FUNCTIONAL CONSTRAINTS AND RBCL EVIDENCE FOR LAND PLANT PHYLOGENY

Although the proportion of ''functional'' DNA in eukaryotic genomes is both debatable and subject to definition, most sequences gathered for phylogenetic purposes are indisputably functional. For example, patterns of variation are likely to be strongly constrained in ribosomal RNAs because of their structural and catalytic roles in protein translation, and in protein-coding genes, because of protein function itself. Although seemingly obvious, these concerns are usually ignored by workers producing gene trees. We have examined the extent of functional constraints in land-plant rbcL sequences. Not only do rbcL sequences appear to change with essentially clocklike regularity, but nucleotide-based cladograms imply that approximately 97.5% of codon changes on internal branches are functionally neutral (i.e., synonymous or functionally labile). From this perspective, rbcL evolution appears to be strongly constrained by function. Transforming nucleotide data into ad hoc string recognitions alters the size of the unit character sufficiently to highlight ''blocks'' of conservative information that may or may not be functionally constrained. Simultaneous cladistic analysis of all available evidence will highlight the proportion of congruent information, despite diverse functional constraints among the characters analyzed. We demonstrate the strength of this approach using different forms of the same rbcL evidence (i.e., nucleotides, strings, or amino acids) in combination with the seed-plant data of Nixon et al.