Phylogenetic utility of the second intron of LEAFY in Neillia and Stephanandra (Rosaceae) and implications for the origin of Stephanandra

A homeotic gene, LEAFY, has been suggested to be a single-copy gene in diploid angiosperms. Nucleotide sequences of the second intron of this gene, along with those of several regions of the chloroplast genome (trnL-trnF, trnD-trn Y-trnE-trnT, and matK-trnK) and nuclear ribosomal ITS, were obtained from the species of Neillia and Stephanandra to examine the phylogenetic utility of the intron and to elucidate the phylogenetic relationships among species of the two genera. PCR amplification of the second intron of LEAFY using universal degenerate primers produced PCR products in sufficient quantity for successful direct sequencing. The length of the intron ranged from 591 to 62 2 base pairs (bp) in Neillia and Stephanandra, except in N. thibetica (ca. 1370 bp), and sequence analysis of this region from multiple accessions revealed low levels of infraspecific variation. Comparison of the LEAFY data with ITS and cpDNA data demonstrated that the LEAFY intron was the most variable and useful for phylogenetic analysis at the species level, providing many more phylogenetically informative characters per 100 bp (7.4) than either ITS (3.2) or cpDNA (0.7). Phylogenetic analyses of LEAFY data using both maximum parsimony and likelihood methods generated well supported and highly resolved gene trees with few homoplasies (CI = 0.97). Stephanandra is monophyletic and is nested within Neillia in both LEAFY and cpDNA trees, while the relationship is poorly resolved by ITS data. LEAFY and cpDNA data, however, strongly conflicted with each other with respect to the position of Stephanandra: LEAFY trees placed Stephanandra as sister to the ((N. affinis, N. gracilis), N. thyrsiflora) clade whereas cpDNA data suggested Stephanandra is sister to N. uekii. Both gene trees, however, are nearly identical to each other when Stephanandra is excluded. A hybrid origin of Stephanandra is suggested as a plausible hypothesis to explain the incongruence between LEAFY and cpDNA data sets, though gene duplication/loss and lineage sorting events cannot be ruled out as possibilities. (C) 2003 Elsevier Science (USA). All rights reserved.