Model-based experimental analysis of enzymatic reaction networks

  • Modellbasierte experimentelle Analyse von enzymkatalysierten Reaktionsnetzwerken

Ohs, Rüdiger Björn Harald; Spieß, Antje (Thesis advisor); Wiechert, Wolfgang (Thesis advisor)

Aachen (2020)
Dissertation / PhD Thesis

Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2020


The kinetic description of enzyme-catalyzed reactions is a core task in biotechnology and biochemical engineering. In particular, mechanistic models help from the discovery of the biocatalyst throughout its application. This thesis aims at advancing both the kinetic identification procedure for enzyme reaction networks in general and the kinetic understanding of Thiamine diphosphate (ThDP)-dependent enzymes in particular. The latter are viable biocatalysts which catalyze a broad range of reactions with excellent enantioselectivity, but their kinetic understanding still falls behind. The development of kinetic models comprises deriving rate equations, performing experiments, estimating parameters, and analyzing the model. Progress curve experiments combined with optimal experimental designs (OED) are an efficient approach to determine enzyme kinetics. As it is hardly possible to verify why specific experiments are suggested for nonlinear enzyme kinetic models, surface and contour plots of different OED criteria were systematically investigated. The analysis improved the understanding of OED and allowed deducing five suggestions for kinetic identification. The developed OED-routines were then applied during the kinetic identification of the benzaldehyde self-carboligation yielding (R)-benzoin catalyzed by benzaldehyde lyase from Pseudomonas fluorescens (Pf BAL). Moreover, benzaldehyde inactivates Pf BAL. Because only few mechanistic models for carboligation and simultaneous inactivation are available today, the reaction kinetics and inactivation of the biocatalyst were simultaneously determined from progress curves. OEDs improved parameter precision significantly while maintaining the necessary number of 13 experiments moderate. These results and previously published experiments on Pf BAL-catalyzed self-ligation for the substrate 3,5-dimethoxybenzaldehyde were then used to investigate both the effect of using different kinetic model equations and the effect of using models with and without enzyme inactivation on the kinetic parameter values. The analysis highlighted possible pitfalls in the interpretation of kinetic parameter estimates and suggests a consistent strategy for data management and validation of kinetic models. Furthermore, ThDP-dependent enzymes catalyze branched reaction networks which undergo two alternative pathways for the release of two different products from a central intermediate. The commonly used Cleland’s notation for linear reactions was extended for branched reaction pathways, and the rate expressions for two coupled ordered bi uni reactions were explicitly derived for the first time. This derivation demonstrated that numerous mathematical operations need to be performed. Therefore, an automated routine was developed to derive micro- and macrokinetic rate expressions for complex reaction networks with up to eight different products. The new rate expressions for branched reaction networks allowed to perform model-based experimental analysisfor the cross-ligation of benzaldehyde and propanal catalyzed by PfBAL yielding (R)-2-hydroxy-1-phenylbutan-1-o ne and to show that such a reaction mechanism can be modeled mechanistically while leading to reasonable kinetic parameters.