Title : Structure and function prediction of a potential ?-glucosidase from Miscanthus lutarioriparius
Abstract:
Plants cannot move to escape threats, so they protect themselves by producing special chemicals. One important group of these chemicals involves enzymes called β-glucosidases, which help create defensive compounds. Finding new β-glucosidases and their matching substrates in nature can be difficult. However, recent advances in DNA sequencing and AI-based protein structure modeling have made it easier to discover new enzymes. In this study, we used an AI tool called Boltz-2 to analyze a candidate β-glucosidase that we found by comparing sequences of known β-glucosidases. This new enzyme was identified in silver grass, a plant species. We named the enzyme MlGh1. Using Boltz-2, we modeled how the enzyme folded and found that its overall structure is consistent with other β-glucosidases. We also docked a potential substrate to MlGh1 and observed that the two key glutamate amino acids in the active site were positioned correctly for enzymatic activity. Our results demonstrate that AI tools like Boltz-2 can help identify new β-glucosidases and predict their substrates. These findings provide a foundation for engineering new plant traits by modifying such enzymes. Using AI in combination with modern sequencing will allow researchers to speed up the discovery of enzymes with potential uses in agriculture and biotechnology. This work highlights the power of AI methods in exploring plant defense biochemistry and opens new avenues for designing enzymes for improved plant protection.
