Abstract
Enzymes are the cornerstone of biocatalysis, biosynthesis and synthetic biology. However, their applicability is often limited by low substrate selectivity. A prime example is the bifunctional linalool/nerolidol synthase (LNS) that can use both geranyl diphosphate (GPP) and farnesyl diphosphate (FPP) to produce linalool and nerolidol, respectively. This bifunctionality can lead to undesired byproducts in synthetic biology applications. To enhance enzyme specificity and create monofunctional linalool synthases, we modified amino acids in the loop between helices C and D of four bifunctional mushroom LNSs. Through these modifications, we successfully shifted the substrate preference of two LNSs (ApLNS from Agrocybe pediades and HsLNS from Hypholoma sublateritium) from FPP towards GPP. Although complete monofunctionality was not achieved, we significantly increased linalool yield by 13 times while minimizing nerolidol production to 1% of the wildtype HsLNS. Docking simulations revealed a substantial reduction in the FPP binding score compared to that of the wildtype. Molecular dynamics simulations suggested that Tyr300 in the apo HsLNS mutant has a greater tendency to adopt an inward orientation. Together with Met77, the inward-facing Tyr300 creates a steric barrier that prevents the longer FPP molecule from entering the substrate binding pocket, thereby greatly reducing its activity towards FPP. This study demonstrates the potential of enzyme engineering to design substrate-specific terpene synthases, which is a challenging task and few successful examples are available. The insights gained can inform future enzyme design efforts, including the development of artificial intelligence models.