基于计算机辅助鱼露寡肽呈味机理和构效关系研究

Fish sauce fermented from low value fish is a traditional Chinese seasoning with unique flavors, among which Umami is the most representative one. In this study, the traditional fermented Shantou fish sauce is selected as raw material to identify fish sauce oligopeptides using ultrafiltration separation, electro-dialysis desalination, HPLC-MS-MS and Edman degradation method, and preparation by solid phase synthesis. Convolutional neural network algorithm is used to establish evaluation system of artificial intelligence group which is sensory analysis oriented in combination of chemical and biological analysis of umami and kokumi peptide receptor, to obtain digital taste data of fish oligopeptide. The three-dimensional crystal structure of flavor peptide is constructed through direct access and homology. Molecular docking model with receptor is established to discover and predict the binding sites between umami/kokumi peptide and receptor as well as investigating the molecular mechanism of taste. According to umami and kokumi data of fish sauce oligopeptide and sequence structure of peptide, the quantitative structure-activity relationship (QSAR) model is constructed by two-dimensional & three-dimensional to verify. The structure-activity relationship and taste mechanism of the fish sauce oligopeptide were studied by using the artificial intelligence sensory evaluation system and molecular simulation. Research on the structure-activity relationship and taste mechanism of fish sauce oligopeptide will provide technology platform and theoretical basis for directional enzymolysis of fish sauce, rapid fermentation production, product quality control, food flavor research and development of new flavor peptide by using evaluation system of artificial intelligence and molecular simulation.

鱼露是海洋低值鱼经发酵形成的具有独特风味的调味品，鲜味是其最具代表的滋味特征。本研究以传统发酵产品潮汕鱼露为原料，采用超滤分离、电渗析脱盐及HPLC-MS-MS、Edman降解法分析鉴定鱼露寡肽，并固相合成制备；利用卷积神经网络算法构建以感官分析为导向结合鲜味肽/浓厚感肽受体化学、生物分析的人工智能感官组学评价体系，获取鱼露寡肽呈味综合表征信息；基于已有呈味肽受体晶体结构和同源建模构建三维晶体结构，建立分子对接模型，寻找和预测鲜味肽/浓厚感肽的受体结合位点以及呈味的分子机理；基于鱼露寡肽的鲜味和浓厚感表征数据，结合肽的序列及三维结构，分别从二维和三维角度构建肽结构与呈味活性间的定量构效关系模型，实验验证模型。运用人工智能感官组学评价体系及分子模拟技术来探究鱼露寡肽的构效关系和呈味机理，为鱼露定向酶解和快速发酵生产、产品质量的控制、食品呈味研究以及新型呈味肽的开发提供理论依据和技术平台。

鱼露是海洋低值鱼经发酵形成的具有独特风味的调味品，鲜味是其最具代 表的滋味特征。本研究团队对传统发酵潮汕鱼露采用超滤分离、电渗析脱盐及凝胶过滤色谱、反相高效液相色谱和毛细管液相色谱-串联质谱(Q Exactive)分析鉴定出多种小肽并分析呈味寡肽结构及特性；采用固相合成了含有脯氨酸和谷氨酸的呈味寡肽；采用电子舌技术、STC-1细胞评价和感官评价相结合的方式构建鱼露鲜味肽鲜味评价体系，在呈味物质浓度和呈味强度间建立相关性；利用分子模拟对鱼露鲜味肽的分子结构进行分析，同时结合鲜味肽的呈鲜强度与阈值来探究鱼露鲜味肽的呈鲜机理与构效关系；构建基于表达鲜味受体T1R1/T1R3细胞系的鱼露鲜味肽鲜味评价方法，为风味评价提供新的途径。