Molecular evolution of the chloroplast trnL-F region in land plants

Quandt D, Müller KF, Stech M, Hilu KW, Frey W, Frahm J-P, Borsch T

Abstract

The trnL group I intron and the intergenic spacer between trnL and trnF are among the most widely utilized noncoding DNA regions in plant systematics. The trnL-F region is probably cotranscribed and located upstream of trnTUGU in the large single copy region of the chloroplast genome. Whereas in algal chloroplast genomes the three tRNA genes are widely scattered, they are found in tandem in land plants. This unique arrangement, therefore, seems to be a synapomorphy for land plants. Secondary structure analysis of the group I trnLUAA intron shows that high sequence variability is confined to certain stem-loop regions (P6, P8), whereas elements of the catalytic core (P, Q, R, S) are highly conserved across land plants. Considerable length differences, particularly in the stem loop P8, as well as in the trnT-L and trnL-F intergenic spacers, were observed among the diverse land plant lineages. Sequences found within mosses and liverworts were generally shorter than in most other lineages. It appears that certain elements in the trnLF spacer and in the most variable P8 stem-loop region of the trnL intron have originated by independent nucleotide additions. These elements do not share a common evolutionary history across all land plant lineages below seed plants because internally repeated sequence segments have further diversified, actually being internal paralogous stretches. Most of the length variation in the {trnL-F} spacer occurs between the putative sigma70-type bacterial promoter motif upstream of {trnF} and the {trnL} 39 exon. Apart from leptosporangiate ferns, lycophytes, and Gnetales this conserved promoter motif is present in most of the land plant lineages sampled as well as in Chaetosphaeridium. Structural data were used to assess the effect that compensating base pair changes (CBC) in helical elements of the intron have for phylogenetic reconstructions. Such changes significantly contribute to the phylogenetic structure of the data set. Although the inclusion of CBCs in general violates the requirement of character independence, their effect is rather comparable to weighting characters. The contribution of CBCs to the total amount of phylogenetic signal is found to be less significant within terminal land plant lineages, such as pleurocarpous mosses or angiosperms, because helical elements are conserved in their primary sequence and variation is lower within these terminal lineages.