Multiple Independent Losses of the rpoC1 Intron in Angiosperm Chloroplast DNA's

Abstract

Previous studies have shown that in Marchantia, black pine, and the vast majority of angiosperms examined to date, the chloroplast gene rpoCl is interrupted by an intron (of about 750 base pairs), but that in the grasses and one of three subfamilies of cacti (Cactoideae) this intron has been lost. DNA's of the complete rpoCl intron region and portions of the flanking exon regions from 107 species (representing 54 families) of angiosperms were amplified with the polymerase chain reaction (PCR) in order to uncover other instances of intron loss. In addition to grass and Cactoideae chloroplast DNA's, we report that the rpoCl intron is missing from the chloroplast genomes of Scaevola and Goodenia (the only two representatives of Goodeniaceae examined), four of ten species of Passiflora (Passifloraceae), two of four genera of Aizoaceae (Delosperma and Faucaria but not Monilaria or Tetragonia), and one of the eight taxa of Medicago examined (M. suffruticosa subsp. leiocarpa; Fabaceae). Mirroring other instances of intron loss from the chloroplast genome, sequence analysis reveals that, for these taxa, the rpoCl gene has undergone a precise deletion of the intron. For those taxa with the intron, minimal size variation is apparent within the region and, in all taxa investigated, the intron lies in precisely the same position in the gene. Parsimony analyses of partial exon and intron rpoCl nucleotide sequences reveal that the latter, owing to considerable conservation of secondary structure in that region of the intron sequenced, can supply useful characters for phylogenetic analysis at high taxonomic levels. The chloroplast rpoC region, homologous to the ' subunit of Escherichia coli (Migula) Castellani & Chalmers RNA polymerase, is divided into two genes, rpoCl and rpoC2, and is located in the large single-copy region of the plastid genome in most angiosperms (Cozens and Walker 1986; Hudson et al. 1988). These genes, along with rpoB, are cotranscribed as a single operon and encode three subunits of the chloroplast RNA polymerase (Hudson et al. 1988). The gene rpoCl is of systematic interest because it is known to be interrupted by a single intron in most, but not all, land plants. DNA sequencing has revealed that this intron is present in the liverwort Marchantia polymorpha L. (Ohyama et al. 1986), black pine (Pinus thunbergii Parl.; Wakasugi et al. 1994), tobacco (Nicotiana tabacum L. 'Bright Yellow 4'; Shinozaki et al. 1986), spinach (Spinacia oleracea L.; Hudson et al. 1988), and numerous species of Apiaceae and relatives (Downie et al. 1996, and unpubl. data) but absent in rice (Oryza sativa L. 'Nihonbare'; Hiratsuka et al. 1989; Shimada et al. 1990), maize (Zea mays L.; Igloi et al. 1990), and from all examined representatives of one subfamily of Cactaceae (Cactoideae; Wallace and Cota 1996). The absence of an intron in rpoCl, based on heterologous filter hybridization and polymerase chain reaction (PCR) experiments, has also been noted in several other grasses (Agropyron, Avena, Bambusa, Cynodon, Danthonia, Dendrocalymus, Lolium, Secale, and Triticum; Katayama and Ogihara 1993; S. Downie and J. Palmer, unpubl. data). Thus, it is clear that the intron was present in the common ancestor of angiosperms and subsequently lost independently in the lineages leading to the grasses