Theoretical and experimental studies in developing rate equations for the enzymatic production of R-phenylacetylcarbinol and by-products N. Leksawasdi 1, M. Breuer 2, B. Hauer 2, P. L. Rogers 1, B. Rosche 1 1 School of Biotechnology and Biomolecular Sciences, University of New South Wales, Sydney, NSW 2052, Australia 2 BASF-AG, Fine Chemicals and Biocatalysis Research, Ludwigshafen, Germany INTRODUCTION The production of the pharmaceutical intermediate, (R) ‑ phenylacetylcarbinol (PAC) is mediated by pyruvate decarboxylase (PDC) in a two substrates reaction involving pyruvate and benzaldehyde. PAC is a valuable precursor for the production of ephedrine and pseudoephedrine, which are used in the treatment of asthma and nasal congestion symptoms. Ligation of benzaldehyde and ‘active acetaldehyde’, produced after CO 2 removal from pyruvate on the active site of PDC, results in the formation of PAC. A proton is consumed in this process (Rosche et al. 2001). Alternatively, decarboxylation of pyruvate can be followed by the release of the by-product acetaldehyde or the formation of the by-product acetoin. The latter species is formed when ‘active acetaldehyde’ undergoes nucleophilic attack (Lobell & Crout 1996) with free acetaldehyde, released earlier from the PDC-acetaldehyde complex. Fig. 1(a) & 1(b) depicts the formation of PAC and its related by-products from PDC in two and three dimensions, respectively. Fig. 1(a) & 1(b): Production of PAC in 2- & 3-D, simplified form of the reaction mechanism shown in 1(b) is used in the derivation of theoretical rate equations MATERIALS AND METHODS PAC and benzaldehyde concentrations were determined by HPLC with UV detection at 283 nm as described by Rosche et al. (2001). Concentrations of pyruvate and acetaldehyde were determined by enzymatic NADH coupled assay with alcohol dehydrogenase and lactate dehydrogenase, respectively (Rosche et al. 2002). Quantification of acetoin was done by GC with flame ionisation detector as described previously (Rosche et al. 2002). To determine changes in PDC activity, gel filtration columns (Micro Bio-Spin ® 6, cat. no ) were used to remove low molecular weight compounds from the reaction solutions (Rosche et al. 2002). PDC was recovered and mixed with collection buffer at the lower end of the column, incubated on ice for 20 min and analysed for carboligase as described by Rosche et al. (2002). OBJECTIVE To develop a set of differential equations that is able to predict the profile of PAC batch biotransformation so that it can be used later in the design of substrate feeding profile to optimize the PAC production in fed-batch system. PARAMETER ESTIMATION PROGRAM The computations of parameter values from initial rate kinetic and batch biotransformation data were carried out using a parameter-searching program described previously (Leksawasdi et al. 2001). The program was written in Microsoft Visual Basic for Application 6.3 under Microsoft EXCEL ® The optimal parameter values were determined when the minimum total residual sum of squares (  RSS) was achieved between predicted profile and experimental data. The method of King and Altman (1956) was selected for the derivation of theoretical rate equations from the simplified rate mechanism of Fig. 1(b). This method had been used by other authors (Wong & Hanes 1962, Volkenstein & Goldstein 1966) as a basis for development of further methods for analysis and derivation of the rate equations. RESULTS (Initial rate & Deactivation by benzaldehyde) RESULTS (Rate equations & Confirmation of model) CONCLUSION In summary, the model provides good prediction of the specified batch biotransformation and has the potential to optimize a fed-batch production process of PAC. REFERENCES King EL, Altman C (1956) A Schematic method of deriving the rate laws for enzyme catalyzed reactions. J. Phys. Chem. 60: 1375–1378. Leksawasdi N, Joachimsthal EL, Rogers PL (2001) Mathemical modelling of ethanol production from glucose/xylose mixtures by recombinant Zymomonas mobilis. Biotechnol. Lett. 23: Lobell M, Crout DHG (1996) Pyruvate decarboxylase: A molecular modelling study of pyruvate decarboxylation and acyloin formation. J. Am. Chem. Soc. 118: Rosche B, Leksawasdi N, Sandford V, Breuer M, Hauer B, Rogers PL (2002) Enzymatic (R)-phenylacetylcarbinol production in benzaldehyde emulsions. Appl. Microbiol. Biotechnol. 60: Rosche B, Sandford V, Breuer M, Hauer B, Rogers PL (2001) Biotransformation of benzaldehyde into (R)-phenylacetylcarbinol by filamentous fungi or their extracts. Appl. Microbiol. Biotechnol. 57: Volkenstein MV, Goldstein BN (1966) A new method for solving the problems of the stationary kinetics of enzymological reactions. Biochim. Biophys. Acta 115: 471. Wong JT, Hanes CS (1962) Kinetic formulations for enzymic reactions involving two substrates. Can. J. Biochem. Physiol. 40: Fig. 2(a)-2(d): Initial rate (enzyme activity, benzaldehyde, & pyruvate effect) and enzyme deactivation studies. Reaction buffer contained 2.5 M MOPS, 1 mM MgSO 4, 1 mM TPP, at 6  C (pH 7.0), 1.5 ml reaction volume: 2(a) both substrates were kept at 100 mM, 2(b) & 2(c) benzaldehyde & pyruvate concentration were 100 mM, 2(d) no pyruvate. The initial enzyme activity was 3 U/ml carboligase for 2(a) – 2(c). Fig. 3(a)-3(e): Combined ODEs from theoretical and experimental studies are shown in Fig. 3(a) with corresponding parameter values in Fig. 3(b). Two batch biotransformation kinetics (Fig. 3(c) & 3(d)) were used in the estimation of parameter k 2, V q, & V r. The confirmation of model prediction was done on independent batch biotransformation kinetics (Fig. 3(e)). The batch biotransformation reaction volume was 10 ml with the same buffer species & composition as Fig. 2. The initial PDC activity was 3.0 U carboligase ml -1.