Predicting the optimum temperature of β-agarase based on the relative solvent accessibility of amino acids

•We analyzed the relative solvent accessibility properties of amino acids related to the thermostability of β-agarase and get 21 significant differences.•We found the electrostatic interactions between Glu and His at specific sites in the internal part of β-agarase is the uppermost contributor to its thermostability.•We proposed a SVM model to predict the optimum temperature of the β-agarases using the 21 significant differences as input vector.•The predicting results were encouraging with the MAEs are 2.2 °C, 2.6 °C and 0.1 °C for our experimentally verified 3 β-agarases, respectively.

It is of great significance to investigate the amino acids in different solvent accessibility parts of β-agarase responsible for its thermal stability, which may help for understanding the structural basis of thermophilic β-agarase and developing practical strategies for reengineering new one. We systematically analyzed the amino acids distributed in the internal, intermediate and external states of β-agarases and found 21 significant differences in the three states by t-test. Among them, Glu, His and charged residues in the internal part and Tyr in the external part are critical for the stability at high temperature. Based on it, we developed a support vector regression model to predict the optimal temperature of β-agarases with the 21 significant factors as input vector and the results are encouraging. The mean-absolute error (MAE) and root-mean-square error (RMSE) in the 10-fold cross-validation are 3.80 °C and 5.46 °C, respectively. In addition, we experimentally verified three β-agarases with the optimal temperature as 45 °C, 50 °C and 50 °C, and the predicted temperatures were 47.2 °C, 47.4 °C, and 49.9 °C, respectively.

Graphical abstractFigure optionsDownload full-size imageDownload as PowerPoint slide