Letter to the Editor Plasmodial Ribosomal RNA as Phylogenetic Probe: A Cautionary Note’ Vladimir Corredor and Vincenzo Enea Department of Medical and Molecular Parasitology, New York University Medical Center If one is to draw reliable taxonomical inferences from the analysis of molecular sequences, the phyletic relations between the genes used to build trees must be or- thologous, i.e., the same as those of the corresponding organisms (Nei 1987, p. 288). Barring xenology, this condition is met when a gene either is single copy or evolves in so concerted a manner as to behave like one. The latter is generally the case for the ribosomal RNA (rRNA) genes, and it is in this sense that one refers to the rRNA sequence of an organism (Nei 1987, p. 129). A partial exception to this state of affairs is found in Plasmodium. These parasites are known to harbor distinct rRNA genes that are developmentally controlled. Type A genes are expressed in the asexual stages, and type B genes are expressed in the sexual stages (Gunderson et al. 1987; Waters et al. 1989a). The available sequences of the small ribosomal RNA subunit (SRS) indicate that the A and B genes do not evolve in a fully concerted manner (Enea and Corredor 199 1). This complicates their use as probes to ascertain phylogenetic relationships within the genus. For example, consider the cladograms that include the complete plasmodial SRS sequences available (seven type A and three type B) (fig. 1) . The B sequences lie close to their cognate A sequences, which might imply that the establishment of develop- mentally controlled SRS occurred three times, once each in P. berghei and P. falciparum and again in the P. cynomolgi/ P. fragile lineage. Although possible, this interpretation is unparsimonious because it requires separate duplications of the structural moieties of the gene, plus de novo establishment of the appropriate stage-specific control se- quences. This entails that the trait of developmentally controlled rRNA isoforms, for all its rarity, is not intrinsic to the plasmodial ontogenetic plan. No such difficulties are met if one views the trait as ancestral and the similarity between the cognate isoforms as stemming from sequence homogenization (gene con- version, nonreciprocal recombination, and the like). If this is the case, it should not be assumed that the gene tree will be isomorphic with the corresponding species tree. Consider three taxa whose phyletic relationship is [ ( 1,2) 3 ] and whose A and B genes originally grouped as [((A 1,A2)A3) ( ( B 1 ,B2) B3)]. If, through the action of rare but relatively recent gene conversions, Bl and B3 came to resemble their A paralogues, whereas A2 came to resemble its B paralogue, the resulting gene tree would be [((A 1 ,B 1 )( A3,B3))( A2,B2)], which is not isomorphic with the species tree. In this light, consider the placement of berA on the tree (fig. 1). On the one hand, it is separated from the genes of the other species by the longest internal branch; on the other hand, it is much closer to its cognate (berB) than cynA and falA are to theirs. Thus berA may be the most divergent of the A genes, either because P. berghei actually is the most divergent of the taxa or because, through conversion, it has come to look more like the berB gene. There seems to be no reason to favor the first possibility: 1. Key words: Plasmodium, rRNA. Address for correspondence and reprints: Dr. Vincenzo Enea, Department of Medical and Molecular Parasitology, New York University Medical Center, 341 East 25th Street, New York, New York 10010. Mol. Biol. Evol. 10(4):924-926. 1993. 0 1993 by The University of Chicago. All rights reserved. 0737-4038/93/1004-00 13$02.00 924 Downloaded from https://academic.oup.com/mbe/article/10/4/924/1011754 by guest on 28 July 2022