Computationally assisted semirational engineering of Nicotiana attenuata α-glucan phosphorylase enhances cellobiose-to-amylose cascade synthesis.

Abstract

Converting non-food cellulosic sugars into programmable amylose offers a route to higher-value carbohydrate materials. α-Glucan phosphorylase (αGP) controls glucose-1-phosphate (G-1-P)-dependent chain elongation in cellobiose-to-amylose cascades. Here, Nicotiana attenuata αGP (NicαGP) was engineered by integrating AlphaFold3-based structural modeling, ProteinMPNN, EVcouplings, structural priors, and DynaMut2 filtering. Of 916 candidate substitutions, 37 were selected for experimental validation, and combinatorial screening identified L226Y/V437L/F751H as the optimal triple mutant, with 2.10-fold higher relative activity than the same-batch wild-type (WT) control and clear positive epistasis. The mutant further shifted the optimal temperature from 40 to 45 °C, increased thermal transition midpoint from 56.2 to 58.8 °C, and enhanced catalytic efficiency toward both maltotetraose and G-1-P. In the Clostridium thermocellum cellobiose phosphorylase (CtCBP)-NicαGP cascade, L226Y/V437L/F751H reached 38.86 ± 2.06 % conversion at 72 h versus the previously reported WT value of 35.84 ± 1.04 % at 84 h, increasing apparent productivity by 26.5 %. Docking, CAVER, and molecular dynamics (MD) analyses suggested that these gains may be associated with optimization of structural regions surrounding the conserved catalytic core. This work provides an engineered αGP for amylose synthesis from cellobiose and supports peripheral regulatory engineering as a practical strategy for improving phosphorylase-based biocatalytic cascades.