The evolutionary history of the white-rayed species of Melampodium (Asteraceae) involved multiple cycles of hybridization and polyploidization

Polyploidy is one of the major processes shaping angiosperm evolution (Wendel, 2000; Hegarty and Hiscock, 2007, 2008; Leitch and Leitch, 2008; Hufton and Panopoulou, 2009). This is evident from early whole-genome duplications in the ancestor of essentially all extant angiosperms (Wendel, 2000; Wolfe, 2001; Adams and Wendel, 2005a, b; De Bodt et al., 2005; Jiao et al., 2011) as well as from more recent poly ploidization events affecting most angiosperms (Masterson, 1994; Leitch and Bennett, 1997; Leitch and Leitch, 2008). Whereas autopolyploids originate from multiplication of the usually identical chromosome sets within a single (sub)species, allopolyploids derive from hybridization and multiplication of the usually differentiated chromosome sets coming from two or more progenitor (sub)species (Comai, 2005; Chen et al., 2007; Otto, 2007). Although the distinction between auto- and allopolyploids is difficult in polyploids exhibiting intermediate chromosome pairing behavior (expected polysomic inheritance in autopolyploids vs. disomic inheritance in allopolyploids; Ramsey and Schemske, 2002; Otto, 2007), both types are unambiguously known to occur frequently in nature (Tate et al., 2005; Soltis et al., 2007), where they often originate recurrently (Soltis et al., 2007; Dixon et al., 2009). Polyploidy plays an important role in race differentiation and eventually speciation. Polyploidization leads to an instantaneous increase in the amount of genetic material upon which evolution can work, potentially leading to the evolution of new functions in duplicated genes (neofunctionalization; Ohno, 1970; Wolfe, 2001; Liu and Wendel, 2002). Furthermore, polyploids often show accelerated genomic evolution expressed as a higher rate of chromosome structural rearrangements and associated pairing behavior in meiosis (Weiss and Maluszynska, 2000; Ramsey and Schemske, 2002; Otto, 2007; Schubert and Lysak, 2011), or genome size changes (Bennetzen and Kellogg, 1997; Bennetzen et al., 2005; Leitch and Leitch, 2008; Hawkins et al., 2008), contributing to the isolation of a polyploid from its lower-ploid ancestor(s). Finally, polyploidization contributes to stabilization of genomes after hybridization, a phenomenon common in plants (Anderson and Stebbins, 1954; Grant, 1981; Mallet, 2005, 2007; Rieseberg and Willis, 2007), by restoring regular meiotic pairing and fertility, thus creating crossing barriers with the parental taxa (allopolyploid speciation: Tate et al., 2005; Hegarty and Hiscock, 2007, 2008; Paun et al., 2009). A prerequisite for studying processes associated with the evolution of polyploid genomes is a sound knowledge of their origin. Molecular sequence data have proven particularly useful for identifying the parental lineages of allopolyploids. Although incongruence between gene trees obtained from plastid DNA (cpDNA; inherited maternally in Asteraceae; Corriveau and Coleman, 1988; Harris and Ingram, 1991), and nuclear DNA (inherited from both parents; Sang et al., 1997; Hughes et al., 2002; Kim et al., 2008) provides evidence for allopolyploidy, this approach is hampered by the often low variation in cpDNA sequences and by concerted evolution of the commonly used nuclear ribosomal internal transcribed spacer (nrITS) region (alvarez and Wendel, 2003). In low-copy nuclear genes, analyses of individual gene copies stemming from the maternal and paternal lineages can be recovered, rendering these genes a powerful and successful tool for inferring the origin of allo polyploids (Ferguson and Sang, 2001; Sang, 2002; Small et al., 2004; Lihova et al., 2006; Fortune et al., 2007; Kim et al., 2008; Shimizu-Inatsugi et al., 2009; Weiss-Schneeweiss et al., 2012). A well-suited group for studying polyploid evolution is the white-rayed complex of Melampodium sect Melampodium ser. Leucantha (Asteraceae; Stuessy, 1972). Characterized by white instead of yellow or orange ray florets, it is a phylogenetically distinct group (Bloch et al., 2009) comprising three xerophytic subshrub species (M. argophyllum, M. cinereum, M. leucanthum) distributed in the arid regions of northern Mexico and the southwestern United States (Fig. 1; Stuessy, 1972; Stuessy et al., 2004). Like the entire sect. Melampodium, species of ser. Leucantha have a basic chromosome number of x = 10 (Weiss-Schneeweiss et al., 2009). The widespread M. cinereum and M. leucanthum comprise both diploid and tetraploid cytotypes (Fig. 1). These originated recurrently via autopolyploidy, as suggested by chromosomal studies and genetic data (Turner and King, 1964; Stuessy et al., 2004; Rebernig et al., 2010a, b). The third species of the complex is the hexaploid M. argophyllum (Stuessy, 1972; Stuessy et al., 2004), restricted to northeastern Mexico (Fig. 1). It shares morphological features with the two other species and has thus been hypothesized to be of allopolyploid origin involving M. cinereum and M. leucanthum (Stuessy et al., 2004). Thus, ser. Leucantha provides an excellent system to compare the type and extent of genomic changes in closely related polyploids with contrasting evolutionary histories. Fig. 1 Distribution map of the analyzed populations of the white-rayed complex of Melampodium. Here, we used molecular sequence (plastid and nuclear DNA, including a low-copy gene), fingerprint (AFLPs), and restriction data in conjunction with karyological, and genome size data to investigate the origin and evolution of ser. Leucantha with special emphasis on M. argophyllum. Specifically, we wanted to (1) infer phylogenetic relationships among M. argophyllum, M. cinereum, and M. leucanthum, in particular to test the hypothesis of an allopolyploid origin for M. argophyllum; and (2) compare karyotype structure, genome size, and rDNA loci number and localization in polyploids of presumably different ages and modes of origin (recent auto- vs. ancient allopolyploidy).