Filling the holes in metabolic pathways through Big Data analysis and experimental validation

Abstract

How can we identify unknown genes encoding proteins that have already known functions in metabolic pathways? This problem occurs in many organisms and therefore also in Homo sapiens despite is one of the most studied. ~25% of the enzymatic reactions in metabolic pathways do not have an assigned gene and are defined as “pathway holes” (PH). Identification of PH genes can have different difficulty levels, from the easiest case when the gene is already identified but not yet confirmed or deposited in a database, to the hardest one, when there is no information for identification of the gene. Fortunately, or not, most PHs are between these two extreme cases, with some information that can be useful but not enough to fill that specific PH. The assignment of genes coding enzymes of all PH will be a long and wasteful work, so we have decided to focus PHs that concern amino acids as reagents or products. In fact, these reactions often require pyridoxal-5’ phosphate (PLP) dependent enzymes, available in a specific database (B6DB) that includes updated information about them and their coding genes, classified by organisms, sequences, family, activities and structures. The first pathway hole that we studied is in human carnitine biosynthesis, where gene coding the second enzyme called hydroxytrimethyllysine aldolase (HTMLA) is still unknown. This kind of pathway is present in almost all mammals, including humans, but even is spread in many others metazoa and in some genus of fungi. The main information of this enzyme that we have exploited is its dependency of pyridoxal phosphate as cofactor. We developed an automatic procedure of reverse docking to screen all PLP enzyme structures of human and mouse using the known substrate of the reaction, hydroxytrimethyllysine (HTML) as ligand. This has allowed us to select valid candidates that we have experimentally assessed, showing promising results. The second pathway hole focuses on Cysteine Lyase (CL) of Gallus gallus, a PLP-dependent enzyme that catalyses a beta substitution of cysteine to form cysteic acid. Its activity was discovered about sixty years ago in the yolk sac of developing chicken eggs, but the gene coding this enzyme is still unknown. Through an in-silico subtraction of the entire set of PLP-dependent enzymes (PLPome) of Homo sapiens and Gallus gallus in B6DB, we identified a specific tandem duplication of Cystathionine beta synthase (CBS) gene only in Gallus and other sauropsids. We have confirmed its cysteine lyase activity through recombinant production and NMR spectroscopy. This enzyme takes part in alternative taurine biosynthesis in sauropsids as confirmed by the co-option of the cysteine sulfinic acid decarboxylase (CSAD) occurs in this vertebrate class, that has shown a specificity for the CL product, cysteic acid. Moreover, the ancestor gene CBS, it has been seen able to recycle toxic hydrogen sulphide produced by CL, constituting a valid mechanism of reduced sulphur in the egg environment.