Differentiation of soy sauce types by HPLC profile pattern recognition. Isolation of novel isoflavones

Nonvolatile minor components in various brands of Japanese fermented soy sauce were analyzed by gradient RP-HPLC and monitored at 280 nm. Chemometric pattern recognition techniques, such as cluster analysis, linear discriminant analysis (LDA), LDA using genetic algorithm (GA-LDA) and soft independent modelling of class analogy (SIMCA), succeeded in differentiating the resulting HPLC profiles according to soy sauce brands. Three components playing key roles in the differentiation were isolated by preparative HPLC and purified by gel-filtration chromatography, or simply repeated preparative HPLC. FAB-MS, 1H-, 13C-NMR and IR spectra suggested that these three components having molecular weights of 386, 402 and 418 were isoflavone derivatives. By applying HMBC spectral analysis, these isoflavones were identified as conjugated ethers of tartaric acid with daidzein, genistein and 8-hydroxygenistein. These new isoflavone derivatives are produced by some strains of Aspergillus fungi.