A comparison of 18s ribosomal RNA and rubisco large subunit sequences for studying angiosperm phylogeny

Partial sequences of 18s rRNA were obtained for 2 gymnosperms and 12 angiosperms from a wide range of families and these were analyzed with 5 other published sequences to form a phylogenetic tree. Using 16 published sequences of the large subunit of rubisco (rbcL), also from a wide range of angiosperm families, another phylogenetic tree was derived and the two approaches were compared. Both phylogenetic trees gave good grouping within families but in neither case was there resolution of the branching order of major taxa. Superficially the long rbcL sequences (whose base composition was homogeneous among all species) seemed very promising, but analysis showed that a large proportion of the variation did not affect the amino acid sequence. Although silent substitution contained some phylogenetic information, at the level required to order major taxa, much of it was random and obfuscating. It was concluded that neither macromolecule alone was likely to yield a solution to the problem of angiosperm phylogeny and therefore that studies of both, at least, will be required. For this reason, a method was described for obtaining both DNA and RNA of good quality from the same preparation and which had been used successfully with a wide range of species including many with pungent leaves.