(444d) Kinetic Model for the Enzymatic Synthesis of Galacto-Oligosaccharides: Describing Galactobiose Formation

Galacto-oligosaccharides (GOS) are high added-value products used in the food industry mainly for infant and senior formulations. GOS are prebiotics composed of a mixture of di-, tri-, and larger oligosaccharides, formed by structures of D-glucose-[β-D-galactose]_nor D-galactose-[β-D-galactose]_n with different glycosidic linkages, where n generally ranges from 1 to 5. GOS can be synthesized enzymatically from lactose rich substrates using β-galactosidase in a reaction known as transgalactosylation. This is a kinetically controlled reaction, where GOS are intermediates in the hydrolysis of lactose. Obtaining a kinetic model that can describe this reaction is critical for process development and optimization. Simple kinetic models proposed in the literature do not fit well to the experimental data. On the other hand, very complex models generate many parameters to be fitted. Moreover, most of the models proposed in the literature take into account only GOS where a glucose residue is present. Although galactobiose (GLB), a disaccharide consisted solely of galactose residues, is also a prebiotic, its formation is mostly neglected in kinetic models. The reason may lie in the small amount of this GLB formed and the difficulty in discerning between this disaccharide and lactose through classical analytical methods such as Liquid Chromatography. Neglecting the GLB in the model, however, leads to systematic errors in the glucose and galactose residues balance, in addition to underestimation of the real GOS production.

In this presentation, a new model is proposed for the enzymatic synthesis of GOS that can also describe GLB formation. Although many reactions are considered, including the formation of tetrasaccharide and enzyme inactivation, the number of kinetic parameters is reduced from 22 to only 10 when some simplifying assumptions are made. Kinetic assays are carried out using a glass-jacketed reactor at 40 ^oC, where free β-galactosidase from Kluyveromyces lactisis is used as biocatalyst. Different initial concentrations of lactose are considered and High Performance Liquid Chromatography is used to measure the components involved in the synthesis. For the modeling effort, the parameters are fitted to the experimental data using an in-house MATLAB algorithm. A bootstrap method, based on the residuals obtained during the parameter estimation, is used to calculate confidence intervals for the parameters and analyze parametric correlation. The model, even with some simplifying assumptions, fits very well to the experimental data. The developed model corresponds to a step forward in inferring the production of GOS consisting solely of galactose residues that is done for the first time. Such model will be subsequently used for state estimation exploring extended Kalman filter (EKF) and moving horizon estimation (MHE) approaches as well as advanced model predictive control (MPC) implementations.

Financial support: grant #2018/04933-5, São Paulo Research Foundation (FAPESP); Finance Code 001, CAPES - Brazil