A phylogenetic approach to the identification of phosphoglucomutase genes.

The expanding molecular database provides unparalleled opportunities for characterizing genes and for studying groups of related genes. We use sequences drawn from the database to construct an evolutionary framework for examining the important glycolytic enzyme phosphoglucomutase (PGM). Phosphoglucomutase plays a pivotal role in the synthesis and utilization of glycogen and is present in all organisms. In humans, there are three well-described isozymes, PGMI, PGM2, and PGM3. PGM1 was cloned 5 years ago; however, repeated attempts using both immunological approaches and molecular probes designed from PGM1 have failed to isolate either PGM2 or PGM3. Using a phylogenetic strategy, we first identified 47 highly divergent prokaryotic and eukaryotic PGM-like sequences from the database. Although overall amino acid identity often fell below 20%, the relative order, position, and sequence of three structural motifs, the active site and the magnesium--and sugar-binding sites, were conserved in all 47 sequences. The phylogenetic history of these sequences was complex and marked by duplications and translocations; two instances of transkingdom horizontal gene transfer were identified. Nonetheless, the sequences fell within six well-defined evolutionary lineages, three of which contained only prokaryotes. Of the two prokaryotic/eukaryotic lineages, one contained bacterial, yeast, slimemold, invertebrate, and vertebrate homologs to human PGM1 and the second contained likely homologs to human PGM2. Indeed, an amino acid sequence, derived from a partial human cDNA, that fell within the second cross-kingdom lineage bears several characteristics expected for PGM2. A third lineage may contain homologs to human PGM3. On a general level, our phylogenetic-based approach shows promise for the further utilization of the extensive molecular database.