Dextranase (EC 3.2.1.11) is a hydrolytic enzyme that specifically recognizes the α-D-1,6 glycosidic bond in dextran and produces low molecular weight dextran (Zhang et al., 2017). Dextranase and its product are widely used in the food, medicine and chemical industries (Reddy Shetty et al., 2021). In the sugar industry, dextranase can be used to hydrolyze dextran in the process of sucrose production to reduce viscosity, the yield and quality of sugar can be increased (Eggleston and Monge, 2005). Dextranase hydrolyzes large molecules of dextran to prepare specific molecular weight products. It has been reported that dextran with relatively low Mw (20, 40, and 70 kDa) is excellent plasma extenders (Bark and Grande, 2014). Isomaltooligosaccharides can also be prepared by dextranase and used in food or drug production. Moreover, compared with the traditional acid hydrolysis method, the reaction conditions are mild and environment-friendly, with high efficiency and good product quality (Gan et al., 2014, Wu et al., 2011). Due to the protective effect of dextranase on enamel, it can be added to toothpaste and mouthwashes to prevent dental caries (Lai et al., 2019).

According to the consistency of the amino acid sequence, dextranase is divided into carbohydrate-active enzyme (CAZy) families (GH) 49, 66 and 31(Liu et al., 2019; Suzuki et al., 2012; Tsutsumi et al., 2020). Many microorganisms can produce dextranase, such as Penicillium, Chaetomium, Streptococcus and Actinomycetes (Ebaya et al., 2020, Jumpei et al., 2012, Larsson et al., 2003). The optimal temperature of fungal-derived dextranase is mostly 30–55 °C, while the optimal temperature of bacterial-derived dextranase is 40–60 °C (Yang et al., 2019, Zohra et al., 2015).

Heat resistant dextranase is more suitable for the application of sugar industry, because there is a high temperature higher than 60 °C the sugar industry process. Most of the dextranase is rapidly deactivated in this environment (Park et al., 2012). Therefore, improving the heat resistance of enzymes is an essential aspect of enzymatic research. In previous studies, the most heat-resistant dextranase was obtained from Thermoanaerobacter pseudethanolicus (TpDex). The half-life of TpDex was 7.4 h at 70 °C (Suzuki et al., 2016). However, the activity of TpDex is too low. Smdex is from Streptococcus mutans, with high activity and short expression cycle. Smdex and TpDex belong to GH66. Their amino acid sequence similarity is 30 % and their three-dimensional structures are very similar. Smdex possess a good basis for improving the heat resistance and a better application prospect. In our laboratory, we cloned the S. mutans-derived dextranase gene on pET-28a (+) and expressed it with E.coli BL21(DE3). As reported previously, the full-size Smdex can be degraded by protease at room temperature, whereas N- and C-terminal truncated SmDex (SmDexTM; bearing Gln100–Ile732) proteins are not (Kim et al., 2011, Klahan et al., 2018). The molecular weight of SmdexTM is 75 kDa. Its optimum temperature is 40 °C and optimum pH is 5.0. The enzyme is stable between pH 4.5–10 at 35 °C. The properties of some of the enzymes belonging to the GH66 family are shown in Table 1. SmdexTM was used as the parent enzyme.

According to reports, many methods have been used to improve thermal stability (Plaks et al., 2020, Wu et al., 2021). Reducing the flexibility of enzymes is an effective way to improve their stability. The B-factor value represents protein flexibility. Generally, the higher the B-factor value is, the lower the thermal stability (David and Wei, 2018). Lin Ge et al. enhanced the thermal stability of α-rhamnosidase based on the B-factor saturation mutagenesis in Aspergillus terreus. The half-life was prolonged by 2.3 times at 70 °C (Ge et al., 2018). Nanyu Han et al. designed mutant enzymes through B-factor and sequence alignment analysis. The xylanase retained 50 % of its enzymatic activity after incubation at 60 °C for 1 h (Han et al., 2017). Based on the semi-rational design of the B-factor, Wei Ren et al. reported the improvement of the thermostability of dextranase derived from Arthrobacter oxydans KQ11(Ren et al., 2019). Calculating the free energy of single-point mutation is a common method for molecular stability modification. Rosetta is a powerful computational design strategy (Kellogg et al., 2011). Cartesian-ΔΔG is a new generation prediction method of Rosetta, the calculation of the skeleton adopts the optimization of the Kadir space to allow the skeleton to move in a small range. Generally, when ΔΔG is positive, it means the point mutation brings an unsteadiness effect to the protein. When ΔΔG is negative, it means the point mutation brings a stabilizing effect. The smaller the value of ΔΔG leads to the higher stability of the protein. In improving the thermostability of transketolase (TK), Haoran Yu et al. show that 9 of 15 variants were predicted correctly by Rosetta, giving a prediction accuracy of 60 %. The best variant had a 3-fold improved half-life at 60 °C and an increase in Tm of 5 °C above the wild type (Yu et al., 2017). Zhao et al. used two structure-oriented design methods, Rosetta and FoldX, to modify the Proteinase K thermostability. The top four variants D260V, T4Y, S216Q, and S219Q showed improved half-lives at 69 °C by 12.4-, 2.6-, 2.3-, and 2.2-fold that of the parent enzyme, respectively (Zhao et al., 2021). Deng Hui et al. reduced the ΔΔG of Thermobifida fusca glucose isomerase by site-directed mutagenesis, and the half-life of the mutant was increased from 14.9 h to 22.3 h at 70 °C (Deng et al., 2013).

In this study, we combined B-factor with Rosetta to identify stabilizing mutations and increase SmdexTM thermostability. Specifically, based on the atomic displacement parameter obtained from X-ray data, we selected the appropriate mutation sites (B-factor). These sites were substituted with more rigid amino acids through Rosetta (ΔΔG).