Terpene synthases (TPSs) are a class of enzymes widely found in plants that synthesize the most diverse family of secondary metabolites in plants: terpenoids. TPSs convert a number of general-purpose substrates (e.g., FPP) into hundreds of stereochemically complex chain, monocyclic, and polycyclic hydrocarbons. The mechanism by which TPSs catalyze terpene formation is quite complex: most TPSs share a similar binding pocket, with product specificity determined by a specific few residues, which are called plasticity residues.
Here, we wanted to obtain a more efficient and better specific (-)-α-bisabolol synthase for our production needs. Therefore, we adopted a rational design scheme with multiple mutations targeting the plasticity residues in TPSs that catalyze the 1,6 cyclization of FPP. And a part of the design was characterized experimentally.
We selected the BOS from artichoke Cynara cardunculus var. scolymus (this protein will be called CcBOS) as the starting point for mutation because it has the highest known catalytic efficiency. First, we modeled the homology of CcBOS based on the known structure of TPSs and plugged the FPP reference Amorpha-4,11-diene Synthase into the catalytic pocket of CcBOS. And this structure was used as the basis for subsequent calculations.
For this structure, we first determined the range of residues to be mutated. The selected positions should not include key catalytic residues performing the chemistry of the catalysis; rather, selected positions should not include key catalytic residues performing the chemistry of the catalysis; rather, residues composing the first and second shell of the active site, or in a(an) (in)direct contact with the substrate, should be selected.
Given the epistatic effects in proteins, multiple mutations cannot simply be expressed as the sum of single mutations, so we used an exhaustive multi-mutant scoring scheme based on: for selected mutation sites, we determined the potential mutation range using the Position-Specific Scoring Matrix(PSSM), and based on the single point mutation free energy (ΔΔG) to determine the final mutation space of each locus. And all possible mutation combinations were exhausted using ROSETTA. We finally obtained 12,288 mutants, of which 9,481 mutants exhibited lower free energy, demonstrating the feasibility of the strategy.
We then performed a clustering analysis of the mutation results, which showed that the mutants with high scores clustered at a specific set of double mutations, further indicating that our design effectively selects sites that play a key activity in the catalytic process. The specific results are not easily presented here due to business secret protection.
Further, we verified the epistatic relationship between individual mutation sites. We analyzed the two triple mutants with the lowest free energy and their corresponding single and double mutants, and found that the single mutations in mut2 and mut3 adversely affected ccBOS, demonstrating the epistatic effect of mut1 on mut2 and mut3, this finding that has key implications for subsequent studies of BOS evolution.
