(485an) Sensitivity Analysis of a Two-Enzyme One-Pot System for Production of Lactobionic Acid

Biocatalysis is now offering a range of new catalytic options for industrial complex syntheses. A further opportunity is also arising in the cases where two or more reaction steps can be carried out by a mixture of enzymes in a single reactor [1,2]. One of these multienzymatic syntheses is the production of lactobionic acid which is obtained by oxidation of lactose in presence of flavocytochrome cellobiose dehydrogenase, CDH, as the first biocatalyst (see figure). During the reaction, the reduced state of CDH is oxidized again by the utilization of an intermediary redox mediator that in this cases is 2,2'-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS). A further action of this redox mediator is to serve as electron donor to obtain the reduction of laccase, which is the second enzyme added to the system. The reduced state of laccase catalyzes the reaction where oxygen is fully reduced to water [3,4] moreover the initial state of laccase is regenerated. In the system, enzymes and redox mediator cycle between their reduced and oxidized states, and are thus in a synergistic manner transforming substrates in products as well as minimizing undesired parallel reactions e.g. formation of hydrogen peroxide that inhibits the enzymes. In biochemistry, most of the systems have a high degree of complexity due to the non-linear interaction between a number of factors. The above-mentioned bi-enzymatic reaction system is such an example, which contains the following factors affecting the process yield: oxygen diffusion, enzyme inhibitors, etc [3]. Often mathematical models are developed to interpret the measurements from these systems thereby helping to understand the dynamic behavior (among others). Typical to any modeling study, however, a number of assumptions and simplifications need to be made to make the modeling exercise tractable. The objective of this contribution is to study the impacts of such assumptions made during modeling for two aims: (i) to evaluate prediction quality of the model (i.e. uncertainty) and (ii) identify critical factors affecting the process yield. To this end Monte-Carlo and Morris Screening techniques are used for uncertainty and sensitivity analysis respectively. The results are used for identifying critical regions affecting process yield in the experimental design space [5]. This information from sensitivity analysis, in turn, is used for better allocating experimental resources, e.g. preventing wastage experimental time investigating factors, which are not highly relevant on the system behavior. Figure shows the ping-pong mechanism for production of lactobionic acid by oxidation of lactose using two enzymes (CDH and Laccase) and an intermediary redox mediator (ABTS). [1] D.Y. Murzin and R. Leino (2008). Sustainable chemical technology through catalytic multistep reactions. ChERD, Vol.86, pag.1002 [2] A. Bruggink, R. Schoevaart and T. Kieboom (2003). Concepts of nature in organic sunthesis: Cascade catalysis and multistep conversions in concert. Org Process Rec Dev, Vol. 4, pag.622 [3] W. Van Hecke, A. Bhagwat, R. Ludwig, J. Dewulft, D. Haltrich and H. Van Langenhove (2009). Kinetic modeling od a bi-enzymatic system for efficient conversión of lactose to lactobionic acid. Biotechnology and Bioengineering, Vol.102(5), pag. 1475 [4] R. Ludwig, M. Ozga, M. Zámocky, C. Peterbauer, K.D. Kulbe and D. Haltrich (2004). Continuous enzymatic regeneration of electron acceptors used by flavoenzymes: Cellubiose dehydrogenase-catalyzed production of lactobionic acid as an example. Biocatalysis and Biotransformation, Vol. 22(2), pag. 97. [5] G. Sin, A.E. Lantz and K.V. Gernaey (2009). Sensitivity analysis of non-linear dynamic models: Prioritizing experimental research. In: Proceedings 10th International Symposium on Process Systems Engineering (PSE'09), August 16-20 2009, Salvador, Bahia, Brazil.