Sequences of Two Highly Divergent Canine Type c Lysozymes: Implications for the Evolutionary Origins of the Lysozyme/α-Lactalbumin Superfamily

The amino acid sequences of two canine lysozymes, from milk and spleen, have been elucidated by direct sequence analyses of the purified proteins and fragments generated from them. The two enzymes are highly divergent, differing from each other by 45% in sequence, but each is closely similar to lysozymes previously obtained from other mammalian species. The milk lysozyme is similar in sequence to equine and donkey milk lysozymes (83% identity) and, like these enzymes, contains a bound Ca2+ ion while the spleen enzyme is most similar in sequence to the majority of previously studied mammalian and avian lysozymes (80 to 83% identity) and, based on its sequence, does not contain a Ca(2+)-binding site. This demonstrates that Ca(2+)-binding lysozymes are expressed in at least two mammalian orders, the carnivores and perissodactyls, as well as confirming that the genes for the Ca(2+)-binding and conventional lysozymes are paralogous. The latter point was further confirmed by the isolation and partial sequence analysis of a conventional lysozyme from equine spleen. The relationships of these new lysozyme sequences to those of other lysozymes and their homologues, the alpha-lactalbumins, were analyzed using different molecular phylogeny algorithms, producing a new model for the evolutionary origins of the superfamily. The most significant conclusion to be drawn from this model is that Ca(2+)-binding activity was an ancient feature of this protein superfamily which was lost during the evolutionary development of the conventional lysozymes. It also supports a previous suggestion that the alpha-lactalbumins and lysozymes diverged at a time earlier than the divergence of the fishes and tetrapods.