From Protein Structure to Function with Bioinformatics.pdf

158 B.H. Dessailly and C.A. OrengoPossible examples of backward evolution have been collected, including that of thetryptophan biosynthesis pathway in which several enzymes that catalyse sequentialsteps are clearly homologous (Todd et al. 2001; Gerlt and Babbitt 2001).However, results from several studies suggest that this hypothetical process hasactually played a marginal role in the evolution of metabolism, which instead,would have resulted mostly from a chemistry-driven recruitment of enzymesbetween pathways (Rison and Thornton 2002). Indeed, superfamilies in which thesubstrate selectivity is conserved seem rare in comparison with those cases wherethe catalytic mechanism is conserved. Interestingly, the TIM-barrel phosphoenolpyruvate-bindingenzymes superfamily, which was the only superfamily withabsolutely conserved substrate specificity in the analysis of Todd et al. (2001),proved to be amongst the superfamilies with most diverse cognate ligands in a morerecent study (Bashton et al. 2006), suggesting that the data used in the previousanalysis may have been misleading due to its scarcity.PROCOGNATE (Bashton et al. 2008) is a very useful tool for the analysisof ligand diversity bound by the different enzymes within a superfamily. ThePROCOGNATE database maps enzymes to their cognate ligands, i.e. the ligandsthat the enzymes bind in vivo. Indeed, ligand data from PDB structures pose aproblem: frequently, non-specific ligands bind to the enzymes in their active sitethus mimicking the real ligand that binds in vivo (Dessailly et al. 2008). Thesecontaminants make it difficult to automatically study ligand diversity in proteinsof known structures as it is not obvious how to distinguish them from the biologicalligands. PROCOGNATE is organised around the superfamilies (CATH, SCOPor PFAM) to which the enzymes belong. It is therefore useful for determining liganddiversity for any given superfamily of interest. For example, searchingPROCOGNATE for the mechanistically diverse haloacid dehalogenase superfamily(CATH code 3.40.50.1000) returns a list of 57 cognate PDB ligands and 17cognate KEGG compounds that bind to enzymes in that superfamily. The ancientand diverse HUP-domain superfamily (CATH code 3.40.50.620) is associated with92 PDB ligands and 29 KEGG ligands in PROCOGNATE. These 29 KEGG ligandsare shown in Fig. 6.3 and illustrate the diversity of molecules that can bebound by evolutionarily related proteins.Functional Changes Due to Changes in the Environmental ContextFunctional changes between duplicated copies of a protein can also arise not so muchfrom changes within the protein itself, but rather from changes in the environmentalconditions in which the different copies are active. For example the recruitment of aprotein in new locations of an organism may theoretically result in its encounter withsmall molecules that were not present in the original environment of the ancestor protein,and the recruited protein may display unexpected ability to bind these newly availableligands. Likewise, the molecular function of a protein may change if other proteinsin its environment undergo mutations which result in the possibility for new interactionsor, on the contrary, in some protein-protein interactions becoming no longer possible.