Systematics of Gagea and Lloydia (Liliaceae) and infrageneric classification of Gagea based on molecular and morphological data

Abstract

Our study represents the first phylogenetic analyses of the genus Gagea Salisb. (Liliaceae), including 58 species of Gagea and 6 species of the closely related genus Lloydia Salisb. ex Rchb. Our molecular results support the infrageneric classification of the genus Gagea in sections according to Levichev and demonstrate that Pascher’s subdivision of this genus into two subgenera can no longer be upheld. Certain Gagea sections (e.g., Gagea, Minimae, and Plecostigma) are well supported by cpDNA and nrDNA data. Gagea sect. Fistulosae is closely related to G. sect. Didymobolbos. Gagea sect. Graminifoliae and G. sect. Incrustatae are closely related to G. sect Platyspermum. The analyses support the monophyly of Gagea and Lloydia collectively. The molecular analyses reveal the basal position of G. graeca in proportion to all other species of Gagea and Lloydia investigated. Minor morphological differences could be established between both genera which support their close relationship.

Introduction

Liliaceae s.s. was identified as being monophyletic by Patterson and Givnish (2002), including the two subfamilies Medeoloideae and Lilioideae (Tamura, 1998). The Lilioideae subfamily can be subdivided into two clades: one clade (Tribe Lilieae including Lilium L., Fritillaria L., Nomocharis Franch., Cardiocrinum Endl. and Notholirion Wall. ex Boiss.) which appears to have diversified in the Himalayas roughly 12 million years ago and a further clade (Tribe Tulipaea including Erythronium L., Tulipa L., Gagea Salisb., Lloydia Salisb. ex Rchb.) which originated in East Asia around the same time. Gagea based on matK sequences (Allen et al., 2003), Lloydia based on rbcL sequences (Vinnersten and Bremer, 2001) and both genera based on matK sequences (Ronsted et al., 2005) were shown to be sister groups to the Amana–Tulipa–Erythronium clade.

Within the Liliaceae, Gagea Salisb. (Salisbury, 1806) represents a relatively unexplored Palaearctic genus of small perennial geophytes. The distribution of the genus is restricted to the temperate and subtropical regions of Eurasia and does not extend into any areas with either a tropical climate or permafrost (Levichev, 1999b). Due to the brief ephemeroide growth phase, there is a lack of herbal vouchers in most herbariums and also of complete monographs. The genus comprises a figure of between 70 and approximately 275 species depending on the author (Stroh, 1937, Uphof, 1958–1960, Melchior, 1964, Willis, 1980, Davlianidze, 1976, Tamura, 1998, Levichev, 1999b, Peruzzi, 2003), and more and more species are being currently described (e.g., Levichev, 1981, Levichev, 1988, Levichev, 2006a, Tison, 2004, Henker, 2005, Levichev and Ali, 2006, Levichev unpubl.). Overlapping primitive and advanced morphological characters, e.g., the number of ground leaves, the formation of subterranean organs (Levichev, 1999a), pollen morphology (e.g., Kosenko and Levichev, 1988, Zarrei and Zarre, 2005) and karyotypes (e.g., Peruzzi, 2003, Peruzzi and Aquaro, 2005) complicate the classification of the species and the infrageneric arrangement. In addition, there are several problems originating from the great variation in vegetative and generative characters during the various ontogenic stages of Gagea under the varying ecological conditions (Levichev, 1990a, Levichev, 1999a). Hypothesized interspecific hybridisation (e.g., Levichev, 1990b, Tison, 1998, Peterson et al., 2004, John et al., 2004, Peterson and Peterson, 2005, Peterson and Peterson, 2006) appears to be common within this genus. All these phenomena mean taxonomic problems in the genus Gagea, the description of a large number of species and a nomenclature which is overloaded with synonyms (Levichev, 1999b). As a result, more than 670 specific and intraspecific combinations have been published.

An initial attempt at infrageneric classification was carried out by Koch (1849) who divided Gagea into two sections (Holobolbos and Didymobolbos); this was followed by the addition of a further two sections (Tribolbos and Platyspermum) by Boissier (1882). At the beginning of the twentieth century, Terracciano, 1905, Terracciano, 1906, Pascher, 1904, Pascher, 1907 constructed independent infrageneric classifications. Pascher (1904) classified the Gagea species into the subgenera Gagea (Eugagea) with four sections and Hornungia with two sections. Pascher (1907) subsequently confirmed this classification but added several new subsections. Terracciano, 1905, Terracciano, 1906 also established two subgenera (Gagea and Gageastrum) both having two sections, but neither Pascher nor Terracciano submitted a complete revision. In the course of time, Pascher’s classification received greater acceptance than that of Terracciano. It can be seen that Stroh, 1937, Uphof, 1958–1960 adhered to Pascher’s infrageneric classification. To date, almost all floras (e.g., Heyn and Dafni, 1971, Heyn and Dafni, 1977, Dasgupta and Deb, 1983, Feinburn-Dotham, 1986, Rechinger, 1986, Wendelbo and Rechinger, 1990, Grubov and Egorova, 2003) have in principle accepted Pascher’s (1907) infrageneric classification, despite the fact that this classification was originally based on the knowledge of a limited number of species. Davlianidze, 1976, Levichev, 1990a, Levichev, 1997 published more recent classifications. Davlianidze (1976) examined 26 Caucasian species. He accepted the two subgenera according to Pascher and additionally established six new sections within each of these subgenera. Levichev (1990a) published a new classification of western Tien Shan species utilising general morphological characters as well as cross-sections of basal leaves and the peduncle and also the constitution of the leaves. Unlike all infrageneric classifications since Terracciano, 1905, Terracciano, 1906, Pascher, 1904, Pascher, 1907, he did not utilise the subgeneric rank, but divided the genus into ten sections, some of which were identical to those of Davlianidze (1976). Currently, Levichev recommends a subgeneric classification of the genus Gagea into 13 sections (see Table 1). The initial molecular studies (Peterson et al., 2004), based on the cpDNA and nrDNA data of seven Gagea species from Germany, all belonging to the subgenus Gagea Pascher, displayed two clades. One clade included members of the G. sect. Gagea Pascher and G. sect. Tribolbos Boiss., the other included members of the G. sect. Didymobolbos Koch and G. sect. Monophyllos Pascher. Interestingly, G. spathacea, which is classified by Pascher (1904) in the G. sect. Monophyllos, was the sister to all other Gagea species and Lloydia serotina in the cpDNA psbA–trnH tree.

The molecular data of several authors showed that Gagea and Lloydia represent sister genera. On the basis of sequence variation in the chloroplast, encoded rbcL and ndhF genes and morphological character-states (Patterson and Givnish, 2002) and also on the basis of matK data (Ronsted et al., 2005), Lloydia was considered to be the sister genus of Gagea, whereas only G. wilczekii and L. serotina were examined for systematic investigation in the core Liliales (discussed in Peterson et al., 2004, Peterson and Peterson, 2006). In our earlier molecular study (Peterson et al., 2004), we found that L. serotina—the type of the genus—was placed within the investigated Gagea species. This was the first molecular evidence which questioned the taxonomic position of Lloydia as a separate sister genus to Gagea. Up to this point, our further molecular studies and analyses including morphological data (Peterson and Peterson, 2005, Peterson and Peterson, 2006) had indicated that Gagea and Lloydia are monophyletic. These studies incorporated L. serotina, L. triflora and also G. graeca the latter was initially described as a member of the genus Lloydia. Lloydia consists however of 12–20 species (Willis, 1980, Hyam and Pankhurst, 1995, Mabberley, 1997). Most species of these small bulbiferous herbs occur in the Himalayan region of Eastern Asia. L. serotina is the most widespread relative of the genus; it is distributed across large regions of Europe and northern Asia as well as in western North America.

The objective of our study was to evaluate the generic and infrageneric circumscription of Lloydia and Gagea with the aid of molecular and morphological data. This study represents the first infrageneric classification of Gagea based on molecular data. We include both cpDNA (trnL–trnF intergenic spacer, psbA–trnH intergenic spacer) and nrDNA (ITS region: ITS1 + 5.8SrDNA + ITS2) data of 58 Gagea species, 6 Lloydia species and a number of outgroup species. The molecular results are discussed in comparison with the classification of Levichev (Levichev, 1990a, Levichev, 2006b; unpublished) and with the common classification of Pascher, 1904, Pascher, 1907. Major morphological characters for the division of Gagea Salisb. into 13 independent sections are discussed. We include important concluding remarks on the taxonomic status of Lloydia within the Liliaceae based on both molecular and morphological data. Our investigation represents a contribution to the understanding of the systematics of the Liliaceae.