Preliminary searches did not reveal significant intra-individual variability within the ITS region of A. atroviolaceum (provided by Kew Gardens, UK), as was the case of other polyploids accessions. Assuming that both accessions were correctly determined, it is possible that there is high genetic variability within A. atroviolaceum across its wide geographical distribution, perhaps due to morphologically cryptic species. Divergent ITS amplicons were also found in A. polyanthum. The majority of amplicons were positioned within Group 2, while the rest clustered next to A. atroviolaceum in the ITS phylogeny. Chloroplast DNA data suggest that A. polyanthum, A. truncatum and A. acutiflorum are closely related. The close relationship between hexaploid GHG and A. polyanthum was previously discussed and rejected by Messiaen et al. (1993). According to our cpDNA results, A. polyanthum is not maternally related to GHG but introgression in either way can not be excluded. In A. truncatum and A. pyrenaicum, one out of ten amplicons was positioned outside the main cluster, possibly as the result of introgression by another species. The close relationship of A. commutatum and A. bourgeaui to GHG in the cpDNA analysis is in agreement with the group defined as ‘‘ampeloprasum complex” by Bothmer (1975). However, A. atro-

496

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497

violaceum and A. polyanthum are not as closely related to this complex as previously proposed. Considering the extremely small genetic distance between A. commutatum and the accession identified as A. bourgeaui ssp. bourgeaui, a misidentification of the latter cannot be completely ruled out. 4.4. Other Taxonomic considerations Although the focus of our studies was on the section Allium, some remarks about representatives of other sections should be made. Sections Avulsea, Pallasia, Brevispatha and Kopetdagia are not monophyletic according to the ITS phylogeny. Similar results concerning non-monophyly, but at the subgeneric level, were obtained in a recently published study focused on the subgenus Amerallium (Nguyen et al., 2008); their results showed essentially the same non-monophyly for the sections mentioned above. Allium turkenistanicum (section Mediasia), which was positioned outside the clade of the subgenus Allium in a previous study (Friesen et al., 2006), clustered within the subgenus Allium, next to A. griffithianum in our study. The position of this species is unstable and sensitive to taxon sampling. 4.5. Ethno-botanical context The geographic distribution of section Allium is predominately Mediterranean but extends east towards Central Asia (Mathew, 1996). As noted by Mathew (1996), the precise habitat details of the species of section Allium is on the whole poorly recorded. To our knowledge, no accurate species distribution map is available for this section. Therefore the details presented in Fig. 4 are merely an interpretation of data gathered from multiple sources (for details see legend of Fig. 4). While some Allium species have a wide distribution area, others are circumscribed to more limited areas (Fig. 4). As demonstrated in Quercus (Whittemore and Schaal, 1991) and Helianthus (Rieseberg, 1991), a strong correlation exists between geographical taxa distribution and phylogenies inferred by cpDNA, probably due to cytoplasmic gene flow and subsequent back crossing. On the contrary, nuclear data correlates with morphological traits (see Rieseberg and Soltis, 1991). Indications of correlation between cpDNA data and geographical distribution can also be recognized in section Allium (Fig. 4). A. polyanthum, A. pyrenaicum and A. acutiflorum share a similar area of distribution and also cluster together in clade I-a (Fig. 2). A similar situation occurs in the Caucasian region with A. tuncelianum, A. leucanthum, A. pseudoampeloprasum and A. sativum (Fig. 4) which cluster together with A. scorodoprasum in clade II (Fig. 2). Despite the fact that three different accessions of A. scorodoprasum were analyzed and that it has a wide range of distribution (Fig. 4), all accessions clustered as a sister group to A. leucanthum (Fig. 2). This could imply that A. scorodoprasum also has its origin in the Caucasian region. Our results on cpDNA regarding A. commutatum, A. bourgeaui, and GHG also support the findings of Bothmer (1975) who proposed that the species described within the ‘‘ampeloprasum complex” might have a secondary centre of origin in the Aegean, more precisely on Crete. The apparent almost strict confinement of weedy GHG-like forms of A. ampeloprasum to highly disrupted areas or ancient orchards seems to satisfactorily prove human activity as the motor of its wide Mediterranean distribution. Human migration might have played a major role on the transfer of Caucasian cytoplasm to SW Europe as it is reflected by the apparently conflictive close relationship between A. truncatum and A. polyanthum and A. acutiflorum (Fig. 2). The close relationship between A. iranicum and A. atroviolaceum and all tetraploid cultivated forms of A. ampeloprasum (Figs. 1 and 2) can be explained also in the same con-

text taking into account that leek is a popular crop in continental Europe (Currah, 1986) but not in the Middle East. Molecular evidence provided in this study, as well as previously published, suggests that at least the tetraploid horticultural groups traditionally included in A. ampeloprasum L.: leek, kurrat, pearl onion, and bulbous leek constitute a distinct taxon. Restoration of A. porrum L. is therefore proposed to address this tetraploid complex. In this sense, A. ampeloprasum L. should refer only to all forms of GHG (including wild taxa) and A. ampeloprasum var. babingtonii. Further studies based on molecular, cytogenetic and morphological evaluations might provide further grounds for this proposal. Acknowledgments We would like to thank Tatjana Gecˇ and Ilia Leitch for their help in gathering literature, Viktorija Dolenc, Valerija Plestenjak and Jozˇe Godeša for their dedication in maintaining the Allium collection, Boris Turk for his help in the taxonomic determination of accessions gathered from the wild and two anonymous reviewers for helping improving this manuscript. We also thank the following for providing specimens used in this study: Charles-Marie Messiaen (France), Royal Botanic Gardens Kew (UK), Leibniz-Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK, Germany), Centre for Genetic Resources (CGN, Netherlands), Botanischer Garten und Botanisches Museum Berlin-Dahlem (Germany) and EEAA INTA La Consulta (Argentina). This work was financially supported by research grant P4-0077 from the Slovenian Research Agency. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ympev.2009.08.030. References Álvarez, I., Wendel, J.F., 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular Phylogenetics and Evolution 29, 417–434. Ariga, T., Kumagai, H., Yoshikawa, M., Kawakami, H., Seki, T., Sakurai, H., Hasegawa, I., Etoh, T., Sumiyoshi, H., Tsuneyoshi, T., Sumi, S., Iwai, K., 2002. Garlic-like but odorless plant Allium ampeloprasum ‘Mushuu-ninniku’. Journal of the Japanese Society for Horticultural Science 71, 362–369. Berninger, E., Buret, P., 1967. Étude des déficients chlorophylliens chez deux especes cultivées du genre Allium: l’oignon A. cepa L. et le poireau A. porrum L.. Annales de l’Amélioration del Plantes 17, 175–194. Bohanec, B., Jakše, M., Sešek, P., Havey, M.J., 2005. Genetic characterization of an unknown Chinese bulbous leek-like accession and its relationship to similar Allium species. Hortscience 40, 1690–1694. Boissier, E., 1884. Flora Orientalis V. Basileae. Geneva & Lugduni. Bothmer, R., 1970. Cytological studies in Allium I. Chromosome numbers and morphology in Allium Sect. Allium from Greece. Botaniska Notiser 123, 518–550. Bothmer, R., 1974. Studies in the Aegean Flora XXI. Biosystematic studies in the Allium ampeloprasum complex. Opera Botanica (Lund) 34, 1–104. Bothmer, R., 1975. The Allium ampeloprasum complex on Crete. Mitteilungen der Botanischen Staatssammlung München 12, 267–288. Currah, L., 1986. Leek breeding: a review. Journal of Horticultural Science 61, 407–415. Demesure, B., Sodzi, N., Petit, R.J., 1995. A set of universal primers for amplification of polymorphic non-coding regions of mitochondrial and chloroplast DNA in plants. Molecular Ecology 4, 129–131. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32, 1792–1797. Feinbrun, N., 1943. Allium section Porrum of Palestine and the neighboring countries. Palestine Journal of Botany 3, 1–21. Friesen, N., Fritsch, R.M., Blattner, F.R., 2006. Phylogeny and new intrageneric classification of Allium (Alliaceae) based on nuclear ribosomal DNA ITS sequences. Aliso 22, 372–395. Friesen, N., Fritsch, R.M., Pollner, S., Blattner, F.R., 2000. Molecular and morphological evidence for an origin of the aberrant genus Milula within Himalayan species of Allium (Alliacae). Molecular Phylogenetics and Evolution 17, 209–218. Fritsch, R., Friesen, N., 2002. Evolution, domestication and taxonomy. In: Rabinowitch, H.D., Currah, L. (Eds.), Allium Crop Science: Recent Advances. CABI Publishing, Wallingford, pp. 5–30. Garcia-Jacas, N., Soltis, P.S., Font, M., Soltis, D., Vilatersana, R., Alfonso, S., 2009. The polyploid series of Centaurea toletana: Glacial migrations and introgression

P. Hirschegger et al. / Molecular Phylogenetics and Evolution 54 (2010) 488–497 revealed by nrDNA and cpDNA sequence analyzes. Molecular Phylogenetics and Evolution 52, 377–394. Ghaffari, S.M., 2006. New cytogenetic information on Allium iranicum (Alliaceae) from Iran. Biologia 61, 375–379. Guern, M., Lecorff, J., Boscher, J., 1991. Comparative karyology of the Allium ampeloprasum complex in France. Bulletin de la Societe Botanique de FranceLettres Botaniques 138, 303–313. Gurushidze, M., Fritsch, R.M., Blattner, F.R., 2008. Phylogenetic analysis of Allium subg. Melanocrommyum infers cryptic species and demands a new sectional classification. Molecular Phylogenetics and Evolution 49, 997–1007. Gurushidze, M., Mashayekhi, S., Blattner, F.R., Friesen, N., Fritsch, R.M., 2007. Phylogenetic relationships of wild and cultivated species of Allium section Cepa inferred by nuclear rDNA ITS sequence analysis. Plant Systematics and Evolution 269, 259–269. Hanelt, P., Schultze-Motel, J., Fritsch, R., Kruse, J., Maass, H., Ohle, H., Pistrick, K., 1992. Infrageneric Grouping of Allium — The Gatersleben approach. In: Hanelt, P., Hammer, K., Knupffer, H. (Eds.), The Genus Allium — Taxonomic Problems and Genetic Resources. IPK, Gatersleben, Germany, pp. 107–123. Hasegawa, M., Kishino, H., Yano, T.A., 1985. Dating of the human ape splitting by a molecular clock of mitochondrial-DNA. Journal of Molecular Evolution 22, 160–174. Havey, M.J., 1991. Phylogenetic relationships among cultivated Allium species from restriction enzyme analysis of the chloroplast genome. Theoretical and Applied Genetics 81, 752–757. Havey, M.J., Leite, D.L., 1999. Toward the identification of cytoplasmic male sterility in leek: evaluation of organellar DNA diversity among cultivated accessions of Allium ampeloprasum. Journal of the American Society for Horticultural Science 124, 163–165. Helm, J., 1956. Die zu Würz- und Speisezwecken kultivierten Arten der Gattung Allium L.. Kulturpflanze 4, 130–180. Hirschegger, P., Galmarini, C., Bohanec, B., 2006. Characterization of a novel form of fertile great headed garlic (Allium sp.). Plant Breeding 125, 635–637. Ipek, M., Ipek, A., Simon, P.W., 2008. Genetic characterization of Allium tuncelianum: an endemic edible Allium species with garlic odor. Scientia Horticulturae 115, 409–415. Jones, H.A., Mann, L.K., 1963. Onions and their Allies — Botany, Cultivation, and Utilization. Leonard Hill Books Limited, London. Kadry, A.E.R., Kamel, S.A., 1955. Cytological studies in two tetraploid species A. kurrat Schweinf. and A. porrum L. and their hybrid. Svensk Botanisk Tidskrift 49, 314–324. Keane, T.M., Creevey, C.J., Pentony, M.M., Naughton, T.J., Mclnerney, J.O., 2006. Assessment of methods for amino acid matrix selection and their use on empirical data shows that ad hoc assumptions for choice of matrix are not justified. BMC Evolutionary Biology 6, 29. Khassanov, F., 1996. Conspectus of the wild growing Allium species of Middle Asia. In: Öztürk, M., Seçmen, Ö., Görk, G. (Eds.), Plant Life in Southwest and Central Asia. EGE University Press, Izmir, Turkey, pp. 141–159. Khazanehdari, K.A., Jones, G.H., 1997. The causes and consequences of meiotic irregularity in the leek (Allium ampeloprasum spp. porrum); implications for fertility, quality and uniformity. Euphytica 93, 313–319. Khazanehdari, K.A., Jones, G.H., Fordlloyd, B.V., 1995. Meiosis in the leek (Allium porrum L.) revisited. 1. Prophase I pairing. Chromosome Research 3, 433–439. Khoshoo, T.N., Atal, C.K., Sharma, V.B., 1960. Cytotaxonomical and chemical investigations on the North-West Indian garlics. Research Bulletin of the Punjab University.. Science 2, 37–47. Kik, C., Samoylov, A.M., Verbeek, W.H.J., van Raamsdonk, L.W.D., 1997. Mitochondrial DNA variation and crossability of leek (Allium porrum) and its wild relatives from the Allium ampeloprasum complex. Theoretical and Applied Genetics 94, 465–471. Kimura, M., 1981. Estimation of evolutionary distances between homologous nucleotide sequences. Proceedings of the National Academy of Sciences of the United States of America 78, 454–458. Kollmann, F., 1971. Allium ampeloprasum — A polyploid complex 1. Ploidy levels. Israel Journal of Botany 20, 13–20. Koul, A.K., Gohil, R.N., 1970. Cytology of tetraploid Allium ampeloprasum with chiasma localization. Chromosoma 29, 12–19. Kump, B., Javornik, B., 1996. Evaluation of genetic variability among common buckwheat (Fagopyrum esculentum Moench) populations by RAPD markers. Plant Science 114, 149–158. Levan, A., 1940. Meiosis in Allium porrum, a tetraploid species with chiasma localization. Hereditas (Lund) 26, 454–462. Liu, Z.P., Chen, Z.Y., Pan, J., Li, X.F., Su, M., Wang, L.J., Li, H.J., Liu, G.S., 2008. Phylogenetic relationships in Leymus (Poaceae: Triticeae) revealed by the

497

nuclear ribosomal internal transcribed spacer and chloroplast trnL-F sequences. Molecular Phylogenetics and Evolution 46, 278–289. Maab, H.I., Klaas, M., 1995. Infraspecific differentiation of garlic (Allium sativum L.) by isozyme and RAPD markers. Theoretical and Applied Genetics 91, 89– 97. Mathew, B., 1996. A Review of Allium section Allium. Royal Botanic Gardens, Kew, Richmond, Surrey, UK. pp. 1–176. Meerow, A.W., Fay, M.F., Guy, C.L., Li, Q.B., Zaman, F.Q., Chase, M.W., 1999. Systematics of Amaryllidaceae based on cladistic analysis of plastid rbcL and trnL-F sequence data. American Journal of Botany 86, 1325–1345. Messiaen, C.M., Cohat, J.H., Leroux, J.P., Pichon, M., Beyries, A., 1993. Les allium alimentaires reproduits par voie végétative.. Institut National de la Recherche Agronomique (INRA), Paris, France. pp. 1–228. Murin, A., 1964. Chromosome study in Allium porrum L.. Caryologia 17, 575–578. Nguyen, N.H., Driscoll, H.E., Specht, C.D., 2008. A molecular phylogeny of the wild onions (Allium; Alliaceae) with a focus on the western North American center of diversity. Molecular Phylogenetics and Evolution 47, 1157–1172. Pospichal, E., 1897. Flora des Oesterreichischen Küstenlandes. Leipzig und Wein, pp. 237–250. Regel, E., 1875. Alliorum adhuc cognitorum monographia. Trudy Imperatorskago S. Peterburgskago Botani Eskago Sada 3, pp.1–266. Rieseberg, L.H., 1991. Homoploid reticulate evolution in Helianthus: evidence from ribosomal genes. American Journal of Botany 78, 1218–1237. Rieseberg, L.H., Soltis, D.E., 1991. Phylogenetic consequences of cytoplasmic gene from in plants. Evolutionary Trends in Plants 5, 65–84. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574. Schweisguth, B., 1970. Preliminary studies for improvement of leek Allium porrum L. Proposal for a breeding method. Annales de l’Amelioration des Plantes 20, 215– 231. Smilde, W.D., van Heusden, A.W., Kik, C., 1999. AFLPs in leek (Allium porrum) are not inherited in large linkage blocks. Euphytica 110, 127–132. Smith, B.M., Crowther, T.C., 1995. Inbreeding depression and single cross hybrids in leeks (Allium ampeloprasum ssp porrum). Euphytica 86, 87–94. Soltis, D.E., Mavrodiev, E.V., Doyle, J.J., Rauscher, J., Soltis, P.S., 2008. ITS and ETS sequence data and phylogeny reconstruction in allopolyploids and hybrids. Systematic Botany 33, 7–20. Stack, S.M., Roelofs, D., 1996. Localized chiasmata and meiotic nodules in the tetraploid onion Allium porrum. Genome 39, 770–783. Stearn, W.T., 1978. European species of Allium and allied genera of Alliaceae: a synonymic enumeration. Annales Musei Goulandris 4, 83–198. Stearn, W.T., 1992. How many species of Allium are known? Kew Magazine 9, 180– 182. Swofford, D.L., 2002. PAUP*. Phylogenetic Analysis Using Parsimony ( and Other Methods), vers. 4 b10. Sinauer Associates, Sunderland, MA, USA. Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for amplification of 3 non-coding regions of chloroplast DNA. Plant Molecular Biology 17, 1105–1109. Tahbaz, F., 1971. Allium Tarée Irani of ampeloprasum L. Group, Cultivated in Iran. Bulletin de la Societe Botanique de France 118, 753–761. Tamura, K., Nei, M., 1993. Estimation of the number of nucleotide substitutions in the control region of mitochondrial-DNA in humans and chimpanzees. Molecular Biology and Evolution 10, 512–526. Tattelman, E., 2005. Health effects of garlic. American Family Physician 72, 103– 106. Treu, R., Holmes, D.S., Smith, B.M., Astley, D., Johnson, M.A.T., Trueman, L.J., 2001. Allium ampeloprasum var. Babingtonii (Alliaceae): an isoclonal plant found across a range of habitats in SW England. Plant Ecology 155, 229–235. van der Meer, Q.P., Hanelt, P., 1990. Leek (Allium ampeloprasum var. porrum). In: Brewster, J.L., Rabinowitch, H.D. (Eds.), Onions and Allied Crops. Biochemistry, Food Science and Minor Crops. CRC Press Inc., Boca Raton, FL, USA, pp. 179–196. Vvedensky, A.I., 1935. Allium L.. In: Komarov, V.L. (Ed.), Flora USSR. Izdatel’stvo Akademi nauk SSSR, Leningrad, Russia, pp. 112–280. Wendelbo, P., 1971. Alliaceae. In: Rechinger, K.H. (Ed.), Flora Iranica 76. Graz, Austria, pp. 1–100. White, T.J., Bruns, T., Lee, S., Taylor, J., 1990. Amplification and Direct Sequencing of Fungal Ribosomal RNA Genes for Phylogenetics. PCR Protocols: A Guide to Methods and Applications. Academy Press Inc.. pp. 315–322. Whittemore, A.T., Schaal, B.A., 1991. Interspecific gene flow in oaks. Proceedings of the National Academy of Sciences of the United States of America 88, 2540– 2544.