Experimentelle Analyse, Modellierung und biochemische Charakterisierung von Ein- und Zweiphasenreaktionen für die technische Biokatalyse

Reaction engineering is a key aspect in the optimisation of chemical reactions. However biocatalysis faces a set of problems while implementing the interdisciplinary methods from biology, thermodynamics, and engineering science. Four issues are of mayor concern: (1) arbitrary simplifications in kinetic modelling, (2) subjective choice of optimal reaction conditions, (3) only a rudimentary knowledge of environmental conditions on the progress curve of enzymatic reactions and (4) lacking consideration of the complex interaction of reaction and mass transfer in two-phase systems. Owing to these problems, a quantitative physical understanding of the kinetic and thermodynamic phenomena dwindles away and impairs high productivities. Thus, a systematic methodology was developed within this thesis in order to rationalise enzymatic reactions with the aim of a process optimisation. Two enzyme reactions were investigated due to their commercial interest in the fine chemical industry: Formate dehydrogenase from Candida boidinii (FDH) and benzaldehyde lyase from Pseudomonas fluorescens Biovar I (BAL). For the first time a mechanistic kinetic model for both FDH and BAL was developed. The FDH model demonstrated a superior fit to experimental data in comparison to the literature model. The model for BAL was developed on the basis of elementary reaction steps as well. As a result a mechanistic kinetic model for BAL was published for the first time and exhibited a prediction ability which was in excellent agreement with all experiments. Subsequently, the influence of the process conditions was modelled as a function of temperature, pH, ionic strength and DMF content, which served as cosolvent for the hydrophobic reactands. These thermodynamic equations were coupled with the kinetic model and the resulting combined model was tested for progress curve prediction under different conditions. Demonstrating good prediction in all experiments the combined model was successfully validated in a wide range of conditions (20-55°C, 25-400 mM ionic strength, pH 7.2-10.4, 5-45% DMF (v/v)). Since two-phase systems facilitate the conversion of hydrophobic compounds, the potential of hexane and methyl isobutyl ketone (MIBK) was investigated. Due to the interaction of enzyme reaction and mass transfer the evaluation of the intrinsic kinetics in two-phase systems are difficult. In order to overcome the problem, reaction conditions were rationally optimised. Under the optimised conditions the error due to mass transfer was successfully lowered to 2.2% and smaller and could be neglected, consequently. Initial rate analysis in the optimised two-phase systems indicated identical intrinsic model parameters and it was concluded, that the enzyme kinetics are not changed as a result of the interphase or dissolved organic solvent molecules in the aqueous phase. This argumentation was supported by four progress curve measurments in buffer-hexan and by analysis of an absolute engineering measure: the Hatta number. A model-assisted optimisation of the reaction conditions (e.g. temperature, pH value, DMF content, interfacial area) was subsequently carried out to maximise space-time-yield (STY). Remarkably, the values depend systematically on reaction time, enzyme concentration and the volume ratio of a two-phase-system. As a result, the identified optimal reaction conditions holds only true as long as the process remains unchanged. Based on the in silico-tests two optimal experiments were performed in buffer-MIBK. In comparison to literare values the productivity could be increased by 4- and 11-fold, respectively. Finally, some of these findings and additional observations associated with this study were discussed from a biochemical view. Three important results are noted: (1) For an efficient synthesis with BAL the lowest possible ionic strength should be chosen. (2) The pH dependency of the enzymatic activity showed to depend on the DMSO concentration. As a result, different pH maxima are observed for varying DMSO concentrations. Now, the seemingly contrary pH maxima (3 units difference) could be elucidated by considering this correlation in respect to the activity of BAL. (3) In this respect the claim could be supported that glutamate in position 50 governs the acidic limb of the pH-activity profile by potentiometric tests