A MOLECULAR REEXAMINATION OF DIPLOID HYBRID SPECIATION OF SOLANUM RAPHANIFOLIUM.

Abstract

Interspecific hybridization has long been the focus of intensive study in many groups of plants. Since the pioneering work of Anderson (1949), hybridization has been implicated in the formation of sporadic, nonviable, accidental hybrids, introgressive hybridization, and the production of new species (Raven, 1976, 1980; Grant, 1981). Hybrid speciation at the diploid level presents a special case where the immediate and potentially strong isolating mechanism provided by allopolyploidy that separates the newly derived hybrid from its parents is absent. Diploid hybridization hypotheses have invoked a number of isolating mechanisms to maintain the integrity of the newly formed species such as ecogeographical, genetic, or mechanical isolation mechanisms (Levin, 1978; Grant, 1981). Most of these hybridization hypotheses were initially inferred on the basis of intermediate morphology and distributional data, but many have been further tested by artificial crosses, cytogenetic data, and secondary chemistry. The strengths and weaknesses of these methods have been pointed out by various workers (Gottlieb, 1972; Gallez and Gottlieb, 1982; Crawford, 1985; Rieseberg et al., 1 988). The major problems with morphological, and to a lesser extent with cytogenetic and secondary chemical data, are our present ignorance of their genetic basis and our resulting inability to reliably distinguish between competing hypotheses of hybridization, convergent evolution, retention of ancestral characters, and phenotypic plasticity (Rieseberg et al., 1988; Sytsma, 1989). Recent work with genetically more precise isozyme and DNA characters have provided new insights into hybridization hypotheses, with some supported (Roose and Gottlieb, 1976; Gallez and Gottlieb, 1982; Palmer et al., 1983; Soltis and Soltis, 1989; Warwick et al., 1989; Rieseberg et al., 1990) and some questioned or rejected (Rieseberg et al., 1988, 1990; Crawford and Ornduff, 1989). Molecular data also have indicated hybridization where it previously has not been suspected (Palmer et al., 1983; Smith, 1988; Doebley, 1989; Smith and Sytsma, 1990). Additionally, molecular data have been useful in documenting that natural hybridization does not always lead to speciation (Rick, 1974; Levin, 1975; Bell and Lester, 1978; Doebley et al., 1984; Doyle et al., 1985; Palmer et al., 1985; Soltis and Soltis, 1986; Doyle and Doyle, 1988; Doyle and Brown, 1989). Solanum sect. Petota, the potato and its wild relatives, is a group where natural hybridization and hybrid speciation has been hypothesized to be prevalent (Hawkes, 1958, 1962; Ugent, 1970a; Hawkes, 1990). Hawkes (1990) includes 232 species in the group, which he divides into 21 taxonomic series. The species are widely distributed throughout the Americas from southwestern Nebraska to southern Chile. Natural hybrids are believed to be quite common in some areas. For example, Hawkes and Hjerting (1969) state that 9.5% of the collections they examined from Argentina, Brazil, Paraguay, and Uruguay represented hybrids. Although there are well-developed sterility barriers between some species (Grun, 1961; Grun et al., 1962, 1977; Hawkes, 1958, 1981; Johnston and Hanneman, 1982), the majority of species can cross naturally or artificially. These hybrids are highly fertile, at least in the F, generation, even in many interseries crosses (Hawkes and Hjerting, 1969, 1989). Solanum raphanifolium (ser. Megistacroloba) was hypothesized by Ugent (1 970b) to be a recent stabilized diploid hybrid species between S. canasense (ser. Tuberosa) and S. megistacrolobum (ser. Megistacroloba). This hypothesis was supported on the basis of a number of features: 1) The morphology of S. raphanifolium is strikingly intermediate between its two putative parents, especially in the features of habit, leaves (Fig. 1), and flower shape (Table 1); 2) Solanum raphanifolium occurs in weedy disturbed habitats unlike the generally undisturbed habitats of its putative parents; 3) S. raphanifolium occurs in a restricted overlap zone between the two species in southern Peru (Fig. 1); and 4) Ugent (1 970b) found many intermediate individuals in this overlap zone that he believed were F1 and later generation hybrids between S. canasense and S. megistacrolobum suggesting that hybridization was an ongoing process. This combination of morphological, ecological, distributional, and field evidence made this a well-documented a very reasonable hypothesis. This study was initiated to test the S. raphanifolium hybridization origin hypothesis with chloroplast DNA (cpDNA) and 18S-25S nuclear ribosomal DNA (nrDNA) characters. Assuming relatively recent or even ongoing hybridization between the two putative parental species, the expectation is for the hybrid to possess an identical, or nearly identical maternally inherited cpDNA pattern to one parent (Hosaka et al., 1984;