The evolution of enzyme function in the isomerases

Abstract

The advent of computational approaches to measure functional similarity between enzymes adds a new dimension to existing evolutionary studies based on sequence and structure. This paper reviews research efforts aiming to understand the evolution of enzyme function in superfamilies, presenting a novel strategy to provide an overview of the evolution of enzymes belonging to an individual EC class, using the isomerases as an exemplar.

Figure 1

Biological importance of isomerases. (a) Core metabolic pathways (the isomerase reactions are emboldened in black). Carbohydrate and terpenoid/polyketide metabolic pathways are highlighted in blue and green squares, (b) Distribution of known enzymes in the human and E. coli genomes, (c) EC classification of isomerases. (d) Bond changes, reaction centres and structure of substrates and products obtained from the reaction catalysed by alanine racemase (EC 5.1.1.1) using EC-BLAST.

Figure 2

(a) Distribution of isomerases in EC classification, UniprotKB, PDB and FunTree. EC exchange matrices representing the changes in function during evolution of isomerases at the EC (b) class and (c) subclass levels. More frequent changes of isomerase function are highlighted in red. Green and blue boxes represent changes within isomerases and with other EC classes, respectively. (d) Frequency of EC changes involving isomerases by superfamily. The 32 superfamilies bearing multiple changes are illustrated.

Figure 3

Sequence and functional similarity of the 145 changes of isomerase function. The three scatterplots represent global sequence identity against overall reaction similarity as calculated using three measures (a) bond change (b) reaction centre and (c) structure similarity of substrate(s) and product(s). Each point represents one change of enzyme function involving two sets of enzymes catalysing two distinct functions each [27]. Average global sequence identities and standard deviations (error bars) from all-against-all pairwise comparisons between sequences corresponding to one function and those corresponding to the second function. Circled in red, the change EC 4.2.1.124→EC 5.4.99.31 (see main text). Pearson's correlation coefficients (r) range from 0.35 to 0.41 and indicate weak but significant linear relationships (p-value < 0.001). (d) Distribution of bond change and structure similarities averaged by CATH superfamily.

Figure 4

The evolution of SDRs acting on NDP-sugars. (a) Overview of the EC changes involving isomerases and domain composition of UDP-glucose 4-epimerases (EC 5.1.3.2). Biochemical reactions are represented in boxes. Black arrows inside boxes denote chemical transformations whereas coloured arrows linking boxes represent EC changes. EC numbers with an asterisk indicate reactions for which we found mechanistic evidence in MACiE [47] or in literature searches. Changing substructures are highlighted in red whereas X corresponds to nucleoside diphosphate moieties (ADP, TDP, GDP, CDP, UDP) in which the base may change, but the ribose diphosphate (or sometimes the 2′-deoxy derivatives) is broadly conserved. Three scatterplots illustrating sequence and functional similarity for this superfamily (b) bond change, (c) reaction centre and (d) structure similarity of substrate(s) and product(s) as in Figure 3.