RESUMO
The amide moiety belongs to the most common motives in pharmaceutical chemistry, present in many prescribed small-molecule pharmaceuticals. Methods for its manufacture are still in high demand, especially using water/buffer as a solvent and avoiding stoichiometric amounts of activation reagents. Herein, we identified from a library of lipases/esterases/acyltransferases and variants thereof a lipase originating from Sphingomonas sp. HXN-200 (SpL) able to form amides in aqueous solution starting from a broad scope of sterically demanding heteroaromatic ethyl esters as well as aliphatic amines, reaching isolated yields up to 99% on preparative scale and space time yields of up to 864 g L-1 d-1; thus, in selected cases, the amide was formed within minutes. The enzyme features an aspartate next to the canonical serine of the catalytic triad, which was essential for amide formation. Furthermore, the enzyme structure revealed two tunnels to the active site, presumably one for the ester and one for the amine, which permit the bringing together of the sterically demanding heteroaromatic esters and the amine in the active site. This work shows that biocatalytic amide formation starting from various five- and six-membered heteroaromatic ethyl esters in the buffer can serve as a platform for preparative amide synthesis.
RESUMO
Enzymatic decarboxylation of biobased hydroxycinnamic acids gives access to phenolic styrenes for adhesive production. Phenolic acid decarboxylases are proficient enzymes that have been applied in aqueous systems, organic solvents, biphasic systems, and deep eutectic solvents, which makes stability a key feature. Stabilization of the enzyme would increase the total turnover number and thus reduce the energy consumption and waste accumulation associated with biocatalyst production. In this study, we used ancestral sequence reconstruction to generate thermostable decarboxylases. Investigation of a set of 16 ancestors resulted in the identification of a variant with an unfolding temperature of 78.1 °C and a half-life time of 45 h at 60 °C. Crystal structures were determined for three selected ancestors. Structural attributes were calculated to fit different regression models for predicting the thermal stability of variants that have not yet been experimentally explored. The models rely on hydrophobic clusters, salt bridges, hydrogen bonds, and surface properties and can identify more stable proteins out of a pool of candidates. Further stabilization was achieved by the application of mixtures of natural deep eutectic solvents and buffers. Our approach is a straightforward option for enhancing the industrial application of the decarboxylation process.