Gene context analysis reveals functional divergence between hypothetically equivalent enzymes of the purine-ureide...

A major problem of genome annotation is the assignment of a function to a large number of genes of known sequences through comparison with a relatively small number of experimentally characterized genes. Because functional divergence is a widespread phenomenon in gene evolution, the transfer of a function to homologous genes is not a trivial exercise. Here, we show that a family of homologous genes which are found in purine catabolism clusters and have hypothetically equivalent functions can be divided into two distinct groups based on the genomic distribution of functionally related genes. One group (UGLYAH) encodes proteins that are able to release ammonia from (S)-ureidoglycine, the enzymatic product of allantoate amidohydrolase (AAH), but are unable to degrade allantoate. The presence of a gene encoding UGLYAH implies the presence of AAH in the same genome. The other group (UGLYAH2) encodes proteins that are able to release ammonia from (S)-ureidoglycine as well as urea from allantoate. The presence of a gene encoding UGLYAH2 implies the absence of AAH in the same genome. Because (S)-ureidoglycine is an unstable compound that is only formed by the AAH reaction, the in vivo function of this group of enzymes must be the release of urea from allantoate (allantoicase activity), while ammonia release from (S)-ureidoglycine is an accessory activity that evolved as a specialized function in a group of genes in which the coexistence with AAH was established. Insights on the active site modifications leading to a change in the enzyme activity were provided by comparison of three-dimensional structures of proteins belonging to the two different groups and by site-directed mutagenesis. Our results indicate that when the neighborhood of uncharacterized genes suggests a role in the same process or pathway of a characterized homologue, a detailed analysis of the gene context is required for the transfer of functional annotations.