MOLECULAR EVOLUTION OF EPENDYMIN AND THE PHYLOGENETIC RESOLUTION OF EARLY DIVERGENCES AMONG EUTELEOST FISHES

The rate and pattern of DNA evolution of ependymin, a single-copy gene coding for a highly expressed glycoprotein in the brain matrix of tel eost fishes, is characterized and its phylogenetic utility for fish sy stematics is assessed. DNA sequences were determined from catfish, ele ctric fish, and characiforms and compared with published ependymin seq uences from cyprinids, salmon, pike, and herring. Among these groups, ependymin amino acid sequences were highly divergent (up to 60% sequen ce difference), but had surprisingly similar hydropathy profiles and i nvariant glycosylation sites, suggesting that functional properties of the proteins are conserved. Comparison of base composition at third c odon positions and introns revealed AT-rich introns and GC-rich third codon positions, suggesting that the biased codon usage observed might not be due to mutational bias. Phylogenetic information content of th ird codon positions was surprisingly high and sufficient to recover th e most basal nodes of the tree, in spite of the observation that pairw ise distances (at third codon positions) were well above the presumed saturation level. This finding can be explained by the high proportion of phylogenetically informative nonsynonymous changes at third codon positions among these highly divergent proteins. Ependymin DNA sequenc es have established the first molecular evidence for the monophyly of a group containing salmonids and esociforms. In addition, ependymin su ggests a sister group relationship of electric fish (Gymnotiformes) an d Characiformes, constituting a significant departure from currently a ccepted classifications. However, relationships among characiform line ages were not completely resolved by ependymin sequences in spite of s eemingly appropriate levels of variation among taxa and considerably l ow levels of homoplasy in the data (consistency index = 0.7). If the d iversification of Characiformes took place in an ''explosive'' manner, over a relatively short period of time this pattern should also be ob served using other phylogenetic markers. Poor conservation of ependymi n's primary structure hinders the design of efficient primers for PCR that could be used in wide-ranging fish systematic studies. However, a lternative methods like PCR amplification from cDNA used here should p rovide promising comparative sequence data for the resolution of phylo genetic relationships among other basal lineages of teleost fishes.