Comprehensive characterization of the complex BAHD acyltransferase family from 25 plant species: phylogenetic analysis and identification of specificity determinant positions

FERNANDO VILLARREAL; AGUSTIN AMALFITANO; HUGO MARCELO ATENCIO; NICOLÁS STOCCHI; ARJEN TEN HAVE

Mendoza

Congreso; X Congreso Argentino de Bioinformática y Biología Computacional.; 2019

A2B2C

BACKGROUNDPlant secondary metabolism constitutes a vast reaction network, which has diversified among land plants in order to produce thousands of compounds with an ample range of functions, associated to the distinct ecological niches plants occupy . This diversification is linked to the increase of the number of homologues in several enzyme superfamilies across this lineage, involved in the biosynthesis of this plethora of compounds. For example, phenylpropanoid metabolism yields multiple metabolites, including flavonoids and anthocyanins. One of the enzyme families involved in this metabolic network is the non-amino acyl transferases known as BAHD acyltransferases, catalyzing the transfer of CoA-activated moieties into acceptor metabolites. It is very difficult to assess functionality to the various homologues in different plant species, due to (i) extended chemical diversity of activated moieties and target metabolites, and (ii) scarce characterization of kinetic data for these enzymes. Here, we perform a deep phylogenetic analysis of the BAHD acyltransferases across the land plants lineage. We aim to understand the evolutionary history of this family, and to identify subfamilies-SDPs (specificity determinant positions), to facilitate function prediction of homologues lacking functional annotation.RESULTSWe identified BAHD transferases homologues in 25 land plant proteomes using HMMER, using the Pfam profile for these enzymes. Then, we eliminated non-homologues, pseudogenes and miss-annotated entries with our in-house built tool SEQrutinator, using Viridiplantae SwissProt annotated sequences as a control dataset. Finally, we ran a sequence-recovery protocol to retrieve back sequences that may have been incorrectly removed, to get a total of 1766 homologues (108 homologues in SwissProt). This dataset was clustered using HMMERCTTER into 20 groups (8 non-clustered sequences). We identify 111 SDPs in the dataset (p < 0.001), using SDPfox. In addition, we investigate the mutual information in the dataset using Mistic. With these data, we substantiate 15 SDPs by building Specificity Determining Networks (SDNs) with all possible SDPs and identifying the network with the highest connectivity level. Most of the validated SDPs are found surrounding the catalytic trench.CONCLUSIONBAHD acyltransferases subfamilies may have evolved through coevolution of amino acids with potential involvement in substrates accomodation and specificity.