Ureidoglycolate hydrolase, amidohydrolase, lyase: how errors in biological databases are incorporated in scientific papers and vice versa

Abstract

An opaque biochemical definition, an insufficient functional characterization, an interpolated database description, and a beautiful 3D structure with a wrong reaction. All these are elements of an exemplar case of misannotation in biological databases and confusion in the scientific literature concerning genes and enzymes acting on ureidoglycolate, an intermediate of purine catabolism. Here we show biochemical evidence for the relocation of genes assigned to EC 3.5.3.19 (ureidoglycolate hydrolase, releasing ammonia), such as allA of Escherichia coli or DAL3 of Saccharomyces cerevisiae, to EC 4.3.2.3 (ureidoglycolate lyase, releasing urea). The EC 3.5.3.19 should be more appropriately named ureidoglycolate amidohydrolase and include genes equivalent to UAH of Arabidopsis thaliana. The distinction between ammonia- or urea-releasing activities from ureidoglycolate is relevant for the understanding of nitrogen metabolism in various organisms and of virulence factors in certain pathogens rather than a nomenclature problem. We trace the original fault in database annotation and provide a rationale for its incorporation and persistence in the scientific literature. Notwithstanding the technological distance, yet not surprising for the constancy of human nature, error categories and mechanisms established in the study of the work of amanuensis monks still apply to the modern curation of biological databases.

Figure 1.

Alternative routes of nitrogen release from ureidoglycolate. The enzymatic activities involved in the reactions are indicated with the corresponding EC numbers.

Figure 2.

Biochemical evidence that AllA catalyses nickel-dependent urea release from ureidoglycolate. (A) Kinetics of ammonia release catalysed by the recombinant AllA protein before and after addition of urease; (B) kinetics of ammonia release with urease present in the reaction mixture; (C) ureidoglycolate-lyase activity of the AllA protein treated with a metal chelator (EDTA) and supplemented with various divalent ions.

Figure 3.

First documentary evidence of functional misassignment of DAL3/AllA proteins in database. The Swiss-Prot record shown was retrieved using the ‘History’ option of Uniprot (32). The text in bold represents interpolated information not present in the referred publication.

Figure 4.

Timeline of DAL3/AllA misclassification as EC 3.5.3.19. Events marked in red represent erroneous interpolation in database or literature; events marked in yellow represent causes or occasions of errors.