2.1 Compactional screening the mutant candidates with better catalysis efficiency and thermostability
α-L-rhamnosidases, a glycoside hydrolase, exhibit a low sequence identity at only 20%–30%, but share a similar (α/α)6-barrel catalysis domain and several β-sandwiches.36-40 Six crystal structures (PDB: 2OKX, 3W5M, 3CIH, 4XHC, 6GSZ and 6I60) from the GH78 family have been determined so far. By homology modelling with the crystal structures of α-L-rhamnosidases 2OKX, 3W5M, and 4XHC as templates, the structure of r-Rha1 was built and used for our research. In order to find out how the non-active-site residues of α-L-rhamnosidase contribute to the catalytic efficiency, bioinformatics analysis was performed. Specifically, docking between r-Rha1 and its substrate p NPR were performed and the interaction between them could offer a guide for site-directed mutagenesis. For example, Lu et al.45 have reported that some amino acids identified from protein docking play key roles in catalysis. Bernardi has reported that hydrogen bonds involved amino acids around the catalytic domain contribute most to the contact between enzyme and substrate, which could be mutated to adjust enzymatic activity.46 In the binding mode of r-Rha1 and substrate p NPR (Fig.1A),9 residues (Trp253, Tyr293, Thr301, Val302, Ser303, Ala355, Ser356, Asp525, Trp640) of r-Rha1 were located around p NPR (Fig, 1B). Notably, some of these residues could be significant for enzyme activity and should not be mutated. According to structure of bound L-rhamnose from Fujimoto37, which was sandwiched between two aromatic residues, Trp640 and Trp747, the corresponding residues in r-Rha1, Trp253 and Trp640, are such essential residues and should not be modified in site-directed mutagenesis. An amino acid sequence alignment of the available α-L-rhamnosidases from different sources revealed that, r-Rha1 shared a low similarity with other α-L-rhamnosidases (Figure S1), but the residues in the catalytic domains are well conserved (Table 1). The model of r-Rha1 exhibited that the side chains of Tyr293 and Trp253 were located near the substrate binding loop which were important in the binding of the substrate. In the predicted catalytic site, the aromatic rings of Tyr293 and Trp253 were parallel to the ring of rhamnose and presumably play roles in fixing the substrate through the pi-pi stacking interaction, which is consistent with what Xu et al 47 have showed, i.e. hydrophobic residues were located around the catalytic pocket. As a consequence, Trp253, Tyr293 and Trp640 should be excluded for mutagenesis. Thus, the rest residues identified from the molecule docking, Thr301, Val302, Ser303, Ala355, Ser356, Asp525, were selected for mutation to test the effect on its catalytic activity. Particularly, 14 mutations, T301S, T301Q, T301G, V302S, V302A, V302N, S303V, S303G, A355N, A355G, S356I, S356Y, D525N, D525G were designed, according results of sequence alignment (Table 1).
To filter out mutants with high thermostability, the thermostability of 14 mutants were evaluated by mutation energy (stable) module of Discovery studio 2019 at 60, 65, 70, 75, 80 ºC, and the calculation results were shown in Fig 1C. As a result, seven mutants were found to be stable at 60, 65, 70, 75, 80ºC. They are D525N, S356Y, D525G, S356I, A355N, S303V, V302N, with the stability from high to low in sequence. While the structure of mutant T301Q, T302S, T301S, V302A, S303G, A355G, T301G were unstable at different temperatures, it means the seven mutants have lower thermostability (see Table S2 for detailed data). Therefore, we obtained seven mutants D525N, S356Y, D525G, S356I, A355N, S303V, V302N through two rounds of screening, which may have both high enzyme activity and good thermostability, and could be verified experimentally.