Artificial intelligence-guided engineering of a thermostable cellulase cocktail enables efficient biomass hydrolysis at elevated temperatures.

Efficient lignocellulosic biomass conversion under industrially relevant high temperatures is limited by the thermolability of commercial cellulases and a lack of synergistic, system-level thermostable cocktails. To address this gap, a thermostable cellulase cocktail comprising the endoglucanase TpEG, the cellobiohydrolase HmCel6A, and the β-glucosidase TnBglB was established, capable of synergistically hydrolyzing cellulose under high-temperature conditions. To further improve system performance, precise component optimization was performed using an advanced artificial intelligence framework. First, using the protein language model-guided enzyme mining pipeline VenusMine, EG5, an ultra-thermostable endoglucanase, was identified from a large sequence space and exhibited excellent thermostability at 90 °C. Concurrently, protein engineering of HmCel6A guided by the protein language model PRIME generated the variant S347P, yielding a 1.8-fold increase in catalytic activity. Both the wild-type (WT) and engineered cocktails showed excellent high-temperature hydrolytic activity. At 90 °C, DNS assay showed that the filter paper hydrolytic activity of the WT cocktail was approximately sevenfold higher than that of Cellic® CTec3 (Novonesis), while the engineered cocktail exhibited approximately 1.15-fold higher activity than the WT cocktail. HPLC analysis showed that the engineered cocktail released significantly more glucose from corn stover than both CTec3 and the WT cocktail. Furthermore, supplementation of CTec3 with these engineered enzymes led to increased reducing sugar release under the tested supplementation conditions. This study demonstrates the power of integrating artificial intelligence-driven enzyme discovery with protein engineering, expanding the design space for thermostable enzymes while delivering a high-performing cellulase cocktail for industrial-scale, high-temperature biomass saccharification.