[2S03m] 07. Biocatalysis Engineering for Sustainable Bioprocessing - Part 1

11:40 AM - 12:00 PM JST(2:40 AM - 3:00 AM UTC)

[2Biocat-05-KL] Bridging Enzyme Engineering with Sustainable Terpene Bioproduction

○Xixian Chen¹, Clement Scipion¹, Jing Sen Ong¹, Leonard Ong¹, Congqiang Zhang¹, Yinghua Yao², Yuangang Pan² (1. Singapore Institute of Food and Biotechnology Innovation, A*STAR (Singapore), 2. Centre for Frontier AI Research, Institute of High Performance Computing, A*STAR (Singapore))

Keywords:

Enzyme engineering,Metabolic engineering,AI and machine learning,High-throughput screening,Terpene and terpenoids

Metabolic engineering has become an attractive method for the efficient production of natural products. It uses enzymes as green catalysts to produce high-quality ingredients from renewable sources. However, many enzymes involved in natural products or secondary metabolite biosynthesis are inefficient, thus limiting the potential of metabolic engineering to achieve industrially viable titer, rate and yield (TRY). Massively increasing the enzymatic concentration—a common strategy—often leads to a marginal gain in product output and sometimes even negatively impacts the TRY, because of the additional metabolic burden to synthesize the extra proteins. An alternative approach is to improve the catalytic efficiencies of enzymes via enzyme engineering. Classical enzyme engineering approach with directed evolution, although effective, is a laborious and trial-and-error process. To accelerate the evolution of enzymes toward a superior performance, we are developing automation-friendly assays to generate high-throughput (HT) data and use the data to develop machine-learning (ML) algorithms to predict the sequence-function relationship and guide mutant design. Combinations of both strategies have led to significant improvement in enzymatic function improvement. In this talk, we demonstrate an artificial pathway to produce cis-α-irone, a premium aroma molecule, by employing a promiscuous methyltransferase (pMT). Using both structure-guided and AI/ML-assisted enzyme engineering strategies, we have improved pMT activity and specificity towards cis-α-irone by >30,000-fold and >1000-fold, respectively. About 800 mg l^-1 cis- α-irone is produced by one-step biotransformation of synthetic psi-ionone. Our work expands our biosynthetic capability beyond natural pathways.