Developing New Antibiotics Using Computationally-Guided Protein Engineering

Streptomyces venezuelae synthesizes pikromycin (pik), a macrolide antibiotic, using a modular system of enzymes. Engineering these enzymes could enable facile biosynthesis of new and diverse antibiotics. New products have proven difficult to access because the pik thioesterase enzyme (pikTE) selectively cyclizes the linear pik polyketide into a macrolactone intermediate. Macrolactones are not formed when pikTE is fed an epimerized substrate. This substrate is cyclized when the active site nucleophile is mutated from serine to cysteine. Density functional theory [M06-2X/6-31+G(d,p)] predicts that the cysteine mutant produces a thioorthoester intermediate with S-C bonds of 1.964 Angstroms and 1.933 Angstroms between pikTE and the acyl carrier protein (pikACP), respectively, facilitating faster handoff of the substrate from pikACP to pikTE. A more exothermic cyclization reaction in the cysteine mutant slows the reversibility of cyclization, limiting the opportunities for premature substrate hydrolysis. Mutating another enzyme in this pathway, pikC hydroxylase, shows up to a 38-fold increase in catalytic activity towards non-natural substrates than the wild type or single mutant pikC(D50N). Simulations revealed that pikC(D50ND176QE246A) has a more compact active site than pikC(D50N) or the wild type enzyme and that mutations improve the stability of polar interactions with the substrate’s anchoring group